DETERMINANTS OF AAV TRANSPORT ACROSS THE BLOOD-BRAIN BARRIER

Blake Harris Albright

A dissertation submitted to the faculty at the University of North Carolina at Chapel Hill in partial

fulfillment of the requirements for the degree of Doctor of Philosophy in the Curriculum of

Genetics and Molecular Biology in the School of Medicine.

Chapel Hill

2018

Approved by:

Aravind Asokan

Ron Swanstrom

Tal Kafri

Juan Song

Saskia Neher

ii

©2018

Blake Harris Albright

ALL RIGHTS RESERVED

iii

ABSTRACT

Blake Harris Albright: Determinants of AAV Transport Across the Blood-Brain Barrier

(Under the direction of Aravind Asokan)

Adeno-associated virus (AAV) is currently the most widely used gene therapy vector for treating

neurological diseases, showing promising results in preclinical and clinical studies. Nonetheless, many

challenges limit effective gene transfer to the central nervous system (CNS) via the vasculature, which

requires that AAV vectors cross the blood-brain barrier (BBB). As a consequence of sub-optimal

transduction efficiency, current vectors must be administered at high dosages to achieve therapeutic CNS

gene transfer and are limited by off-target tissue transduction/sequestration. Thus, developing vectors

with improved efficiency and specificity is critical towards achieving widespread and therapeutic gene

transfer to the CNS. This requires a better understanding of the structural determinants for AAV tropism

and neurovascular transport – which is the primary aim for this dissertation. To achieve this, we

generated a chimeric capsid library between two highly homologous serotypes – AAVrh.10, which

crosses the BBB, and AAV1, which is limited to the vasculature. Through screening individual variants in

vivo, computational analyses, and rational design, we mapped a footprint from the AAVrh.10 capsid

which confers BBB transport and widespread CNS transduction when grafted onto AAV1. In this way,

we engineered the novel and neurotropic AAV1RX capsid, which mediates robust gene transfer

throughout the brain, with reduced glial and endothelial transduction, and is detargeted from peripheral

tissues (i.e. liver).

We hypothesized that the 1RX footprint may alter capsid-sialic acid (SIA) interactions in a way

that promotes BBB traversal and CNS transduction. We tested this through functional characterization of

several capsid variants with alterations in their SIA binding site. These capsids were functionally grouped

according to their differential dependencies on cell-surface SIA for in vitro transduction and binding.

iv

Further evaluation in vivo revealed an inverse correlation between capsid-SIA interactions and BBB

transport/CNS transduction. These experiments suggest that more moderate capsid-SIA interactions may

facilitate, though only partially explain, the 1RX phenotype. Nonetheless, this led us to propose a

Goldilocks model wherein the ideal glycan binding affinity, in conjunction with other contributing

factors, impacts BBB traversal and CNS transduction. Thus, this dissertation sheds light on AAV

determinants for BBB transport, providing a structure-guided platform for engineering improved CNS-

targeted AAV vectors.

v

To my fiancée, Allison Napier

&

My parents, Bill and Sharon

vi

ACKNOWLEDGEMENTS

First and foremost, I want to thank my advisor, Aravind Asokan. I am grateful for his mentorship,

for investing in me, and for his patience, wisdom, and guidance. More so than an advisor or boss, I have

always thought of him as my coach – consistently pushing me to ensure I grow and to expand my

perspective. Above all, he’s taught me to set my sights high and to never let an opportunity pass, focusing

not on my perceived limitations, or on the blood, sweat, and tears which come from the path forward, but

to see only the fire ahead, focused only on how bright it burns. I want to thank him for training me to be

an independent and well-rounded scientist and person. I consider myself very fortunate to have him as a

mentor, and it’s something that I will always cherish.

I would like to thank each member of my thesis committee, Drs. Ron Swanstrom, Juan Song,

Saskia Neher, and Tal Kafri, for their time and valuable input. Additionally, I want to thank the Genetics

& Molecular Biology curriculum, especially John Cornett, for providing support and resources, and for

making everything run smoothly throughout graduate school.

I would especially like to thank all of the members, past and present, of the Asokan lab family.

My lab mates have been less like coworkers and more like a family. Thank you all for your support,

friendship, and for all the laughs, through the many ups and downs of graduate school. You have all been

such a strong supportive force in my life and I am truly honored to have spent these years with you. I am

genuinely honored to have had the privilege of being a part of the Asokan lab and am thankful for the

training and mentorship I have received, which has sculpted me into both the scientist and person that I

am today. Briefly, I would like to mention a few of you here. Giridhar Murlidharan, Garrett Berry, and

Erin Borchardt, thank you for welcoming me into your family, for your support, and for mentoring me

through the early years of graduate school. Victor Long Ping Tse and Sven Moller-Tank, thank you for

your countless hours of entertainment, advice, and mentorship. Victoria Madigan, Patrick Havlik, Rita

vii

Meganck, Ruth Castellanos, Lavanya Rao, and Claire Storey, and to every other person who has been a

part of the Asokan lab, thank you for your support, your friendship, for all the fun times, and for playing

an important part in my life.

Throughout graduate school, I have had the honor of meeting many wonderful people and

developing lasting friendships with a large number of companions through UNC, especially those from

my BBSP cohort. You’ve all made grad school an incredibly fun time and I’m thankful for your

companionship along the journey. While there are many more, there are a few that I would like to

especially thank for their friendship and support through the years - Andy Chan, Temperence Rowell,

Amanda Raimer, Rachel McMullan, Anthony Arceci, Kevin Santa Maria, Anne Beall, Heather Vincent,

Zach Nash, Chris Holmquist, Raulie Raulerson, and Ray Morales. And a special thanks to those listed in

my weekly DnD group for providing some healthy escapism, ridiculousness, and a weekly night full of

laughs and adventure. I want to acknowledge and thank Eric Lange and Houston Fullerton, my two best

friends of nearly a decade, for their friendship, support, and inspiration. I’m grateful to be able to

maintain our friendship despite the distance between us.

I am especially grateful to Ray Morales and Jen Hernandez, as well as their furry companions –

Garfield, Masha, and Felix – for their support, friendship, and for generously opening up their home to

Allison and me during my last month of grad school, as we’re in between apartments and preparing to

move to New York. I can’t thank you enough for both providing us with a temporary home and for

providing me with a workspace for writing this dissertation.

I want to thank my family for their bottomless love and support throughout not just graduate

school, but my entire life. I am incredibly fortunate to have parents, Sharon and Bill, as well a sister,

Rachel, who are so supportive and compassionate. I have always been able to count on them to help get

me through all the rough times in my life and am thankful to have been able to share all of the good times

with them as well. Thank you for always supporting me in any way you can, for listening, and for trying

your very best to understand the basics of how graduate school works as well as the science. Thank you

for being my rock, for your inspiration, and for raising me to be the person I am today.

viii

Lastly, I want to thank my loving fiancé, Allison. I have acknowledged a lot of people for their

support in this dissertation, but you are by far my greatest source of support, strength, and inspiration.

Thank you so much for being my best friend, for your perpetual love and support – whether a simple hug

after a long and stressful day in lab or coming with me back to lab at night or on a weekend and only

minimally complaining, as what I promised would only be a quick 10 minutes in lab somehow turns into

two hours. Thanks for the support from afar during the first few years of grad school, as we were long

distance, and thank you for facilitating all the long hours of work by taking an unfair amount of

responsibilities on your shoulders, and for picking up all my slack during the last couple months of

graduate school and while writing this dissertation. I am constantly amazed at what a caring and

thoughtful person you are and you inspire me to be a better person every single day. Finally, I am

thankful for all of your support, patience, and understanding through all the times where I may have been

difficult to deal with due to being stressed, anxious and overwhelmed by work. Your love and support

over the years has enabled me to dedicate myself towards achieving these goals and has allowed me to

complete graduate school with my sanity intact.

ix

TABLE OF CONTENTS

LIST OF FIGURES ..................................................................................................................................... xi

LIST OF ABBREVIATIONS .................................................................................................................... xiii

CHAPTER 1: INTRODUCTION – AAV BIOLOGY AND USE AS CNS-TARGETED

GENE THERAPY VECTOR ....................................................................................................................... 1

1.1. AAV Biology Overview ........................................................................................................... 1

1.2. AAV Genome & Replication .................................................................................................... 1

1.3. AAV Lifecycle .......................................................................................................................... 3

1.4. AAV Capsid Structure .............................................................................................................. 4

1.5. Background on Natural AAV Serotypes ................................................................................... 7

1.6. Glycan Interactions and Receptor Usage .................................................................................. 8

1.7. Gene Therapy and AAV Vectors ............................................................................................ 11

1.8. Gene Therapy for the Central Nervous System Using AAV Vectors ..................................... 17

1.9. AAV in the Brain – Phenotypes, Routes of Delivery, & Transport ........................................ 18

1.10. CNS Gene Transfer Across the Blood-Brain Barrier Using AAV Vectors .......................... 20

1.11. Engineering AAV Capsids for Gene Transfer to the Brain .................................................. 22

CHAPTER 2: MAPPING THE STRUCTURAL DETERMINANTS REQUIRED FOR

AAVRH.10 TRANSPORT ACROSS THE BLOOD-BRAIN BARRIER ................................................. 26

2.1. Overview ................................................................................................................................. 26

2.2. Introduction ............................................................................................................................. 27

2.3. Methods .................................................................................................................................. 28

2.4. Results ..................................................................................................................................... 32

2.5. Discussion ............................................................................................................................... 41

CHAPTER 3: CAPSID-SIALIC ACID INTERACTIONS INFLUENCE ADENO-

ASSOCIATED VIRAL TRANSPORT ACROSS THE BLOOD-BRAIN BARRIER .............................. 56

x

3.1. Overview ................................................................................................................................. 56

3.2. Introduction ............................................................................................................................. 57

3.3. Methods .................................................................................................................................. 59

3.4. Results ..................................................................................................................................... 63

3.5. Discussion ............................................................................................................................... 72

CHAPTER 4: CONCLUSIONS AND FUTURE DIRECTIONS .............................................................. 83

4.1. Overview ................................................................................................................................. 83

4.2. Mapping Structural Determinants for Transport of AAV across the Blood-

Brain Barrier. ................................................................................................................................. 84

4.3. Influence of Capsid-Glycan Interactions on Blood-Brain Barrier Transport,

Neurotropism, and Transduction ................................................................................................... 85

4.4. Future Directions .................................................................................................................... 87

REFERENCES ........................................................................................................................................... 97

xi

LIST OF FIGURES

Figure 1.1. Structure of the AAV Viral Protein monomer .......................................................................... 24

Figure 1.2. AAV capsid structure and axes of symmetry ........................................................................... 25

Figure 2.1. Phylogenetic and structural analyses of the AAV1/rh.10 domain swap capsid

library .......................................................................................................................................................... 45

Figure 2.2. In vivo screen yields AAV1/rh.10 chimeras capable of crossing the blood-

brain barrier following intravenous administration..................................................................................... 46

Figure 2.3. In vivo screen of AAV1/rh.10 chimeras for their ability to cross the blood-

brain barrier and transduce various brain regions following intravascular administration ......................... 47

Figure 2.4. CNS transduction profile of AAV1R6 and AAV1R7 compared to parental

AAV1 and AAVrh.10 in the brain .............................................................................................................. 48

Figure 2.5. Quantitative comparison of neuronal and glial transduction levels for parental

and chimeric capsid variants ....................................................................................................................... 49

Figure 2.6. Relative cardiac and liver transduction by AAV1R6 and AAV1R7 compared

to parental capsids ....................................................................................................................................... 50

Figure 2.7. Structural analysis of the AAV1R6 chimeric capsid variant .................................................... 51

Figure 2.8. Rational design and functional mapping of a minimal AAVrh.10 footprint for

crossing the BBB ........................................................................................................................................ 52

Figure 2.9. Quantification of neuronal and glial transduction levels mediated by

AAV1RX compared to parental and other chimeric capsid variants .......................................................... 53

Figure 2.10. Peripheral tissue transduction and biodistribution of AAV1R6, AAV1R7 and

AAV1RX following intravenous administration ........................................................................................ 54

Figure 2.11. Sequence alignment of 1RX footprint, and Variable Region I (VR-I) across

common AAV serotypes ............................................................................................................................. 55

Figure 3.1. Structural models of AAV capsids used to interrogate the role of sialic acid

interactions upon blood-brain barrier transport & CNS tropism ................................................................ 75

Figure 3.2. Selected capsid variants display intermediate transduction profiles in a cell

type-specific manner in vitro ...................................................................................................................... 76

Figure 3.3. Capsid variants display differential sensitivities to removal of cell-surface

sialic acid for both transduction and binding .............................................................................................. 77

Figure 3.4. Evaluating the impact of Sialic Acid binding on blood-brain barrier traversal

and CNS transduction in vivo ..................................................................................................................... 78

Figure 3.5. Quantitation of neuronal and glial transduction profiles for capsid variants

displaying differential dependencies on Sialic Acid ................................................................................... 79

xii

Figure 3.6. Relative liver transduction mediated by capsid variants with differential sialic

acid sensitivities .......................................................................................................................................... 80

Figure 3.7. Structural analysis of the 1RX footprint within the context of the AAV1 Sialic

acid binding pocket ..................................................................................................................................... 81

Figure 3.8. Goldilocks model for the influence of capsid-sialic acid interactions on blood-

brain barrier transport and CNS transduction ............................................................................................. 82

Figure 4.1. Transduction of the choroid plexus and paraventricular nucleus of the

thalamus following intravenous administration .......................................................................................... 96

xiii

LIST OF ABBREVIATIONS

AADC Amino acid decarboxylase

AAP Assembly activating protein

AAT Alpha-1-antitrypsin

AAV Adeno-Associated Virus

AAVR Adeno-associated Virus Receptor (aka KIAA0319L)

AAVS1 AAV integration site 1

Ad Adenovirus

AD Alzheimer’s Disease

AMG Amygdala

ANOVA Analysis of variation

BBB Blood-Brain Barrier

bp base pairs

CA1/2/3 Cornu Ammonis, pyramidal neuron layers 1/2/3 of the

hippocampus

cap capsid gene

CBA Chicken beta actin

CBh Chicken beta actin hybrid

cDNA Complementary DNA

CMAH cytidine monophosphate-N-acetylneuraminic acid

hydroxylase CMV cytomegalovirus

CNS Central Nervous System

CREATE Cre Recombinase-based AAV targeted evolution

Cryo-EM Cryo-electron microscopy

xiv

CSF Cerebrospinal fluid

CTX Cortex

DAPI 4’6-diamidino-2-phenylindole

DG Dentate Gyrus

DMD Duchenne muscular dystrophy

DMEM Dulbecco’s Modified Eagle Media

DMSO Dimethyl sulfoxide

DNA Deoxynucleic acid

dsDNA double-stranded DNA

E1a/b Adenovirus early protein 1a/b (transactivator)

E2 Adenovirus early protein 2 (DNA-binding protein)

E4 Adenovirus early protein 4

ECM extracellular matrix

EM Electron Microscopy

FA Friedrich’s ataxia

FBS Fetal bovine serum

FDA Food and drug administration

g gram

GABA Gamma-aminobutyric acid

GAL Galactose

GAPDH Glyceraldehyde-3-phosphate dehydrogenase

GFP Green Fluorescent Protein

GMP Good manufacturing practice

HC Hippocampus

xv

HEK293 Human embryonic kidney cell line

HIV Human immunodeficiency virus

hr Hour

HS Heparan Sulfate

HSPG Heparan sulfate proteoglycan

HSV Herpes Simplex Virus

hSyn human synapsin

HY Hypothalamus

i.v. Intravenous

IACUC Institutional Animal Care and Use Committee

IC Intracranial/Intracisternal

ICV Intracerebroventricular

IM Intramuscular

IT Intrathecal

ITR Inverted Terminal Repeats

IV Intravenous

kb kilobases

kg kilogram

LacNac N-acetyl-lactosamine

LSD Lysosomal Storage Disease

Luc Luciferase

MB114 Mouse Brain Microvascular Endothelial Cell Line

MCT Motor Cortex

MEM Minimal essential media

xvi

Neuro2a Mouse neuroblastoma cell line

NIH National Institutes of Health

nm nanometer

ns Not significant

NVU Neurovascular unit

PBS Phosphate Buffered Saline

PCR Polymerase chain Reaction

PD Parkinson’s Disease

PEG Polyethylene glycol

PEI Polyethylenimine

PFA Paraformaldehyde

PSA Polysialic acid

qPCR quantitative PCR

rAAV recombinant Adeno-Associated Virus

Rep Replication gene

RMS Rostral migratory stream

RNA Ribonucleic acid

scGFP Self-complementary Green Fluorescent Protein

SCT Somatosensory Cortex

SD Standard Deviation

SEM Standard error of the mean

Sf9 Insect cell line

SIA Sialic Acid

Sia Sialic acid

xvii

SMA Spinal Muscular Atrophy

SMN Survival of motor neuron

ssDNA single-stranded DNA

ssGFP single-stranded Green Fluorescent Protein

ssLuc single-stranded luciferase

STR Striatum

TH Thalamus

TR Terminal Repeat

U87 Human astrocytoma cell line

UNC University of North Carolina

VA RNA Viral associated RNA

Vg viral genomes

VP Viral protein

VR (I-IX) Variable regions (I-IX)

WT wildtype

αA α-helix A

βB-H β-sheets B-through-H on the AAV monomer

1

CHAPTER 1: INTRODUCTION – AAV BIOLOGY AND USE AS CNS-TARGETED GENE

THERAPY VECTOR

1.1. AAV Biology Overview

In the famous words of Bruce Lee, “Simplicity is brilliance1.” When it comes to viruses, Adeno-

associated virus (AAV) is as simple as they come. Originally discovered as a contaminant in Adenoviral

preparations in the 1960s, by Atchison and Rowe, AAV is a small (~20 nm) virus characterized by a non-

enveloped icosahedral capsid which contains a 4.7kb single-stranded DNA (ssDNA) genome2–4,. As a

helper-dependent member of the Parvoviridae family, AAV requires helper gene functions provided by

co-infection with a helper virus such as Adenovirus or various herpesviruses5–8. In addition to its

replication-defective nature, AAV is nonpathogenic, yet is able to persist within the host cell, and several

serotypes are prevalent in the human population8–12. These attributes, in part, make it an ideal candidate

for gene therapy applications, but also suggest that AAV may have formed an ideal relationship with the

human host. Furthermore, because AAV can replicate up to 1 x 106 vg copies per cell, which generally

causes cell lysis, only during co-infection with a helper virus, it has been theorized that AAV may

possibly have a beneficial effect on the host by providing protection from other viral infections, i.e.

Adenovirus or herpesviruses8.

1.2. AAV Genome & Replication

As mentioned above, AAV is a remarkably simple virus, which is reflected in the minimalist

nature of its 4.7 kb ssDNA genome. With a limited coding capacity, the AAV genome consists of the

nonstructural rep (replication) gene and the structural Cap (capsid) gene, which are flanked by 145 bp

inverted terminal repeats (ITRs) that form base paired hairpin structures at both ends of the genome8.

These serve as the only cis requirement for packaging the genome into the capsid13,14. Without a

repertoire of genes for manipulating the cellular environment, AAV is reliant upon cellular components

2

(i.e. transcriptional machinery) in addition to helper gene functions, for replication and transcription of its

genome8,15.

Despite its simplicity, the AAV genome possesses additional complexity in its transcripts,

accomplished through the use of different promoters, alternative start sites, and alternative splicing8,16,17.

There are 3 promoters within the AAV genome, two of which control transcription of the rep gene, with

additional variation due to alternative splicing of a single intron to generate a total of 4 transcripts -

Rep78, Rep58, Rep52, and Rep408,16,18. Rep78/58 are both transcribed from the p5 promoter and are the

longest of the transcripts, with the sole difference between the two variants being inclusion and exclusion,

respectively, of an intron at the 3’-end. The two smaller transcripts, Rep52/40, are transcribed by the p19

promoter and are similarly differentiated by alternative splicing.

The cap gene is transcribed by the p40 promoter and generates three transcripts via alternative

start sites, which in turn produce the structural viral proteins VP1, VP2, and VP319–21. Each of these

variants possess a shared C terminus, while VP2 and VP1 additionally possess unique N termini, with

VP1 being the largest of the three21. The AAV capsid is composed of 60 copies of the three capsid

proteins (CAP) – VP1, VP2, and VP3, assembled in a 1:1:10 stoichiometric ratio. The VP3 protein serves

as the predominant capsid protein while VP2 is nonessential for capsid assembly; however, the VP1

unique region (VP1u) possesses phospholipaseA2 activity, which is important for endosomal escape and

infectivity. Recently, the cap gene of AAV was found to encode an additional (~ 200 amino acid) protein

encoded by an alternative open reading frame resulting from a +1 frameshift21–26. It was discovered that

this protein plays a role in capsid assembly, and was accordingly named Assembly-activating protein

(AAP). Interestingly, AAP is found to be essential for capsid assembly, in a serotype-dependent manner.

While the mechanism for AAP-mediated capsid assembly is unknown, it possesses an NLS and recent

studies have suggested that it may play a role as a chaperone to bring AAV capsid proteins to the

nucleus25–27. Very recently, our lab has mapped the functional domains required for AAP activity,

identifying linker domains which can be replaced with alternative domains to facilitate engineering new

functions onto AAP28.

3

The first step for replication of the AAV genome is second-strand synthesis, which converts the

ssDNA genome into a transcriptionally competent double-stranded DNA (dsDNA) template8,29,30. The

next step is transcription of the rep genes. Transcriptional repression by Rep78/58 is lifted when Ad

helper genes allow for Rep-mediated activation of AAV promoters8,31,32. The ITRs serve as the origin of

replication for the AAV genome, containing a rep binding site for Rep-mediated activation, and

additionally providing a base-paired free 3’-hydroxyl group which allows DNA synthesis to occur8,32,33.

Subsequent replication of the ITR occurs, mediated by binding of Rep and nicking at the terminal

resolution site within the ITR sequence, to facilitate fidelity of ITR replication and to generate a new free

3’-hydroxyl group for subsequent rounds of replication8,29,30. Replication of the AAV genome is known to

produce two products due to resolution of the ITR – a dsDNA molecule and a displaced ssDNA molecule.

Productive AAV infection requires helper gene functions which aid in AAV replication, from

helper viruses such as Ad and HSV6,8. These essential helper genes are E1a/b, E2a, E4, and the viral

associated RNA (VAI RNA)8,15,34. The E1a gene product is a transactivator which relieves repression for

the AAV p5 promoter8,15. Briefly, the proteins encoded by these helper genes function to activate,

promote, and ensure processivity of replication and transcription of the AAV genome, while the VA RNA

is thought to promote translation of AAV transcripts35. Importantly, these helper functions can be

provided in trans for virus production15.

1.3. AAV Lifecycle

The first step in the lifecycle of AAV is binding to the cell surface, which is facilitated by

recognition of primary glycan receptors19,36. The identity of the glycan recognized varies depending on

the serotype. After binding to the cell surface, uptake occurs, generally via endocytosis through clathrin-

coated pits22,37. Importantly, several secondary co-receptors have been identified, which differ for each

serotype and enhance transduction, likely through facilitating uptake. Examples include integrins for

AAV2 & 9, as well as various growth factor receptors (i.e. Fibroblast growth factor receptor for AAV2 &

4, hepatocyte growth factor receptor (for AAV2 & 3), platelet-derived growth factor receptor (for AAV5)

and epidermal growth-factor receptor for AAV6)21,38–46. AAV virions then traffic through the endosomal

4

pathway before escaping from early endosomes into the cytosol, a process dependent on the pH-

dependent phospholipase A2 domain on VP122–24,47,48. The virion is then thought to traffic through the

trans-golgi network before undergoing nuclear import and subsequent trafficking to the nucleolus, where

capsid uncoating and viral genome release occurs8,22,37,46,49. Second-strand synthesis, a rate-limiting step in

the pathway, then occurs in order to generate a transcriptionally competent template for subsequent

transduction and replication steps22. Following replication, both + and – strands are packaged into newly

formed AAV capsids at equal efficiencies8. It is also important to note that AAV depends upon the host

cell machinery for both transcription and translation.

AAV is capable of persisting in the host cell as an episome, enabling long-term gene expression,

and is also able to enter a latent phase which is characterized by targeted genomic integration50,51. The

AAV genome integrates into a specific locus on the short arm of chromosome 19, designated the AAVS1

site52. Importantly, integration is dependent on the Rep protein and occurs only for WT virus in the

presence of Rep and helper genes, and thus does not occur for recombinant viral vectors53.

There are also several rate-limiting steps along the AAV transduction pathway that have been

identified. Briefly, these include trafficking of the capsid to the nucleus22, the rate of capsid uncoating and

subsequent release of the viral genome, and second-strand synthesis to generate a dsDNA molecule that

can be transcribed8,22,29. The requirement for second-strand synthesis can be overcome through production

of self-complementary AAV vectors. These are generated through deletion of the terminal resolution site,

a sequence in the ITR, which is required for resolving replication. This deletion enables formation of a

partially dsDNA genome packaged into the capsid, bypassing the second-strand synthesis step required

for the usual single-stranded genome29,30. Lastly, stability of the vector genome is also known to affect

transduction8.

1.4. AAV Capsid Structure

The AAV capsid is a small (~20nm) icosahedron composed of 60 VP monomers - VP1, VP2 &

VP3, assembled in a 1:1:10 stoichiometric ratio19. The capsid has a T=1 symmetry (where T is the

triangulation number), meaning the icosahedron consists of 12 pentameric clusters of capsid subunits. The

5

predominant VP3 monomer is composed of a highly conserved core β-barrel jellyroll structure which

consists of 8 antiparallel βsheets (βB-H) and a similarly conserved α-helix (αA) (Fig. 1.1.)19,20,54. These

strands are connected by interdigitating loop regions that are highly variable in sequence identity and are

identified based upon the strands they connect. This basic structure is conserved across both

dependoviruses as well as the autonomous parvoviruses19,55,56. The loops are surface-exposed on the full

60mer particle and their interactions dictate the surface topology on the capsid. Consequently, these

surface loops are critical determinants of tropism, antigenicity, and receptor usage across capsid

serotypes19,57–59.

Furthermore, nine highly variable regions (VRs) are detected in sequence alignments of the

primary amino acid sequences across various capsid proteins from various AAV serotypes19,57. They are

easily demarcated from the more conserved regions in these alignments. These VR regions, labeled VR-I

through VR-IX, vary in length from just a few to a couple hundred amino acids in length. The residues in

these VRs overlap with the loop regions between β-strands, and diversity in these VRs dictates nearly all

phenotypic differences between various AAV capsid serotypes.

The full length amino acid sequence of VP1 is ~736 amino acids, varying based on relative

deletions and/or insertions present across different serotypes. VP1, VP2 and VP3 possess a shared C-

terminus; however, VP2 and VP3 are differentially truncated and thus possess different N termini. The

VP1 unique N-terminal region is ~140 residues, which is absent in the truncated VP2 sequence (full VP2

length is ~600 amino acids)19,21,23,47,60,61. The further truncated VP3 sequence begins around residue 217 of

the VP1 sequence, and consists of the latter ~520 residues of the VP1 sequence, which again varies across

serotypes.

Much of the work dissecting AAV capsid structure and correlating structure with function has

been done through genetic modification of capsids and through solving capsid structures using X-ray

crystallography and/or cryo-EM. Structures have been solved for many of the common serotypes (i.e.

AAV1-AAV9), and even for some capsid-receptor pairs19,60,69–71,61–68. As mentioned before, the AAV

capsid is composed primarily of the VP3 monomer. This is reflected in the inability to obtain

6

crystallographic information for the asymmetrically incorporated VP1 and VP2 monomers, which appear

disordered on the inside of the capsid19,68. Thus, structures have been solved for the VP3 capsid

interactions, neglecting the VP1 and VP2 molecules. It is also of note, that the VP1/2 unique regions are

not known to directly impact features such as tropism, antigenicity, and receptor usage. Rather, these

determinants seem to reside on the VP3 sequence. Nonetheless, the phospholipase A2 domain of the

VP1u region is critical for infectivity21,23,47,54,72.

The interactions between the loops of each monomer manifest as various landmark structures that

are recognizable on the various axes of symmetry at the capsid surface (Fig. 1.2.). The 5-fold axis of

symmetry is a pentamer of VP3 monomers characterized by a cylindrical pore, formed by interactions

between the DE loops (corresponding to VR-II) of monomers, and is encircled by a local depression

formed largely by interactions involving the HI loop21,73–75 This region is known to be important for

capsid assembly and for packaging of the viral genome into the capsid, with the viral genome thought to

be inserted through the 5-fold pore54,76.

The 3-fold axis of symmetry is formed by a VP3 trimer and is characterized by three large surface

protrusions, formed from interactions between residues within the BC, EF and GH loops (corresponding

to VR-IV, VR-V, VR-VI, and VR-VIII) on VP monomers, with the GH loops specifically contributing to

the three-fold protrusions19,55,57,77,78. These protrusions, or spikes, have been identified as perhaps the most

significant region on the capsid for affecting capsid-host interactions, dictating tropism, transduction and

antigenicity. Likewise, the clustered VR-IV, VR-V, VR-VI, and VR-VIII localized to the spikes are key

targets for mutagenesis to affect these phenotypes and dictate differences between serotypes. For

example, structural studies have demonstrated that while most of the common AAV serotypes have

relatively sharp protrusions jutting out at the 3-fold axis, some serotypes (i.e. AAV4) display more

rounded protrusions, which can be accounted for by sequence variations at these regions21,57,59,79,80.

Recent structural studies have revealed the glycan receptor binding sites (aka footprints) for

several AAV capsid-glycan pairs. For instance, the precise residues in such receptor footprints have been

identified by co-crystallography studies for the AAV1/6-SIA complexes, for the AAV4-mucin complex,

7

for the AAV5-SIA complex, and for the AAV9-GAL complex12,36,38,64,68,81–85. Each of these capsids have

been found to bind their respective glycan receptors either on or near the three-fold protrusions. The

binding pocket for each pair accordingly involves many of the same amino acid residues, most of which

are present in VR-IV, VR-V, VR-VI, and VR-VIII, as well as several residues in VR-I. Nonetheless, there

is some variation in the precise residues involved, directly and indirectly, in glycan binding.

The 2-fold axis of symmetry is characterized by a topological depression at the interface between

two trimers19,20,55,72. This 2-fold channel, also known as the 2-fold dimple, is formed by interactions

between VR-I and VR-IX on adjacent monomers, with the α-A helix contributing to formation of the wall

lining the 2-fold depression19,57,72,86. The residues composing VR-I, from the B-E loop on the monomer,

are also clustered at the shoulder near the base of the three-fold protrusions.

1.5. Background on Natural AAV Serotypes

A large array of AAV variants exist, both natural and synthetic12,20,87–89. The more common of the

natural serotypes discovered include AAV1-AAV9. Out of these, AAV2 is by far the most highly

characterized and has historically served as the representative model for most of the studies that have

dissected the molecular biology of AAV, elucidating replication and transduction pathways amongst other

steps in the AAV lifecycle90. The numerous natural serotypes discovered thus far display broad tropisms

at the species, tissue, and cellular levels in addition to broad antigenic profiles12.

In recent years, hundreds of AAV sequences have been isolated from various species. For the

most part, these variants were isolated from mostly human and non-human primates (i.e. rhesus

macaques), but new AAV sequences have also been isolated from many other species, including bovine,

avian, and reptilian species88,89,91–93. AAV2/3/5/6/9 were all isolated from human hosts, while

AAV1/7/8/9, as well as AAVrh.10/rh.8 were all isolated from non-human primates11,88,94,95. The numerous

natural serotypes discovered thus far have been categorized phylogenetically based on their capsid amino

acid sequences and have been organized into several clades along this phylogeny88. While each of the

main clades is represented by a relatively common AAV serotype (AAV1/6 in clade A, AAV2 in clade B,

AAV2/3 in clade C, AAV7 in Clade D, AAV8 in clade E, and AAV9 in clade F), each also includes

8

varying numbers of human-derived and non-human primate-derived variants. AAV4 and AAV5,

however, diverge significantly from most of the other capsids and fall outside of these clades. Due to their

low sequence homology to other AAV serotypes, pre-existing immunity to the AAV4 and AAV5

serotypes is rare in the human population74,96. Additionally, AAV4 and AAV5 have more rounded

protrusions at their 3-fold axes of symmetry, compared to the spikier versions of most other capsids57,71,96.

