function, therapeutic potential and cell biology of bace proteases: current status and future...

, , , ,

, ,

, , ,

*Department of Cell and Molecular Biology, Feinberg University School of Medicine, Northwestern

University, Chicago, Illinois, USA

†German Center for Neurodegenerative Diseases (DZNE), Munich, Germany

‡Neuroproteomics, Klinikum rechts der Isar, Technische Universit€at M€unchen, Munich, Germany

§Institute for Advanced Study, Technische Universit€at M€unchen, Garching, Germany

¶Munich Cluster of Systems Neurology (SyNergy), Munich, Germany

**Adolf-Butenandt Institute, Biochemistry, Ludwig-Maximilians University Munich, Munich, Germany

††Neurosciences, Merck Research Labs, Boston, Massachusetts, USA

‡‡Division of Psychiatry Research, University of Zurich, Zurich, Switzerland

§§Systems and Cell Biology of Neurodegeneration, Division of Psychiatry Research, University of

Zurich, Zurich, Switzerland

¶¶Graduate programs of the Zurich Center for Integrative Human Physiology and Zurich Neuroscience

Center, University of Zurich, Zurich, Switzerland

***Departments of Pathology and Neuroscience, Johns Hopkins University School of Medicine,

Baltimore, Maryland, USA

AbstractThe b-site APP cleaving enzymes 1 and 2 (BACE1 and BACE2)were initially identified as transmembrane aspartyl proteasescleaving the amyloid precursor protein (APP). BACE1 is amajordrug target for Alzheimer’s disease because BACE1-mediatedcleavage of APP is the first step in the generation of thepathogenic amyloid-b peptides. BACE1, which is highlyexpressed in thenervoussystem, isalso required formyelinationby cleaving neuregulin 1. Several recent proteomic and in vivostudiesusingBACE1-andBACE2-deficientmicedemonstrateamuch wider range of physiological substrates and functions forboth proteases within and outside of the nervous system. For

BACE1 this includes axon guidance, neurogenesis, musclespindle formation, and neuronal network functions, whereasBACE2wasshowntobe involved inpigmentationandpancreaticb-cell function. This review highlights the recent progress inunderstanding cell biology, substrates, and functions of BACEproteases and discusses the therapeutic options and potentialmechanism-based liabilities, in particular for BACE inhibitors inAlzheimer’s disease.Keywords: Alzheimer’s disease, BACE1, BACE2, protease,regulated intramembrane proteolysis, secretase.J. Neurochem. (2014) 130, 4–28.

In October 2013 the first international meeting focusing onsubstrates, functions, trafficking and therapeutic potential ofb-site APP cleaving enzyme (BACE) proteases took place inGermany at Kloster Seeon near Munich. The new andexciting data presented at the meeting prompted the presentreview. We start with a general introduction to BACEproteases followed by their cell biology and the complexintracellular trafficking of BACE1, which appears to be

Received February 2, 2014; revised manuscript received March 12,2014; accepted March 14, 2014.Address correspondence and reprint requests to Stefan F. Lichtent-

haler, German Center for Neurodegenerative Diseases (DZNE), 81377Munich, Germany. E-mail: [email protected] used: AD, Alzheimer’s disease; APP, amyloid precursor

protein; BACE, beta-site amyloid cleaving enzyme; CHL1, close homologof L1; EGF, epidermal growth factor; ER, endoplasmic reticulum; IPB,infrapyramidal bundle; NRG1, neuregulin 1; NTF, N-terminal fragment;OSNs, olfactory sensory neurons; TGN, trans-Golgi network.