Here, we will briefly discuss the various tropisms and transduction profiles for some of these

common serotypes. When delivered systemically, AAV1 transduces mostly the vasculature in various

tissues and is generally unable to cross the vascular wall97–99. Nonetheless, AAV1 efficiently transduces

skeletal muscle and can also transduce neurons and glia in the CNS, following direct intramuscular or

intracranial injections, respectively100–103. AAV2 is similar in that it transduces cells in various tissues,

particularly photoreceptors in the eye, with moderate efficiency104–106, but generally relies on direct

injections into tissue, performing poorly when administered systemically. AAV2 performs exceptionally

well in cell culture, however. AAV3 and AAV8 both transduce the liver at high levels, and are being

pursued for liver-targeted gene therapy applications107–110. In humanized murine models, however, AAV3

preferentially transduces human hepatocytes, while AAV8 prefers murine hepatocytes111. AAV8

additionally transduces skeletal muscle as well as pancreatic and cardiac tissues109. Similar to AAV1,

AAV6/7 both transduce skeletal muscle at high levels following intramuscular injections; however,

AAV7 is also capable of transducing the CNS following systemic administration98,112. AAV4 exhibits a

pronounced cardiopulmonary tropism following systemic administration85,112,113. AAV9 is interesting, as

it transduces nearly all tissues very efficiently following systemic administration, outperforming nearly all

of the other common serotypes in any given tissue112,114,115. Thus, AAV9 is a great example of the tradeoff

between efficiency and specificity. Likewise, another serotype, AAVrh.10 (isolated from a rhesus

macaque) behaves very similarly to AAV9 in this regard98.

1.6. Glycan Interactions and Receptor Usage

As previously stated, cell-surface attachment is the first and most crucial step in AAV

infection/transduction. Binding to the cell surface largely relies on recognition of primary glycan

9

receptors. Thus, receptor usage is a key determinant for tropism as well as transduction, and AAV capsids

are known to utilize several different glycans for attachment19,56. For instance, AAV2 and AAV3 both

recognize heparan sulfate proteoglycan (HSPG), and the AAV2-HSPG interaction is historically the most

characterized capsid-glycan interaction63,116.

Sialic acid (SIA), in various linkages, is a common receptor for multiple AAV serotypes,

including AAV1, AAV4, AAV5, and AAV6. AAV1 recognizes α2,3- and α2,6-N-linked SIA, whereas

AAV5 recognizes only α2,3-N-linked SIA117,118. The AAV6 capsid recognizes both HSPG as well as

α2,3- and α2,6-N-linked SIA117,118. While these capsid variants all recognize SIA in varying N-linkages,

AAV4 is the only variant so far that has been shown to utilize α-2,3-O-linked SIA (also known as

mucin)85. Previous studies from our lab have identified galactose (GAL) as the primary receptor for

AAV981. A sulfated N-acetyllactosamine (LacNac) was recently identified as the receptor for AAVrh.10,

a variant which shares many attributes with AAV9119. Lastly, contrary to the glycoprotein receptors

discussed so far, a glycolipid (ganglioside) was found to be the receptor for bovine AAV (BAAV)120,121.

Nonetheless, the primary receptors for many other AAV variants remain unknown, i.e. AAV7/8 for

instance.

Structural studies involving cryo-EM and X-ray crystallography, along with mutational analyses,

have additionally identified the precise residues of the receptor footprints for several capsids38. For

instance, we know that the binding site for HSPG on AAV2 is located at the 3-fold spikes, utilizing

several basic residues (R484, R585, R487, K532, and R588)116. Additionally, critical residues for binding

HSPG have been identified for AAV3 (R594) and for AAV6 (K459 and K531), which do not overlap

with the AAV2-HS 63,116,122. The receptor footprint for AAV4 has been determined and found to be

located at the tip of the 3-fold spikes, where K492, K503, M523, G581, Q583, and N585 were found to be

the critical residues for binding O-SIA57,85,96,113. The receptor footprint for AAV5 is uniquely located at

the depression between the spikes at the center of the 3-fold axis of symmetry and includes residues

M569, A570, T571, G583, T584, Y585, N586, and L58783.

10

The most important receptor footprint in regards to this dissertation is the SIA binding pocket

shared by AAV1 and AAV6, which is located at the base of the three-fold protrusions117. This footprint is

comprised of 11 critical amino acid residues, including S268, D270, N271, Y445, G470, S472, V473,

N500, T502, and W503. Importantly, only 6/11 of these residues are involved in directly contacting the

SIA (N447S G470, S472, V473, N500 T502, and W503). Interestingly, this footprint is analogous to the

GAL binding pocket on AAV9. The GAL footprint on AAV9 is localized to the same pocket at the base

of the 3-fold protrusions and includes the equivalent residues82. However, the precise amino acid residues

of the AAV9 sequence differ at several of these positions (S269, D271, N272, Y446, S448, T471, A473,

N474, N501, A503, and W504). Both of these binding pockets involve residues derived from VR-I, VR-

IV, and VR-V regions of the VP3 monomer, and the pocket is formed by interactions between loops of

two different monomers.

For each of the capsid-glycan interactions described here, the residues and capsid regions

involved in binding glycan receptors had also previously been implicated as determinants for

transduction, tropism, and/or antigenicity38,123. This further stresses the role glycan receptor usage has on

determining these attributes and reinforces these regions as key structure-function correlates.

Many other, autonomous, members of the parvovirus family also utilize various forms and

linkages of SIA as receptors. For instance, Bocavirus recognizes α2,3-N-linked and α2,3-O-linked SIA,

minute virus of mice (MVM) recognizes α2,3- and α2,8-N-linked SIA, and both canine (CPV) and feline

(FPV) parvoviruses recognize SIA variants as well56. Furthermore, AAV shares several structural

features in common with the autonomous parvoviruses involved in capsid-SIA interactions. For example,

several parvoviruses (i.e. MVM & CPV) are known to bind SIA in a pocket localized to the 2-fold

depression, while others, such as the B19 parvovirus bind SIA in a pocket located at the three-fold

depressions38,55,72,123. While the SIA binding site of AAV is located at the base of the 3-fold protrusions, it

should be noted that this site is adjacent to the 3-fold depression and in close proximity to the 2-fold

depression19,56. Thus, despite some variation, the SIA binding sites across these parvovirus variants are

located at nearby regions of the capsid surface.

11

In addition to primary glycan receptors, subsequent binding to secondary co-receptors has also

been identified as another important step for transduction8,21,38. The nature of such factors have been

identified for several AAV variants. These interactions are thought to further facilitate cellular uptake,

and generally involve various growth factor receptors or integrins. For example, secondary co-receptors

include fibroblast growth factor receptor (FGFR) for AAV2, hepatocyte growth factor receptor (HGFR)

for both AAV2 and AAV3, platelet-derived growth factor receptor (PDGFR) for AAV5, and epidermal

growth factor receptor (EGFR) for AAV640–42,107,124. Additionally, α5β5 integrins have been identified as

important co-receptors for AAV2 and AAV9, while the 37/67-kDa laminin receptor has been found to be

a co-receptor for AAV2/3/8/943,46,125. However, no co-receptor has been identified for some serotypes, like

AAV4 and AAV7.

In addition to these primary glycan receptors and secondary co-receptors, a previously

uncharacterized transmembrane protein, (KIAA0319L), was recently identified as a universal receptor for

AAV. This protein has thus been dubbed the AAV receptor (AAVR)126. While the precise mechanism of

how AAVR impacts transduction remains to be clarified, it is thought to facilitate endocytosis and

subsequent cellular trafficking. Furthermore, the relative role played by AAVR alongside primary glycan

and secondary co-receptors remains contentious.

Lastly, elucidation of these structure-function correlates for glycan receptor usage has afforded

our lab and others the ability to rationally engineer these footprints in order to generate novel AAV

capsids for both gene therapy applications as well as for basic studies aimed at further expanding our

understanding of AAV biology and capsid-glycan interactions38. Historically, such modification of

receptor footprints focused on the AAV2 HSPG binding site, but our lab has generated novel AAV4 and

AAV9-based vectors with altered mucin and GAL interactions that result in unique phenotypes, which

will be discussed in a later section.

1.7. Gene Therapy and AAV Vectors

Background on Gene Therapy Using AAV Vectors. The principal concept behind gene therapy is

to deliver and express a therapeutic transgene in target tissues and cell types in order to treat a disease

12

characterized by an underlying genetic component. For example, in patients afflicted with a disease

caused by loss-of-function mutations in a specific gene, the disease may be corrected by delivery of a

functional copy of the respective gene. While there are many ways to achieve in vivo gene delivery (i.e.

directly administering cDNA, using lipid nanoparticles, etc.), viral vectors are often the most effective

means by which to deliver therapeutic transgenes11,12,127–131. Many other viruses have also been evaluated

as gene therapy vectors, such as Adenovirus, paramyxoviruses, and the more prominent lentivirus132–137.

For the most part, these other viral vectors have been met with concern regarding their safety profile and

efficacy; however, lentiviral vectors remain a promising tool for gene therapy applications in which

chromosomal integration is required – particularly for ex vivo applications through genetically modifying

patient-derived cells for combination gene and cell therapies, such as for hematopoietic stem cells.

Nonetheless, AAV is the leading gene therapy vector.

Several of its favorable attributes make AAV an ideal viral vector for gene therapy applications.

Chief amongst these is its inherent lack of pathogenicity and replication-defective nature12,138,139.

Additionally, AAV infects both dividing and non-dividing cells, and is able to sustain long-term gene

expression in vivo by persisting as an episome within host cells. AAV has an exceptional safety profile,

with a lack of integration for the recombinant virus and a favorable immune response8,140. The large array

of serotypes which display broad tropisms further provide a large toolbox of capsid reagents for targeting

different tissues12. Furthermore, ability to generate high titers of recombinant virus carrying therapeutic

transgene cassettes and the relative scalability of production methods further make AAV an ideal viral

vector, providing the potential to treat a vast number of monogenetic and familial diseases, or any disease

which is treatable by delivery of a therapeutic transgene, afflicting nearly any tissue11,12,141.

In 2018, AAV is currently the premiere viral vector for gene therapy applications. To date, there

have been over 2500 gene therapy clinical trials and two AAV gene therapy products have made it to the

marketplace, one in Europe and the second in the United States142. Additionally, the past several years

have seen a mushrooming of biotech companies emerging and/or incorporating gene therapy-focused

pipelines using AAV vectors to target numerous indications. AAV made its commercial debut in

13

European markets with the release of Glybera® (aka Alipogene tiparvovec) a gene therapy for the

treatment of familial lipoprotein lipase deficiency, which was developed by uniQure and released in

European markets in 2012. Glybera utilizes a single intramuscular injection of an AAV1 capsid to deliver

the human lipoprotein lipase gene to muscle cells in order to treat lipoprotein lipase deficiency142–144. Due

to current challenges with vector production, it is also important to note that Glybera is currently the most

expensive drug in the world, fetching a price around $1,000,000. Furthermore, 2017 became a landmark

year for the gene therapy industry, as Luxturna became the first FDA-approved gene therapy using AAV.

Luxturna (aka voretigene neparvovec-ryzl) was developed by Spark Therapeutics to treat Leber’s

congenital amaurosis, utilizing subretinal injections of an AAV2-based vector to deliver a functional copy

of the RPE65 gene to retinal cells with biallelic RPE65 mutations142,145. Luxturna offers incredible

restoration of vision for patients with this inherited disease; however, like Glybera, is incredibly costly,

amounting to over $4,000 per eye146.

Gene therapy using AAV vectors is being pursued for numerous indications which are currently

being evaluated in preclinical and clinical studies142. AAV-mediated gene therapy has shown remarkable

potential towards the treatment of numerous diseases with a genetic basis. Examples range from

neuromuscular disorders, such as Duchenne muscular dystrophy (DMD) and spinal muscular atrophy

(SMA), to neurocardiovascular diseases such as Friedrich’s ataxia, and cardiovascular diseases11,147–151.

For example, phase III clinical trials are underway for an AAV9 vector developed by Avexis to treat

SMA by delivering a functional copy of the SMN gene to motor neurons following both intravenous and

intrathecal administration152–154. AAV vectors are being developed and evaluated by multiple companies

to treat Friedrich’s Ataxia (FA) and FA-associated cardiomyopathy by using an AAV vector to deliver a

therapeutic frataxin transgene via systemic administration. AAV vectors have also been evaluated in

phase I/II clinical trials for the treatment of Alpha-1 antitrypsin (AAT) deficiency, using an AAV1 vector

delivered via intramuscular injections to deliver a functioning copy of the Alpha-1 antitrypsin gene155.

Additionally, AAV gene therapies are currently being developed for treating numerous

neurological and neurodegenerative diseases characterized by an underlying genetic component. Two of

14

the most predominant indications being pursued in the CNS space are for Parkinson’s disease and

Alzheimer’s disease142. As an example, Voyager has recently developed a drug using an AAV vector to

deliver amino acid decarboxylase to improve L-Dopa levels for the treatment of Parkinson’s, which has

shown promising safety profiles improvement in patient motor functions following phase 1 clinical

studies156. Similarly, gene therapies using AAV vectors could be developed to target Parkinson’s by

delivering various other therapeutic genes to replace those identified as disease-associated mutations,

such as those involved in lysosomal trafficking defects. Less prominent neurological diseases with a

genetic component, such as Gauche’s disease, Pompe disease, and Canavan disease are also being

targeted with AAV-based pipelines157–160.

Production and Purification of Recombinant AAV Vectors. Recombinant AAV (rAAV) vectors

were first generated in the 1980s using AAV2161. Production of rAAV vectors can be done by replacing

the rep & cap genes with a transgene driven by a promoter of choice along with whatever regulatory

elements desired, as long as the total size of the transgene cassette is under the 4.5 kb packaging capacity.

This transgene cassette is inserted between the flanking ITRs, replacing the rep and cap genes of the WT

genome. These ITRs serve as the only cis elements required for packaging. The triple plasmid

transfection method is generally used to produce rAAV vectors. A producer cell line is used to produce

the virus. Generally, HEK293 cells are used for this; however, alternative methods such as insect cell-

based methods (i.e. using baculovirus/sf9 cells) are also used.

In the triple transfection method, the producer cells are transiently transfected with three

plasmids15,141,162. The first is the aforementioned transgene packaging cassette in a TR backbone. The

second plasmid provides the rep and cap genes of the WT AAV genome in trans to facilitate production.

For producing rAAV in the lab, a cap gene of choice is provided alongside the AAV2 rep gene, which

generally remains constant. The third and final plasmid provides helper genes from Ad, which are

required to facilitate replication of the AAV genome15. These plasmids are transfected using PEI and

virus is harvested 2-5 days post transfection, depending on the protocol. Depending on the serotypes

15

being produced, virus may be harvested from the supernatant and/or the cell pellet, depending on the

serotype being made.

Once produced and harvested, a variety of methods are employed to purify rAAV preps141,163–167.

In the lab, small-scale preps are purified using PEG precipitation followed by loading onto a density

gradient (either cesium chloride or iodixanol) and purified by ultracentrifugation. Depending on the

serotype, sonication may be used prior to ultracentrifugation in order to isolate virus from the cells.

Following ultracentrifugation, further purification steps such as dialysis or buffer exchange are used;

additionally, further purification using a sucrose gradient is employed if more pure virus is needed.

Lastly, purified virus preps may be concentrated via centrifugation using size-exclusion columns.

Finally, purified virus preps are titered using qPCR as a gold standard (generally with primers

specific for either the transgene or universal primers targeting the ITR) or dot blots. Subsequent

biochemical methods can be employed as quality control, including Western blots and electron

microscope (to assess purity and ratio of empty:full capsids, respectively). While production of small-

scale virus preps at high titer are relatively easy to perform in an academic laboratory setting, scaling up

to large-scale good manufacturing practice (GMP) virus production is both challenging and costly.

Various methods have been employed to improve these methods, including replacement of adherent

HEK293 producer cells with suspension cultures, and further using larger bioreactors and cell stacks to

process larger volumes166,168–176. Furthermore, a substantial amount of virus is lost in purification steps,

and large-scale manufacturers are employing alternative techniques, such as tangential flow filtration and

various forms of chromatography177–179.

Challenges Facing AAV-Mediated Gene Therapy. Despite its success, there are still several

obstacles to overcome for successful AAV-mediated gene therapy. First, there is the prevalence of pre-

existing neutralizing antibodies in the patient population, which range up to ~80% of the population for

some of the more common AAV serotypes10,180,181. This may preclude patients from treatment/clinical

trials. Both the use of less prevalent capsid serotypes (which are less likely to be recognized by pre-

existing neutralizing antibodies in the patient population) as well as the use of transient immune

16

suppression are two strategies which have shown promise in circumventing pre-existing immunity. Both

the use of less prevalent AAV serotyped capsids and capsid engineering efforts are continually being

pursued in order to generate AAV vectors capable of escaping pre-existing immunity while still mediating

therapeutic expression levels 57,96,158,182–185.

Second is the limited packaging capacity (4.5 kb) of rAAV vectors, which prevents larger

transgenes from being eligible for AAV-mediated gene delivery. Importantly, efforts have been made

utilizing dual AAV vectors to deliver these larger transgenes as split versions, packaged within different

capsids, which rely on co-infection of the same cell with both vectors and subsequent recombination

events, or similarly, dual vectors in which the transcriptional elements are delivered separately from the

transgene186–188. These efforts have been met with limited results, but there has been some success in

engineering smaller, minimal versions of large transgenes, such as dystrophin. For instance, a smaller

micro dystrophin has been engineered which is small enough to be packaged and delivered using rAAV

vectors187.

A third obstacle facing rAAV vectors is the requirement for high dosages in order to achieve

therapeutic expression levels of transgenes in human patients. Many gene therapy applications require

extremely high virus titers to be administered, i.e. 2.5 x 1014 vg/kg for a recent SMA clinical trial, in order

to achieve therapeutic transgene expression levels189. This need to administer such high vector loads can

in order to achieve therapeutic levels of expression can result in unfortunate consequences. Firstly is the

difficulty associated with GMP production of AAV vectors at high enough titers for clinical use in human

patients. Production of such high viral titers is extremely difficult and results in considerable scale-up

costs. Innovations in production methods (i.e. suspension 293 cell systems and the use of baculovirus/sf9

insect cell protocols) as well as purification schemes (i.e. moving from density ultracentrifugation to

methods such as tangential flow filtration and various forms of column chromatography) are continually

being made to improve manufacturing168,169. Secondly, administration of such high vector loads has the

potential to result in unfortunate side effects such as dose-related hepatotoxicity, as has been seen lately

for systemically administered AAV9-based vectors190–192.

17

This need for improved efficiency of viral vectors goes hand in hand with a need to improve the

(tissue) specificity of vectors. Thus, there is an inherent struggle between optimizing and balancing both

efficiency and specificity (i.e. transduction & tropism). For example, vectors that transduce cells in

specific target tissues (i.e. the CNS) at a high level (i.e. AAV9 and AAVrh.10) also transduce peripheral

tissues, such as the liver, at high levels112,157,193,194. When delivered systemically, these vectors suffer from

liver sequestration. Even when these vectors are delivered directly into the CNS through

intracerebroventricular injections, they suffer from systemic leakage, and still transduce the liver195.

Efforts of have been made to place transgenes under the control of a cell-type specific promoter in order

to optimize specificity, such as the human synapsin (hSyn) promoter, which expresses only in neurons;

however, this results in a dramatic drop in transduction efficiency, compared to expression using a more

ubiquitous promoter, such as the chicken beta actin (CBA) or CMV promoter196. Thus, this issue is

sometimes thought of in terms of an equation in which efficiency plus specificity equals 1, where the two

are at odds in a zero-sum game. Nonetheless, capsid engineering efforts in our lab and others, are

continually focused on improving both efficiency and specificity of rAAV vectors, with the goal of

achieving therapeutic expression levels at a lower dose.

1.8. Gene Therapy for the Central Nervous System Using AAV Vectors

CNS Gene Therapy and AAV. There has been a lot of success so far in using rAAV vectors to

achieve therapeutic CNS gene transfer for targeting multiple neurological diseases, in both preclinical and

clinical studies157. AAV possesses many attributes which make it an ideal vector for gene transfer to the

CNS and there are many natural and synthetic serotypes with varied phenotypes in regards to CNS

transduction and tropism105. Also detailed in previous sections, these attributes are largely the result of

structural variation and differential glycan receptor usage. Gene therapy has emerged as an attractive

means to treat numerous neurological disorders of the CNS that are characterized by an underlying

genetic basis. Indications include examples such as Parkinson’s disease and Alzheimer’s disease. For

example, neurodegenerative disorders such as Huntington’s disease and Parkinson’s disease,

characterized by loss of GABAergic and dopaminergic neuron populations, respectively, can be treated

18

through gene replacement using rAAV vectors to restore cellular functions and prevent

neurodegeneration157,159.

Examples of CNS Gene Therapy. Additionally, CNS gene therapy is being pursued to treat

numerous lysosomal storage disorders (LSDs) such as Krabbe disease, Gaucher disease, Pompe disease,

and various mucopolysaccharidoses disorders (MPS)157. These diseases are often caused by loss-of-

function mutations which lead to insufficient enzymatic activities, altered metabolism and the

accumulation of toxic products within the cell. This buildup leads to cell death and/or dysfunction, and

the resulting neurodegeneration ultimately causes the cognitive and/or motor defects associated with the

respective diseases197–199. Similar genetic defects resulting in altered lysosomal storage dysfunctions have

also been implicated in causing Parkinson’s disease as well200. Similarly, gene therapy products are

currently being developed to treat Parkinson’s disease through delivery of an amino acid decarboxylase

(AADC) transgene to increase dopamine levels and are currently undergoing clinical trials156.

1.9. AAV in the Brain – Phenotypes, Routes of Delivery, & Transport

AAV Transport within the CNS. Within the CNS, AAV undergoes several modes of transport,

which can also be affected by capsid structure/receptor usage, route of delivery, and the animal models

used157. Once in the CNS, AAV vectors can undergo paravascular transport through the CSF and

interstitial fluid, through the recently characterized glymphatic pathway201,202. Recently, studies from our

lab demonstrated a role for glymphatics, and aquaporin 4-mediated water transport, in the paravascular

clearance of AAV203. AAV spread in the CNS is known to occur via axonal transport subsequent to

neuronal infection, where the virus can travel to other brain regions through one or both directions204.

Anterograde transport occurs across the synapse and the virus is taken up by dendrites of the receiving

neuron, facilitating transduction of that cell or release of the virion to transduce nearby cells. Conversely,

retrograde transport occurs when viruses are taken up at an axon and transported back to the neural cell

body. The direction of axonal transport, and whether it is unidirectional or bidirectional depends on the

capsid204. For instance, AAV1 and AAV6 have both been shown to display exclusively retrograde

19

transport, while AAV2 is only transported in the anterograde direction204,205. AAV8/9, however, can be

transported bidirectionally206.

Routes of administration for CNS Gene Delivery. CNS transduction is largely influenced by the

tropism of the AAV capsid being used; however, the route of injection also plays a large role. Several

routes of administration exist for AAV vectors targeting the CNS, both directly and indirectly, and have

been performed in different animal models157,207,208. Direct CNS administration can be achieved through

intraparenchymal injections directly into the brain tissue and through administering directly to the

cerebrospinal fluid (CSF). Intra-CSF routes include intracerebroventricular (ICV), intracisternal (IC), and

intrathecal (IT) injections into the spinal cord. Intracranial routes generally use stereotaxic injections.

Intraparenchymal injections. Direct injections into the brain parenchyma are the most invasive

and generally result in punctate transduction profiles. While many serotypes can effectively transduce a

given brain region upon direct intraparenchymal injections at that site, their transduction is limited to the

site of injection209. Generally these injections have been performed targeting specific brain regions, such

as the thalamus, striatum, hippocampus, etc. Serotypes like AAV2, AAV4, and to a slightly lesser degree,

AAV1 and AAV5, show limited spread in the CNS, punctate expression localized to the region injected,

and generally show a preferentially neuronal tropism104,210. Thus intraparenchymal injections are desirable

for applications in which highly localized and limited gene transfer is desired, but less suited for

applications requiring more global gene transfer.

Intra-CSF administration. Intra-CSF administration, while being similarly invasive, provides the

ability to transduce larger areas of the brain. Nonetheless, spread of the virus away from the direct site of

injection (i.e. the specific ventricle or cisterna magna) is variable, depending on the capsid. Intra-CSF

administration of AAV vectors has been used to successfully achieve CNS gene transfer104,157,211.

Intrathecal (IT) administration via lumbar puncture has proven successful as well157,184,212. AAV1/5/6/8/9

efficiently transduce the spinal cord and brain stem following IT administration, with AAV9

outperforming the others213,214. AAV2/4 generally show the least spread, with AAV1/5/6 generally

showing slightly more spread and moderately higher transduction levels in the CNS. AAV8/9/rh.8/rh.10

20

have all been shown to mediate the highest transduction levels and spread more effectively in the CNS

than other natural serotypes, following intra-CSF administration through the various routes discussed

above105,215,216. Furthermore, AAV1/2/6 capsids mediate preferentially neuronal transduction following

intra-CSF or intraparenchymal injections157,210. AAV5 transduces both neurons and glia at low levels

while AAV8/9/rh.8/rh.10 transduce both cell types much more efficiently105,216. AAV4, following ICV

administration, preferentially transduces ependymal cells, and transduces astrocytes upon

intraparenchymal injection into the subventricular zone and rostral migratory stream (RMS)217,218.

However, our lab recently engineered an AAV4 variant, AAV4.18, which was found to have switched

receptor usage from O-SIA to α-2,8-polysialic acid (PSA)219. This switch resulted in improved spread in

the CNS in addition to selective transduction of migrating neural progenitor cells in the RMS.

Furthermore recent studies from the Song lab at UNC have demonstrated that, following direct

microinjections in the dentate gyrus, AAV4 selectively transduces quiescent neural stem cells in this

region of the hippocampus220.

1.10. CNS Gene Transfer Across the Blood-Brain Barrier Using AAV Vectors

Systemic Administration. Lastly, AAV vectors can be administered systemically via intravenous

injections (IV). This is by far the least invasive route of administration and has the added benefit of

promoting a more global distribution of gene transfer across the brain, rather than the more localized

expression observed for other routes. Nevertheless, it is also one of the most difficult routes for achieving

therapeutic transduction levels, requiring both optimal capsid selection as well as high vector

dosages157,160. This is due to either poor specificity and/or efficiency of vectors targeting the CNS,

resulting from tropism for peripheral tissues and sequestration within these tissues, such as the liver221.

Effective gene delivery to the CNS is also largely prohibited by the blood-brain barrier157.

The Blood-Brain Barrier. The blood-brain barrier (BBB) prevents large and/or hydrophilic

molecules in circulation from entering the brain parenchyma through the vasculature. Structurally, the

BBB is composed primarily of the tight junctions formed between endothelial cells in the brain

vasculature222. Additional components of this neurovascular unit include pericytes and smooth muscle

21

cells which line the endothelium, astrocytic end-feet, which provide biochemical support from the brain

tissue, and the extracellular matrix (ECM) in which these interact223. The BBB precludes entry of >98%

of drugs from entering the brain and the vast majority of all viruses, including AAV. Several pathogenic

viruses do possess the ability to cross the BBB and infect cells in the CNS through a few clever

mechanisms224. For instance, HIV can hitchhike along extravasating immune cells225. Others, like Rabies,

infect peripheral nerve cells and cross into the brain though axonal transport226. Contrarily, some

pathogenic viruses utilize mechanisms which disrupt the integrity of the BBB in order to gain entry into

the CNS, such as the induction of hemorrhages by mouse adenovirus227.

AAV & the BBB. As a very simple virus with a limited coding capacity, both wildtype AAV as

well as rAAV vectors are unable to use the more complicated mechanisms employed by pathogenic

viruses, as discussed above. Thus, most AAV serotypes are unable to cross the BBB and enter the brain

parenchyma following systemic administration. This includes most of the common serotypes (i.e. AAV1-

6 and AAV8)98. As an example, AAV1, one of the strains most relevant to this dissertation, effectively

transduces and is sequestered within the brain microvasculature following intravenous injection. In the

past, strategies have been employed to circumvent the BBB, such as administering vectors in preclinical

animal models with a seizure-compromised BBB or through the use of drugs (i.e. mannitol) to disrupt the

BBB228,229.

Notably, several of the more recently discovered serotypes, many of which were isolated from

nonhuman primates, effectively cross the intact BBB following systemic administration and transduce

cells within the brain parenchyma98. These neurotropic capsids include AAV7, AAV9, AAVrh.8,

AAVrh.10, AAVrh.39, and AAVrh.43, and have further expanded the vector toolkit. In a direct

comparison following facial vein injections in neonatal mice, AAV9, AAVrh.8, and AAVrh.10 were

found to exhibit the highest CNS transduction levels. Nonetheless, they all mediated widespread CNS

gene transfer. Of these three “top tier” capsids, AAVrh.10 and AAVrh.8 both outperformed AAV9, in

terms of CNS transduction levels. This is noteworthy, as AAV9 is currently regarded as the gold-standard

for CNS-targeted gene therapy. In a follow-up study by the same lab, AAV9, AAVrh.8 and AAVrh.10

22

underwent a direct head-to-head comparison evaluating CNS transduction following systemic

administration in adult marmosets230. Here, all three exhibited high transduction levels compared to other

serotypes, but AAVrh.8 outperformed both AAV9 and AAVrh.10.

1.11. Engineering AAV Capsids for Gene Transfer to the Brain

Significant efforts in recent years have focused on engineering the AAV capsid in order to

achieve improved transduction and tropism for tissues of interest, including the CNS. Strategies to

engineer the AAV capsid include peptide insertion, chemical conjugation and mutagenesis231–233.

Combinatorial library-based approaches have seen increased use in recent years to generate new synthetic

variants. Often, such libraries are created using either DNA shuffling to recombine cap genes of different

serotypes or through targeted mutagenesis strategies228,234. Subsequently, such libraries generally undergo

either simple screening or directed evolution in order to identify optimal capsids for the phenotype of

interest235–237. An example of a simple screen, from our lab, is the generation of liver-detargeted AAV9

capsid mutants238. Additionally, engraftment of structural features from one capsid to another can

generate variants with novel phenotypes. For example, engraftment of the galactose receptor footprint

from AAV9 onto the AAV2 capsid generated the AAV2g9 variant, which demonstrated improved

neuronal transduction, increased spread, and reduced systemic leakage following ICV injections239. Our

lab has recently developed a structure-guided evolution platform, utilizing random mutagenesis of

specific capsid regions, informed by structure-function studies, and targeted evolution to generate AAV

variants for immune invasion while retaining tropism and transduction properties182. This platform can be

extended to carry out structure-guided evolution of glycan receptor footprints as well, which is the focus

of future studies.

Capsid strategies such as these have been employed and are continually iterated upon in order to

expand the vector toolkit for improving CNS gene transfer. In particular, many efforts have been made to

generate capsid variants with improved CNS gene transfer following systemic administration, to expand

upon the aforementioned naturally neurotropic isolates. These generally involve library-based strategies

including directed evolution and screening for CNS gene transfer following systemic administration. For

23

example, directed evolution of capsid libraries has generated a couple capsid variants which are able to

cross the seizure-compromised BBB; however, these are unlikely to cross an intact BBB228. A similar

approach was used to generate the synthetic AAV-B1 capsid, which showed improved transduction of the

motor cortex and thalamus, as well as the spinal cord, relative to AAV9 following IV injection240. Similar

region-focused improvements in gene transfer were observed for the synthetic AAV-AS variant,

generated through peptide-display241. The AAV2-retro vector was engineered through a combination of

peptide display and directed evolution and generated a vector with a strong propensity for retrograde

axonal transport242. Recently, a Cre recombinase-based AAV targeted evolution (CREATE) method was

developed to generate capsids with improved transduction of specific cellular populations in the CNS

after IV injections in mice243. This method employed evolution of a peptide-insertion motif inserted into

the C-terminus of the AAV9 VP3, and resulted in generation of the AAV-PHP.B vector. AAV-PHP-b

was found to improve gene transfer to the astrocytes and neurons in the brain by 40-fold relative to

AAV9. Subsequent iterations of this method generated further improved capsids, AAV-PHP.eB and

AAV-PHP.S, which exhibit vast improvements relative to AAV9 in terms of CNS and peripheral nervous

systems, respectively, following IV injection244. While the AAV-PHP capsids have garnered a lot of

attention due to their extremely robust CNS transduction profile, recent studies have shown that these

CNS transduction properties are limited to the C57BL/6J mouse background upon which they were

selected, and that they fail to transduce BALB/cJ or nonhuman primates245.

This dissertation highlights the efforts I have made towards furthering our understanding of AAV

biology in regards to neurovascular transport and CNS tropism/transduction following systemic

administration. In Chapter 2, I set out to dissect the structure-function correlates for the transport of AAV

across the BBB. Subsequently, in Chapter 3, I expand upon this by attempting to dissect the role of

capsid-sialic acid interactions on BBB traversal and CNS transduction.

24

Figure 1.1. Structure of the AAV Viral Protein monomer. Cartoon diagram of the AAV1 VP3 (PDB

ID # 3NG9) monomer. The monomer has a β-barrel jellyroll structure with 8 antiparallel β-sheets (βB-βI,

indicated as arrows on the diagram) and a single α helix (αA). This diagram depicts the variable regions

(VR) I-IX which are located on surface-exposed loops connecting the various β-strands, and are critical

determinants for tropism antigenicity and receptor usage. The variable regions are color-coded as follows:

VR-I is red, VR-II is orange, VR-III is yellow, VR-IV is light green, VR-V is dark green, VR-VI is teal,

VR-VII is blue, VR-VIII is purple, VR-IX is magenta, and the HI loop is brown.

25

Figure 1.2. AAV capsid structure and axes of symmetry. Surface-rendered homology models

generated in PyMoL showing the full T=1 icosahedral 60mer AAV1 capsid (PDB ID # 3NG9, left)

alongside the various axes of symmetry – the two-fold axis of symmetry (i.e. VP3 trimer dimer, top),

five-fold axis of symmetry (VP3 pentamer, middle), and the three-fold axis of symmetry (i.e. VP3 trimer,

bottom).

26

CHAPTER 2: MAPPING THE STRUCTURAL DETERMINANTS REQUIRED FOR AAVRH.10

TRANSPORT ACROSS THE BLOOD-BRAIN BARRIER1

2.1. Overview

Effective gene delivery to the central nervous system (CNS) by intravenously administered

adeno-associated viral (AAV) vectors requires crossing the blood-brain barrier (BBB). In order to achieve

therapeutic CNS transgene expression, high systemic vector doses are often required which pose

challenges such as scale up costs and dose-dependent hepatotoxicity. In order to improve the specificity

and efficiency of CNS gene transfer, a better understanding of the structural features that enable AAV

transit across the BBB is needed. We generated a combinatorial domain swap library using AAV1, a

serotype that does not traverse the vasculature, and AAVrh.10, which crosses the BBB in mice. We then

screened individual variants by phylogenetic and structural analyses and subsequently conducted systemic

characterization in mice. Using this approach, we identified key clusters of residues on the AAVrh.10

capsid that enabled transport across the brain vasculature and widespread neuronal transduction in mice.

Through rational design, we mapped a minimal footprint from AAVrh.10, which, when grafted onto

AAV1, confers the aforementioned CNS phenotype, while diminishing vascular and hepatic transduction

through an unknown mechanism. Functional mapping of this capsid surface footprint provides a roadmap

for engineering synthetic AAV capsids for efficient CNS gene transfer with an improved safety profile.

1This chapter previously appeared as an article published in the journal Molecular Therapy and is reprinted

with permission with some modifications. The original citation is as follows: Albright BH, Storey CM,

Murlidharan G, Castellanos Rivera RM, Berry GE, Madigan VJ, Asokan A. Mapping the Structural

Determinants Required for AAVrh.10 Transport across the Blood-Brain Barrier. Molecular Therapy 26:

510-523, 2018.

27

2.2. Introduction

Adeno-associated viruses (AAV) are non-pathogenic parvoviruses composed of a small, 25 nm

icosahedral capsid packaging an approximately 4.7 kb single-stranded DNA genome55. A wide array of

AAV capsid sequences have been isolated from human and primate tissues, which have been categorized

into several distinct clades based upon sequence and structural diversity246,247. Across these clades,

different serotypes display broad tropism at the species, tissue and cellular levels19,246,247. These diverse

phenotypes are determined by capsid structure19. The AAV capsid is assembled from 60 viral protein

(VP) subunits. The core VP monomer (VP3) has a jellyroll, beta barrel structure comprised of 7 anti-

parallel beta strands connected by interdigitating loop regions. Portions of these highly variable loops are

surface exposed and define the topology of the AAV capsid, which in turn determines tissue tropism,

antigenicity and receptor usage across the various AAV serotypes19. The surface loop residues on the

AAV capsid are highly plastic and amenable to modification, affording control over antigenicity,

transduction profile and tissue tropism248.

The first step in the AAV lifecycle is recognition of and attachment to cell surface glycan

receptors19. These include heparan sulfate (HS) for AAV2, AAV3 and AAV6, α2,3- and α2,6- N-linked

Sialic acid (SIA) for AAV1, AAV5 and AAV6, O-linked sialic acid for AAV4, and galactose (Gal) for

AAV956. Secondary to glycan binding, cellular uptake of AAV implicates secondary co-receptors,

including various growth factor receptors as well as integrins40,42,44,249. Recently, a transmembrane protein,

KIAA0319L (AAVR) has been identified as a universal receptor for multiple AAV serotypes250. These

factors, along with tissue glycosylation patterns, contribute to the varying tissue tropisms of different

AAV serotypes. Particularly with regard to the CNS, different AAV serotypes display a spectrum of

transduction profiles and cellular tropisms, depending on the route of administration157. For example,

when directly administered into the mouse CNS through either direct intra-parenchymal or intra-

cerebrospinal fluid (CSF) injections, AAV capsids undergo axonal transport and transynaptic spread in

the anterograde and/or retrograde directions, depending on the serotype157. Furthermore, our lab has

28

recently shown that glial-associated lymphatic (glymphatic) transport of CSF influences AAV spread

within the mouse brain parenchyma and clearance from the CNS203.

To achieve CNS gene transfer, intravenously administered AAV vectors must first cross the

blood-brain barrier (BBB) in order to gain entry into the brain. Comprised of endothelial cell tight

junctions along with associated astrocytic end-feet and pericytes, the BBB blocks the diffusion and

paracellular flux of macromolecules/particles and regulates the transport of other molecules251. Most

viruses which infect the brain do so through disrupting or weakening the BBB; however, some viruses

have devised strategies to gain entry into the CNS by methods such as hitchhiking within host immune

cells (HIV), by infecting brain endothelial cells or by infecting peripheral nerves and exploiting axonal

transport (e.g., Rabies)252,253. In the case of AAV, the BBB prevents entry of most serotypes into the brain

with a few notable exceptions. For instance, intravascular administration of AAV serotypes 1-6 and 8

results in poor CNS transduction, while isolates AAV9, AAVrh.8 and AAVrh.10 have been shown,

amongst others, to efficiently traverse the BBB in different animal models254–257. In the current study, we

generated a chimeric capsid library by shuffling the Cap genes of two isolates, AAV1 and AAVrh.10,

which, despite a high degree of structural similarity, display distinct CNS phenotypes. While AAV1 does

not appear to cross the BBB efficiently and predominantly transduces the brain endothelium, AAVrh.10

transduces the brain parenchyma by crossing the murine and non-human primate BBB254,258. Using a

combination of structural analyses and bioinformatics followed by in vivo screening, we identified a

structural footprint on the AAVrh.10 capsid which, when grafted onto AAV1, imparts the ability to

traverse the BBB and preferentially transduce neurons within the brain parenchyma.

2.3. Methods

Generation of a chimeric AAV capsid panel. An AAV1/rh.10 domain swap capsid library was

generated through DNA shuffling. Briefly, the Cap genes of AAV1 and AAVrh.10 were randomly

fragmented by brief DNase digestion and reassembled using primerless PCR with Phusion High-Fidelity

DNA polymerase (NEB Cat #M0530L) in which the partial homology of the short (< 400 bp) fragments

allows for self-priming of the fragments. A secondary PCR step using specific conserved primers flanking

29

the Cap gene were then used to amplify the library of reassembled full-length Cap sequences and

simultaneously insert flanking restriction sites to facilitate subsequent cloning into a pTR plasmid

backbone used for virus production.

Phylogenetic and sequence analyses. The amino acid sequences of different AAV capsid isolates

were aligned using ClustalW and phylogenetic trees were generated using the MEGAv7.0.21 software

package259. The phylogeny was produced using the neighbor-joining algorithm and amino acid distances

were calculated using a Poisson correction260. Statistical testing was done by bootstrapping with 1,000

replicates to test the confidence of the phylogenetic analysis and to generate the bootstrap consensus

tree261. Branches corresponding to partitions reproduced in less than 50% of bootstrap replicates are

collapsed. The percentage of replicate trees in which associated taxa clustered together in the bootstrap

test is displayed next to the branches. All sequence alignments were performed using Invitrogen’s Vector

NTI Advance 11.5.2 software.

Virus production and titers. An updated triple plasmid transfection protocol was used to produce

recombinant AAV vectors195. Specifically, the transfected plasmids include (i) a capsid-specific pXR

helper plasmid (i.e. pXR1, pXRrh.10, or various plasmids encoding the various chimeric Cap genes used

in this study), (ii) the adenoviral helper plasmid pXX680, and (iii) either pTR-CBh-scGFP or pTR-CBA-

Luc plasmids (encoding either a self-complementary green fluorescent protein (GFP) reporter transgene

driven by the chicken beta actin hybrid (CBh) promoter or a luciferase reporter transgene (Luc) driven by

the chicken beta actin promoter (CBA), respectively, flanked by inverted terminal repeats (TRs) derived

from the AAV2 genome. Viral vectors were purified using iodixanol density gradient ultracentrifugation.

Vectors packaging a CBh-scGFP transgene were subsequently subjected to buffer exchange and

concentration using Sartorius Vivaspin2 100 kDa molecular weight cut-off (MWCO) centrifugation

columns (F-2731-100 Bioexpress, Kaysville, UT). Vectors packaging a CBA-Luc transgene were

subjected to buffer exchange and de-salting using Zeba Spin Desalting Columns, 40K MWCO (Thermo

Scientific, Cat # 87770). Following purification, viral genome titers were determined via quantitative

PCR using a Roche Lightcycler 480 (Roche Applied Sciences, Pleasanton, CA). Quantitative PCR

30

primers were designed to specifically recognize the AAV2 inverted terminal repeats (forward, 5’-

AACATGCTACGCAGAGAGGGAGTGG –3’; reverse, 5’-

CATGAGACAAGGAACCCCTAGTGATGGAG -3’) (IDT Technologies, Ames IA).

Animal studies. All animal experiments were performed using 6-to-8-week-old female C57/BL6

mice purchased from Jackson Laboratories (BAR Harbor, ME). These mice were maintained and treated

in compliance with NIH guidelines and as approved by the UNC Institutional Animal Care and Use

Committee (IACUC). To investigate the ability of AAV vectors to cross the BBB and transduce CNS

cell populations, AAV vectors packaging a CBh-scGFP transgene or 1x PBS (as mock treatment) were

administered intravenously (i.v.) via tail vein injection at a dose of 5 x 1011 vg. To assay for GFP reporter

transgene expression, animals were sacrificed 21 days post injection with tribromoethanol (Avertin) (0.2

ml of 1.25% solution) followed by transcardial perfusion with 30 ml of 1x PBS followed by 30 ml of 4%

paraformaldehyde in PBS. Tissues including the brain, heart and liver were removed and post fixed for 24

h, and 50-μm-thick sections were obtained for each tissue using a Leica VT 1200S vibrating blade

microtome (Leica Biosystems, IL). The mouse brain sections were then immunostained as described

below. For in vivo luciferase transduction and viral genome biodistribution experiments, mice were

injected with either 1x PBS or viral vectors packaging a CBA-Luciferase transgene at a dose of 1 x 1011

vg. Mice were sacrificed, as described above, at 14 days post injection and various tissues were removed.

For these experiments, no fixation with 4% paraformaldehyde in 1x PBS was performed and instead

tissues were dissected and frozen at -80⁰C prior to use.

To quantify luciferase expression, mice injected with 1 x 1011 viral genomes packaging a CBA-

Luc transgene were sacrificed at 14 days post injection and tissues were harvested and frozen at -80 ⁰C.

Tissues were later thawed, weighed and lysed by adding 150 μl of 2x passive lysis buffer (Promega,

Madison WI) prior to mechanical lysis using a Tissue Lyser II 352 instrument (Qiagen, Valencia, CA)

followed by centrifugation to remove any remaining tissue debris. To measure luciferase transgene

expression, 50 μl of supernatant from each lysate was then loaded onto an assay plate along with 50 μl of

luciferin and luminometric analysis was performed using a Victor2 luminometer (PerkinElmer, Waltham,

31

MA). The relative light units obtained for each sample were then normalized to the input tissue weight for

each sample, measured in grams. Data was graphed and statistical analyses carried out using an unpaired

two-tailed T-test with Welch’s correction as well as ANOVA followed by Tukey’s multiple comparisons

test where indicated. These statistical analyses were performed in GraphPad Prism 6® software.

Tissue processing and histological analysis. For mouse experiments using virus packaging a GFP

reporter transgene, free-floating 50-μm-thick coronal brain sections were stained in 24-well plates.

Sections were incubated in blocking buffer containing 10% goat serum and 1% Triton X (Sigma-Aldrich)

in 1x PBS for 1 hour at room temperature. The sections were then incubated at 4⁰C overnight with a

primary monoclonal rabbit α-GFP antibody (Life-Technologies-G10, 362 1:750) diluted in blocking

buffer. The next day, three 10 minute washes were performed with 1x PBS. The subsequent histochemical

analysis of GFP expression was performed using a Vectastain ABC kit (Rabbit IgG PK-4001 kit, Vector

biolabs, Burlingame, CA) and tissues were mounted onto microscopy slides. The immunostained sections

were digitally imaged in brightfield (20x objective) using an Aperio ScanScope XT instrument (Aperio

Technologies, Vista, CA) by the UNC Translational Pathology Laboratory and images were obtained

using Leica eSlide Manager (centralized image storage and data management software) and analyzed

using Aperio ImageScope and WebViewer software. Quantifications were calculated by counting the

number of GFP+ neuronal or glial cells, determined based upon morphology, per 50 μm coronal brain

section. Data was graphed and statistical analyses were performed as outlined above. Specific brain

regions were identified based upon comparison to a coronal mouse brain reference obtained from the

Allen Mouse Brain Atlas262. To assay for GFP expression in the heart and liver, tissues were stained for

GFP with the anti-GFP primary antibody as described above; however, an anti-rabbit goat antibody

conjugated to Alexa-488 was used as the secondary antibody at a dilution of 1:500 (anti-rabbit Abcam-

96,883). Immunostained GFP in these tissues was then imaged using an EVOS FL epifluorescence cell

imaging system (AMC/Life Technologies) using the GFP light cube (excitation 470 nm, emission 510

nm). Statistical analyses were carried out as outlined earlier.

32

Vector genome biodistribution. Animal studies were performed as described above. At 21 days

post injection, mice were sacrificed and tissues were frozen at -80⁰C. Tissues were later thawed and viral

genomes were extracted from the tissue lysates using the DNeasy kit (Qiagen, Valencia, CA). Viral

genome copy numbers were then determined for each tissue using quantitative PCR with primers specific

to the luciferase transgene (forward, 5′- AAAAGCACTCTGATTGACAAATAC -3′; and reverse, 5′-

CCTTCGCTTCAAAAAATGGAAC -3′). These viral genome copy numbers were then normalized to the

mouse lamin B2 housekeeping gene using the primers (forward, 5′-

GGACCCAAGGACTACCTCAAGGG -3′; and reverse, 5′- AGGGCACCTCCATCTCGGAAAC -3′).

The biodistribution of viral genomes are represented as the ratio of vector genomes per cell recovered for

each tissue. Data was graphed and statistical analyses carried out as described earlier.

Molecular modeling. Previously published coordinates (PDB ID, 3NG9) were used to generate

three-dimensional structures of the AAV1 VP3 trimer/three-fold axis of symmetry.263 Homology models

of AAVrh.10 and various AAV1/rh.10 chimeric capsid structures were obtained using the SWISS-Model

server (http://swissmodel.expasy.org/)264, with the crystal structure of AAV8 VP3 (PDB ID, 2QA0) used

as a template and a structure-based alignment was generated using the secondary structure matching

(SSM) application in the WinCoot software59,205,265, with the AAV1 VP3 (PDB ID 3NG9) monomer being

used as a template. The VP3 trimers/3-fold symmetry axes, VP3 trimer dimers/2-fold symmetry axes,

VP3 pentamers/5-fold symmetry axes, and full capsids were generated using the VIPERdb oligomer

generator utility (http://viperdb.scripps.edu/oligomer_multi.php)266. Surface rendered depictions of these

models were visualized using PyMOL (the PyMOL Molecular Graphics System, Schrödinger LLC,

http://www.pymol.org/). Stereographic roadmap projections of the AAV1RX capsid surface highlighting

surface-exposed amino acid residues within the AAVrh.10-derived neurotropic footprint were generated

using RIVEM (Radial Interpretation of Viral Electron Density Maps) software267.

2.4. Results

Generation of an AAV1/rh.10 domain swap library and isolation of chimeric capsid variants. We

hypothesized that the comparative analysis of different capsid domains will allow us to ascertain

33

structure-function correlates for traversing the BBB. Correspondingly, we generated an AAV1/rh.10

domain swap library through DNA shuffling. We selected AAV1 and AAVrh.10 as parental capsid

sequences for DNA shuffling since they differ markedly in their abilities to cross the BBB and because of

the sequence homology (85%) shared by their capsid (Cap) genes. Thirty-six chimeric capsid sequences

were then clonally isolated and sequenced. The variants generated from this library displayed substantial

diversity at the DNA and amino acid level. Sequence alignment revealed a spectrum of domain swaps,

which we then organized in order of increasing homology from AAV1 to AAVrh.10 (Fig. 2.1c), top to

bottom). This panel of clones was further characterized phylogenetically by constructing a neighbor-

joining tree, which broadly categorized these variants as either more AAV1-like (Clade A) or more

AAVrh.10-like (Clade E) (Fig. 2.1b). Small-scale vector production was then used to establish relative

titers to exclude capsids defective in assembly or packaging from the study. Homology-based structural

models of the parental and representative chimeric capsid trimers highlighting key surface

domains/residues at the three-fold axis of symmetry (Fig. 2.1c) were generated to further narrow the list

of chimeric capsid variants for initial screening in vivo. Ten chimeric capsid variants were selected based

on structural analyses, out of which six yielded recombinant vectors (packaging scGFP or ssLuc

transgene cassettes) at titers similar to parental AAV1 and AAVrh.10 vectors. These variants were then

screened further.

In vivo screening identifies two AAV1/rh.10 chimeric capsids capable of crossing the BBB after

intravenous administration in adult mice. We hypothesized that, based on their structural diversity, the

selected panel of capsid variants would differ in their ability to cross the BBB and transduce the CNS

after I.V. administration. It is important to note that our approach does not involve directed evolution, as

this strategy is generally applicable for selecting optimal capsids and less suited for studying structure-

function relationships. We injected 6-8-week-old mice with a dose of 5 x 1011 viral genomes (vg) per

mouse of AAV1, AAVrh.10, or one of six different chimeric AAV vectors packaging a self-

complementary CBh-GFP reporter cassette via tail vein injection. Immunostaining of coronal brain

sections at three weeks post-injection revealed that AAV1 transduction in the cerebral cortex was limited

34

to the vasculature while AAVrh.10 demonstrated robust transduction of various cell populations,

including neurons, glia and endothelial cells, as determined morphologically (Fig. 2.2). Further

morphological assessment of GFP+ cells indicates differing phenotypes for the chimeric variants.

AAV1R19.1 and AAV1R20 transduce predominantly microvascular endothelial cells in the cortex, while

low to modest transduction of neuronal and glial cells is evident for AAV1R8 and AAV1R19d vectors

(Fig. 2.2). In contrast, AAV1R6 and AAV1R7 demonstrate robust transduction of cortical neurons with

modest transduction of glia and scarce, if any, transduction of the vasculature within the cortex.

Representative images of the somatosensory area of the cortex at high magnification are shown. A similar

trend for these chimeras was observed consistently across other regions of the brain (Fig. 2.3). These

observations indicate that the chimeric capsids AAV1R6 and AAV1R7 likely possess the ability to cross

the BBB, similar to the parental AAVrh.10 vector, although the mechanism is unknown. However, unlike

either parent, AAV1 or AAVrh.10, neither chimera efficiently transduces the vasculature.

Intravenously delivered AAV1R6 and AAV1R7 chimeric capsids cross the BBB and transduce

neurons preferentially throughout the brain. Having identified the AAV1R6 and AAV1R7 chimeras as

variants able to cross the BBB and transduce the CNS following systemic injection, we next sought to

further characterize the transduction profiles of these variants across multiple functionally relevant

structures of the brain. For each of the brain regions discussed below, representative images at higher

magnification were obtained from mouse brain coronal sections immunostained for GFP and scanned in

bright field at 20x magnification. Additionally, quantitative data for neuronal and glial transduction for

each region were determined by counting GFP+ neuronal and glial cells, respectively, based on cellular

morphology.

Cerebral Cortex. Previous studies in mice have established that AAVrh.10 displays robust

transduction of neuronal, glial and vascular endothelial cells throughout the cortex254. In the case of

intravascularly delivered AAV1, vascular transduction accompanied by the absence of neuronal

expression and low expression levels in glia has been reported257. Our analyses are consistent with these

prior reports (Figs. 2.4 and 2.5). We also observed that intravenously administered AAV1R6 and

35

AAV1R7 (AAV1R6/7) vectors mediate robust GFP expression in cortical neurons and reduced

expression in glia throughout the amygdalar, piriform, entorhinal, temporal association, auditory,

somatosensory, motor, posterior parietal and retrosplenial areas of the cortex at varied yet consistently

trending levels across the sections analyzed. Representative images of cortical regions transduced are

shown for the motor cortex (Fig. 2.4a) and the somatosensory area of the cortex (Fig. 2.4b). Furthermore,

AAV1R6/7 seem to display a preferentially neuronal cellular tropism across the entire cortex, showing

moderate glial transduction and remarkably reduced vascular transduction. This profile is in stark contrast

to the vasculotropic AAV1 as well as AAVrh.10, which transduces neuronal, glial, and vascular

endothelial cells with high efficiency (Fig. 2.4a and 2.4b). While AAV1R6/7 mediate robust neuronal and

moderate glial expression throughout the cortex, their expression levels are nonetheless lower than those

achieved by AAVrh.10. These trends are corroborated by quantitation of our morphological analysis of

GFP+ cells (Figs. 2.5a and 2.5b.), which show a statistically significant difference in cortical transduction

relative to AAV1 for AAVrh.10, AAV1R6 and AAV1R7. One last difference to note is that AAVrh.10

transduction across the cortex has a tendency to be concentrated at the peripheries of the tissue, around

cortical layers 1-3, particularly within the retrosplenial, motor, somatosensory and visual cortical areas.

While this can be attributed to differential immunostaining, we observed this phenomenon with high

consistency across the stained AAVrh.10 sections in our analysis. This trend does not hold true for

AAV1R6/7, which appear to mediate more uniformly distributed expression across the cortical layers.

Hippocampus. Following intravenous delivery, the chimeric AAV1R6/7 vectors appear to

mediate robust GFP expression in neurons throughout the hippocampus, bilaterally across brain

hemispheres. GFP+ hippocampal neurons are observed within the CA1, CA2 and CA3 pyramidal layers

(Fig. 2.4c, CA2 and partial CA1 shown). AAV1R6/7 also appear to be effective at transducing neurons

within the dentate gyrus, showing a large number of GFP+ neurons which appear to be granular cells

based on morphology (Fig. 2.4d). The neuronal transduction profile displayed by AAV1R6/7 in the

hippocampus is similar to that seen for AAVrh.10. In contrast to AAVrh.10, however, AAV1R6/7 do not

appear to transduce either glia or endothelial cells in either region at any appreciable level, with no

36

significant difference in GFP+ glial cells observed between AAV1 and AAV1R6/7 (Fig. 2.5c and 2.5d).

AAV1 transduction observed in the hippocampus is again limited to the vasculature. Furthermore, these

qualitative trends are corroborated by quantitative data for neuronal and glial cell transduction in the

hippocampus (Fig. 2.5c) and dentate gyrus (Fig. 2.5d).

Thalamus. GFP+ neurons were detected in the thalamus for AAV1R6/7 at levels comparable to

AAVrh.10 (Fig. 3e). Furthermore, AAV1R6/7 demonstrate minimal transduction of the endothelium and

dramatically reduced levels of glial transduction in the thalamus compared to AAVrh.10. These trends are

further corroborated by quantitative data (Fig. 2.5e) in which transduction of thalamic neurons was found

to be significantly different for AAVrh.10, AAV1R6 and 1R7, relative to AAV1. Contrarily, no

significant difference was found for thalamic glial transduction for AAV1R6/7 relative to AAV1.

Hypothalamus. Although somewhat variable, we observed consistently low levels of GFP

expression, regardless of cell type, within the hypothalamus for parental and chimeric vectors when

administered systemically (Fig. 2.4f). Variation in neuronal transduction levels in the hypothalamus was

observed for AAVrh.10, revealing high as well as low numbers of GFP+ neurons across mice. This

transduction profile is illustrated by the deviation shown for AAVrh.10 transduction in our quantitative

data (Fig. 2.5f). AAV1 displays moderate hypothalamic transduction that is restricted to the vasculature.

AAV1R6/7 vectors demonstrate low numbers of GFP+ neurons and glial cells in the hypothalamus, and a

small number of GFP+ endothelial cells (Fig. 2.4f and Fig. 2.5f).

Striatum. AAV1R6/7 transduce neurons in the striatum (specifically, the caudate putamen) as

efficiently as AAVrh.10 (Fig. 2.4g), as corroborated further by our quantitative data demonstrating a

significant difference for each relative to AAV1; however, these vectors transduce ~2-fold fewer glial

cells and lower numbers of endothelial cells in the striatum (Fig. 2.5g) despite some observed variation.

Amygdala. The amygdalar transduction profiles for AAV1R6/7 show robust neuronal GFP

expression, significantly different compared to AAV1. Few GFP+ glial cells were detected, which were

not significantly different compared to AAV1 (Fig. 2.4h and Fig. 2.5h), and barely detectable GFP+

endothelial cells were observed, similarly to other brain regions such as the striatum and thalamus.

37

Taken together, the morphological assessment of immunostained brain regions derived from mice

following intravenous delivery of AAV1R6 and AAV1R7 demonstrates robust and selective neuronal

transduction comparable to parental AAVrh.10. Furthermore, these chimeras display reduced glial

transduction and their ability to transduce endothelial cells of the brain microvasculature appears

diminished. Furthermore, these results appear to be consistent across various brain regions, with the

exception of the hypothalamus, where low transduction levels in general are observed.

AAV1R6 and AAV1R7 are de-targeted from the liver while retaining parental cardiac

transduction profiles. Since, AAVrh.10 transduces multiple tissues, including the heart and liver,

following systemic administration254, we analyzed the relative cardiac and liver transduction of these

variants compared to parental serotypes by immunostaining cardiac and liver sections (Fig. 2.6a and

2.6b). GFP expression was then quantified by averaging relative fluorescence for multiple images per

mouse using ImageJ software (Fig. 2.6). Both chimeric vectors demonstrate comparable levels of GFP

expression in the heart relative to the parental serotypes, AAV1 and AAVrh.10, with no significant

difference found for AAVrh.10, AAV1R6 or AAV1R7, relative to AAV1; however, AAVrh.10

expression levels in cardiac tissue showed substantial variation across mice (Fig. 2.6a and 2.6b). In the

liver, AAV1 demonstrated moderate transduction levels while AAVrh.10 performed exceptionally well,

demonstrating a log-fold higher transduction. In contrast, AAV1R6 and AAV1R7 mediated negligible

GFP expression in the liver at background levels comparable to mock-treated mice (Fig. 2.6b) that were

significantly reduced relative to AAVrh.10. Thus, we conclude that AAV1R6 and AAV1R7 are liver-de-

targeted relative to their parental serotypes.

Structural analysis of the AAV1R6 chimeric capsid identifies three potential domains from

AAVrh.10 that might enable crossing the BBB. Sequence analyses revealed that AAV1R6 is 97-98%

identical to AAV1 with 18 unique amino acid residues derived from AAVrh.10. AAV1R7 is also largely

identical to AAV1, but with a total of 22 AAVrh.10-derived residues, including the 18 present in

AAV1R6. Of the 4 additional residues unique to AAV1R7, two are located within the VP1 unique

(VP1u) N-terminal region (189I, 206A) and the other two are located at the buried VP3 N-terminal region

38

(224S and 225S). Since these residues are not exposed on the capsid surface and since the in vivo data

suggests that the transduction profiles displayed by AAV1R6 and AAV1R7 are equivalent (Figs. 2.4 and

2.5), we excluded AAV1R7 from the remainder of our analyses and focused on AAV1R6 alone.

AAV1R6 possesses 3 stretches of non-consecutive residues derived from the AAVrh.10 parental

strain. The first group of residues (group i) includes three amino acids (148P, 152R and 153S) within the

VP1u region (Fig. 2.7a). Specifically, one residue is located adjacent to basic region 1 (BR1), which

contains the first nuclear localization signal (NLS) of VP148,54. Additionally, group i contains two residues

(158T and 163K) also near the NLS located within the VP2 N-terminal region (Fig. 2.7a) As mentioned

earlier, these residues are not surface-exposed and instead remain internalized within the capsid until later

in the intracellular trafficking pathway54.

The second group of AAVrh.10-derived residues on AAV1R6 (group ii) consists of 8 amino

acids comprising the BC loop, located within variable region I (VR-I) (Fig. 2.7a and 2.7b). This region

has previously been implicated as being important in determining the tropism and antigenicity of

AAV1147,268. These residues are exposed on the surface of the capsid at the base of the protrusions at the

3-fold axis of symmetry (Fig. 2.7d and 2.7f) and are also localized to the depression at the 2-fold

symmetry axis (Fig. 2.7c and 2.7f). It is also important to note that this group of amino acid residues on

AAVrh.10 is substituted with AAV1 residues in the chimeric AAV1R8/19/20 capsids that are unable to

penetrate the CNS after systemic administration.

The third group of AAVrh.10-derived residues present in AAV1R6 (group iii) includes a total of

6 amino acids. Four of these residues (328Q, 330E, 332T and 333K) are located at VR-II, within the DE

loop, which contributes to formation of the pore at the 5-fold axis of symmetry (Fig. 2.7a-b and 2.7e-f).

This region on the capsid is important for viral genome packaging and release76. The last two remaining

residues within this group (343I and 347T) are located within β-strand E, positioned within the interior of

the capsid (Fig. 2.7a and 2.7b). Although these residues differ between AAV1 and AAVrh.10, it should

be noted that they are relatively conserved across different AAV serotypes.

39

Rational design of AAV1RX, a chimeric capsid with a minimal footprint from AAVrh.10 for

crossing the BBB. Using a rational approach, we narrowed down the minimum number of AAVrh.10-

derived amino acid residues in the 1R6 footprint that enable crossing the BBB and impart CNS tropism.

We first excluded any amino acids that were not exposed on the capsid surface, eliminating stretches of

residues located within the VP1/2 N-terminal regions of the capsid sequence (148P, 152R, 153S, 158T

and 163K). Using the same rationale, we also excluded residues within the conserved β-strand E (343I

and 347T) as well as those residing within VR-II, which is the pore-forming loop connecting β-strands D

and E (328Q, 330E, 332T and 333K). Through this approach, the footprint containing 8 amino acids

(263N, 264G, 265T, 266S, 268G (an insertion relative to the AAV1 sequence), 269S, 270T and 274T)

found within VR-I, the loop bridging β-strands B and C, on AAV1R6 was chosen for further evaluation

(Fig. 2.8a and 2.8b). These surface-exposed residues are located near the depression at the 2-fold axis of

symmetry (Fig. 2.8c and 2.8f) and at the base of the three-fold protrusions (Fig. 2.8d and 2.8f). The

stereographic roadmap projection of surface-exposed residues at the 3-fold axis of symmetry highlights

the topological orientation of these 8 residues in relation to the surrounding amino acids at the capsid

surface (Fig. 2.8g). As mentioned earlier, we also took into account that these amino acids were absent in

the AAV1R8/19/20 variants that were unable to transduce the brain parenchyma following systemic

administration.