4 © 2014 International Society for Neurochemistry, J. Neurochem. (2014) 130, 4--28

JOURNAL OF NEUROCHEMISTRY | 2014 | 130 | 4–28 doi: 10.1111/jnc.12715

tightly linked to its ability to cleave the membrane-boundsubstrates. Subsequently, we describe our current knowledgeabout substrates and functions of BACE1, in particular in theperipheral and central nervous system (CNS), and then aboutsubstrates and functions of the homologous protease BACE2.Finally, we discuss the therapeutic potential of both prote-ases, highlight the current therapeutic development forBACE1 in Alzheimer’s disease (AD) and give an outlookon future BACE protease research. Regulation of BACEprotease expression and activity has been partly reviewedelsewhere (Dislich and Lichtenthaler 2012; Rossner et al.2006) and is not the topic of this review article.

b-amyloid and Alzheimer’s disease

AD is a devastating neurodegenerative disease characterizedby the cerebral accumulation of two hallmark brain lesions:amyloid plaques and neurofibrillary tangles. Amyloidplaques are extracellular deposits of short 38 to 43residue-long peptides called b-amyloid (Ab), whereas neu-rofibrillary tangles are intracellular aggregates of aberrantlyprocessed hyperphosphorylated tau, a microtubule-associ-ated protein. Amyloid is a generic term referring to differentproteins that mis-fold and self-aggregate into b-pleated sheetstructures that deposit in various tissues thus causingdisease, the so-called peripheral amyloidoses. Amyloidplaques define AD as an amyloidosis disease of the brainand suggest the amyloid cascade hypothesis of AD, whichposits cerebral Ab accumulation as a critical early step inAD pathogenesis that leads to neurofibrillary tangle forma-tion, neuroinflammation, synaptic loss, neuron death, andultimately dementia (Hardy and Selkoe 2002). If the amyloidhypothesis is true, then inhibition of cerebral Ab accumu-lation should be efficacious for AD, if given early enough inthe disease process.Ab is a normal metabolite made and secreted by most cell

types, although neurons are the major producers of Ab in thebrain. Ab is generated by endoproteolysis of the type Imembrane protein amyloid precursor protein (APP; Fig. 1a).Two proteases called b- and c-secretases cleave APPsequentially to liberate Ab. APP is first cut by the b-secretase thus creating the amino (N)-terminus of Ab andyielding a membrane bound carboxy (C)-terminal fragmentcalled C99; a secreted APP ectodomain, sAPPb is alsogenerated (Vassar et al. 2009). Alternatively, a differentprotease called a-secretase may cut within the Ab domain ofAPP, generating the soluble ectodomain sAPPa andthe membrane bound C83 fragment, thus precluding Abformation. After b-secretase or a-secretase cleavages, thec-secretase enzyme then cuts C99 or C83 to release Ab or thenon-toxic p3 fragment into the lumen of the endosome,respectively. The c-secretase is a multi-subunit complexcomposed of four transmembrane proteins: presenilin,nicastrin, Pen2, and Aph1 (Sisodia and St George-Hyslop

2002; De Strooper et al. 2010). Ab subsequently undergoesexocytosis and is secreted into the interstitial fluid of thebrain. As both b- and c-secretases are necessary for Abformation, these enzymes are prime drug targets for reducingcerebral Ab levels for AD and therapeutic strategies to inhibitthem are being intensely pursued. Conversely, activation ofa-secretase should also lower Ab levels, although approachesto accomplish this goal are less clear.Human genetics has taught much about the underlying