We then engineered a chimeric capsid by grafting the 8 amino acid residues from AAVrh.10 onto

the AAV1 serotype, naming this chimera ‘AAV1RX’. We administered 6-8 week-old female C57/Bl6

mice with AAV1RX-CBh-scGFP vectors at a dose of 5 x 1011 vg per mouse via tail vein injection. GFP

reporter expression in the brain was assessed by immunostaining at 21 days post injection. As seen in Fig.

2.8h, AAV1RX transduces neurons within the motor cortex and cortical neurons across the entire cortex

while demonstrating limited transduction of glia and reduced vascular transduction, as seen earlier with

the chimeric capsids AAV1R6 and AAV1R7. Continuing this trend, numerous GFP+ pyramidal neurons

are observed in the hippocampus along with an abundance of GFP+ granular cells in the dentate gyrus.

Notably, GFP+ glial and endothelial cells in these regions are largely absent. In the thalamus, AAV1RX

40

demonstrates robust neuronal transduction at levels comparable to AAV1R6 and AAV1R7 with modest

transduction of glial and endothelial cells, albeit at levels higher than for other brain regions. In the

hypothalamus, GFP+ neurons are sparse and few GFP+ glia and vascular endothelial cells are observed.

Lastly, the preferentially neuronal transduction profile displayed by AAV1RX is also seen within the

striatum (specifically, the caudate putamen) and amygdala (Fig. 2.8h). Quantitative analyses of these

brain regions and comparisons to those seen for AAV1, AAVrh.10 and the chimeric AAV1R6 and

AAV1R7 vectors suggest that this minimal footprint of 8 amino acid residues from AAVrh.10 is critical

for crossing the BBB (Fig. 2.9).

AAV1R6, AAV1R7 and AAV1RX mediate low transduction levels and biodistribution in

peripheral tissues. To carry out a comparative analysis of different chimeric capsids with parental

serotypes, 6-8 week old female C57/Bl6 mice were injected via tail vein with either AAV1, AAVrh.10,

AAV1R6, AAV1R7 or AAV1RX vectors packaging a single-stranded luciferase reporter transgene driven

by a chicken beta actin promoter (ssCBA-Luc) at a dose of 1 x 1011 vg per animal. At 2 weeks post

injection, mice were sacrificed and tissues were harvested. Luciferase activity assays on tissue lysates as

well as qPCR analyses to determine vector biodistribution were carried out (Fig. 2.10). All three chimeric

vectors mediate higher luciferase transgene expression and a corresponding increase in viral genome

copies, for each, is observed within the brain compared to AAV1 (Fig. 2.10a and 2.10b). However, their

transgene expression levels and viral genome copy numbers were ~2-to-3-fold lower in the brain

compared to AAVrh.10, which were all found to be statistically significant, with the exception of

AAV1RX biodistribution (Fig. 2.10a and 2.10b). Similar transduction levels and viral genome copy

numbers were observed for AAV1R6, AAV1R7 and AAV1RX in the heart, comparable to those seen for

AAV1, despite some variation in cardiac transduction levels observed for AAV1RX (Fig. 2.10c and

2.10d). AAVrh.10 mediates ~2-4-fold higher cardiac luciferase expression and correspondingly ~2-fold

higher viral genome copies in the heart compared to the other vectors (Fig. 2.10c and 2.10d). As

demonstrated by other groups, AAVrh.10 displayed several fold higher luciferase transgene expression

levels and consistently high viral genome copies in the liver257. In contrast, low to background levels of

41

luciferase expression and significantly reduced viral genome copies were detected in the liver for AAV1,

1R6, 1R7 and 1RX (Fig. 2.10e and 2.10f). It is noteworthy to mention that luciferase expression and a

corresponding trend in viral genome copy numbers were detected for AAVrh.10 in the kidney (Fig. 2.10g

and 2.10h). Although the three chimeric vectors showed similar viral genome copy number levels,

luciferase expression in the kidney was absent. Taken together, these data seem to suggest that the

chimeric AAV1R6, 1R7 and 1RX vectors are de-targeted from the liver and may be cleared from the

blood circulation by the kidney.

2.5. Discussion

AAV1 and AAVrh.10, the parental serotypes used to generate the chimeric capsid library in this

study, are both non-human primate isolates that belong to two different clades - AAV1 belongs to clade A

while AAVrh.10 falls within Clade E246. Despite the evolutionary distance between them, they share

~85% sequence homology between their Cap genes. AAV1 is known to recognize N-linked sialic acid

(SIA) and the key residues composing the SIA recognition footprint on the AAV1 capsid have been

recently reported117,118. The current study identified 262N, 263G, 264T, 265S, 267G, 268S, 269T and

273T as key residues on the chimeric AAV1RX capsid, derived from AAVrh.10, that are essential for

crossing the BBB. While none of the SIA contact residues for AAV1 are located within the chimeric 1RX

footprint, the backbone carbonyl oxygen atoms of residues S268, D271, and N272 have been implicated

in possibly stabilizing capsid-glycan interactions and increasing SIA binding affinity117. Furthermore,

residues at this position (D271 and N272) are located within the galactose-binding footprint on AAV9269.

Thus, although further studies will be required to identify and understand the exact nature of capsid-

glycan interactions for AAVrh.10 and the chimeric AAV1RX capsid, the current study supports the

notion that residues within VR-I are critical.

Alignment of the VP3 sequences of natural AAV isolates reveals that the 8 amino acid residues

that confer the CNS phenotype to AAV1RX are conserved in AAVrh.8 and AAVrh.39, other neurotropic

capsids in Clade E (Fig. 2.11). Similarly, AAV8 and AAVrh.43, also located within Clade E, demonstrate

the ability to cross the BBB, albeit far less efficiently257. Both possess 7 out of the 8 residues in this

42

footprint; therefore, it is tempting to speculate that introducing the complete footprint onto these capsids

will enhance their ability to cross the BBB. Although it is beyond the scope of the current study, we

postulate that this footprint can potentially be grafted onto other serotypes such as AAV2, AAV6 and

AAV7, amongst others, to confer the transvascular and neurotropic phenotype. Nonetheless, it should be

noted that this minimal footprint may be less compatible with more divergent serotypes such as AAV4,

AAV5 and AAV12.

Phenotypically, AAV1 and AAVrh.10 differ markedly in their ability to cross the BBB and

transduce neurons and glia in the brain. Several groups have now established that AAVrh.10 is also

superior to AAV1 in terms of vector spread and transduction in peripheral tissues, when injected

intravenously or within the CNS after direct intracranial injection257,258,270,271. Specifically, following

direct tissue injections, such as intraparenchymal administration in the brain, AAV1 transduction is

generally limited to the injection site272. When delivered intravenously, AAV1 demonstrates preferential

uptake and transduction in vascular endothelial cells257,273. In contrast, AAVrh.10 appears not only to be

taken up by and transduce vascular endothelial cells, but also to traverse the vasculature and transduce

underlying tissue such as the CNS or skeletal muscle254,257. The current study identifies a minimum set of

residues on the AAVrh.10 capsid that partially confers these traits to AAV1 capsids. While this minimal

footprint on the AAV1RX chimera appears essential for traversing the BBB and transducing neurons,

transduction in peripheral tissues such as the liver is markedly reduced, while cardiac expression remains

similar to that of AAV1. These differences suggest that other structural domains on the AAVrh.10 capsid

may play a role in imparting a systemic transduction profile. With regard to the brain parenchyma, it is

important to note that after CNS entry, the post-entry transduction profile of AAV1RX is preferentially

neuronal with decreased glial transduction compared to AAVrh.10. This can likely be attributed to the

fact that the majority of the capsid is derived from AAV1, which is known to mediate a preferentially

neuronal transduction following intracerebroventricular injections. Although speculative, it is also

possible that the unknown mechanism by which 1RX crosses the BBB mediates preferential neuronal

uptake or that additional motifs/amino acid residues on the AAVrh.10 capsid may mediate post-entry

43

steps in glial transduction. Ongoing work includes an incremental dose response study to evaluate

vascular transduction mediated by AAV1RX and its apparent liver-detargeting, relative to AAVrh.10, and

its relative efficiency across multiple doses.

The precise mechanism of virus-host interactions utilized by the AAV1RX footprint to gain entry

into the brain is currently unknown and is the focus of ongoing studies. As outlined earlier, it is plausible

that capsid-glycan interactions on the endothelial cell surface determine whether viral capsids undergo

transcytosis across the BBB. Another possible scenario is that AAV capsids can bind soluble, heavily

glycosylated host factors in serum, which could confer the ability to traverse the BBB through interaction

with cognate receptors. Our lab has shown that reduced glycan binding avidity is generally associated

with prolonged blood circulation and the ability to traverse the vasculature. For instance, AAV2i8 and

AAV9.45 are examples of AAV capsid variants that are liver de-targeted and capable of traversing the

vasculature to transduce skeletal muscle274,275. It is therefore likely that the chimeric AAV1RX capsid has

an altered affinity for SIA and recognizes other glycans that could confer the novel phenotype observed in

this study. At this writing, it is unclear whether AAVrh.10 or AAV1R6/7/X, the novel capsid variants

described in this study, cross the BBB by transiently disrupting the vasculature or by exploiting

transcytosis mechanisms. It is also noteworthy to mention our observation that the liver-detargeted

AAV1R variants display increased uptake in the kidney suggests a possible systemic AAV capsid

clearance mechanism that warrants further investigation. Substantial biochemical and biological analyses

will also be required to dissect the mechanism(s) underlying transvascular AAV transport.

From a clinical perspective, capsids that can traverse the BBB, preferentially transduce neurons

within the brain and that are simultaneously de-targeted from the liver could demonstrate an improved

safety profile for CNS-targeted gene transfer applications from a systemic route of administration.

Decreased vector uptake in peripheral organs, particularly the liver, could reduce the potential threat of

hepatotoxicity, as evidenced by the transient elevation in transaminases in patient serum following

systemic injection of certain AAV serotypes157,276,277. Additionally, our study indicates that systemic AAV

vector clearance could potentially occur through the kidney, which might warrant further investigation

44

when evaluating the pharmacokinetics of AAV vectors in different animal models. The AAV1R6/7/X

vectors could be broadly applicable for global CNS gene transfer and in particular, for developing gene

therapy strategies to treat neurological disorders such as Friedreich’s Ataxia or spinal muscular atrophy.

45

Figure 2.1. Phylogenetic and structural analyses of the AAV1/rh.10 domain swap capsid library.

(A) Schematic of the panel of representative AAV1/rh.10 chimeric capsids isolated from the library

demonstrating sequence diversity, at the amino acid level, of domain swaps obtained through shuffling.

Individual parental or chimeric capsids are listed vertically, while the VP1 capsid sequence with different

domain swaps is displayed horizontally, from the N-terminus on the left to the C-terminus on the right.

AAV1-derived residues are shown in gray, AAVrh.10 residues in green-cyan and consensus residues

between the two in black. (B) Neighbor-joining phylogeny of the VP1 capsid sequences of AAV1/rh.10

chimeric capsids. Capsid amino acid sequences were aligned with ClustalW and the phylogeny was

generated using a neighbor-joining algorithm and a Poisson correction was used to calculate amino acid

distances, represented as units of the number of amino acid substitutions per site. The tree is drawn to

scale, with branch lengths in the same units as those of the evolutionary distances used to infer the tree.

Bootstrap values were calculated with 1000 replicates and the percentage of replicate trees in which the

associated taxa clustered together are shown next to the branches. (C) Three-dimensional surface models

of capsid subunit trimer/3-fold symmetry axis regions of parental and representative chimeric AAV

capsid mutants selected for in vivo screening. Amino acid residues derived from AAV1 are shown in gray

while residues derived from AAVrh.10 are shown in green-cyan. Structural models were visualized and

generated using PyMol.

46

Figure 2.2. In vivo screen yields AAV1/rh.10 chimeras capable of crossing the blood-brain barrier

following intravenous administration. Scans of immunostained brain sections showing GFP expression

in the cerebral cortex mediated by parental or chimeric vectors packaging a CBh-scGFP transgene at three

weeks post administration via tail vein injection at a dose of 5 x 1011 vg (or PBS in the case of mock

treatment). The global coronal brain section (top) indicates the boxed region of the cortex seen in the

insets shown for each parental or chimeric vector, representing their individual transduction profiles.

Scale bar = 100μm.

47

Figure 2.3. In vivo screen of AAV1/rh.10 chimeras for their ability to cross the blood-brain barrier

and transduce various brain regions following intravascular administration. Mice received a dose of

5 x 1011 vg of either parental (AAV1 or AAVrh.10) or chimeric vector packaging a CBh-scGFP

transgene via tail vein injection. Immunostained and imaged mouse brain sections show GFP expression,

as black/dark brown staining, at 21 days post injection Scale bar = 100 μm. The global coronal brain

section at the top of the figure indicates the various brain regions shown in the insets: (A) MCT, motor

cortex; (B) HC, hippocampus; (C) DG, dentate gyrus; (D) TH, thalamus; (E) HY, hypothalamus; (F)

STR, striatum; (G) Amg, amygdala (n = 3 for each).

48

Figure 2.4. CNS transduction profile of AAV1R6 and AAV1R7 compared to parental AAV1 and

AAVrh.10 in the brain. Transduction profiles at three weeks post tail vein injection of AAV vectors

packaging a CBh-scGFP transgene at a dose of 5 x 1011 vg for parental serotypes, AAV1 and AAVrh.10

(left columns), and for AAV1R6 and AAV1R7 (right columns) across various brain regions are shown,

including the (A) motor cortex, (B) cerebral (somatosensory) cortex, (C) hippocampus, (D) dentate gyrus,

(E) thalamus, (F) hypothalamus, (G) striatum, and (H) amygdala. Scale bar = 100μm.

49

Figure 2.5. Quantitative comparison of neuronal and glial transduction levels for parental and

chimeric capsid variants. Relative neuronal transduction levels at 3 weeks post administration of vectors

packaging a CBh-scGFP transgene via tail vein injection at a dose of 5 x 1011 vg were quantified by

counting the number of transduced neurons and glia, identified based on morphology, for each brain

region, per 50 μm coronal section. Transduction of parental serotypes, AAV1 and AAVrh.10, are shown

alongside the chimeric variants, AAV1R6 and AAV1R7, across various regions of the brain, including

the (A) motor cortex, (B) cerebral (somatosensory) cortex, (C) hippocampus, (D) dentate gyrus, (E)

thalamus, (F) hypothalamus, (G) striatum, and (H) amygdala. Error bars represent standard deviation (n >

or = 3). An unpaired two-tailed T-test with Welch’s correction and one-way ANOVA were used to

demonstrate statistical significance for neuronal and glial transduction by each group relative to AAV1,

and that differences between the means were statistically significant, respectively. ns, not significant; *, P

< 0.05.

50

Figure 2.6. Relative cardiac and liver transduction by AAV1R6 and AAV1R7 compared to parental

capsids. Either parental (AAV1 or AAVrh.10) or chimeric (AAV1R6 or AAV1R7) vectors packaging a

CBh-scGFP transgene were administered at a dose of 5 x 1011 vg via tail vein injection. At 21 days post

injection, heart and liver sections were immunostained for GFP and imaged. (A) GFP fluorescence for

cardiac (top panel) and liver (bottom panel) tissues in mice treated with either PBS (Mock), AAV1,

AAVrh.10, AAV1R6 or AAV1R7. (B) Transduction levels for cardiac (top) and liver (bottom) measured

by quantifying relative fluorescence across multiple images taken for each treatment. The bars represent

the range from lowest to highest values with the center line representing the average across samples.

Relative fluorescence was normalized to mock treated tissues. Error bars represent standard deviation (n =

3). One-way ANOVA and unpaired two-tailed T-test with Welch’s correction for each group were

carried out and significance relative to AAVrh.10 is shown. ns, not significant. *, P < 0.05.

51

Figure 2.7. Structural analysis of the AAV1R6 chimeric capsid variant. (A) The sequence

alignment of AAV1R6 with AAV1 & AAVrh.10 highlights the amino acid residues (AAV1R6

numbering) that are uniquely derived from AAVrh.10 (shown in black lettering and highlighted in blue).

Conserved residues are lettered red and highlighted in yellow, while the non-conserved equivalent

residues on AAV1 are shown in black or green text and are highlighted in either green or white. The

Three-dimensional structural model of the chimeric AAV1R6 VP3 subunit (B) monomer was generated

using SWISS-MODEL with the crystal structure of AAV8 serving as the template for homology

modeling (PDB ID: 2QAO). Surface models of the AAV1R6 VP3 subunit (C) trimer dimer/2-fold axis of

symmetry, (D) trimer/three-fold axis of symmetry, (E) pentamer/5-fold axis of symmetry and (F) Surface

rendering of an intact AAV1R6 capsid (60mer) were generated via the VIPERdb oligomer generator and

were visualized using PyMol. Residues derived from AAV1 are colored in gray while residues derived

from AAVrh.10, clustered at the base of the protrusions at the 3-fold axis of symmetry, depression at the

2-fold axis of symmetry and at the 5-fold pore, are colored in green cyan.

52

Figure 2.8. Rational design and functional mapping of a minimal AAVrh.10 footprint for crossing

the BBB. (A) The sequence alignment of AAV1RX with AAV1 and AAVrh.10 shows the numbered

amino acid residues within and adjacent to the neurotropic footprint. Conserved residues are highlighted

in yellow, the residues composing the footprint are highlighted in cyan and non-conserved residues are

highlighted in either green or white. Structural models of the engineered AAV1RX chimeric capsid were

generated using SWISS-MODEL software for homology-based modelling and VIPERdb was used to

generate oligomers. PyMol was used to generate surface rendered models of the AAV1RX VP3 subunit

(B) monomer, (C) trimer dimer / 2-fold axis of symmetry, (D) trimer / three-fold axis of symmetry, (E)

pentamer / 5-fold axis of symmetry and (F) the full capsid (60-mer). AAV1-derived amino acids and

those homologous between AAV1 and AAVrh.10 are represented in gray while the residues comprising

the neurotropic footprint from AAVrh.10 for crossing the BBB are depicted in green cyan. (G) The

stereographic roadmap projection shows the neurotropic footprint as viewed down the threefold symmetry

axis on the AAV1RX capsid and was generated using RIVEM. Only surface-exposed amino acid residues

are shown, with the boundaries between each delineated with black lines. The green regions depict the

topological protrusions at the three-fold axis of symmetry while the blue regions represent topological

depressions. The key amino acid residues of this footprint are colored magenta (AAVrh.10 VP1

numbering). (H) CNS transduction profile of AAV1RX packaging a CBh-scGFP transgene administered

via tail vein injection at a dose of 5 x 1011 vg. Sections taken at 21 days post injection were

immunostained and imaged. Scale bar = 100μm.

53

Figure 2.9. Quantification of neuronal and glial transduction levels mediated by AAV1RX

compared to parental and other chimeric capsid variants. Mice received a dose of 5 x 1011 vg of

either parental (AAV1 or AAVrh.10) or chimeric vector packaging a CBh-scGFP transgene via tail vein

injection. At 3 weeks post injection, brains were taken, sectioned, immunostained for GFP and imaged.

Relative neuronal and glial transduction levels were quantified by counting the number of transduced

neurons and glia, identified based on morphology, for each region per 50 μm coronal section. AAV1RX

transduction levels are shown alongside those for parental serotypes, AAV1 & AAVrh.10, as well as the

chimeric variants, AAV1R6 and AAV1R7, (same data as in Fig. 4 of the main text), across various

regions, including the (A) motor cortex, (B) cerebral cortex, (C) hippocampus, (D) dentate gyrus, (E)

thalamus, (F) hypothalamus, (G) striatum, and (H) amygdala. Error bars represent standard deviation (n =

3). An unpaired two-tailed T-test with Welch’s correction was used to demonstrate statistical significance

of each group relative to AAV1. ns, not statistically significant; *, P < 0.05. One-way ANOVA was also

used to demonstrate that differences among means are statistically significant. *, P < 0.05 for neuronal

and glial transduction across all brain regions shown here.

54

Figure 2.10. Peripheral tissue transduction and biodistribution of AAV1R6, AAV1R7 and

AAV1RX following intravenous administration. Mice were injected via the tail vein with PBS or with

either AAV1, AAVrh.10, AAV1R6, AAV1R7 or AAV1RXvectors packaging a CBA-Luc reporter

transgene at a dose of 1 x 1011 vg. At 2 weeks post-injection, (A) luciferase reporter transgene expression

levels and (B) biodistribution of viral genome copies across various tissues were quantified. Luciferase

expression levels were normalized to grams of tissue lysate and are represented as relative light units.

Vector genome (vg) copies are normalized to mouse lamin B (mLamB) as an endogenous housekeeping

gene and are represented as vg copies per cell. Error bars represent standard deviation (n = 3 for luciferase

assays; n = 4 for biodistribution). One-way ANOVA and an unpaired two-tailed T-test with Welch’s

correction for each group were carried out and significance relative to AAVrh.10 is shown. ns, not

significant. *, P < 0.05.

55

Figure 2.11. Sequence alignment of 1RX footprint, and Variable Region I (VR-I) across common

AAV serotypes. The sequence alignment of AAV1RX with common AAV serotypes highlights the

amino acid residues (VP1 numbering) of the 1RX footprint, within the larger context of variable region I

(VR-I), compared across common natural AAV serotypes. Black lines above and below demarcate the

specific residues within the 1RX footprint. Key: Residues that are identical across serotypes are

represented by red letters on a yellow background; conservative residues, blue letters on a cyan

background; similar residues, black letters on a green background; weakly similar residues, green letters

on a white background; non-similar residues, black letters on a white background. This sequence

alignment was generated using Vector NTI Advance 11.5.2 software.

56

CHAPTER 3: CAPSID-SIALIC ACID INTERACTIONS INFLUENCE ADENO-ASSOCIATED

VIRAL TRANSPORT ACROSS THE BLOOD-BRAIN BARRIER2

3.1. Overview

Central nervous system (CNS) transduction by systemically administered recombinant adeno-

associated viral (AAV) vectors requires crossing the blood brain barrier (BBB). We recently mapped a

structural footprint on the AAVrh.10 capsid, which when grafted onto the AAV1 capsid (AAV1RX)

enables viral transport across the BBB; however, the mechanisms involved remain unknown. Here, we

establish through structural modeling that this footprint overlaps in part with the sialic acid (SIA) receptor

footprint on AAV1. In the current study, we test the hypothesis that altered SIA-capsid interactions may

partly explain the ability of AAV1RX to traverse the BBB and subsequently transduce the CNS. We

focused on a panel of AAV1 capsid variants with altered SIA footprints, which can be divided into three

functional subgroups based on their relative SIA dependence. Specifically, capsids with ablated SIA

binding show an improved ability to cross the BBB, although transduce the CNS with low-to-moderate

efficiency. At the other end of the spectrum, the AAV1 capsid has a strong SIA dependency and does not

cross the BBB after systemic administration; instead, AAV1 shows high levels of vascular transduction

and is sequestered within the liver. The AAV1RX variant, which shows an intermediate SIA binding

phenotype, is not only effective at crossing the BBB, but also transduces the CNS at levels comparable to

AAVrh.10. In corollary, the reciprocal swap of the AAV1RX footprint onto AAVrh.10 (AAVRX1)

attenuated CNS transduction relative to AAVrh.10. We conclude that the composition of residues within

the capsid variable region I (VR-I) of AAV1 and AAVrh.10 appears to profoundly influence transport

across the BBB, with altered SIA interactions playing a partial role in this phenotype. Further, we

2Contributing authors: Blake H. Albright, Kate Simon, Minakshi Pillai, Garth Devlin, and Aravind

Asokan. K.S. and M.P. assisted in conducting animal experiments. G.D. performed tail vein injections.

57

postulate a “Goldilocks” model, wherein optimal glycan receptor interactions can influence BBB

transport and CNS transduction of AAV capsids.

3.2. Introduction

Adeno-associated viral (AAV) vectors for gene therapy applications are being clinically

evaluated for a range of genetic disorders affecting different tissue types11,130,140,142. One prime target for

delivering therapeutic transgenes using AAV is the central nervous system (CNS). In particular, CNS-

targeted AAV vectors are being evaluated for treating a broad range of monogenic and neurodegenerative

diseases afflicting the CNS, e.g., Spinal Muscular Atrophy (SMA), Parkinson’s, Alzheimer’s, Friedrich’s

Ataxia and several lysosomal storage diseases, to name a few142,156,157,189,278. Understanding AAV-host

interactions, particularly within the context of the CNS at the mechanistic level can help guide optimal

design of recombinant AAV vectors for such therapeutic applications.

AAV is a small, non-pathogenic, helper-dependent virus of the family Parvoviridae with a 4.7 kb

ssDNA genome8. There have been a wide array of both natural and synthetic serotypes that have been

discovered, which display broad tropisms across the host species, tissue and cellular levels11. These

differential tropisms are ultimately determined by capsid structural features and capsid-host interactions

afforded by that structure21. The AAV capsid is a small (~25 nm) T=1 icosahedron composed of 60 viral

protein (VP) monomers (VP1, VP2 and VP3 variants, assembled in a ratio of 1:1:10, respectively). The

VP monomer is characterized by a conserved jellyroll, β-barrel structure that is comprised of 7 anti-

parallel β-strands. The interdigitating loop regions connecting these strands are highly variable regions

(VRs) in sequence and highly plastic, being amenable to modification/mutation21. Furthermore, these

surface-exposed loops determine the topology of the capsid surface, and have a profound effect on virus-

host interactions, dictating receptor usage, antigenicity, tissue tropism, and species specificity279.

Principal amongst these virus-host interactions is the recognition of specific glycans, which serve

as primary receptors for attachment8,20,36,55,248. Binding of the AAV capsid to primary glycan receptors at

the cell surface serves as the initial step in the lifecycle of AAV. This is then followed by subsequent

interactions with host factors such as AAVR leading to endocytic uptake, intracellular trafficking steps

58

and nuclear entry leading to transduction22,126,280. The glycan receptors for numerous AAV serotypes have

been identified. Examples include recognition of heparan sulfate (AAV2/3), sialic acid (SIA,

AAV1/4/5/6), galactose (GAL, AAV9), and N-acetyl lactosamine (LacNAc,

AAVrh.10)68,85,113,116,119,281,282. Specifically, AAV1 & AAV6 recognize both α2,3 and α2,6 N-linked SIA.

AAV5 also recognizes N-linked SIA, while AAV4 uniquely recognizes α2,3 O-linked SIA (mucin)85.

From a structural perspective, the structures of various AAV capsids complexed with their cognate glycan

receptors have further pinpointed glycan binding sites to the protrusions at the 3-fold axis of symmetry on

the capsid38. The residues and locations on the AAV capsid surface that influence such interactions,

referred to as receptor footprints, are now known for AAV1, 2, 3, 4, 5, 6, and 968,113,116,283.

Different AAV serotypes exhibit various CNS tropisms, and our understanding of the role of

glycans in dictating these cellular tropisms is expanding. CNS tropism and transduction profiles can vary

drastically based on the route of AAV administration in different hosts (e.g., intravascular,

intracerebrospinal fluid (CSF), or intraparenchymal)157. Notably, certain serotypes, AAV9, AAVrh.8, and

AAVrh.10, are able to traverse the BBB following systemic administration and enter the brain

parenchyma to efficiently transduce cells within the CNS98,230. However, the capsid-glycan interactions

that might influence the mechanisms underlying AAV capsid transport across the BBB into the CNS are

not well understood. Towards furthering this understanding, our lab recently identified structural

determinants on the AAVrh.10 capsid, which when grafted onto the AAV1 capsid permits BBB

transport284. The engineered AAV1RX variant demonstrates marked transduction of neurons across

multiple brain regions and reduced tropism for glial and vascular endothelial cells in the brain. To further

understand the mechanisms underlying such transendothelial transport into the CNS, we extensively

characterized several AAV capsid variants displaying a range of SIA interactions in the current study. We

postulate a model, wherein the glycan binding affinity of AAV capsids inversely correlates with their

ability to traverse the blood-brain barrier.

59

3.3. Methods

Generation of mutant AAV plasmid constructs. Mutant AAV plasmid constructs were cloned

using site-directed mutagenesis via the QuikChange Lightning site-directed mutagenesis kit (Agilent

Technologies, catalog no. 200518) according to the manufacturer’s protocol. The following primer pairs

were used to generate the single-amino acid changes onto the pXR1 (AAV1) backbone for AAV1-N447S

(forward, 5′-CCAATACCTGTATTACCTGAGCAGAACTCAAAATCAG -3′; and reverse, 5′-

CTGATTTTGAGTTCTGCTCAGGTAATACAGGTATTGG -3′) and for AAV1-W503A (forward, 5 ′-

CAACAGCAATTTTACCGCTACTGGTGCTTCAAAATATAACC-3′; and reverse, 5′-

GGTTATATTTTGAAGCACCAGTAGCGGTAAAATTGCTGTTG-3′). The following primer pair was

used to generate the AAVRX1 mutant onto the pXRrh.10 (AAVrh.10) backbone: (forward, 5′-

CAGCAACGACAACCACTACTTCGGCTACAGCACCCCCTGG -3′; and reverse, 5′-

GCCCCCGTTGAAGCACTGGAGATTTGCTTGTAGAGGTGGTTGTTGTAGG -3′; IDT

Technologies, Ames IA).

Cell culture. HEK293 cells were cultured in Dulbecco’s modified eagle medium (DMEM)

supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher, Waltham, MA) and 1% penicillin,

streptomycin (P/S, Thermo Fisher), in 5% CO2 at 37⁰C. MB114 (mouse brain microvascular endothelial),

U87 (human astrocytoma), and Neuro2a (mouse neuroblastoma) cell lines were all cultured as described

above, except they were cultured with DMEM supplemented with 5% FBS and 1% P/S. Gratefully, we

received the MB114 cell line as a gift from Linda Van Dyk, University of Colorado.

Virus production and titers. Recombinant AAV vectors were produced using an updated triple

plasmid transfection method described previously15,285. A capsid-specific helper plasmid (i.e. pXR1,

pXRrh.10, pXR1RX, pXRRX1, pXR1-N447S, or pXR1-W503A) an adenoviral helper plasmid

(pXX680), and with a transgene packaging cassette, encoding either self-complimentary green fluorescent

protein driven by a chicken beta actin hybrid promoter (pTR-CBh-scGFP) or luciferase driven by the

chicken beta actin promoter (pTR-CBA-Luc), flanked by AAV2 inverted terminal repeat (ITR)

sequences. Viral vectors were harvested and were purified via iodixinol density gradient

60

ultracentrifugation. Viral vectors were further subjected to buffer exchange and concentration using

Sartorius Vivaspin6 100 kDa molecular weight cut-off (MWCO) centrifugation columns (F-2731-100

Bioexpress, Kaysville, UT). Purified virus preparations were titered by quantitative PCR with primers

amplifying the AAV2 ITR regions. (forward, 5’- AACATGCTACGCAGAGAGGGAGTGG –3’; reverse,

5’- ATGAGACAAGGAACCCCTAGTGATGGAG -3’; IDT Technologies, Ames IA).

In vitro transduction assays. Various cell lines were seeded in 24-well plates at a density of 1 x

105 cells per well and allowed to adhere overnight at 37⁰C. The following day, cells were incubated with

virus at a multiplicity of infection (MOI) of 10,000 vector genomes (vg) per cell and were incubated in

5% CO2 at 37⁰C. At 48 h post infection, cells were lysed using 150 μl of 1x passive lysis buffer

(Promega) for 20 minutes at room temperature. 50 μl of each cell lysate was then transferred to an assay

plate and luciferase reporter transgene expression levels were quantified using a Victor 3 multilabel plate

reader (Perkin-Elmer) immediately after adding 50 μl of luciferin (Promega) to each well.

Enzymatic desialylation assays. MB114 mouse brain microvascular endothelial cells were seeded

in 24-well plates at a density of 1 x 105 cells per well and allowed to adhere overnight at 37⁰C. The

following day, cells were pretreated with 50 mU/ml neuraminidase type III (from Vibrio cholera, Sigma

N7885) or with 1x PBS as a negative control, in serum-free DMEM at 37⁰C for 2 h. Cells were then

washed three times with 1x PBS, were given fresh DMEM containing 5% FBS and 1% P/S and were

incubated with virus at an MOI of 10,000 vg per cell at 37⁰C. At 24 h post infection, cells were lysed and

luciferase transgene expression levels were assayed as described earlier for the in vitro transduction

assays.

Cell surface binding assays following enzymatic desialylation. MB114 mouse brain

microvascular endothelial cells were seeded in 24-well plates at a density of 1 x 105 cells per well and

allowed to adhere overnight at 37⁰C. The following day, cells were pretreated with 50 mU/ml

neuraminidase type III (from Vibrio cholera, Sigma N7885) in serum-free DMEM at 37⁰C for 2 h. Cells

were then washed three times with 1x PBS, were given fresh DMEM containing 5% FBS and 1% P/S.

Cells were then pre-chilled at 4⁰C for 30 min prior to incubation with virus at an MOI of 10,000 vg/cell

61

for 1 h at 4 ⁰C. Cells were then washed three times with ice-cold 1x PBS to remove unbound virions

before adding 200 μl of molecular biology grade water to each well. Cells were subjected to three freeze-

thaw prior to collection of cell lysates along with cell surface-bound virions. Viral genome copy numbers

of cell-surface bound virions were then quantified using quantitative PCR (qPCR) with primers specific

to the luciferase transgene (forward, 5′- AAAAGCACTCTGATTGACAAATAC -3′; and reverse, 5′-

CCTTCGCTTCAAAAAATGGAAC -3′). Viral genome copy numbers were then normalized to the

mouse lamin B2 housekeeping gene using the following primer pair: forward, 5′-

GGACCCAAGGACTACCTCAAGGG -3′; and reverse, 5′- AGGGCACCTCCATCTCGGAAAC -3′.

The number of viral genomes are represented as the ratio of vg/cell and were plotted as fold-change

relative to the respective untreated control samples.

Animal studies. All animal experiments were performed using 6-to-8-week-old female C57/BL6

mice purchased from Jackson Laboratories (BAR Harbor, ME). These mice were maintained and treated

in compliance with NIH guidelines and as approved by the UNC Institutional Animal Care and Use

Committee (IACUC). Recombinant AAV vectors packaging a CBh-scGFP transgene or 1x PBS (as mock

treatment) were administered at a dose of 5 x 1011 vg via tail vein injection. Animals were sacrificed three

weeks post injection with an intraperitoneal injection of tribromoethanol (Avertin) (0.2 ml of 1.25%

solution) followed by transcardial perfusion with 30 ml of 1x PBS followed by 30 ml of 4%

paraformaldehyde in PBS. Tissues were extracted, were post fixed for 48 h, and brain and liver tissues

were sectioned (generating 50-μm-thick sections) using a Leica VT 1200S vibrating blade microtome

(Leica Biosystems, IL). Brain sections were then immunostained as described below.

Tissue processing and histological analysis. Free-floating 50-μm-thick coronal brain sections

were stained in 24-well plates. First, sections were blocked with buffer containing 10% goat serum and

1% Triton X (Sigma-Aldrich) in 1x PBS for 1 hour at room temperature. The primary antibody incubation

was then performed at 4⁰C overnight with a primary monoclonal rabbit α-GFP antibody (Life-

Technologies-G10, 362 1:750) diluted in blocking buffer. The next day, three 10 minute washes were

performed with 1x PBS and subsequent immunostaining was performed using the Vectastain ABC kit

62

(Rabbit IgG PK-4001 kit, Vector biolabs, Burlingame, CA). Immunostained brain sections were finally

mounted onto microscopy slides and were digitally imaged in bright field (20x objective) using an Aperio

ScanScope XT instrument (Aperio Technologies, Vista, CA) by the UNC Translational Pathology

Laboratory. Images were obtained using Leica eSlide Manager (centralized image storage and data

management software) and analyzed using Aperio ImageScope and WebViewer software. Quantifications

of cellular transduction were calculated by counting the number of GFP+ neuronal or glial cells,

determined based upon morphology, per 50 μm coronal brain section. Data was graphed and statistical

analyses were performed as outlined above. Specific brain regions were identified based upon comparison

to a coronal mouse brain reference obtained from the Allen Mouse Brain Atlas. To assay for GFP

expression in the liver, sections were stained with the anti-GFP primary antibody as described above;

however, an anti-rabbit goat antibody conjugated to Alexa-488 was instead used as the secondary

antibody at a dilution of 1:500 (anti-rabbit Abcam-96,883). These sections were then imaged with an

EVOS FL epifluorescence cell imaging system (AMC/Life Technologies) using the GFP light cube

(excitation 470 nm, emission 510 nm). Relative fluorescence was quantified from these images using

ImageJ software and data was plotted and statistical analyses were carried out as described earlier, using

GraphPad Prism7® software.

Vector genome biodistribution. Animal studies were performed as described above. DNA was

extracted and purified from paraformaldehyde-fixed 50-μm-thick tissue sections using the QIAamp DNA

FFPE Tissue Kit (QIAGEN, Cat No. 56404). Viral genome copy numbers were then determined for each

tissue using quantitative PCR with primers specific to the chicken beta actin (CBA) promoter/CMV

enhancer element sequence driving the CBh-GFP transgene (forward, 5′-

TGTTCCCATAGTAACGCCAA -3′; and reverse, 5′- TGCCAAGTAGGAAAGTCCCAT -3′). These

viral genome copy numbers were then normalized to the mouse lamin B2 housekeeping gene using the

primers (forward, 5′- GGACCCAAGGACTACCTCAAGGG -3′; and reverse, 5′-

AGGGCACCTCCATCTCGGAAAC -3′). The biodistribution of viral genomes are represented as the

ratio of vector genomes per cell recovered for each tissue.

63

Structural modeling. Previously published amino acid sequences and coordinates for the AAV1

VP3 structure (PDB ID, 3NG9) were used to generate three-dimensional homology models for the capsid

variants seen here263. SWISS-MODEL (http://swissmodel.expasy.org/) was used for structural prediction

based on homology to the AAV1 capsid264. Subsequently, a structure-based alignment was generated

using the secondary structure matching (SSM) application in the WinCoot software, again using AAV1

VP3 monomer as a template205,265,286. The VIPERdb oligomer generator utility

(http://viperdb.scripps.edu/oligomer_multi.php) was then used to generate oligomeric structural models of

the VP3 trimer/3-fold symmetry axes, the VP3 trimer dimers/2-fold symmetry axes, and the full capsid

(60mers)266. Lastly, PyMOL (Molecular Graphics System, Schrödinger LLC, http://www.pymol.org/) was

used visualize surface-rendered depictions of these models.

Data and Statistical Analysis. All quantitative data was plotted and statistical analyses were

carried out using the GraphPad Prism 7® software. One-way Anova was used to demonstrate statistical

significance. ns, not significant; *, p < 0.05.

3.4. Results

Characterization of AAV capsid variants spanning the SIA sensitivity spectrum. We hypothesized

that altered SIA interactions might be involved in the mechanism by which AAV1RX is able to cross the

BBB and transduce cells within the brain parenchyma. To investigate such a relationship, we first selected

several AAV capsid variants possessing differential affinities for binding α2,3- and 2,6- N-linked Sialic

acid (SIA), the primary glycan receptor for AAV1. First, we included AAV1RX, as well as its parental

serotypes – AAV1, which binds SIA with high affinity, as a positive control, and AAVrh.10, which does

not bind SIA, as a negative control (Fig. 3.1a). We also generated the AAVRX1 capsid variant to further

interrogate the 1RX footprint, by performing the reciprocal swap relative to AAV1RX (Fig. 3.1b).

Briefly, AAV1RX was generated by grafting a stretch of 8 residues from the variable 1 region (VR1) of

AAVrh.10 onto the AAV1 capsid. In contrast, AAVRX1 was generated by grafting the corresponding

residues from AAV1 onto the AAVrh.10 backbone. Additionally, we generated two AAV1-based capsid

mutants harboring single amino acid substitutions – AAV1-N447S and AAV1-W503A, in which the

64

asparagine at position 447 (AAV1 VP1 numbering) has been substituted for serine or in which the

tryptophan at position 503 has been substituted for an alanine, respectively (Fig. 3.1c & 3.1d). These

mutations have been previously published and both the asparagine at position 447 and the tryptophan at

position 503 have been shown to be located within the AAV1 SIA binding pocket and play a direct role in

binding SIA as direct contact residues283. Accordingly, these mutations have both been shown to result in

diminished SIA binding for AAV1. Structural models of these capsid variants, showing both the full

capsid (60mer) as well as the three-fold axis of symmetry/trimer are shown (Fig. 3.1).

We first evaluated the infectivity of these capsid variants by performing in vitro transduction

assays on three biologically relevant CNS-derived cell lines – the Neuro2a mouse neuroblastoma cell line,

the U87 human astrocytoma cell line, and the MB114 mouse brain microvascular endothelial cell line

(Fig. 3.2). These cells were infected with each capsid variant at a multiplicity of infection of 10,000

vg/cell and infection was carried out for 48 h. In the case of the Neuro2a cell line, transduction by most

capsid variants seemed to phenocopy either AAV1, producing relatively high transduction levels, or they

phenocopied AAVrh.10, demonstrating transduction levels that were roughly a log-fold lower (Fig. 3.2a).

Despite some variation in transduction levels, the overall transduction profiles observed for the U87 cell

line were similar, again yielding two groups, behaving either more AAV1-like or more AAVrh.10-like

(Fig. 3.2b). Only in the case of the MB114 mouse brain microvascular cell line did we see these capsid

variants demonstrate a more intermediate phenotype, showing substantial (~ 0.5 - 1.5 log-fold)

differences, relative to either AAV1 or AAVrh.10. Based upon this observation, as well as upon the

physiological relevance in terms of the BBB, we selected the MB114 brain microvascular cell line for

further experiments to assess the dependency of these capsids on sialic acid for cell surface attachment

and transduction.

Capsid variants display differential sensitivities to removal of cell-surface SIA. We next wanted

to evaluate the dependency of these capsids on the presence of cell surface SIA. To do so, we performed

in vitro transduction assays on MB114 cells following pretreatment with either serum-free media (as a

negative control) or with 50 mU/ml of neuraminidase (type-III, from Vibrio cholera), which

65

enzymatically removes terminal sialic acid residues from the cell surface (Fig. 3.3a). As expected, AAV1

is highly sensitive to removal of cell-surface SIA, resulting in a roughly 1.5-log decrease in transduction

following treatment. In contrast, AAVrh.10, which binds LacNAc, does not appear to depend upon cell

surface SIA for transduction, as treatment with neuraminidase had no effect on transduction. Similarly,

both AAVRX1 as well as AAV1-N447S transduction appeared to be SIA-independent, as they were

unaffected by neuraminidase treatment. However, a third functional group can be observed for other

capsid variants. Both AAV1RX, and, to a slightly lesser extent, AAV1-W503A demonstrated an

intermediate sensitivity to enzymatic removal of cell surface SIA with moderate, yet significant decrease

in transduction levels (roughly 2-fold, or 1.5-fold, respectively).

Having observed differential sensitivities of these capsid mutants to enzymatic removal of cell

surface SIA in terms of transduction levels, we next wanted to determine whether this finding would

correlate to cell surface binding of virions. Thus, we performed binding assays measuring viral genomes

(vg) bound per cell and plotted the data as fold-change following desialylation using neuraminidase

treatment (Fig. 3.3b). Consistent with the transduction profile, AAV1 binding was found to be highly

dependent on the presence of SIA, as desialylation results in a nearly 2-log decrease in bound viral

genomes per cell. Further, AAVrh.10 binding appears to be independent of SIA, as desialylation did not

result in any decrease in the number of bound virions detected per cell. Likewise, cell surface attachment

for both AAV1-N447S and AAVRX1 does not depend on the presence of SIA. Lastly, the intermediate

functional group identified in the transduction assays is recapitulated here as well – AAV1RX, and again,

to a lesser degree, the AAV1-W503A mutant both show moderate sensitivities to removal of cell surface

SIA, demonstrating roughly a 1.25 log-fold and near log-fold decrease in the number of bound vg

following neuraminidase treatment. To summarize, AAV1, the sole capsid in functional group I, binds

SIA with relatively high affinity while the capsids in group III, AAVrh.10, AAVRX1, and AAV1-N447S

do not possess any capacity to bind SIA. Furthermore, this data suggests that the capsids in functional

group II, AAV1RX and AAV1-W503A both retain the ability from their parental AAV1 to recognize

66

SIA, however their affinity for SIA is reduced, with AAV1-W503A showing an even greater reduction in

SIA binding affinity than seen for AAV1RX.

SIA binding affinity inversely correlates with BBB transport and CNS transduction. Having

observed these different functional groups in terms of sensitivities to enzymatic removal of cell-surface

terminal SIA residues from in vitro transduction and binding assays, we next wanted to see if these results

would correlate with in vivo transduction assays. In particular, we hypothesized that differential SIA

binding affinities displayed by these capsid mutants would correlate with differential capacities for BBB

transport and CNS tropism/transduction following systemic administration. Thus, we administered

AAV1, AAVrh.10, AAV1RX, or one of the two mutant capsids, AAV1-N447S and AAV1-W503A,

packaging a CBh-GFP reporter cassette at a dose of 5 x 1011 vg per mouse (~ 2.5 x 1013 vg/kg) via tail

vein injections in C57/Bl6 mice. Additionally, we also administered the reciprocal swap variant,

AAVRX1, at the same dose in order to evaluate the 1RX footprint outside the context of SIA interactions

via ablation of the footprint on AAVrh.10. At 21 days post injection, mice were sacrificed, tissues were

removed, processed, and 50 μm coronal brain sections were immunostained for GFP. These sections were

then scanned in bright field at 20x magnification and representative images were taken from multiple

brain regions in order to assess GFP transgene expression throughout the brain parenchyma. We have

included representative images for multiple brain regions – including the cerebral cortex, showing both

the somatosensory cortex (SCT, Fig. 3.4a) and the motor cortex (MCT, Fig. 3.4b). We also evaluated

transduction profiles in the hippocampus, visualizing CA1/2 pyramidal neuron layers (Fig. 3.4c) and the

dentate gyrus (Fig. 3.4d). We assessed transduction in the striatum (caudoputamen) region of coronal

lateral ventricle sections (Fig. 3.4e). We also visualized transduction in the thalamus (Fig. 3.4f) and

hypothalamus (Fig. 3.4g). Qualitative trends in these regions are further corroborated by quantitations of

GFP+ neurons and glia, in the cortex (CTX, Fig. 3.5a), the hippocampus (HC, Fig. 3.5b), and the striatum

(STR, Fig. 3.5c).

AAV1, which is a capsid highly dependent on SIA interactions, fails to undergo transvascular

transport across the BBB. Consistent with prior reports, we observed that, following intravascular

67

administration, AAV1 transduction is limited to microvascular endothelial cells across each region of the

brain that we evaluated. Negligible numbers of GFP+ neurons and glial cells were observed, regardless of

brain region, emphasizing the known vascular sequestration phenotype for AAV1.

AAVrh.10 does not bind SIA and is a neurotropic capsid characterized by its ability to efficiently

cross the BBB and mediate high levels of gene transfer in the CNS. Due to its known neurotropic

phenotype, AAVrh.10 was included as a positive control for BBB transport and CNS transduction.

Accordingly, AAVrh.10 demonstrates robust transduction of neurons, glia and vascular endothelial cells

across the cerebral cortex, including the SCT (Fig. 3.4a) and the MCT, as well as for overall quantitations

of cortical expression (Fig. 3.5a). In the case of AAVrh.10, we also observe substantial variation in terms

of the intensity of immunostained GFP+ cells, contributing to the variation we observe measurements in

number of GFP+ cells across brain sections from different mice. Interestingly, we also frequently observe

a pronounced gradient in terms of transduced neurons in the cortex (Fig. 3.4a and 3.4b). Specifically we

observe a high density of GFP+ neurons in cortical layers 1-3, which lie at the periphery of the

parenchyma. This density in GFP+ cortical neurons decreases across layers 4-6. This observation is

consistent with previous studies performed by our lab. Similarly, AAVrh.10 transduces neurons, glia and

endothelial cells in the hippocampus, including pyramidal neuron layers (Fig. 3.4c) and the DG (Fig.

3.4d), as well as for overall levels of hippocampal neuron and glial expression (Fig. 3.5b). AAVrh.10

transduces all cell types with high efficiency, indicating efficient BBB transport within the STR (Fig. 3.4d

and Fig. 3.5c), as well as within both the thalamus (Fig. 3.4e) and the hypothalamus (Fig. 3.4f).

AAV1RX is more moderately dependent on SIA for binding, efficiently traverses the BBB, is

neurotropic, and mediates high transduction levels in the brain parenchyma. Upon systemic

administration, we observe high levels of neuronal transduction across the somatosensory (Fig. 3.4a) and

motor cortices (Fig. 3.4b). Neuronal expression levels are only marginally lower than those observed for

AAVrh.10 across the entire cortex, while AAV1RX demonstrates a 2-fold decrease in glial transduction

alongside greatly diminished vascular transduction, relative AVrh.10 (Fig. 3.5a). Again, AAV1RX

recapitulates previous data, showing robust neuronal transduction at levels only marginally lower than

68

those for AAVrh.10 in the hippocampus while transducing ~5-fold fewer glial cells than AAVrh.10 (Fig.

3.5b). This trend is evident for both pyramidal neuronal layers (Fig. 3.4c) and granular layers in the

dentate gyrus (Fig. 3.4d). We see a similar trend in other brain regions, where we observe a robust

number of GFP+ neurons with relatively fewer GFP+ glia in the striatum (Fig. 3.4e) and thalamus (Fig.

3.4f). In the hypothalamus; however, AAV1RX seems to transduce lower levels of both neurons and glia

(Fig. 3.4g). Whereas in other brain regions we do see a modest decrease in AAV1RX neuronal

transduction levels compared to AAVrh.10, AAV1RX transduces striatal neurons at levels equal to

AAVrh.10 while simultaneously showing a greater than 2-fold decrease in glial transduction (Fig. 3.5c).

AAV1-N447S and AAV1-W503A are capsid mutants with defective SIA binding and demonstrate

inefficient BBB transport and low CNS transduction levels. We next wanted to evaluate our two AAV1-

based mutants. AAV1-W503A demonstrated a moderate (although further diminished relative to

AAV1RX) dependency on SIA interactions for in vitro transduction and binding. The other capsid

mutant, AAV1-N447S, which does not use SIA, was evaluated in tandem. Both mutants, when injected

intravenously, demonstrate increased uptake across the BBB and equivalent increases in neuronal and

glial transduction levels, relative to AAV1, as evidenced by roughly 20-fold and 2-fold increases in the

numbers of GFP+ cortical neurons and glia detected, respectively (Fig. 3.5a). Interestingly, we also

observed lower levels of vascular transduction for AAV1-W503A; however, AAV1-N447S mediated

considerable GFP expression in the vasculature (Fig. 4a and 4b). AAV1-N447S additionally produced a

gradient of transduction across the cortical neuron layers and in some paraventricular regions, similarly to

AAVrh.10. However, this was not observed in all samples, particularly for those exhibiting lower overall

expression. Despite an observable increase in the amount of GFP+ neurons and glia in the cortex relative

to AAV1, mean AAV1-W503A and AAV1-N447S, transduction levels averaged only ~ 5-8% and ~ 22-

25% of those observed for AAV1RX, in regards to neurons and glia, respectively. Throughout the

hippocampus, the SIA-binding defective mutants show the same general trend, across pyramidal neuron

layers (Fig. 3.4c) and the dentate gyrus (Fig. 3.4d). Both mutants mediate roughly equivalent neuronal

transduction levels, ~3-8-fold higher than AAV1, and glial transduction levels equivalent to AAV1 (Fig.

69

3.5b). Nonetheless, there is a significant difference in the number of GFP+ hippocampal neurons

observed for both mutants relative to AAV1RX, which demonstrated an approximately log-fold higher

neuronal transduction level and a roughly half-log difference in the number of transduced glia. In the

striatum, both mutants mediate roughly equivalent transduction profiles compared to other brain regions,

although more sample-to-sample variation is seen for AAV1-W503A, consistent, with its greater, yet

modest, SIA affinity (Fig. 3.4e). Here, we see a modest increase in transduction levels for both mutants

relative to AAV1, with ~3-4-fold higher numbers of GFP+ striatal neurons and ~1.5-fold more GFP+ glia

(Fig. 3.5c). Nonetheless, a significant log-fold decrease in the number of GFP+ striatal neurons, is

observed for AAV1-N447S and AAV1-W503A, compared to AAV1RX (Fig. 3.5c). Consistent with the

results described above, few GFP+ neurons and glia are seen for either mutant in the thalamus indicating

only marginal improvement in BBB transport and/or transduction, relative to their wildtype AAV1 capsid

(Fig. 3.4f). Lastly, few GFP+ neurons or glia are seen in the hypothalamus for either mutant, while low-

to-moderate levels of vascular transduction are seen in the hypothalamus for all capsid variants. In

general, we observed a trend in which SIA binding affinity seems to inversely correlate with transvascular

transport across the BBB and transduction of the brain parenchyma.

AAVRX1, the reciprocal swap relative to AAV1RX, demonstrates attenuated BBB transport and

CNS transduction relative to AAVrh.10. Lastly, we evaluated AAVRX1, the other remaining capsid

variant that does not bind SIA. Across the cortex, we see that intravascularly administered AAVRX1 does

transduce neurons and glia in the brain, as observed in the somatosensory (Fig. 3.4a) and motor (Fig.

3.4b) regions of the cortex; however, these transduction levels are markedly reduced, with ~22% and

~26% of the number of GFP+ cortical neurons and glia, respectively, compared to those observed for

AAVrh.10 (Fig. 3.5a). AAVRX1 also transduced the brain microvasculature similarly to AAVrh.10,

assessed visually (Fig. 3.4). Interestingly, we found that AAVRX1 seems to phenocopy AAVrh.10 in

terms of the aforementioned gradient in the density of GFP+ cells across cortical neuron layers;

furthermore, we see an even greater number of more faintly stained GFP+ cells relative to AAVrh.10.

AAVRX1 exhibited a significant log-fold decrease in neuronal transduction and a 2-fold decrease in glial

70

transduction, relative to AAVrh.10 in both the hippocampus (Fig. 3.4c-d and Fig. 3.5b) and in the

striatum (Fig. 3.4e and Fig. 3.5c). Lastly, while AAVRX1 continues to show noticeably fewer GFP+

neurons and glia relative to AAVRh.10 in the thalamus (Fig. 3.4f), it appears to behave similarly to

AAVrh.10 in the hypothalamus, showing similar albeit slightly lower numbers of GFP+ hypothalamic

neurons and similarly sparse GFP+ glia (Fig. 3.4g).

Effect of SIA binding on liver tropism and transduction. In comparison with AAV1, which

transduces the liver with moderate efficiency, AAVrh.10 is known to exhibit pronounced hepatotropism

following intravenous administration and mediates very high liver transduction. Our lab recently showed

that upon intravenous injection in mice, AAV1RX is detargeted from the liver despite maintaining robust

CNS transduction. Having found that AAV1RX demonstrates reduced SIA interactions in vitro, which

correlate with its CNS phenotype, we hypothesized that interference with capsid-SIA interactions could

also explain the decreased liver tropism and transduction observed in the case of AAV1RX. Briefly, we

evaluated GFP reporter expression in liver samples from mice that were intravenously injected with either

AAV1, AAVrh.10, AAV1RX, AAVRX1, AAV1-N447S or AAV1-W503A, each packaging a CBh-GFP

reporter cassette, at a dose of 5 x 1011 vg/mouse. Liver sections were immunostained for GFP reporter

expression and were imaged using a fluorescent microscope to assess GFP transgene expression (Fig. 3.6a

and 6c). GFP expression was then quantified using ImageJ software in order to measure average

fluorescence for multiple images per mouse (Fig.3.6b and 3.6d). Consistent with previous studies, we

observed moderate GFP expression levels for AAV1 and extremely low GFP expression for AAV1RX in

the liver. Next, we evaluated AAV1-W503A, which displays further reduced, yet still moderate SIA

binding affinity relative to AAV1RX. Consistently, AAV1-W503A also exhibits extremely low GFP

expression in the liver and is similarly liver-detargeted, relative to AAV1 (Fig. 3.6a and 3.6b).

Interestingly, the AAV1-N447S mutant, which unlike the W503A mutation, exhibits complete

independence from SIA interactions for in vitro transduction and binding, is not liver-detargeted. AAV1-

N447S actually shows increased GFP expression in the liver relative to AAV1, although substantial

variability in GFP expression levels was observed across different tissue samples (Fig. 3.6b).

71

We also evaluated the transduction profile for AAVRX1 in the liver alongside AAV1, AAV1RX

and AAVrh.10. Consistent with previous studies, the hepatotropic AAVrh.10 capsid exhibited extremely

high GFP expression in the liver. Interestingly, despite having exhibited attenuated CNS transduction,

AAVRX1 was found to retain the liver tropism of its parental AAVrh.10 capsid. Equivalent, if not higher,

GFP expression levels were observed for AAVRX1 in the liver, compared to AAVrh.10 (Fig. 3.6c and

3.6d). Because of oversaturation of the fluorescent signal due to the extremely high intensity of GFP

expression observed for AAVrh.10 and AAVRX1, we captured images using a lower (1%) intensity

setting on the microscope and these representative images are shown in a separate panel (Fig. 3.6c)

compared to the images taken at a higher intensity setting (10%) shown in Fig. 3.6a. All quantifications of

relative fluorescence, however, were measured using images captured at the 10% intensity setting.

Structural analysis of the 1RX footprint within the context of the AAV1 SIA binding pocket. We

recently mapped the coordinates of the AAV1RX surface footprint consisting of 8 key amino acid

residues [263N, 264G, 265T, 266S, 268G (an insertion relative to wildtype AAV1), 269S, 270T, and

274T (Vp1 numbering, relative to AAV1)] located within variable region I (VR-I), the loop connecting β

strands, B and C in the VP monomer. As previously described, these residues are exposed on the surface

of the full 60mer capsid and manifest as a cluster located on a protruding shoulder region located near the

base of the larger protrusions at the three-fold axis of symmetry. This capsid structural feature serves as a

hotspot for determining species, tissue and cellular tropism as well as antigenicity. This shoulder region

containing the residues of the 1RX footprint are also located in close proximity to and at the edge of the

depression at the two-fold axis of symmetry.

Based on recent structural studies, we modeled the SIA receptor binding pocket on AAV1,

containing eleven amino acid residues (S268, D270, N271, Y445, N447, G470, S472, V473, N500, T502,

& W503, AAV1 Vp1 numbering) and located at the base of the three-fold spikes. On the capsid

monomer, these residues span VR-I, VR-IV, and VR-V regions. We generated models of the AAV1RX

capsid on which we overlaid the AAV1 SIA binding pocket residues (magenta) with the AAV1RX

footprint residues (greencyan) (Fig. 3.7). From this overlay, we see that a single amino acid residue

72

overlaps between the two footprints (pale blue), a serine at position 268 (S268) in the AAV1 sequence,

which has been mutated to a threonine in the AAV1RX footprint (270T, AAV1RX VP1 numbering). Out

of the aforementioned eleven residues of the AAV1 SIA binding pocket, only six residues have been

implicated as direct contact residues for binding the glycan – these are N447, S472, V473, N500, T502,

and W503 (again, AAV1 VP1 numbering). The other five residues (S268, D270, N271, Y445, and G470)

play an indirect role in mediating capsid-SIA interactions. When overlaid with the 1RX footprint, we see

that the single overlapping residue that has been mutated in the SIA binding pocket, S268T (AAV1, VP1

numbering), is not directly involved in binding SIA, but still contributes to capsid-SIA interactions.

Furthermore, we see in our structural model that this overlapping (blue) residue lies nearby, but not

directly at the mouth of the hole that forms the binding pocket upon the three-fold spike (Fig. 3.7, top

inset). Additionally, we highlighted the six direct contact residues for SIA in red (Fig. 3.7, bottom inset).

Within this region are residues directly involved in binding SIA were mutated in the AAV1-W503A and

AAV1-N447S substitutions that were evaluated in this study. Consistent with previous studies, we found

that single amino acid substitutions of these direct SIA contact residues results in greatly diminished and

abolished SIA-binding ability for the W503A and N447S mutations, respectively. In conclusion, the

AAV1RX footprint appears to have optimally engineered SIA recognition, which contributes to improved

transport across the BBB.

3.5. Discussion

AAV serotypes utilize a broad range of cell surface glycans for attachment, which are, in part,

key determinants of tropism. In particular, several AAV serotypes recognize different forms of SIA, with

AAV1 binding α-2,3- and α-2,6- N-linked SIA within a pocket of residues located at the base of the three-

fold capsid surface protrusions283. Systemic administration of AAV1 results in a largely vascular and

hepatic transduction profile, with negligible transport across the BBB and CNS transduction. However,

engraftment of the 1RX footprint from AAVrh.10 onto the AAV1 capsid enables efficient BBB transport

and robust neurotropism/CNS transduction, along with reduced vascular and peripheral tissue

transduction.

73

Structural analyses revealed that the 8 residues of the 1RX footprint lay on a shoulder which is

adjacent to the AAV1 SIA binding pocket located at the base of the three-fold protrusions. By modifying

residues adjacent to the SIA footprint rather than direct contact residues, our data supports the notion that

partially attenuated SIA interactions might contribute to capsid transport across the BBB. Accordingly,

we propose a “Goldilocks” model, wherein the ideal SIA binding affinity, in conjunction with other

contributing factors can influence CNS tropism of AAV capsids (Fig. 3.8). Specifically, capsids with

strong SIA interactions show high levels of vascular sequestration and do not cross the BBB following

systemic administration. Additionally, these capsids also transduce the liver efficiently. At the other end

of the spectrum, capsids with either severely diminished or abolished SIA binding cross the BBB and

transduce the CNS with only low-to-moderate efficiency, either due to reduced transport and/or

transduction. The AAV1RX capsid displays an intermediate SIA binding phenotype, where it is not only

effective at crossing the BBB, but also transduces cells in the brain parenchyma at a high level.

It is important to note that our proposed model, while sufficient to explain the phenotypes of

AAV1 based variants (and possibly those attributed to other SIA binding serotypes), cannot directly

address the mechanism utilized by AAVrh.10 in crossing the BBB. As mentioned earlier, AAVrh.10 has

recently been shown to recognize sulfated LacNac119. It is likely that LacNAc interactions influence the

CNS tropism of AAVrh.10, and that mutations that interfere with these interactions could effectively

attenuate BBB transport and/or CNS transduction. Such a scenario is corroborated in part by the

attenuated phenotype displayed by AAVRX1 (reciprocal swap of AAV1 VR-1 residues onto AAVrh.10).

Interestingly, however, AAVRX1 appears to retain the extremely high liver transduction phenotype of its

parent. Another deviation from our model is the AAV1-N447S mutant, which despite abolished SIA

binding and unable to traverse the BBB, displays improved AAVrh.10-like transduction in the liver and

vasculature. One possible explanation is that the AAV1-N447S change constitutes an amino acid change

identical to that observed on AAVrh.10 at that position, potentially imparting AAVrh.10-like features.

However, whether AAVrh.10 utilizes a footprint similar to AAV1-SIA capsids to recognize the LacNAc

receptor remains to be seen. Further, we note that capsid-glycan interactions only offer a partial

74

explanation for AAV tropism in general and acknowledge that other host factors such as AAVR or

soluble serum factors might play a critical role in this regard.

Our model is corroborated by other studies demonstrating that affinity for SIA is likewise

inversely correlated with both virulence and spread for other viruses, such as other Parvoviruses and

Polyomaviruses. This has been shown to be the case for the Minute Virus of Mice (MVM), another

parvovirus, which also recognizes α-2,3-sialylated glycans as a critical step for infection287. Studies with

SIA binding MVM mutants implicated differential SIA receptor affinity as a key determinant of virulence

in mice288. Similarly, differential SIA receptor recognition and affinity has been reported as a key factor

influencing CNS tropism and pathogenicity of Theiler’s virus, a Picornavirus as well as the systemic

spread of Polyomaviruses in mouse289–291. Another facet of the aforementioned studies that support our

conclusions is that similar SIA binding structural footprints, localized to the depression at the 2-fold axes

of symmetry appear to be employed by other Parvoviruses, including MVM, CPV, and FPV36,55. This

region has now been established as a key determinant of glycan receptor binding, antigenicity, tissue

tropism and BBB transport by a multitude of studies in the literature.