mechanisms of disease pathogenesis. In the 1970s, Brownand Goldstein discovered mutations in the LDL receptor thatcause early on-set familial hypercholesterolemia, thus reveal-ing the pathological role of high serum cholesterol levels incardiovascular disease. These seminal studies provided thefoundation for the development of one of the most widelyprescribed drug classes, the statins that inhibit HMG-CoAreductase and thus lower serum cholesterol for the treatmentof heart disease (Goldstein and Brown 2009). In a similarfashion, studies in the human genetics of AD have revealedthat cerebral accumulation of Ab plays a critical early role inAD pathogenesis (Tanzi 2012). Several lines of evidencedraw this conclusion. First, over 200 autosomal dominantmutations that cause familial AD (FAD) have been found inthe genes for APP and presenilin, the active subunit ofc-secretase. These mutations invariably lead to either increasedAb42 : Ab40 ratio or over-production of total Ab. Notably,the FAD mutations in APP are found near the b- andc-secretase cleavage sites and make APP a more efficientsubstrate for endoproteolysis by the secretases. Of particularrelevance are the K670N; M671L (Swedish) double mutation(Mullan et al. 1992a) and the A673V mutation (Di Fedeet al. 2009) that are adjacent to the b-secretase cleavage siteand cause FAD by increasing b-secretase processing andtotal Ab production (Fig. 1b). Duplication of the APP genein Down syndrome/trisomy 21 and rare APP locus duplica-tions also cause FAD because of APP over-expression andtotal Ab over-production. Moreover, the epsilon 4 allele ofapolipoprotein E is the major genetic risk factor for late-onsetAD (LOAD) and is associated with increased accumulation ofcerebral Ab. Finally, the major a-secretase enzyme in thebrain, a disintegrin and metalloprotease 10 (ADAM10) (Kuhnet al. 2010; Jorissen et al. 2010), has recently been shown toharbor rare mutations in the prodomain that attenuatea-secretase activity, thereby causing increased b-secretasecleavage of APP, Ab over-production, and LOAD (Suh et al.2013). Together, the evidence demonstrating that AD isassociated with mutations in at least five different genes thatall lead to increased cerebral Ab accumulation stronglysuggests that Ab plays a central role in AD pathogenesis.The occurrence of mutations that cause AD implies that

the converse might be true as well, namely that geneticvariants exist that protect against AD. Indeed, a low-frequency mutation in APP, the A673T coding substitution,was recently shown to be associated with decreased risk of

© 2014 International Society for Neurochemistry, J. Neurochem. (2014) 130, 4--28

BACE proteases 5

AD and reduced cognitive decline in the elderly (Jonssonet al. 2012). The A673T substitution occurs only two aminoacids C-terminal to the b-secretase cleavage site (Fig. 1b)and is at the identical position as the A673V mutation thatcauses FAD (Di Fede et al. 2009). However, unlike A673V,APP harboring A673T is less efficiently cleaved by b-secretase, leading to a ~ 40% reduction in Ab production invitro. These results suggest that heterozygous carriers of theA673T mutation should have a life-long ~ 20% decrease inAb generation, thus protecting them from AD. Mostimportantly, the A673T mutation serves as proof-of-principlethat modest inhibition of b-secretase cleavage of APP mayprevent AD.

BACE1: the b -secretase enzyme

Soon after the characterization of APP processing and Abproduction in the early 1990s, intense efforts in academia andindustry were under way to identify the b- and c-secretase, asit was clear these enzymes are prime therapeutic targets forAD. Genetic, biochemical, and cellular assays were devel-oped specifically for secretase identification. In 1999, fivegroups independently discovered the molecular identity ofthe b-secretase enzyme and named it BACE, Asp2, ormemapsin 2 (Vassar et al. 1999; Yan et al. 1999; Sinhaet al. 1999; Hussain et al. 1999; Lin et al. 2000). The groupsused disparate experimental methodologies, including

C99 C59

APP sAPPβ

Aβ

Lumen

Cytosol

N

A

A

C

C

β

γ

1

23

48

228

421

478

501

S--S

S--S

P498

N153N172

N223

N354

R275

R279

R285

R299

R300

R307

R126

Signal Peptide

PeptidePro-

Loop454

N

C

Lumen

Cytosol Ub501

DTGS

DSGT

domainCatalytic

Cytosolicdomain

Trans-membrane

S--S

C-term

Lumen

Cytosol

N-term

Swedish K670N/M671L

A673T

APP

K

A673V

PROTECTIVE

FAD

dish

FAD

SweK670N

C t

α

67

A

A

AM

β

γCC

N

(a) (b) (c)