In summary, our studies establish that capsid-glycan interactions can influence AAV capsid

transport across the BBB. While the current study hinges on AAV1 capsid variants and sialylated glycans

as an example, it is plausible that other AAV serotypes and their cognate glycan receptors exhibit similar

relationships that impact their CNS tropism. These findings could provide a structure-guided platform for

engineering capsid-glycan interactions that enable optimal BBB transport and CNS transduction by AAV

vectors for gene therapy of neurological disorders.

75

Figure 3.1. Structural models of AAV capsids used to interrogate the role of sialic acid interactions

upon blood-brain barrier transport & CNS tropism. Three-dimensional molecular models of the

whole 60mer capsid and of the trimer/three-fold axis of symmetry were generated for each capsid chosen

for evaluating the impact on sialic acid interactions on crossing the blood-brain barrier: (A) AAV1RX,

footprint for crossing the BBB from AAVrh.10 grafted onto AAV1, (B) AAVRX1, the corresponding

residues from AAV1 grafted onto the AAVrh.10 backbone, (C) AAV1-N447S and (D) AAV1-W503A,

two AAV1-based single amino acid changes previously shown to display defective capsid-sialic acid

interactions. Amino acid residues derived from AAV1 are shown in gray, residues derived from

AAVrh.10 are depicted in green-cyan, and single amino acid substitutions (AAV1-N447S & AAV1-

W503A) are shown in red. These structural models were visualized and generated using PyMOL.

76

Figure 3.2. Selected capsid variants display intermediate transduction profiles in a cell type-specific

manner in vitro. Infectivity of capsid variants selected for differential sialic acid interaction potentials

was tested by incubating each capsid variant virions on various CNS-derived cell lines at an MOI =

10,000 vg/cell. Cells were lysed 48 h post infection to measure luciferase reporter expression. Cell lines

evaluated include (A) Neuro2a mouse neuroblastoma cell line, (B) U87 human astrocytoma cells, and (C)

MB114 mouse brain microvascular endothelial cells. Data are presented as means and error bars represent

standard deviation (n = 3).

77

Figure 3.3. Capsid variants display differential sensitivities to removal of cell-surface sialic acid for

both transduction and binding. (A) Transduction of capsid variants on MB114 cells following a 2 hour

pre-treatment with either serum-free media or serum-free media containing 50 mU/mL Neuraminidase

(Type-III, from Vibrio Cholerae). Viruses packaging a CBA-Luc reporter gene were incubated on MB114

cells (MOI = 10,000 vg/cell) immediately following treatment. Cells were lysed 48 h post infection to

measure relative luciferase transgene expression. Solid bars represent transduction in absence of enzyme

while dotted bars represent transduction following treatment with neuraminidase. (B) Binding on MB114

cells following treatment with or without neuraminidase. Cells were treated as described above with

neuraminidase and were pre-chilled prior to incubation with virus for 1 h at 4 ⁰C. Cells were then lysed

after three washes, DNA was extracted and the number bound viral genomes was quantified by qPCR.

Binding data is shown and fold change. Data are presented as means ± SEMs (n > or = 3). For

transduction data, an unpaired two-tailed t-test with Welch’s correction was performed to demonstrate

statistical significance. ns, not significant; *, p < 0.05.

78

Figure 3.4. Evaluating the impact of Sialic Acid binding on blood-brain barrier traversal and CNS transduction in vivo. Viruses displaying

differential dependencies on sialic acid were produced packaging a CBh-scGFP transgene and were administered to mice at a dose of 5 x 1011 via

tail vein injection and Immunostained brain sections were Immunostained for GFP expression 21 days later. Images from scans show transduction

profiles for each variant (organized as higher to lower dependency on sialic acid, from left to right: AAV1, AAV1RX, AAV1-W503A, AAV1-

N447S, AAVRX1, and AAVrh.10) across multiple brain regions: (A) somatosensory cortex, (B) motor cortex, (C) pyramidal neurons of the

hippocampus, (D) dentate gyrus, (E) striatum (caudoputamen), (F) thalamus, and the (G) hypothalamus. Images are representative of n > or = 3

mice. Scale bars, 100 μm.

79

Figure 3.5. Quantitation of neuronal and glial transduction profiles for capsid variants displaying

differential dependencies on Sialic Acid. Three weeks after tail vein injections of vectors packaging a

CBh-scGFP transgene at a dose of 5 x 1011 vg, brain sections were immunostained for GFP transgene

expression, imaged, and relative neuronal and glial transduction levels were quantified by counting the

number of GFP+ cells, as determined by cellular morphology. Relative transduction levels were quantified

for (A) the cerebral cortex, (B) the hippocampus, and (C) the striatum. Data is shown as means and error

bars represent standard deviation (n > or = 3).

80

Figure 3.6. Relative liver transduction mediated by capsid variants with differential sialic acid

sensitivities. Three weeks after tail vein injections of vectors packaging a CBh-scGFP transgene at a dose

of 5 x 1011 vg, liver sections were immunostained for GFP transgene expression. (A) GFP expression was

imaged at 10% light intensity for mock treatment, AAV1, ARX, AAV1-W503A, and AAV1-N447S

variants and (B) relative fluorescence was quantified with ImageJ and averaged across multiple images

taken for multiple sections per liver sample. (C) GFP expression was imaged at 1% light intensity for

mock treatment, AAV1, AAV1RX, AAVRX1, and AAVrh.10 (due to oversaturation of sensors by

AAVRX1 & AAVrh.10). (D) Relative fluorescence levels were quantified as described above. Images

taken at 10% light intensity were used for all quantifications of relative fluorescence. Data represents

means ± SD (n = or > 3). One-way Anova followed by Tukey’s multiple-comparison test were carried out

to demonstrate statistical significance relative to either AAV1 or AAVrh.10, for (B) and (C), respectively.

ns, not significant. *, p < 0.05.

81

Figure 3.7. Structural analysis of the 1RX footprint within the context of the AAV1 Sialic acid

binding pocket. Structural surface models overlaying the AAV1 sialic acid (SIA) binding pocket with on

top AAV1RX. Homology models were generated using SWISS-MODEL and oligomers were

subsequently generated using VIPERdb. PyMol was used to generate surface-rendered models of the

AAV1-sialic acid binding pocket overlaid on top of (A) the AAV1RX capsid (60mer). (B) AAV1RX

trimer-dimer showing the 1RX-SIA receptor footprint overlay within the context of both the three-fold

and two-fold axes of symmetry. (C) Zoom-in of the AAV1RX-SIA binding pocket overlay within the

context of a representative protrusion at the three-fold axis of symmetry. AAV1-derived residues are

depicted in grey, the 8 residues of the 1RX footprint are depicted in greencyan, the residues of the AAV1

SIA binding pocket are shown in magenta, residues present in both the SIA receptor footprint and the

1RX region are shown in blue, and the residues in the SIA binding pocket that actually make direct

contact with the SIA are shown in red.

82

Figure 3.8. Goldilocks model for the influence of capsid-sialic acid interactions on blood-brain

barrier transport and CNS transduction. Here we suggest a goldilocks model which describes how

capsid-SIA interactions influence the ability of systemically administered AAV capsids to cross the

blood-brain barrier and transduce the CNS. On one end of the spectrum, capsids with strong SIA

interactions (i.e. AAV1) undergo vascular sequestration and are prevented from crossing into the brain

parenchyma and also transduce the liver. On the other end of the spectrum, capsids with severely

diminished or abolished SIA-binding are either cross the BBB inefficiently and/or transduce the CNS

with low-to-moderate efficiency, due to either reduced transport and/or transduction capabilities. The

AAV1RX capsid, which shows an intermediate SIA-binding phenotype, however, seems to fall within a

goldilocks zone, where it is not only effective at crossing the BBB but also at transducing cells in the

brain parenchyma, at levels only slightly lower than seen for AAVrh.10.

83

CHAPTER 4: CONCLUSIONS AND FUTURE DIRECTIONS

4.1. Overview

The overarching goal of this dissertation was to better understand structure-function correlates on

the AAV capsid for blood-brain barrier (BBB) transport as well as central nervous system (CNS)

transduction and tropism. Previous work elucidating structural determinants on the AAV capsid for

tropism and immunogenicity have provided structural insights that have paved the way for efforts to

engineer new AAV vectors with improved safety and efficacy21,55,66,72,75. A recent study made the pivotal

discovery that several natural AAV serotyped capsids are able to traverse the BBB in mice and opened up

the possibility for treating CNS diseases afflicting multiple brain regions with AAV vectors administered

conveniently through intravascular injections257,292. Nonetheless, the precise capsid features that allowed

for BBB transport and CNS transduction remained unknown.

AAV is currently a leading vector for CNS gene therapy applications, bolstered by the success of

many preclinical and clinical studies90,142,144,278,293. Substantial challenges nevertheless limit the

therapeutic potential of CNS gene therapy. Foremost amongst these are insufficient transduction

efficiencies and/or a lack of specificity in regards to tissue tropism for the current leading AAV vectors

for CNS applications, namely AAV9 (and to a lesser extent AAVrh.10). These issues often necessitate

extremely high dosages in order to achieve therapeutic expression, increasing scale up costs for vector

production and can cause dose-related hepatotoxicity189,192,245. Thus, there is still a critical need to develop

improved capsids for targeting the CNS via systemic administration, which is more convenient, less

invasive, and better suited for achieving global CNS gene transfer, compared to other routes of

administration, which often result in more punctate or limited localization of transgenes.

Many application-based capsid engineering efforts, generally involving high-throughput shotgun

approaches such as library generation combined with directed evolution, are continually being employed

84

in an attempt to develop improved vectors. Despite this, studies focused on expanding our understanding

of AAV capsid biology and capsid-host interactions are currently lacking. The rational design of new

capsid variants with improved CNS gene transfer following systemic administration requires a better

understanding of structural features on the AAV capsid involved in virus-host interactions, such as those

which occur at the neurovascular unit. Therefore, we focused on understanding capsid structure,

dissecting structural determinants and capsid-host interactions important for AAV transport across the

BBB and CNS gene transfer.

4.2. Mapping Structural Determinants for Transport of AAV across the Blood-Brain Barrier.

In chapter 2, we set out to identify the structural features on the capsid that confer the ability of

some AAV serotypes to cross the BBB and penetrate the brain parenchyma. To address this question, we

first generated a chimeric capsid library through DNA shuffling, using AAVrh.10 and AAV1 as parental

capsids, which either can or cannot cross the BBB, respectively. In contrast to other studies using high-

throughput strategies such as library generation and directed evolution to evolve ideal vectors, we chose

not to subject our library to selective pressures. Instead, we isolated individual chimeric variants from our

library and screened them for their ability to cross the BBB and transduce the CNS in vivo. We found two

variants, AAV1R6 and AAV1R7, which crossed the BBB and transduced multiple regions in the brain

following intravenous injections. These capsids were also found to preferentially transduce neurons while

showing reduced transduction of glial and endothelial cells in the parenchyma.

Using structural cues, excluding non-surface-exposed residues and residues implicated in genome

packaging or assembly rather than for tropism or transduction, we were able to identify a minimal

footprint consisting of 8 residues from AAVrh.10 that are important for crossing the BBB. Through

grafting this 8-residue footprint onto AAV1, we were able to rationally engineer the AAV1RX capsid,

which phenocopied AAV1R6/7 in terms of BBB transport and the ability to preferentially transduce

neurons throughout the brain. Lastly, we found that AAV1RX (in addition to AAV1R6/7) retains the

cardiac transduction profile shared by its parental capsids and is detargeted from nearly all other

peripheral tissues, including the liver, as assessed by in vivo transduction assays and biodistribution of

85

viral genomes. Lastly, we found that this footprint was shared, to varying degrees, across other naturally

neurotropic capsids (i.e. AAVrh.8, and partially in AAV7, AAV8, AAVrh.39, and AAVrh.43). This

further supports the importance of these 8 residues and suggests that completion of the footprint may

enhance BBB transport and CNS transduction for some serotypes.

While the mechanism of action for how the 1RX footprint affords transport across the BBB is

unknown, these findings simultaneously advance our understanding of AAV biology and provide a

structure-guided platform for rationally engineering novel CNS-targeted AAV vectors with the potential

to provide therapeutic gene expression at a lower dose, thus improving both safety and efficacy.

Additionally, AAV1RX is a novel AAV capsid with therapeutic potential that is ideally suited for CNS

indications, due to its ability to traverse the neurovascular unit and efficiently transduce the brain

parenchyma, with simultaneous detargeting from peripheral tissues such as the liver, following systemic

administration.

Furthermore, these findings highlight the synergy between using high-throughput techniques such

as library generation combined with structure-guided design. Together, they provide the potential to

simultaneously elucidate AAV biology while generating improved AAV vectors. By providing a

structural clue in the mystery of how AAV capsids cross the BBB, the discovery of the 1RX footprint will

hopefully serve as a springboard for further dissection of capsid-host interactions which occur to promote

or restrict transvascular transport across the neurovascular unit.

4.3. Influence of Capsid-Glycan Interactions on Blood-Brain Barrier Transport, Neurotropism, and

Transduction

Following up the discovery of the AAV1RX footprint that enables efficient BBB transport, we

wanted to shed light on the mechanisms and potential capsid-host interactions involved in this phenotype.

Structural analyses mapped the residues of the 1RX footprint to a shoulder adjacent to the AAV1 SIA

binding pocket located at the base of the 3-fold protrusions. Thus, in chapter 3 of this dissertation, we set

out to further understand the influence of capsid-SIA interactions on AAV transport across the BBB as

well as CNS tropism and transduction. We characterized several AAV capsid variants which were found

86

to exhibit differential sensitivities to enzymatic removal of cell surface SIA in vitro and in vivo. We found

an inverse correlation between capsid-SIA interactions and BBB transport as well as CNS

transduction/tropism. We additionally observed a mostly direct correlation between SIA dependency and

both vascular and hepatic transduction phenotypes, although some variation was seen.

Structural analyses in which the 1RX and AAV1-SIA footprints were overlaid revealed that the

mutations we made through engraftment of the 1RX footprint are adjacent to the direct contact residues

for SIA. We found that mutating direct contact residues for SIA results in either greatly weakening or

ablating capsid-SIA interactions. By modifying adjacent residues in the 1RX footprint, rather than the

direct contact residues themselves, we seem to have instead only partially attenuated SIA interactions.

Thus, the 1RX residues seem to tweak capsid-SIA interactions to promote, while not entirely accounting

for, the phenotype observed for AAV1RX, in terms of BBB transport and neurotropism.

We concluded chapter 3 by proposing a “Goldilocks” model, in which the ideal glycan binding

affinity, in conjunction with other factors, can impact the ability of AAV capsids to cross the BBB.

Briefly, we found that capsids with a high SIA dependence are restricted to the vasculature (as well as the

liver). On the other side of the spectrum, capsids with very low SIA dependence cross the BBB and

transduce the CNS with very low efficiency, with mostly reduced vascular and liver transduction. We

found that AAV1RX, which displays more moderate SIA dependency seems to lie in a “Goldilocks

zone”, where it is able to efficiently cross the BBB and efficiently transduce the parenchyma, while being

simultaneously detargeted from vasculature and the liver. Nonetheless, this model is insufficient to

describe capsids which do not recognize SIA, such as AAVrh.10 and AARX1, which is the reciprocal

swap relative to AAV1RX. Moreover, we found that grafting the corresponding residues from AAV1

onto the AAVrh.10 capsid attenuated BBB transport efficiency and CNS transduction compared to

AAVrh.10. Thus, the AAV1RX phenotype cannot be solely explained by altered capsid-SIA interactions.

Rather, our data suggests the involvement of other mechanisms important for BBB transport, such as

interactions with the sulfated LacNac receptor for AAVrh.10 or interactions between the capsid and other

host factors such as a soluble serum protein.

87

In conclusion, our findings shed light on the mechanisms governing AAV CNS tropism,

providing a structure-guided platform for fine-tuning capsid-glycan interactions in order to effect tropism

and transduction profiles, such as to promote BBB traversal for CNS gene therapy. Furthermore, while

most structural-guided rational design efforts for capsid engineering focus on modifications to known

hotspot determinants of tropism, such as the three-fold protrusions themselves, our data demonstrates the

potential for fine-tuning precise interactions to effect the aforementioned phenotypes by modifying

contextual landscapes surrounding these regions, pointing towards the 2-fold axis of symmetry and

regions adjacent to the three-fold protrusions as areas for investigation that are amenable to mutation in

order to modulate, but not abolish desirable phenotypes, such as transvascular transport.

4.4. Future Directions

In this dissertation, we sought to enhance our understanding of structure-function correlates on

the AAV capsid for BBB transport, neurotropism, and transduction of the brain parenchyma following

systemic (intravascular) administration. We set this as our goal in an effort to guide rational and structure-

guided design of improved AAV vectors for neurological gene therapy applications (particularly for those

which require widespread gene transfer throughout the brain). Given the discovery of several naturally

neurotropic AAV serotypes, along with current limitations of CNS gene therapy due to the BBB, we

resolved to identify structural determinants for transvascular transport. We utilized a high-throughput

library based strategy in tandem with structural analyses to identify structural determinants for these CNS

phenotypes and used the discovery of a structural footprint for BBB transport to engineer the synthetic

and neurotropic AAV1RX capsid, an ideal candidate for neurological gene therapy applications in the

future. We further expanded on this in our subsequent study where we drew a link between SIA binding

and BBB transport, glimpsing a partial mechanism through which the 1RX footprint may afford BBB

transport. These findings suggested a model for the influence of capsid-glycan interactions on BBB

traversal and CNS transduction. Further, these studies gave us a better understanding of capsid-host

interactions and identified key targets, on the capsid side (1RX footprint) and the host side (glycan

interactions), that are involved in the mechanism by which AAV capsids cross the BBB. In this way, the

88

identification of these key targets for evaluation may act as a stepping stone to inform future studies

which will hopefully build upon these findings and further unravel this mysterious mechanism.

Mechanism of transvascular transport for AAV across the blood-brain barrier and defining

interactions at the neurovascular unit. Mapping of the 1RX footprint for BBB transport from AAVrh.10

provides a clue as to the structural elements involved in BBB transport; however, the actual mechanism of

AAV transport across the neurovascular unit remains a mystery. Perhaps the largest remaining question -

how do AAV1RX and other AAV capsids engage the neurovascular unit, and are these interactions

serotype specific? Furthermore, experiments using intracerebroventricular (ICV) injections of AAV1RX

should be performed to assess the transduction properties and cellular tropism of AAV1RX in the brain

outside the context of BBB transport, and will also assess the ability of AAV1RX to spread within the

brain parenchyma. Future studies should be aimed at uncovering the mechanism(s) through which

AAV1RX, as well as other neurotropic capsids (i.e. AAV9, AAVrh.8, AAVrh.10, etc.), cross the BBB

and should be focused on defining how these capsids engage the neurovascular unit and whether this

engagement is capsid-specific.

We cannot rule out the possibility that AAV crosses through disrupting BBB integrity, i.e.

mediated through inflammatory effects which may weaken tight junction complexes between endothelial

cells (i.e. through weakening tight junctions between endothelial cells); however, this scenario seems

unlikely due to the favorable immune profile of AAV, lack of supporting data, and the lack of viral genes

to facilitate such a process. In a more likely theory, AAV transport across the BBB is thought to occur via

receptor-mediated transcytosis157. While current in vitro cell culture-based BBB models fail to faithfully

recapitulate the in vivo neurovascular unit, future studies involving cultured endothelial cell monolayers

(and perhaps other co-cultures, such as astrocytic end-feet, pericytes, etc.) could play a role in pinpointing

differences in cellular uptake and trafficking routes between AAV1RX and AAV1/rh.10, as well as more

defined serotypes like AAV2 & AAV9. The abilities of AAV2, AAV9 and bovine AAV (BAAV) to

undergo transcytosis across monolayers of both endothelial and epithelial cells in vitro has been

characterized. It was found to be an active process, is serotype-dependent and that vectors are still

89

transduction-competent following apical-to-basolateral transport294,295. Furthermore, these studies

provided evidence that transduction and transcytosis may occur through independent and possibly

competitive cellular routes296. This theory makes sense in the context of AAV1/rh.10/1RX, where it is

tempting to speculate that AAV1 may exclusively utilize a transduction pathway in the endothelium. In

contrast, it is possible that AAV1RX preferentially undergoes trafficking routes targeted for transcytosis,

while AAVrh.10 may use both routes ubiquitously. Future experiments could support or deny this theory

and could focus on ascertaining capsid-host factor interactions that may be involved in these intracellular

pathways. Additionally, future directions should include dissecting the 1RX mechanism through co-

administering AAV with various pharmacological agents which block various steps in the

transduction/transcytosis/trafficking pathways and to look for differential effects between AAV1RX &

AAV1/rh.10. We know that glymphatics affect the spread of AAV within and clearance from the CNS;

however, it remains to be determined whether BBB transport and/or CNS transduction are also impacted

by glymphatics239. Moreover, we do not know whether AAV1RX/rh.8/rh.10 are capable of undergoing

axonal transport, either unidirectional or bidirectional, which should be addressed in future studies.

Many questions still remain regarding the precise mechanism of how AAV capsids cross the BBB

and the interactions involved. While thought to undergo receptor-mediated endocytosis, it is still

unknown precisely how AAV1RX gains entry into the brain. It is possible that AAV1RX may bind

another host factor as part of its mechanism, and we still do not know how AAV1RX interacts with the

individual components of the BBB. If it does undergo receptor-mediated endocytosis across the

vasculature, is its interaction limited to vascular endothelial cells, or is this transport impacted by

interactions between the AAV1RX capsid and other components of the neurovascular unit? Vascular

endothelial cells and their tight junctions are not the only components present at the BBB. Rather, the

neurovascular unit also consists of vascular smooth muscle cells and pericytes, which both line the

endothelium, as well as astrocytic end feet, the basement membrane, and the highly glycosylated

extracellular matrix224,297. Interactions likely occur between AAV1RX (as well as for other capsids) and

90

these components, which may either facilitate or hinder transvascular transport at the BBB. Dissection of

these interactions should be the focus of future studies, as they are critical gaps in our knowledge.

Translatability of the 1RX footprint to enhance blood-brain barrier transport and CNS

transduction for other AAV serotypes. We know that engraftment of the 8 key residues of the 1RX

footprint from AAVrh.10 onto the AAV1 capsid affords BBB transport and CNS transduction; however,

we don’t know if this footprint can be grafted onto other serotypes and whether it will be sufficient to

confer this CNS phenotype. Firstly, we do not know whether other residues in the AAV1 (or AAVrh.10)

VP3 capsid sequence(s) act synergistically with the 1RX footprint to enable robust BBB transport/CNS

transduction. Sequence alignments comparing multiple naturally occurring neurotropic capsids reveal that

this footprint is either entirely conserved (in the case of AAVrh.8/rh.39) or mostly conserved (i.e. for

AAV7/8/9/rh.43)284. It is tempting to speculate that completion of this footprint (i.e. mutation of

mismatched residues to the equivalent in AAVrh.10) may enhance transvascular transport and CNS

transduction for these capsids. Nonetheless, we do not know whether engraftment of the 1RX footprint is

compatible with more divergent serotypes (i.e. AAV2/3/4/5). Studies in the future should focus on

answering these questions regarding the translatability of the 1RX footprint to other capsid variants.

Determining the role of the blood-cerebrospinal fluid and cerebrospinal fluid-brain barriers, and

the choroid plexus in regards to brain penetration by systemically administered AAV vectors. While the

BBB is a critical interface between the brain and the vasculature, it is not the only way that a virus could

enter the brain from the bloodstream. It’s also possible that AAV could gain access to the brain at regions

without a strong BBB, at circumventricular organs such as the subfornical organ and the medial regions

of the hippocampus. Here, AAV may travel through the extracellular spaces and the interstitial fluid.

Furthermore, while we have demonstrated that the 1RX footprint affords efficient BBB transport, we

cannot rule out the possibility that AAV1RX may be capable of crossing the blood-cerebrospinal fluid

barrier as well (BCSFB), although our data shows low transduction of the choroid plexus and no

indication of involvement with a CSF intermediate for AAV1RX. The BCSFB predominantly consists of

a monolayer of choroid plexus epithelial cells, which separates the two fluids. The choroid plexus is the

91

brain region which generates CSF, and leaky capillaries within the choroid plexus are involved in

regulating the transport of substances from the bloodstream into the CSF298. We observed robust

transduction of choroid plexus epithelial cells mediated by AAVrh.10, and to a lesser degree, AAVRX1

(and to an even lesser degree AAV1-N447S). None of the other capsid variants showed a similar capacity

to transduce the choroid plexus, suggesting that capsid-SIA interactions do not mediate these phenotypes.

These observations are further corroborated by other studies showing similar transduction of the choroid

plexus for AAVrh.10 and transduction gradients at the periventricular spaces98,254.

Our preliminary data (Fig. 4.1) shows that AAVrh.10 and AAVRX1 (as well as AAV1-N447S,

to a lesser degree consistently mediate a gradient-like transduction pattern in specific areas of the brain

located nearby the CSF-containing ventricles, such as at the periphery of the somatosensory cortex

(decreasing in density of GFP+ cells as you move away from the periphery, from cortical neuron layers 1-

5). We observe a similar transduction gradient at the paraventricular nucleus of the thalamus and the

medial habenula of the hypothalamus, with a high density of GFP+ cells present in these regions adjacent

to the third ventricle, with the density of GFP+ cells showing a sharp decrease as you move away from

the third ventricle and into the mediodorsal, intermediodorsal, and central lateral nuclei of the thalamus as

well as the lateral habenula (Fig. 4.1). Together, these data indicate that AAVrh.10 may be capable of

utilizing a CSF intermediate in order to achieve its robust global CNS transduction. Whether AAVrh.10 is

able to cross both the BBB as well as the blood-CSF barrier (and the CSF-brain barrier) and/or is

transported from the parenchyma into the CSF and undergoes CSF-mediated spread in the brain in order

to transduce the CNS are currently unknown and should be the focus of future studies298.

Expanding our understanding of the relationship between capsid-glycan interactions and both

BBB transport and CNS transduction. In Chapter 3, we demonstrated an inverse correlation between the

dependence of capsids on SIA for binding and transduction and the ability to cross the BBB. Furthermore,

we demonstrated a direct correlation between SIA dependence and both vascular and liver

transduction/sequestration, although we observed some deviation observed for this trend. Nonetheless,

future directions should expand on this analysis, through methods such as characterizing additional capsid

92

mutants across the SIA sensitivity spectrum, in order to strengthen this evidence. Additionally, more

comprehensive biochemical analyses, such as saturation binding experiments and surface plasmon

resonance should be carried out in order to more precisely determine binding affinities and correlate these

with BBB transport and CNS transduction.

A further hypothesis is that the 1RX footprint may mediate other interactions which depend upon

involvement of SIA interactions in the context of AAV1. Thus, it is possible that glycan interactions for

other capsids may play either a synergistic or antagonistic role in enabling a similar CNS phenotype.

Thus, it is possible that the 1RX footprint may not work at all for some serotypes, or that it may only

work for some serotypes in the context of altered (i.e. attenuated) glycan interactions. As mentioned

before, a sulfated N-acetyllactosamine (LacNac) has recently been identified as the receptor for

AAVrh.10, and we speculate (in chapter 3) that interfering with these interactions attenuates BBB

transport and/or CNS transduction. Similar studies focused on the AAVrh.10 LacNac interaction may

help further our understanding of the neurotropic properties of AAVrh.10 within the context of BBB

traversal and biology in the CNS as well as for peripheral tissues such as the liver.

Future studies should be focused on illuminating these relationships, and should also focus on

determining to what extent this Goldilocks model for capsid-SIA interactions can be extended to fine-tune

various other capsid-glycan interactions, such as those for AAV9 and galactose, AAV2 and HS, and

AAVrh.10 and LacNac. It is tempting to speculate on the possibility of generating a matrix across the

spectrum of various capsid-glycan receptor pairs in order to find similar goldilocks zones for BBB

transport, specific tropisms, transduction profiles, or any other phenotype of interest.

Effect of sialic acid linkage and species-specific glycosylation patterns on blood-brain barrier

transport and central nervous system transduction. Multiple other variables, such as the different forms,

linkages, and patterns of glycans add a further layer of complexity to our model for capsid glycan

interactions. All of these variables are likely to play a role in determining these CNS phenotypes and

further studies will be needed to shed light on these complex carbohydrate interactions. Excluding other

capsid-glycan pairs, this complexity can be appreciated looking just at SIA. For example, what are the

93

effects of altering the preference of a given capsid for either specific linkages of SIA (i.e. α-2,3- or α-2,6-

N-linkages) or specific sugar content of SIA (i.e. skewing preference towards either the mammalian

Neu5GC or the human-specific Neu5AC variant of SIA)? For example, preliminary data from our lab has

shown evidence of differential transduction profiles for AAV vectors in a mouse model of human

glycosylation. Specifically, we observed decreased liver transduction for AAV6 and muscle transduction

for AAV1, corroborated by decreased in vitro transduction for both following metabolic incorporation of

the human variant Neu5AC (unpublished data). These phenomena should be investigated in future

studies which could take advantage of the well-characterized CMAH (cytidine monophosphate-N-

acetylneuraminic acid hydroxylase) knockout mouse model which recapitulates human glycosylation

patterns. Furthermore, can these specific interactions be manipulated in order to modulate certain

phenotypes? And does AAV1RX demonstrate skewed preferences for SIA identity and/or linkage?

Identification of potential host factor interactions for the 1RX footprint. In proposing our

“Goldilocks” model to describe the effect of capsid-SIA interactions, we have provide a partial

explanation for the mechanism by which the 1RX footprint affords BBB traversal and CNS transduction;

however, other contributing factors likely also play a role. We suspect that BBB transport is mediated

through interactions between AAV1RX and currently unknown host factors, such as soluble proteins in

circulation or alternatively recognition of (likely glycosylated) cell surface transmembrane proteins. For

instance, does AAV1 bind a sialylated surface protein on brain microvascular endothelial cells which

shuttles the virion down a transduction route, whereas perhaps AAVrh.10 and AAV1RX bind a

transmembrane protein receptor which destines the virion for transcytosis across the vasculature?

Furthermore, it is possible that the 1RX footprint mediates an interaction that is entirely separate from

those used by either parent. Similarly, what role, if any is played by secondary receptors, i.e. the recently

discovered AAV receptor (AAVR), integrins and various co-receptor interactions in effecting BBB

transport and CNS transduction?

Future studies should focus on identifying interaction partners for the 1RX footprint in order to

elucidate the mechanism(s) that result in its CNS phenotype. For instance, experiments could be

94

performed involving immunoprecipitation to pull-down host factors which bind the AAV1RX capsid

exclusively versus control (AAV1/rh.10) capsids followed by mass-spectrometry to identity these factors.

For example, these types of experiments could be performed in vitro on primary cultures of brain

microvascular endothelial cells and potentially in vivo as well (i.e. using brain lysates). Additionally, pull-

down studies such as these could be carried out to identify host factors interacting exclusively with the

1RX motif in serum or CSF samples. Along these lines, I have performed preliminary pulldown

experiments using beads attached to lectins which bind either sialylated or galactosylated glycoproteins in

murine CSF samples, to discover potential binding partners for either SIA recognizing capsids (i.e.

AAV1/4/5/6/1RX) or GAL recognizing capsids (i.e. AAV9). This pull-down was followed up by mass

spectrometry and the analysis of resulting peptides (based on number of reads) identified lactotransferrin

and SPARC-like protein 1 (SPARCL1) proteins as potential binding partners to be validated in future

experiments. Future studies such as these may discover a receptor involved in transvascular transport

processes, such as the well-characterized transferrin receptor or alternatively discover that 1RX may bind

a soluble serum factor which facilitates uptake and transport across the vasculature, such as transferrin.

Concluding remarks. Overall, the means by which AAV1RX and naturally neurotropic AAV

capsids engage the neurovascular unit at the BBB and gain entry into the brain parenchyma following

systemic administration remains elusive. The mechanism(s) responsible and many unknown variables

remain to be resolved - i.e. unidentified capsid-host factor interactions, the possibility of capsid-dependent

methods, and the relative roles of capsid-glycan interactions. Whether AAV capsids are capable of

employing routes other than BBB transport, such as transport through the BCSFB, the choroid plexus,

and/or circumventricular organs with a more permeable BBB, as well as the influence of glymphatics,

must all be resolved through future studies. Similarly, the determinants of CNS tropism at the tissue,

cellular, and subcellular population (i.e. different neuronal populations) levels remain to be discovered.