Fig. 1 APP processing, FAD mutations, and b-site APP cleavingenzyme (BACE)1. (a) APP is a type-I membrane protein that issequentially cleaved by two aspartic proteases to generate Ab. First,

the b-secretase enzyme (b) cuts APP (1) to create the N-terminus ofAb. Two APP fragments are produced: membrane-bound C99 andsecreted sAPPb ectodomain (yellow). Second, C99 is cleaved by thec-secretase enzyme (c) to generate the C-terminus of Ab. Ab (orange)

is then released into the lumen of the endosome and secreted into theextracellular medium. An intracellular domain, C59 (green), is alsoproduced. (b) The membrane-bound APP polypeptide is represented

by the gray string. APP residues that affect b-secretase processing ofAPP in humans are represented by gray circles, within which the wild-type residue is identified by the single-letter amino acid code. The

K670N/M671L (Swedish) and A673V mutations cause FAD byincreasing the rate of b-secretase cleavage and Ab production,

whereas the A673T mutation protects against Alzheimer’s disease(AD) by doing the opposite. All three mutations occur at or within oneamino acid of the b-secretase cleavage site. Red, blue, and lavender

notched ellipses represent a, b, and c-secretases, respectively, cuttingat their respective cleavage sites in APP. (c) BACE1 is a 501 aminoacid type-I transmembrane aspartic protease. The various subdomainsof BACE1 are indicated to the right of the structure. Numbers and

letters refer to amino acid positions and single-letter code, respectively.The two signature aspartic protease active site motifs at positions 93and 289 are shaded black. S–S denote positions of disulfide bridges

within the catalytic domain; N represents positions of N-linkedglycosylation sites; R indicates positions of acetylated arginineresidues; C marks positions of S-palmitoylated cysteine residues; P

indicates phosphorylation of serine 498; Ub denotes ubiquitination oflysine 501.


6 R. Vassar et al.

expression cloning, genomic strategies, and biochemicalpurification, to discover the b-secretase (henceforth referredto as BACE1). All arrived at the same polypeptide sequencefrom these different approaches, thus raising confidence thatthe authentic b-secretase enzyme had been identified.Initial characterization of BACE1 demonstrated that the

enzyme exhibits all the molecular and cellular propertiesexpected for the b-secretase (Vassar et al. 2009). BACE1 isa 501 amino acid type I transmembrane aspartic protease thatis related to the pepsin family and the retroviral asparticproteases (Fig. 1c). The BACE1 catalytic domain containstwo signature aspartic acid active site motifs (DT/SGS/T)that are spaced approximately 200 residues apart, similar tothe pepsins. The active site of the enzyme is topologicallyorientated on the same side of the membrane as theb-secretase cleavage site of APP, as is required for b-secretase.Moreover, BACE1 is localized within acidic intracellularcompartments including endosomes and trans-Golgi network(TGN) and has optimal enzyme activity at acidic pH, aspredicted for b-secretase. Both BACE1 and b-secretaseactivities are relatively insensitive to the pan-aspartic prote-ase inhibitor pepstatin. BACE1 is expressed at low levels inmost cell types of the body, although it is more highlyexpressed in neurons, again as expected. Finally, BACE1 hasthe correct cleavage specificity predicted for b-secretase:BACE1 cleaves at Asp+1 and Glu+11 of the Ab sequence;BACE1 over-expression and knockdown increases anddecreases sAPPb, C99, and Ab production, respectively;BACE1 knockdown increases sAPPa and p3 production;BACE1 sequence specificity at P1 is Leu>>Met>Val. Takentogether, these features of BACE1 made it an extremelystrong b-secretase candidate.

BACE1: post-translational modifications

Similar to other aspartic proteases, BACE1 undergoes anumber of post-translational modifications. BACE1 is syn-thesized as a zymogen in which its pre- and pro-peptidedomains are removed in the endoplasmic reticulum (ER) andTGN by signal peptidase and pro-protein convertase,respectively (Benjannet et al. 2001; Bennett et al. 2000b).BACE1 also undergoes N-glycosylation at four asparagineresidues (N153, N172, N223, N354) in the ER and Golgiapparatus (Haniu et al. 2000; Capell et al. 2000). During itstransit through the ER, the BACE1 catalytic domain is foldedand cross-linked with three disulfide bonds (C216-C420,C278-C443, C330-C380) (Haniu et al. 2000). This atypicaldisulfide structure, together with its transmembrane domain,makes BACE1 an unusual aspartic protease. The cytosolicdomain of BACE1 can undergo phosphorylation at serineresidue 498, an event that influences BACE1 trafficking inthe endosomal–lysosomal system (Pastorino et al. 2002). Inaddition, lysine 501 in the cytosolic domain can beubiquitinated, a process that affects BACE1 trafficking and