In this dissertation, we focused on expanding our understanding of structure-function correlates

on the AAV capsid for BBB transport. We did this in chapter 2 by using high throughput library-based

methods and an in vivo screen combined with structural analyses to map a structural footprint from

95

AAVrh.10 that is important in BBB traversal. Through providing this structural insight, we rationally

engineered the neurotropic AAV1RX capsid, which is detargeted from peripheral tissues such as the liver,

and serves as an idyllic vector for achieving therapeutic gene transfer to the CNS through systemic

administration. Hopefully, future preclinical studies will be carried out in large animal models, such as

non-human primates, to validate the safety and efficacy of AAV1RX as a gene therapy vector.

Subsequently performing a comparative analysis of capsids with differential dependencies on

SIA, in chapter 3, I found that the ability to cross the BBB was inversely correlated with the SIA

dependence of a given capsid. These results prompted me to propose a model to describe the influence of

capsid-glycan interactions on BBB transport and CNS transduction. Future studies will hopefully provide

additional evidence to either disprove or confirm this model and possibly expand upon it, especially

within the context of other capsid-glycan interactions.

These findings are a small but important step towards a more complete understanding of the

principal mechanisms and capsid-host interactions which promote AAV transport across the BBB, and

provide clues as to some of the players involved. Other studies to dissect the potential involvement of

other host-factors, and identification of such factors, will be critical in elucidating this mechanism.

Perhaps some of these lessons can be applied to improve transvascular transport and obtain higher

transduction levels for other capsids in other tissues, such as skeletal muscle or cardiac tissues.

Engineering improved gene therapy vectors to treat neurological diseases, which can be administered

systemically and target the entire brain parenchyma is critical. Additionally, such efforts require a better

understanding of AAV capsid-host interactions, such as how AAV engages the neurovascular unit.

Therefore, it is imperative that future studies build upon these findings in order to resolve the mystery of

how AAV capsids cross the BBB to transduce the brain parenchyma.

96

Figure 4.1. Transduction of the choroid plexus and paraventricular nucleus of the thalamus

following intravenous administration. Viruses displaying differential dependencies on sialic acid were

produced packaging a CBh-scGFP transgene and were administered to mice at a dose of 5 x 1011 viral

genomes via tail vein injection. Brain sections were immunostained for GFP expression 21 days later.

Images from scans show transduction of the choroid plexus/paraventricular nucleus of the thalamus for

each variant (organized as higher to lower dependency on sialic acid, from left to right: AAV1,

AAV1RX, AAV1-W503A, AAV1-N447S, AAVRX1, and AAVrh.10). Scale bars represent 100 μm.

Images are representative of n > or = 3.

97

REFERENCES

1. Lee, B. Tao of Jeet Kune Do. (Ohara Publications, 1975).

2. Atchison, R. W., Casto, B. C. & Hammon, W. M. Adenovirus-associated defective virus particle.

Science (1965). doi:10.1126/science.149.3685.754

3. Hoggan, M. D., Blacklow, N. R. & Rowe, W. P. Studies of small DNA viruses found in various

adenovirus preparations: physical, biological, and immunological characteristics. Proc. Natl. Acad.

Sci. U. S. A. (1966). doi:10.1073/pnas.55.6.1467

4. Rose, J. A., Hoggan, M. D. & Shatkin, A. J. Nucleic acid from an adeno-associated virus:

chemical and physical studies. Proc. Natl. Acad. Sci. (1966).

5. Phillips, J. L., Hegge, J., Wolff, J. A., Samulski, R. J. & Asokan, A. in Methods in molecular

biology (Clifton, N.J.) 709, 141–151 (2011).

6. Geoffroy, M. & Salvetti, A. Helper functions required for wild type and recombinant adeno-

associated virus growth. Curr. Gene Ther. (2005). doi:10.2174/1566523054064977

7. Alazard-Dany, N. et al. Definition of herpes simplex virus type 1 helper activities for adeno-

associated virus early replication events. PLoS Pathog. (2009). doi:10.1371/journal.ppat.1000340

8. Weitzman, M. D. & Linden, R. M. Adeno-associated virus biology. Methods in Molecular Biology

807, 1–23 (2011).

9. Walker, J. M. Adeno-associated virus methods and protocols. Life Sciences (2009).

doi:10.1007/978-1-62703-239-1_1

10. Tse, L. V, Moller-Tank, S. & Asokan, A. Strategies to circumvent humoral immunity to adeno-

associated viral vectors. Expert Opin. Biol. Ther. (2015). doi:10.1517/14712598.2015.1035645

11. Asokan, A., Schaffer, D. V. & Samulski, R. J. The AAV vector toolkit: Poised at the clinical

crossroads. Molecular Therapy (2012). doi:10.1038/mt.2011.287

12. Wu, Z., Asokan, A. & Samulski, R. J. Adeno-associated Virus Serotypes: Vector Toolkit for

Human Gene Therapy. Molecular Therapy (2006). doi:10.1016/j.ymthe.2006.05.009

13. Wu, P. et al. Mutational analysis of the adeno-associated virus type 2 (AAV2) capsid gene and

construction of AAV2 vectors with altered tropism. J. Virol. 74, 8635–47 (2000).

14. Xiao, X., Xiao, W., Li, J. & Samulski, R. J. A novel 165-base-pair terminal repeat sequence is the

sole cis requirement for the adeno-associated virus life cycle. J. Virol. (1997).

15. Xiao, X., Li, J. & Samulski, R. J. Production of high-titer recombinant adeno-associated virus

vectors in the absence of helper adenovirus. J. Virol. (1998). doi:10.1073/pnas.1201800109

16. Srivastava, A., Lusby, E. W. & Berns, K. I. Nucleotide sequence and organization of the adeno-

associated virus 2 genome. J. Virol. (1983).

98

17. Cassinotti, P., Weitzand, M. & Tratschin, J. D. Organization of the adeno-associated virus (AAV)

capsid gene: Mapping of a minor spliced mRNA coding for virus capsid protein. Virology (1988).

doi:10.1016/0042-6822(88)90067-0

18. Green, M. R. & Roeder, R. G. Transcripts of the adeno-associated virus genome: mapping of the

major RNAs. J. Virol. (1980).

19. Agbandje-McKenna, M. & Kleinschmidt, J. AAV capsid structure and cell interactions. Methods

Mol. Biol. 807, 47–92 (2011).

20. Van Vliet, K. M., Blouin, V., Brument, N., Agbandje-McKenna, M. & Snyder, R. O. in Methods

in molecular biology (Clifton, N.J.) (2008). doi:10.1007/978-1-59745-210-6_2

21. Agbandje-McKenna, M. & Kleinschmidt, J. in Methods in molecular biology (Clifton, N.J.) 807,

47–92 (2012).

22. Berry, G. E. & Asokan, A. Cellular transduction mechanisms of adeno-associated viral vectors.

Curr. Opin. Virol. 21, 54–60 (2016).

23. Girod, A. et al. The VP1 capsid protein of adeno-associated virus type 2 is carrying a

phospholipase A2 domain required for virus infectivity. J. Gen. Virol. (2002). doi:10.1099/0022-

1317-83-5-973

24. Stahnke, S. et al. Intrinsic phospholipase A2 activity of adeno-associated virus is involved in

endosomal escape of incoming particles. Virology (2011). doi:10.1016/j.virol.2010.09.025

25. Sonntag, F., Schmidt, K. & Kleinschmidt, J. A. A viral assembly factor promotes AAV2 capsid

formation in the nucleolus. Proc. Natl. Acad. Sci. (2010). doi:10.1073/pnas.1001673107

26. Maurer, A. C. et al. The Assembly-Activating Protein Promotes Stability and Interactions between

AAV’s Viral Proteins to Nucleate Capsid Assembly. Cell Rep. (2018).

doi:10.1016/j.celrep.2018.04.026

27. Sonntag, F. et al. The Assembly-Activating Protein Promotes Capsid Assembly of Different

Adeno-Associated Virus Serotypes. J. Virol. (2011). doi:10.1128/JVI.05359-11

28. Tse, L. V, Moller-Tank, S., Meganck, R. M. & Asokan, A. Mapping and Engineering Functional

Domains of the Assembly Activating Protein of Adeno-Associated Viruses. J. Virol. (2018).

doi:10.1128/JVI.00393-18

29. McCarty, D. M. et al. Adeno-associated virus terminal repeat (TR) mutant generates self-

complementary vectors to overcome the rate-limiting step to transduction in vivo. Gene Ther.

(2003). doi:10.1038/sj.gt.3302134

30. McCarty, D. M., Monahan, P. E. & Samulski, R. J. Self-complementary recombinant adeno-

associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis.

Gene Ther. (2001). doi:10.1038/sj.gt.3301514

31. Pereira, D. J., McCarty, D. M. & Muzyczka, N. The adeno-associated virus (AAV) Rep protein

acts as both a repressor and an activator to regulate AAV transcription during a productive

infection. J Virol (1997).

99

32. McCarty, D. M., Ryan, J. H., Zolotukhin, S., Zhou, X. & Muzyczka, N. Interaction of the adeno-

associated virus Rep protein with a sequence within the A palindrome of the viral terminal repeat.

J. Virol. (1994).

33. McCarty, D. M. et al. Identification of linear DNA sequences that specifically bind the adeno-

associated virus Rep protein. J Virol (1994).

34. Gonçalves, M. A. et al. Adeno-associated virus: from defective virus to effective vector. Virol. J.

(2005). doi:10.1186/1743-422X-2-43

35. Nayak, R. & Pintel, D. J. Adeno-Associated Viruses Can Induce Phosphorylation of eIF2 via

PKR Activation, Which Can Be Overcome by Helper Adenovirus Type 5 Virus-Associated RNA.

J. Virol. (2007). doi:10.1128/JVI.01132-07

36. Huang, L.-Y., Halder, S. & Agbandje-McKenna, M. Parvovirus glycan interactions. Curr. Opin.

Virol. 7, 108–118 (2014).

37. Bartlett, J. S., Wilcher, R. & Samulski, R. J. Infectious Entry Pathway of Adeno-Associated Virus

and Adeno-Associated Virus Vectors. J. Virol. (2000). doi:10.1128/JVI.74.6.2777-2785.2000

38. Madigan, V. J. & Asokan, A. Engineering AAV receptor footprints for gene therapy. Curr. Opin.

Virol. 18, 89–96 (2016).

39. Ling, C. et al. Human hepatocyte growth factor receptor is a cellular coreceptor for adeno-

associated virus serotype 3. Hum. Gene Ther. 21, 1741–7 (2010).

40. Kashiwakura, Y. et al. Hepatocyte Growth Factor Receptor Is a Coreceptor for Adeno-Associated

Virus Type 2 Infection. J. Virol. 79, 609–614 (2005).

41. Weller, M. L. et al. Epidermal growth factor receptor is a co-receptor for adeno-associated virus

serotype 6. Nat. Med. 16, 662–664 (2010).

42. Di Pasquale, G. et al. Identification of PDGFR as a receptor for AAV-5 transduction. Nat. Med. 9,

1306–12 (2003).

43. Summerford, C., Bartlett, J. S. & Samulski, R. J. AlphaVbeta5 integrin: a co-receptor for adeno-

associated virus type 2 infection. Nat. Med. 5, 78–82 (1999).

44. Asokan, A., Hamra, J. B., Govindasamy, L., Agbandje-McKenna, M. & Samulski, R. J. Adeno-

associated virus type 2 contains an integrin alpha5beta1 binding domain essential for viral cell

entry. J. Virol. 80, 8961–9 (2006).

45. Kaminsky, P. M., Keiser, N. W., Yan, Z., Lei-Butters, D. C. M. & Engelhardt, J. F. Directing

integrin-linked endocytosis of recombinant AAV enhances productive FAK-dependent

transduction. Mol. Ther. 20, 972–83 (2012).

46. Shen, S. et al. Functional analysis of the putative integrin recognition motif on adeno-associated

virus 9. J. Biol. Chem. 290, 1496–504 (2015).

47. Venkatakrishnan, B. et al. Structure and Dynamics of Adeno-Associated Virus Serotype 1 VP1-

Unique N-Terminal Domain and Its Role in Capsid Trafficking. J. Virol. 87, 4974–4984 (2013).

100

48. Grieger, J. C., Snowdy, S. & Samulski, R. J. Separate Basic Region Motifs within the Adeno-

Associated Virus Capsid Proteins Are Essential for Infectivity and Assembly. J. Virol. 80, 5199–

5210 (2006).

49. Nonnenmacher, M. & Weber, T. Intracellular transport of recombinant adeno-associated virus

vectors. Gene Therapy (2012). doi:10.1038/gt.2012.6

50. Penaud-Budloo, M. et al. Adeno-Associated Virus Vector Genomes Persist as Episomal

Chromatin in Primate Muscle. J. Virol. (2008). doi:10.1128/JVI.00649-08

51. Samulski, R. J. et al. Targeted integration of adeno-associated virus (AAV) into human

chromosome 19. EMBO J. (1991). doi:10.1523/JNEUROSCI.4323-08.2008

52. Linden, R. M., Ward, P., Giraud, C., Winocour, E. & Berns, K. I. Site-specific integration by

adeno-associated virus. Proc Natl Acad Sci U S A (1996). doi:10.1073/pnas.93.21.11288

53. McCarty, D. M., Young, S. M. & Samulski, R. J. Integration of Adeno-Associated Virus (AAV)

and Recombinant AAV Vectors. Annu. Rev. Genet. (2004).

doi:10.1146/annurev.genet.37.110801.143717

54. Sonntag, F., Bleker, S., Leuchs, B., Fischer, R. & Kleinschmidt, J. A. Adeno-associated virus type

2 capsids with externalized VP1/VP2 trafficking domains are generated prior to passage through

the cytoplasm and are maintained until uncoating occurs in the nucleus. J. Virol. 80, 11040–11054

(2006).

55. Berns, K. & Parrish, C. in Knipe DM, Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B,

Straus SE (ed), Fields virology, 5th ed, vol II. 2437–2477 (Lippincott Williams & Wilkins, 2007).

56. Huang, L. Y., Halder, S. & Agbandje-McKenna, M. Parvovirus glycan interactions. Curr. Opin.

Virol. 7C, 108–118 (2014).

57. Padron, E. et al. Structure of Adeno-Associated Virus Type 4. J. Virol. (2005).

doi:10.1128/JVI.79.8.5047

58. DiPrimio, N., Asokan, A., Govindasamy, L., Agbandje-McKenna, M. & Samulski, R. J. Surface

Loop Dynamics in Adeno-Associated Virus Capsid Assembly. J. Virol. 82, 5178–5189 (2008).

59. Nam, H.-J. et al. Structure of Adeno-Associated Virus Serotype 8, a Gene Therapy Vector. J.

Virol. 81, 12260–12271 (2007).

60. Padron, E. et al. Structure of Adeno-Associated Virus Type 4. J. Virol. 79, 5047–5058 (2005).

61. Padron, E. et al. Structure of adeno-associated virus type 4. J. Virol. (2005).

doi:10.1128/JVI.79.8.5047-5058.2005

62. Xie, Q., Lerch, T. F., Meyer, N. L. & Chapman, M. S. Structure-function analysis of receptor-

binding in adeno-associated virus serotype 6 (AAV-6). Virology (2011).

doi:10.1016/j.virol.2011.08.011

101

63. Lerch, T. F., Xie, Q. & Chapman, M. S. The structure of adeno-associated virus serotype 3B

(AAV-3B): Insights into receptor binding and immune evasion. Virology (2010).

doi:10.1016/j.virol.2010.03.027

64. Kern, A. et al. Identification of a Heparin-Binding Motif on Adeno-Associated Virus Type 2

Capsids Identification of a Heparin-Binding Motif on Adeno-Associated Virus Type 2 Capsids. J.

Virol. (2003). doi:10.1128/JVI.77.20.11072

65. DiMattia, M. A. et al. Structural Insight into the Unique Properties of Adeno-Associated Virus

Serotype 9. J. Virol. (2012). doi:10.1128/JVI.07232-11

66. Halder, S. et al. Structure of neurotropic adeno-associated virus AAVrh.8. J. Struct. Biol. (2015).

doi:10.1016/j.jsb.2015.08.017

67. Kern, A. et al. Identification of a Heparin-Binding Motif on Adeno-Associated Virus Type 2

Capsids. J. Virol. (2003). doi:10.1128/JVI.77.20.11072-11081.2003

68. Ng, R. et al. Structural Characterization of the Dual Glycan Binding Adeno-Associated Virus

Serotype 6. J. Virol. (2010). doi:10.1128/JVI.01235-10

69. DiMattia, M. A. et al. Structural insight into the unique properties of adeno-associated virus

serotype 9. J Virol (2012). doi:10.1128/JVI.07232-11

70. Kronenberg, S., Kleinschmidt, J. A. & Böttcher, B. Electron cryo-microscopy and image

reconstruction of adeno-associated virus type 2 empty capsids. EMBO Rep. (2001).

doi:10.1093/embo-reports/kve234

71. Govindasamy, L. et al. Structural insights into adeno-associated virus serotype 5. J. Virol. (2013).

doi:10.1128/JVI.00867-13

72. Huang, L.-Y., Halder, S. & Agbandje-McKenna, M. Parvovirus glycan interactions. Curr. Opin.

Virol. 7, 108–118 (2014).

73. Girod, A. et al. Genetic capsid modifications allow efficient re-targeting of adeno- associated virus

type 2. Nat. Med. (1999). doi:10.1038/12491

74. Padron, E. et al. Structure of adeno-associated virus type 4. J. Virol. (2005).

doi:10.1128/JVI.79.8.5047-5058.2005

75. Chapman, M. S. & Agbandja-McKenna, M. Atomic structure of viral particles. Parvoviruses

(2006).

76. Bleker, S., Sonntag, F. & Kleinschmidt, J. A. Mutational analysis of narrow pores at the fivefold

symmetry axes of adeno-associated virus type 2 capsids reveals a dual role in genome packaging

and activation of phospholipase A2 activity. J. Virol. 79, 2528–2540 (2005).

77. Raupp, C. et al. The Threefold Protrusions of Adeno-Associated Virus Type 8 Are Involved in

Cell Surface Targeting as Well as Postattachment Processing. J. Virol. (2012).

doi:10.1128/JVI.00209-12

102

78. Shanker, S. et al. Structural features of glycan recognition among viral pathogens. Curr. Opin.

Struct. Biol. 44, 211–218 (2017).

79. Walters, R. W. et al. Structure of adeno-associated virus serotype 5. J. Virol. (2004).

doi:10.1128/jvi.78.7.3361-3371.2004

80. James, J. A. et al. Crystal structure of the SF3 helicase from adeno-associated virus type 2.

Structure (2003). doi:10.1016/S0969-2126(03)00152-7

81. Shen, S., Bryant, K. D., Brown, S. M., Randell, S. H. & Asokan, A. Terminal n-linked galactose is

the primary receptor for adeno-associated virus. J. Biol. Chem. 286, 13532–13540 (2011).

82. Bell, C. L., Gurda, B. L., Van Vliet, K., Agbandje-McKenna, M. & Wilson, J. M. Identification of

the Galactose Binding Domain of the Adeno-Associated Virus Serotype 9 Capsid. J. Virol. (2012).

doi:10.1128/JVI.00448-12

83. Afione, S. et al. Identification and Mutagenesis of the Adeno-Associated Virus 5 Sialic Acid

Binding Region. J. Virol. (2015). doi:10.1128/JVI.02503-14

84. Walters, R. W. et al. Binding of Adeno-associated Virus Type 5 to 2,3-Linked Sialic Acid is

Required for Gene Transfer. J. Biol. Chem. (2001). doi:10.1074/jbc.M101559200

85. Kaludov, N., Brown, K. E., Walters, R. W., Zabner, J. & Chiorini, J. A. Adeno-associated virus

serotype 4 (AAV4) and AAV5 both require sialic acid binding for hemagglutination and efficient

transduction but differ in sialic acid linkage specificity. J. Virol. 75, 6884–93 (2001).

86. Bleker, S., Pawlita, M. & Kleinschmidt, J. A. Impact of Capsid Conformation and Rep-Capsid

Interactions on Adeno-Associated Virus Type 2 Genome Packaging. J. Virol. (2006).

doi:10.1128/JVI.80.2.810

87. Gao, G. et al. Adeno-associated viruses undergo substantial evolution in primates during natural

infections. Proc. Natl. Acad. Sci. (2003). doi:10.1073/pnas.0937739100

88. Gao, G. et al. Clades of Adeno-Associated Viruses Are Widely Disseminated in Human Tissues.

J. Virol. (2004). doi:10.1128/JVI.78.12.6381-6388.2004

89. Gao, G., Vandenberghe, L. H. & Wilson, J. M. New recombinant serotypes of AAV vectors. Curr.

Gene Ther. (2005). doi:10.2174/1566523054065057

90. Hastie, E. & Samulski, R. J. Adeno-Associated Virus at 50: A Golden Anniversary of Discovery,

Research, and Gene Therapy Success—A Personal Perspective. Hum. Gene Ther. (2015).

doi:10.1089/hum.2015.025

91. Bossis, I. & Chiorini, J. A. Cloning of an avian adeno-associated virus (AAAV) and generation of

recombinant AAAV particles. J. Virol. 77, 6799–810 (2003).

92. Farkas, S. L. et al. A parvovirus isolated from royal python (Phyton regius) is a member of the

genus Dependovirus. J. Gen. Virol. (2004). doi:10.1099/vir.0.19616-0

93. Chiorini, J. A., Kim, F., Yang, L. & Kotin, R. M. Cloning and characterization of adeno-associated

virus type 5. J. Virol. (1999).

103

94. Gao, G., Vandenberghe, L. & Wilson, J. New Recombinant Serotypes of AAV Vectors. Curr.

Gene Ther. 5, 285–297 (2005).

95. Mori, S., Wang, L., Takeuchi, T. & Kanda, T. Two novel adeno-associated viruses from

cynomolgus monkey: pseudotyping characterization of capsid protein. Virology 330, 375–83

(2004).

96. Govindasamy, L. et al. Structurally Mapping the Diverse Phenotype of Adeno-Associated Virus

Serotype 4. J. Virol. (2006). doi:10.1128/JVI.01536-06

97. Asokan, A. & Samulski, R. J. An emerging adeno-associated viral vector pipeline for cardiac gene

therapy. Hum. Gene Ther. 24, 906–13 (2013).

98. Zhang, H. et al. Several rAAV Vectors Efficiently Cross the Blood–brain Barrier and Transduce

Neurons and Astrocytes in the Neonatal Mouse Central Nervous System. Mol. Ther. 19, 1440–

1448 (2011).

99. Bartlett, J. S., Samulski, R. J. & McCown, T. J. Selective and Rapid Uptake of Adeno-Associated

Virus Type 2 in Brain. Hum. Gene Ther. (1998). doi:10.1089/hum.1998.9.8-1181

100. Hauck, B. & Xiao, W. Characterization of tissue tropism determinants of adeno-associated virus

type 1. J. Virol. (2003). doi:10.1128/JVI.77.4.2768

101. Mount, J. D. et al. Sustained phenotypic correction of hemophilia B dogs with a factor IX null

mutation by liver-directed gene therapy. Blood (2002). doi:10.1182/blood.V99.8.2670

102. Ferreira, V., Petry, H. & Salmon, F. Immune responses to AAV-vectors, The Glybera example

from bench to bedside. Front. Immunol. (2014). doi:10.3389/fimmu.2014.00082

103. Gaudet, D., Méthot, J. & Kastelein, J. Gene therapy for lipoprotein lipase deficiency. Current

Opinion in Lipidology (2012). doi:10.1097/MOL.0b013e3283555a7e

104. Davidson, B. L. et al. Recombinant adeno-associated virus type 2, 4, and 5 vectors: transduction of

variant cell types and regions in the mammalian central nervous system. Proc. Natl. Acad. Sci. U.

S. A. 97, 3428–32 (2000).

105. Burger, C. et al. Comparative analysis of adeno-associated viral vector serotypes 1, 2, 5, 7, and 8

in mouse brain. Gene Ther. (2004). doi:10.1038/sj.gt.3302815

106. Pang, J. jing et al. Comparative analysis of in vivo and in vitro AAV vector transduction in the

neonatal mouse retina: Effects of serotype and site of administration. Vision Res. (2008).

doi:10.1016/j.visres.2007.08.009

107. Ling, C. et al. Human Hepatocyte Growth Factor Receptor Is a Cellular Coreceptor for Adeno-

Associated Virus Serotype 3. Hum. Gene Ther. 21, 1741–1747 (2010).

108. Vercauteren, K. et al. Superior in vivo Transduction of Human Hepatocytes Using Engineered

AAV3 Capsid. Mol. Ther. (2016). doi:10.1038/mt.2016.61

109. Wang, Z. et al. Adeno-associated virus serotype 8 efficiently delivers genes to muscle and heart.

Nat. Biotechnol. (2005). doi:10.1038/nbt1073

104

110. Sands, M. S. AAV-mediated liver-directed gene therapy. Methods Mol. Biol. (2011).

doi:10.1007/978-1-61779-370-7_6

111. Y., W. et al. Transduction of primary human hepatocytes in vitro and in humanized murine livers

in vivo by recombinant AAV3 vectors. Molecular Therapy (2014).

112. Zincarelli, C., Soltys, S., Rengo, G. & Rabinowitz, J. E. Analysis of AAV serotypes 1-9 mediated

gene expression and tropism in mice after systemic injection. Mol. Ther. (2008).

doi:10.1038/mt.2008.76

113. Shen, S., Troupes, A. N., Pulicherla, N. & Asokan, A. Multiple Roles for Sialylated Glycans in

Determining the Cardiopulmonary Tropism of Adeno-Associated Virus 4. J. Virol. 87, 13206–

13213 (2013).

114. Aschauer, D. F., Kreuz, S. & Rumpel, S. Analysis of Transduction Efficiency, Tropism and

Axonal Transport of AAV Serotypes 1, 2, 5, 6, 8 and 9 in the Mouse Brain. PLoS One (2013).

doi:10.1371/journal.pone.0076310

115. Saraiva, J., Nobre, R. J. & Pereira de Almeida, L. Gene therapy for the CNS using AAVs: The

impact of systemic delivery by AAV9. Journal of Controlled Release (2016).

doi:10.1016/j.jconrel.2016.09.011

116. Opie, S. R., Warrington, K. H., Agbandje-McKenna, M., Zolotukhin, S. & Muzyczka, N.

Identification of amino acid residues in the capsid proteins of adeno-associated virus type 2 that

contribute to heparan sulfate proteoglycan binding. J. Virol. (2003). doi:10.1128/JVI.77.12.6995

117. Huang, L.-Y. et al. Characterization of the Adeno-Associated Virus 1 and 6 Sialic Acid Binding

Site. doi:10.1128/JVI.00161-16

118. Wu, Z., Miller, E., Agbandje-McKenna, M. & Samulski, R. J. Alpha2,3 and alpha2,6 N-linked

sialic acids facilitate efficient binding and transduction by adeno-associated virus types 1 and 6. J.

Virol. 80, 9093–103 (2006).

119. Hahm, H. S. et al. Automated Glycan Assembly of Oligo-N-Acetyllactosamine and Keratan

Sulfate Probes to Study Virus-Glycan Interactions. Chem (2017).

doi:10.1016/j.chempr.2016.12.004

120. Schmidt, M. & Chiorini, J. A. Gangliosides are essential for bovine adeno-associated virus entry. J

Virol (2006). doi:10.1128/JVI.02393-05

121. Schmidt, M., Katano, H., Bossis, I. & Chiorini, J. A. Cloning and characterization of a bovine

adeno-associated virus. J. Virol. (2004). doi:10.1128/JVI.78.12.6509-6516.2004

122. Lerch, T. F. & Chapman, M. S. Identification of the heparin binding site on adeno-associated virus

serotype 3B (AAV-3B). Virology (2012). doi:10.1016/j.virol.2011.10.007

123. Ströh, L. J. & Stehle, T. Glycan Engagement by Viruses: Receptor Switches and Specificity. Annu.

Rev. Virol. 1, 285–306 (2014).

105

124. Messina, E. L. et al. Adeno-Associated Viral Vectors Based on Serotype 3b Use Components of

the Fibroblast Growth Factor Receptor Signaling Complex for Efficient Transduction. Hum. Gene

Ther. (2012). doi:10.1089/hum.2012.066

125. Akache, B. et al. The 37/67-Kilodalton Laminin Receptor Is a Receptor for Adeno-Associated

Virus Serotypes 8, 2, 3, and 9. J. Virol. (2006). doi:10.1128/JVI.00878-06

126. Pillay, S. et al. An essential receptor for adeno-associated virus infection. Nature (2016).

doi:10.1038/nature16465

127. Viruses, M. A. & Chen, H. Viral Vectors for Gene Therapy. Methods (2011). doi:10.1007/978-1-

61779-095-9

128. Lentz, T. B., Gray, S. J. & Samulski, R. J. Viral vectors for gene delivery to the central nervous

system. Neurobiol. Dis. 48, 179–88 (2012).

129. Robbins, P. D. et al. Viral vectors for gene therapy. Trends Biotechnol. (1998).

doi:10.1016/S0167-7799(97)01137-2

130. Mueller, C. & Flotte, T. R. Clinical gene therapy using recombinant adeno-associated virus

vectors. Gene Therapy (2008). doi:10.1038/gt.2008.68

131. Lehtonen, E. & Tenenbaum, L. Virus Vectors for use in the Central Nervous System: adeno-

associated viral vectors. Int. Rev. Neurobiol. (2003). doi:10.1016/S0074-7742(03)01002-X

132. Engeland, C. E. et al. A Tupaia paramyxovirus vector system for targeting and transgene

expression. J. Gen. Virol. (2017). doi:10.1099/jgv.0.000887

133. Sakuma, T., Barry, M. A. & Ikeda, Y. Lentiviral vectors: basic to translational. Biochem. J.

(2012). doi:10.1042/BJ20120146

134. Yáñez-Muñoz, R. J. et al. Effective gene therapy with nonintegrating lentiviral vectors. Nat. Med.

(2006). doi:10.1038/nm1365

135. Yin, H. et al. Non-viral vectors for gene-based therapy. Nature Reviews Genetics (2014).

doi:10.1038/nrg3763

136. Volpers, C. & Kochanek, S. Adenoviral vectors for gene transfer and therapy. J. Gene Med.

(2004). doi:10.1002/jgm.496

137. Khare, R., Y. Chen, C., A. Weaver, E. & A. Barry, M. Advances and Future Challenges in

Adenoviral Vector Pharmacology and Targeting. Curr. Gene Ther. (2011).

doi:10.2174/156652311796150363

138. Naso, M. F., Tomkowicz, B., Perry, W. L. & Strohl, W. R. Adeno-Associated Virus (AAV) as a

Vector for Gene Therapy. BioDrugs (2017). doi:10.1007/s40259-017-0234-5

139. Daya, S. & Berns, K. I. Gene therapy using adeno-associated virus vectors. Clinical Microbiology

Reviews (2008). doi:10.1128/CMR.00008-08

140. Samulski, R. J. & Muzyczka, N. AAV-Mediated Gene Therapy for Research and Therapeutic

Purposes. Annu. Rev. Virol. 1, 427–51 (2014).

106

141. Grieger, J. C., Choi, V. W. & Samulski, R. J. Production and characterization of adeno-associated

viral vectors. Nature Protocols (2006). doi:10.1038/nprot.2006.207

142. Ginn, S. L., Amaya, A. K., Abedi, M. R., Alexander, I. E. & Edelstein, M. Gene therapy clinical

trials worldwide to 2017 : An update. 1–16 (2018). doi:10.1002/jgm.3015

143. Stroes, E. S. et al. Intramuscular administration of AAV1-lipoprotein lipaseS447Xlowers

triglycerides in lipoprotein lipase-deficient patients. Arterioscler. Thromb. Vasc. Biol. 28, 2303–

2304 (2008).

144. Scott, L. Alipogene Tiparvovec: A Review of Its Use in Adults with Familial Lipoprotein Lipase

Deficiency. Drugs 75, 175–182 (2015).

145. US Food and Drug Administration. FDA Approves Novel Gene Therapy to Treat Patients with a

Rare Form of Inherited Vision Loss. US Food and Drug Administration Newsroom Website: Press

Release Available at:

https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm589467.htm. (Accessed:

21st August 2018)

146. Tirrell, M. "A US drugmaker offers to cure rare blindness for $850,000. CNBC

147. Bowles, D. E. et al. Phase 1 gene therapy for Duchenne muscular dystrophy using a translational

optimized AAV vector. Mol. Ther. 20, 443–455 (2012).

148. Bowles, D. E. et al. Phase 1 Gene Therapy for Duchenne Muscular Dystrophy Using a

Translational Optimized AAV Vector. Mol. Ther. (2012). doi:10.1038/mt.2011.237

149. E.P., R. et al. Gene therapy for wet-AMD: Progress report on a phase I/II clinical trial. Mol. Ther.

(2013). doi:http://dx.doi.org/10.1038/mt.2013.82

150. Chamberlain JR, C. J. Progress toward gene therapy for Duchenne muscular dystrophy. Mol. Ther.

1125–1131 (2017).

151. Aartsma‐Rus A. FDA Approval of nusinersen for spinal muscular atrophy makes 2016 the year of

splice modulating oligonucleotides. Nucleic Acid Ther. 67–69 (2017).

152. Avexis. Pre-Symptomatic Study of Intravenous AVXS-101 in Spinal Muscular Atrophy (SMA)

for Patients With Multiple Copies of SMN2 (SPR1NT). ClinicalTrials.gov Identifier:

NCT03505099 (2018). Available at: https://clinicaltrials.gov/ct2/show/NCT03505099.

153. Avexis. Study of Intrathecal Administration of AVXS-101 for Spinal Muscular Atrophy

(STRONG). NIH ClinicalTrials.gov Identifier: NCT03381729

154. Avexis. ene Replacement Therapy Clinical Trial for Patients With Spinal Muscular Atrophy Type

1 (STR1VE). ClinicalTrials.gov ClinicalTrials.gov Identifier: NCT03306277

155. Loring HS, F. T. Current status of gene therapy for α-1 antitrypsin deficiency. Expert Opin Biol

Ther. 15, 329–336 (2015).

107

156. Voyager Therapeutics. Safety Study of AADC Gene Therapy (VY-AADC01) for Parkinson’s

Disease (AADC). ClinicalTrials.gov ClinicalTrials.gov Identifier: NCT01973543 (2018).

Available at: https://clinicaltrials.gov/ct2/show/NCT01973543?term=AADC.

157. Murlidharan, G., Samulski, R. J. & Asokan, A. Biology of adeno-associated viral vectors in the

central nervous system. Front. Mol. Neurosci. 7, 1–9 (2014).

158. Li, C. et al. Development of Patient-specific AAV Vectors After Neutralizing Antibody Selection

for Enhanced Muscle Gene Transfer. Mol. Ther. 24, 53–65 (2016).

159. Weinberg, M. S., Samulski, R. J. & McCown, T. J. Adeno-associated virus (AAV) gene therapy

for neurological disease. Neuropharmacology (2013). doi:10.1016/j.neuropharm.2012.03.004

160. Ojala, D. S., Amara, D. P. & Schaffer, D. V. Adeno-Associated Virus Vectors and Neurological

Gene Therapy. Neurosci. (2015). doi:10.1177/1073858414521870

161. Samulski, R. J., Berns, K. I., Tan, M. & Muzyczka, N. Cloning of adeno-associated virus into

pBR322: rescue of intact virus from the recombinant plasmid in human cells. Proc. Natl. Acad.

Sci. (1982). doi:10.1073/pnas.79.6.2077

162. Potter, M. et al. A simplified purification protocol for recombinant adeno-associated virus vectors.

Mol. Ther. - Methods Clin. Dev. (2014). doi:10.1038/mtm.2014.34

163. Ayuso, E., Mingozzi, F. & Bosch, F. Production, Purification and Characterization of Adeno-

Associated Vectors. Curr. Gene Ther. (2010). doi:10.2174/156652310793797685

164. Strobel, B., Miller, F. D., Rist, W. & Lamla, T. Comparative Analysis of Cesium Chloride- and

Iodixanol-Based Purification of Recombinant Adeno-Associated Viral Vectors for Preclinical

Applications. Hum. Gene Ther. Methods (2015). doi:10.1089/hgtb.2015.051

165. Wang, Q. et al. Identification of an adeno-associated virus binding epitope for AVB sepharose

affinity resin. Mol. Ther. - Methods Clin. Dev. (2015). doi:10.1038/mtm.2015.40

166. Smith, R. H., Levy, J. R. & Kotin, R. M. A simplified baculovirus-AAV expression vector system

coupled with one-step affinity purification yields high-titer rAAV stocks from insect cells. Mol.

Ther. (2009). doi:10.1038/mt.2009.128

167. Ayuso, E. et al. High AAV vector purity results in serotype- and tissue-independent enhancement

of transduction efficiency. Gene Ther. (2010). doi:10.1038/gt.2009.157

168. Galibert, L. & Merten, O.-W. Latest developments in the large-scale production of adeno-

associated virus vectors in insect cells toward the treatment of neuromuscular diseases. J.

Invertebr. Pathol. (2011). doi:10.1016/j.jip.2011.05.008

169. Kotin, R. M. Large-scale recombinant adeno-associated virus production. Hum. Mol. Genet.

(2011). doi:10.1093/hmg/ddr141

170. Urabe, M., Ding, C. & Kotin, R. M. Insect Cells as a Factory to Produce Adeno-Associated Virus

Type 2 Vectors. Hum. Gene Ther. (2002). doi:10.1089/10430340260355347

108

171. Grimm, D. et al. Titration of AAV-2 particles via a novel capsid ELISA: Packaging of genomes

can limit production of recombinant AAV-2. Gene Ther. (1999). doi:10.1038/sj.gt.3300946

172. McClure, C., Cole, K. L. H., Wulff, P., Klugmann, M. & Murray, A. J. Production and titering of

recombinant adeno-associated viral vectors. J. Vis. Exp. (2011). doi:10.3791/3348

173. Burova, E. & Ioffe, E. Chromatographic purification of recombinant adenoviral and adeno-

associated viral vectors: Methods and implications. Gene Ther. (2005). doi:10.1038/sj.gt.3302611

174. Reed, S. E., Staley, E. M., Mayginnes, J. P., Pintel, D. J. & Tullis, G. E. Transfection of

mammalian cells using linear polyethylenimine is a simple and effective means of producing

recombinant adeno-associated virus vectors. J. Virol. Methods (2006).

doi:10.1016/j.jviromet.2006.07.024

175. al Yacoub, N., Romanowska, M., Haritonova, N. & Foerster, J. Optimized production and

concentration of lentiviral vectors containing large inserts. J. Gene Med. (2007).

doi:10.1002/jgm.1052

176. Grimm, D., Kay, M. A. & Kleinschmidt, J. A. Helper virus-free, optically controllable, and two-

plasmid-based production of adeno-associated virus vectors of serotypes 1 to 6. Mol. Ther. (2003).

doi:10.1016/S1525-0016(03)00095-9

177. Smith, R. H., Yang, L. & Kotin, R. M. Chromatography-based purification of adeno-associated

virus. Methods Mol. Biol. (2008). doi:10.1007/978-1-60327-248-3_4

178. Pulicherla, N. & Asokan, A. Peptide affinity reagents for AAV capsid recognition and

purification. Gene Ther. (2011). doi:10.1038/gt.2011.46

179. GE-Healthcare-Life-Sciences. AVB Sepharose High Performance. Prod. Man. (2007).

180. Calcedo, R., Vandenberghe, L. H., Gao, G., Lin, J. & Wilson, J. M. Worldwide Epidemiology of

Neutralizing Antibodies to Adeno‐Associated Viruses. J. Infect. Dis. (2009). doi:10.1086/595830

181. Calcedo, R. & Wilson, J. M. Humoral Immune Response to AAV. Front. Immunol. (2013).

doi:10.3389/fimmu.2013.00341

182. Tse, L. V. et al. Structure-guided evolution of antigenically distinct adeno-associated virus

variants for immune evasion. Proc. Natl. Acad. Sci. 201704766 (2017).

doi:10.1073/pnas.1704766114

183. Ellsworth, J. L., O’Callaghan, M., Rubin, H. & Seymour, A. Low Seroprevalence of Neutralizing

Antibodies Targeting Two Clade F AAV in Humans. Hum. Gene Ther. Clin. Dev. (2018).

doi:10.1089/humc.2017.239

184. Gray, S. J., Nagabhushan Kalburgi, S., McCown, T. J. & Jude Samulski, R. Global CNS gene

delivery and evasion of anti-AAV-neutralizing antibodies by intrathecal AAV administration in

non-human primates. Gene Ther. (2013). doi:10.1038/gt.2012.101

185. Govindasamy, L. et al. Structural Insights into Adeno-Associated Virus Serotype 5. J. Virol.

(2013). doi:10.1128/JVI.00867-13

109

186. Chamberlain, K., Riyad, J. M. & Weber, T. Expressing Transgenes That Exceed the Packaging

Capacity of Adeno-Associated Virus Capsids. Hum. Gene Ther. Methods (2016).

doi:10.1089/hgtb.2015.140

187. Zhang, Y. & Duan, D. Novel Mini–Dystrophin Gene Dual Adeno-Associated Virus Vectors

Restore Neuronal Nitric Oxide Synthase Expression at the Sarcolemma. Hum. Gene Ther. (2012).

doi:10.1089/hum.2011.131

188. McClements, M. E. & Maclaren, R. E. Adeno-associated virus (AAV) dual vector strategies for

gene therapy encoding large transgenes. Yale Journal of Biology and Medicine (2017).

189. Avexis. Gene Transfer Clinical Trial for Spinal Muscular Atrophy Type 1. NIH -

ClinicalTrials.gov ClinicalTrials.gov Identifier: NCT02122952 Available at:

https://clinicaltrials.gov/ct2/show/NCT02122952.

190. Vandendriessche, T. et al. Efficacy and safety of adeno-associated viral vectors based on serotype

8 and 9 vs. lentiviral vectors for hemophilia B gene therapy. J. Thromb. Haemost. (2007).

doi:10.1111/j.1538-7836.2006.02220.x

191. High, K. Human immune responses in AAV-mediated gene transfer: Implications for safety and

efficacy. Hum. Gene Ther. (2010).

192. Hinderer, C. et al. Severe toxicity in nonhuman primates and piglets following high‐ dose

intravenous administration of an AAV vector expressing human SMN. Hum. Gene Ther. (2018).

doi:10.1089/hum.2018.015

193. Inagaki, K. et al. Robust systemic transduction with AAV9 vectors in mice: efficient global

cardiac gene transfer superior to that of AAV8. Mol. Ther. (2006).

doi:10.1016/j.ymthe.2006.03.014

194. Schuster, D. J. et al. Biodistribution of adeno-associated virus serotype 9 (AAV9) vector after

intrathecal and intravenous delivery in mouse. Front. Neuroanat. (2014).

doi:10.3389/fnana.2014.00042

195. Murlidharan, G. et al. CNS-restricted Transduction and CRISPR/Cas9-mediated Gene Deletion

with an Engineered AAV Vector. Mol. Ther. - Nucleic Acids 5, e338 (2016).

196. Gray, S. J. et al. Optimizing Promoters for Recombinant Adeno-Associated Virus-Mediated Gene

Expression in the Peripheral and Central Nervous System Using Self-Complementary Vectors.

Hum. Gene Ther. (2011). doi:10.1089/hum.2010.245

197. Rincon, M. Y. et al. Widespread transduction of astrocytes and neurons in the mouse central

nervous system after systemic delivery of a self-complementary AAV-PHP.B vector. Gene Ther.

(2018). doi:10.1038/s41434-018-0005-z

198. Bradbury, A. M. et al. AAVrh10 gene therapy ameliorates central and peripheral nervous system

disease in canine globoid cell leukodystrophy (Krabbe disease). Hum. Gene Ther. (2018).

doi:10.1089/hum.2017.151

199. Boustany, R.-M. N. Lysosomal storage diseases—the horizon expands. Nat. Rev. Neurol. (2013).

doi:10.1038/nrneurol.2013.163

110

200. Abeliovich, A. & Gitler, A. D. Defects in trafficking bridge Parkinson’s disease pathology and

genetics. Nature 539, 207–216 (2016).

201. Nedergaard, M. Garbage truck of the brain. Science (2013). doi:10.1126/science.1240514

202. Plog, B. A. & Nedergaard, M. The Glymphatic System in Central Nervous System Health and

Disease: Past, Present, and Future. Annu. Rev. Pathol. Mech. Dis. (2018). doi:10.1146/annurev-

pathol-051217-111018

203. Murlidharan, G., Crowther, A., Reardon, R. A., Song, J. & Asokan, A. Glymphatic fluid transport

controls paravascular clearance of AAV vectors from the brain. JCI Insight 1, 1–11 (2016).

204. Hollis, E. R., Kadoya, K., Hirsch, M., Samulski, R. J. & Tuszynski, M. H. Efficient retrograde

neuronal transduction utilizing self-complementary AAV1. Mol. Ther. (2008).

doi:10.1038/sj.mt.6300367

205. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta

Crystallogr. D. Biol. Crystallogr. 66, 486–501 (2010).

206. Castle, M. J., Gershenson, Z. T., Giles, A. R., Holzbaur, E. L. F. & Wolfe, J. H. Adeno-associated

virus serotypes 1, 8, and 9 share conserved mechanisms for anterograde and retrograde axonal

transport. Hum. Gene Ther. (2014). doi:10.1089/hum.2013.189

207. Tenenbaum, L. et al. Recombinant AAV-mediated gene delivery to the central nervous system.

Journal of Gene Medicine (2004). doi:10.1002/jgm.506

208. Bockstael, O., Foust, K. D., Kaspar, B. & Tenenbaum, L. Recombinant AAV delivery to the

central nervous system. Methods Mol. Biol. (2011). doi:10.1007/978-1-61779-370-7_7

209. Murlidharan, G., Samulski, R. J. & Asokan, A. Biology of adeno-associated viral vectors in the

central nervous system. Front. Mol. Neurosci. (2014). doi:10.3389/fnmol.2014.00076

210. Burger, C. et al. Recombinant AAV viral vectors pseudotyped with viral capsids from serotypes 1,

2, and 5 display differential efficiency and cell tropism after delivery to different regions of the

central nervous system. Mol. Ther. (2004). doi:10.1016/j.ymthe.2004.05.024

211. Cearley, C. N. & Wolfe, J. H. Transduction characteristics of adeno-associated virus vectors

expressing cap serotypes 7, 8, 9, and Rh10 in the mouse brain. Mol. Ther. (2006).

doi:10.1016/j.ymthe.2005.11.015

212. Hinderer, C. et al. Intrathecal gene therapy corrects cns pathology in a feline model of

mucopolysaccharidosis i. Mol. Ther. (2014). doi:10.1038/mt.2014.135

213. Passini, M. A. et al. Translational Fidelity of Intrathecal Delivery of Self-Complementary AAV9–

Survival Motor Neuron 1 for Spinal Muscular Atrophy. Hum. Gene Ther. (2014).

doi:10.1089/hum.2014.011

214. Gray, S. J. Gene therapy and neurodevelopmental disorders. Neuropharmacology (2013).

doi:10.1016/j.neuropharm.2012.06.024

111

215. Glascock, J. J., George, K. L. & Lorson, C. L. Determining the therapeutic window: Central

nervous system administration for vector mediated gene replacement in a severe model of spinal

muscular atrophy. Mol. Ther. (2012). doi:http://dx.doi.org/10.1038/mt.2012.86

216. Rafi, M. A., Rao, H. Z., Luzi, P., Curtis, M. T. & Wenger, D. A. Extended Normal Life After

AAVrh10-mediated Gene Therapy in the Mouse Model of Krabbe Disease. Mol. Ther. 20, 2031–

2042 (2012).

217. Liu, G., Martins, I. H., Chiorini, J. A. & Davidson, B. L. Adeno-associated virus type 4 (AAV4)

targets ependyma and astrocytes in the subventricular zone and RMS. Gene Ther. (2005).

doi:10.1038/sj.gt.3302554

218. Liu, G. Functional Correction of CNS Phenotypes in a Lysosomal Storage Disease Model Using

Adeno-Associated Virus Type 4 Vectors. J. Neurosci. (2005). doi:10.1523/JNEUROSCI.2936-

05.2005

219. Murlidharan, G., Corriher, T., Ghashghaei, H. T. & Asokan, A. Unique Glycan Signatures

Regulate Adeno-Associated Virus Tropism in the Developing Brain. J. Virol. 89, 3976–3987

(2015).

220. Crowther, A. J. et al. An Adeno-Associated Virus-Based Toolkit for Preferential Targeting and

Manipulating Quiescent Neural Stem Cells in the Adult Hippocampus. Stem Cell Reports (2018).

doi:10.1016/j.stemcr.2018.01.018

221. Gray, S. J. et al. Preclinical differences of intravascular aav9 delivery to neurons and glia: A

comparative study of adult mice and nonhuman primates. Mol. Ther. (2011).

doi:10.1038/mt.2011.72

222. Ballabh, P., Braun, A. & Nedergaard, M. The blood-brain barrier: An overview: Structure,

regulation, and clinical implications. Neurobiology of Disease (2004).

doi:10.1016/j.nbd.2003.12.016

223. Zlokovic, B. V. The Blood-Brain Barrier in Health and Chronic Neurodegenerative Disorders.

Neuron (2008). doi:10.1016/j.neuron.2008.01.003

224. Bell, R. D. & Zlokovic, B. V. Neurovascular mechanisms and blood-brain barrier disorder in

Alzheimer’s disease. Acta Neuropathologica (2009). doi:10.1007/s00401-009-0522-3

225. Williams, D. W. et al. Monocytes mediate HIV neuropathogenesis: mechanisms that contribute to

HIV associated neurocognitive disorders. Curr. HIV Res. 12, 85–96 (2014).

226. Salinas, S., Schiavo, G. & Kremer, E. J. A hitchhiker’s guide to the nervous system: the complex

journey of viruses and toxins. Nat. Rev. Microbiol. 8, 645–655 (2010).

227. Gralinski, L. E., Ashley, S. L., Dixon, S. D. & Spindler, K. R. Mouse adenovirus type 1-induced

breakdown of the blood-brain barrier. J. Virol. (2009). doi:10.1128/JVI.00954-09

228. Gray, S. J. et al. Directed evolution of a novel adeno-associated virus (AAV) vector that crosses

the seizure-compromised blood-brain barrier (BBB). Mol. Ther. (2010). doi:10.1038/mt.2009.292

112

229. Foley, C. P. et al. Intra-arterial delivery of AAV vectors to the mouse brain after mannitol

mediated blood brain barrier disruption. J. Control. Release (2014).

doi:10.1016/j.jconrel.2014.09.018

230. Yang, B. et al. Global CNS Transduction of Adult Mice by Intravenously Delivered rAAVrh.8

and rAAVrh.10 and Nonhuman Primates by rAAVrh.10. Mol. Ther. 22, 1299–1309 (2014).

231. Zhong, L. et al. Tyrosine-phosphorylation of AAV2 vectors and its consequences on viral

intracellular trafficking and transgene expression. Virology (2008).

doi:10.1016/j.virol.2008.08.027

232. Rabinowitz, J. E. & Samulski, R. J. Building a better vector: The manipulation of AAV virions.

Virology (2000). doi:10.1006/viro.2000.0707

233. Rabinowitz, J. E., Xiao, W. & Samulski, R. J. Insertional mutagenesis of AAV2 capsid and the

production of recombinant virus. Virology (1999). doi:10.1006/viro.1999.0045

234. Koerber, J. T., Jang, J. H. & Schaffer, D. V. DNA shuffling of adeno-associated virus yields

functionally diverse viral progeny. Mol. Ther. (2008). doi:10.1038/mt.2008.167

235. Kwon, I. & Schaffer, D. V. Designer gene delivery vectors: Molecular engineering and evolution

of adeno-associated viral vectors for enhanced gene transfer. Pharmaceutical Research (2008).

doi:10.1007/s11095-007-9431-0

236. Bartel, M. A., Weinstein, J. R. & Schaffer, D. V. Directed evolution of novel adeno-associated

viruses for therapeutic gene delivery. Gene Therapy (2012). doi:10.1038/gt.2012.20

237. Asokan, A. & Samulski, R. J. AAV does the shuffle. Nature Biotechnology (2006).

doi:10.1038/nbt0206-158

238. Pulicherla, N. et al. Engineering liver-detargeted AAV9 vectors for cardiac and musculoskeletal

gene transfer. Mol. Ther. (2011). doi:10.1038/mt.2011.22

239. Murlidharan, G., Crowther, A., Reardon, R. A., Song, J. & Asokan, A. Glymphatic fluid transport

controls paravascular clearance of AAV vectors from the brain. JCI Insight (2016).

doi:10.1172/jci.insight.88034

240. Choudhury, S. R. et al. In vivo selection yields AAV-B1 capsid for central nervous system and

muscle gene therapy. Mol. Ther. (2016). doi:10.1038/mt.2016.84

241. Choudhury, S. R. et al. Widespread CNS gene transfer and silencing after systemic delivery of

novel AAV-AS vector. Mol. Ther. (2015). doi:10.1038/mt.2015.231

242. Tervo, D. G. R. et al. A Designer AAV Variant Permits Efficient Retrograde Access to Projection

Neurons. Neuron 92, 372–382 (2016).

243. Deverman, B. E. et al. Cre-dependent selection yields AAV variants for widespread gene transfer

to the adult brain. Nat. Biotechnol. (2016). doi:10.1038/nbt.3440

244. Chan, K. Y. et al. Engineered AAVs for efficient noninvasive gene delivery to the central and

peripheral nervous systems. Nat. Neurosci. (2017). doi:10.1038/nn.4593

113

245. Hordeaux, J. et al. The Neurotropic Properties of AAV-PHP.B Are Limited to C57BL/6J Mice.

Mol. Ther. 26, 664–668 (2018).

246. Gao, G. et al. Clades of Adeno-associated viruses are widely disseminated in human tissues. J.

Virol. 78, 6381–8 (2004).

247. Gao, G. et al. Adeno-associated viruses undergo substantial evolution in primates during natural

infections. Proc. Natl. Acad. Sci. U. S. A. 100, 6081–6086 (2003).

248. Madigan, V. J. & Asokan, A. Engineering AAV receptor footprints for gene therapy. Curr. Opin.

Virol. 18, 89–96 (2016).

249. Weller, M. L. et al. Epidermal growth factor receptor is a co-receptor for adeno-associated virus

serotype 6. Nat. Med. 16, 662–4 (2010).

250. Pillay, S. et al. An essential receptor for adeno-associated virus infection. Nature 530, 108–112

(2016).

251. Ballabh, P., Braun, A. & Nedergaard, M. The blood-brain barrier: An overview: Structure,

regulation, and clinical implications. Neurobiol. Dis. 16, 1–13 (2004).

252. Williams, D. W., Eugenin, E. A., Calderon, T. M. & Berman, J. W. Monocyte maturation, HIV

susceptibility, and transmigration across the blood brain barrier are critical in HIV

neuropathogenesis. J. Leukoc. Biol. 91, 401–15 (2012).

253. Salinas, S., Schiavo, G. & Kremer, E. J. A hitchhiker’s guide to the nervous system: the complex

journey of viruses and toxins. Nat. Rev. Microbiol. 8, 645–655 (2010).

254. Yang, B. et al. Global CNS transduction of adult mice by intravenously delivered rAAVrh.8 and

rAAVrh.10 and nonhuman primates by rAAVrh.10. Mol. Ther. 22, 1299–1309 (2014).

255. Rosenberg, J. B. et al. Comparative Efficacy and Safety of Multiple Routes of Direct CNS

Administration of Adeno-Associated Virus Gene Transfer Vector Serotype rh.10 Expressing the

Human Arylsulfatase A cDNA to Nonhuman Primates. Hum. gene Ther. Dev. (2014).

doi:10.1089/humc.2013.239 [doi]

256. Gray, S. J. et al. Preclinical differences of intravascular AAV9 delivery to neurons and glia: a

comparative study of adult mice and nonhuman primates. Mol. Ther. 19, 1058–1069 (2011).

257. Zhang, H. et al. Several rAAV vectors efficiently cross the blood-brain barrier and transduce

neurons and astrocytes in the neonatal mouse central nervous system. Mol. Ther. 19, 1440–1448

(2011).

258. Cearley, C. N. et al. Expanded repertoire of AAV vector serotypes mediate unique patterns of

transduction in mouse brain. Mol. Ther. 16, 1710–8 (2008).

259. Kumar, S., Stecher, G. & Tamura, K. MEGA7: Molecular Evolutionary Genetics Analysis Version

7.0 for Bigger Datasets. Mol. Biol. Evol. 33, 1870–4 (2016).

260. Saitou, N. & Nei, M. The neighbor-joining method: a new method for reconstructing phylogenetic

trees. Mol. Biol. Evol. 4, 406–25 (1987).

114

261. Felsenstein, J. CONFIDENCE LIMITS ON PHYLOGENIES: AN APPROACH USING THE

BOOTSTRAP. Evolution (N. Y). 39, 783–791 (1985).

262. Lein, E. S. et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature 445,

168–176 (2007).

263. Miller, E. B. et al. Production, purification and preliminary X-ray crystallographic studies of

adeno-associated virus serotype 1. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 62,

1271–1274 (2006).

264. Bordoli, L. et al. Protein structure homology modeling using SWISS-MODEL workspace. Nat.

Protoc. 4, 1–13 (2008).

265. Krissinel, E. & Henrick, K. Secondary-structure matching (SSM), a new tool for fast protein

structure alignment in three dimensions. Acta Crystallogr. Sect. D Biol. Crystallogr. 60, 2256–

2268 (2004).

266. Carrillo-Tripp, M. et al. VIPERdb2: an enhanced and web API enabled relational database for

structural virology. Nucleic Acids Res. 37, D436-42 (2009).

267. Xiao, C. & Rossmann, M. G. Interpretation of electron density with stereographic roadmap

projections. J. Struct. Biol. 158, 182–187 (2007).

268. Li, C. et al. Single amino acid modification of adeno-associated virus capsid changes transduction

and humoral immune profiles. J. Virol. 86, 7752–7759 (2012).

269. Bell, C. L., Gurda, B. L., Vliet, K. Van, Agbandje-McKenna, M. & Wilson, J. M. Identification of

the galactose binding domain of the AAV9 capsid. J. Virol. 86, 7326–7333 (2012).

270. Rosenberg, J. B. et al. Comparative efficacy and safety of multiple routes of direct CNS

administration of adeno-associated virus gene transfer vector serotype rh.10 expressing the human

arylsulfatase A cDNA to nonhuman primates. Hum. Gene Ther. Clin. Dev. 25, 164–77 (2014).

271. Cearley, C. N. & Wolfe, J. H. Transduction characteristics of adeno-associated virus vectors

expressing cap serotypes 7, 8, 9, and Rh10 in the mouse brain. Mol. Ther. 13, 528–37 (2006).

272. Hadaczek, P. et al. Transduction of nonhuman primate brain with adeno-associated virus serotype

1: vector trafficking and immune response. Hum. Gene Ther. 20, 225–237 (2009).

273. Chen, S. et al. Efficient Transduction of Vascular Endothelial Cells with Recombinant Adeno-

Associated Virus Serotype 1 and 5 Vectors. Hum. Gene Ther. 16, 235–247 (2005).

274. Pulicherla, N. et al. Engineering Liver-detargeted AAV9 Vectors for Cardiac and Musculoskeletal

Gene Transfer. Mol. Ther. 19, 1070–1078 (2011).

275. Asokan, A. et al. Reengineering a receptor footprint of adeno-associated virus enables selective

and systemic gene transfer to muscle. Nat. Biotechnol. 28, 79–82 (2010).

276. Nathwani, A. C. et al. Long-term safety and efficacy of factor IX gene therapy in hemophilia B. N.

Engl. J. Med. 371, 1994–2004 (2014).

115

277. Mingozzi, F. & High, K. A. Immune responses to AAV vectors: overcoming barriers to successful

gene therapy. Blood 122, 23–36 (2013).

278. Burger, C. & Nash, K. R. Gene Therapy for Neurological Disorders. Methods in molecular

biology (Clifton, N.J.) (2016). doi:10.1007/978-1-4939-3271-9

279. DiPrimio, N., Asokan, A., Govindasamy, L., Agbandje-McKenna, M. & Samulski, R. J. Surface

Loop Dynamics in Adeno-Associated Virus Capsid Assembly. J. Virol. 82, 5178–5189 (2008).

280. Pillay, S. & Carette, J. E. Host determinants of adeno-associated viral vector entry. Current

Opinion in Virology (2017). doi:10.1016/j.coviro.2017.06.003

281. Seiler, M. P., Miller, A. D., Zabner, J. & Halbert, C. L. Adeno-Associated Virus Types 5 and 6

Use Distinct Receptors for Cell Entry. Hum. Gene Ther. (2006). doi:10.1089/hum.2006.17.10

282. Shen, S., Bryant, K. D., Brown, S. M., Randell, S. H. & Asokan, A. Terminal N-linked galactose

is the primary receptor for adeno-associated virus 9. J. Biol. Chem. 286, 13532–40 (2011).

283. Huang, L.-Y. et al. Characterization of the Adeno-Associated Virus 1 and 6 Sialic Acid Binding

Site. J. Virol. 90, 5219–5230 (2016).

284. Albright, B. H. et al. Mapping the Structural Determinants Required for AAVrh.10 Transport

across the Blood-Brain Barrier. Mol. Ther. (2017). doi:10.1016/j.ymthe.2017.10.017

285. Grieger, J. C., Choi, V. W. & Samulski, R. J. Production and characterization of adeno-associated

viral vectors. Nat. Protoc. (2006). doi:10.1038/nprot.2006.207

286. Nam, H.-J. et al. Structure of adeno-associated virus serotype 8, a gene therapy vector. J. Virol.

(2007). doi:10.1128/JVI.01304-07

287. López-Bueno, A. et al. Host-selected amino acid changes at the sialic acid binding pocket of the

parvovirus capsid modulate cell binding affinity and determine virulence. J. Virol. (2006).

doi:10.1128/JVI.80.3.1563

288. Nam, H. J. et al. Identification of the sialic acid structures recognized by minute virus of mice and

the role of binding affinity in virulence adaptation. J. Biol. Chem. (2006).

doi:10.1074/jbc.M604421200

289. Lipton, H. L., Kumar, A. S. M., Hertzler, S. & Reddi, H. V. Differential usage of carbohydrate co-

receptors influences cellular tropism of Theiler’s murine encephalomyelitis virus infection of the

central nervous system. Glycoconjugate Journal (2006). doi:10.1007/s10719-006-5436-x

290. Bauer, P. H. et al. Discrimination between sialic acid-containing receptors and pseudoreceptors

regulates polyomavirus spread in the mouse. J. Virol. (1999). doi:10.1128/JVI.74.12.5746-

5746.2000

291. Jnaoui, K., Minet, M. & Michiels, T. Mutations that affect the tropism of DA and GDVII strains of

Theiler’s virus in vitro influence sialic acid binding and pathogenicity. J. Virol. (2002).

doi:10.1128/JVI.76.16.8138-8147.2002

116

292. Yang, B. et al. Global CNS transduction of adult mice by intravenously delivered rAAVrh.8 and

rAAVrh.10 and nonhuman primates by rAAVrh.10. in Molecular Therapy (2014).

doi:10.1038/mt.2014.68

293. Hocquemiller, M., Giersch, L., Audrain, M., Parker, S. & Cartier, N. Adeno-Associated Virus-

Based Gene Therapy for CNS Diseases. Hum. Gene Ther. 27, 478–96 (2016).

294. Di Pasquale, G. & Chiorini, J. A. AAV transcytosis through barrier epithelia and endothelium.

Mol. Ther. (2006). doi:10.1016/j.ymthe.2005.11.007

295. Di Pasquale, G., Kaludov, N., Agbandje-McKenna, M. & Chiorini, J. A. BAAV transcytosis

requires an interaction with β-1-4 linked- glucosamine and gp96. PLoS One (2010).

doi:10.1371/journal.pone.0009336

296. Di Pasquale, G. et al. Bovine AAV transcytosis inhibition by tannic acid results in functional

expression of CFTR in vitro and altered biodistribution in vivo. Gene Ther. (2012).

doi:10.1038/gt.2011.138

297. Bell, R. D. & Ehlers, M. D. Breaching the Blood-Brain Barrier for Drug Delivery. Neuron (2014).

doi:10.1016/j.neuron.2013.12.023

298. Engelhardt, B. & Sorokin, L. The blood-brain and the blood-cerebrospinal fluid barriers: Function

and dysfunction. Seminars in Immunopathology (2009). doi:10.1007/s00281-009-0177-0