degradation (Kang et al. 2012, 2010). In addition, BACE1exhibits S-palmitoylation at four cysteine residues (C474,C478, C482, C485) at the junction of the transmembrane andcytosolic domains that determines localization to lipid rafts(Vetrivel et al. 2009). Finally, BACE1 may be acetylated atseveral arginine residues in the catalytic domain, modifica-tions that appear to affect the regulation of the enzyme (Koand Puglielli 2009).

Cell biology of BACE1

Cellular localization of b-secretase cleavageBACE1 localizes to lipid rafts and this localization correlateswith its b-cleavage activity toward APP. Interestingly invitro, specific lipids that are constituents of lipid raftsstimulate BACE1 activity (Kalvodova et al. 2005). AlthoughBACE1 is localized in various organelles, its activity isreported to be at a maximum in endosomes and to a lowerextent also in the TGN. BACE1, being an aspartyl protease,requires low pH for its activity and this correlates with itslocalization in these two organelles that have a lumenalacidic pH (4.5–6.0) (Rajendran and Simons 2008; Kalvodovaet al. 2005; Ehehalt et al. 2003). While the Swedish mutantof APP – which has a higher affinity to BACE1 than wild-type APP – can be cleaved by BACE1 in biosyntheticcompartments (likely in the TGN)(Haass et al. 1995), mostof the wild-type APP gets cleaved in the endocyticcompartment. Besides the low endosomal pH several addi-tional lines of evidence point to this compartment, inparticular to the early endosomes (Kinoshita et al. 2003;Rajendran et al. 2006; Sannerud et al. 2011). First, theYENPTY motif in the APP cytoplasmic domain and thecorresponding dileucine motif of the BACE1 cytoplasmicdomain play active roles in their sorting into endosomes andinhibition of endocytosis reduces b-cleavage of APP.Second, both APP and BACE1 intermolecularly interact inendosomes. Third, retromer and retromer-associated proteinsthat regulate the sorting of APP from early/late endosomes,regulate BACE1-mediated cleavage of APP; other proteinsthat regulate the sorting of BACE1 from endosomesincluding Golgi-localized, gamma-ear containing, ADP-ribosylation factor binding 1 (GGA1) and GGA3 playcritical roles in BACE1 cleavage. Fourth, synaptic activitythat increases endocytosis also increases b-cleavage of APP.Sixth, targeting a transition-state BACE1 inhibitor to endo-somes via membrane anchoring or using compounds raisingthe membrane-proximal endosomal pH effectively inhibit Abproduction in cultured cells and in animals (Rajendran et al.2008; Mitterreiter et al. 2010). All this clearly suggests thatBACE1 cleavage of APP mostly occurs in endosomes.

Intracellular trafficking of BACE1

The molecular events that underlie the initial sorting ofBACE1 and APP to early endosomes are not fully understood.


BACE proteases 7

Whether BACE1 and APP are co-internalized from the cellsurface or whether they have differential requirements forsorting into distinct clathrin-coated pits is still not clear.However, there is some evidence that supports the latterclaim. While APP is internalized similar to tetanus toxinusing cholesterol-dependent clustering and the adaptorprotein-2 (Perez et al. 1999; Schneider et al. 2008), BACE1internalization is remarkably similar to that of TfR. Inaddition, BACE1 sorting to early endosomes requires Arf6,again dependent on its cytosolic sorting motif pathway(Prabhu et al. 2012; Sannerud et al. 2011). It is plausible thatAPP and BACE1 are internalized through separate endocyticroutes and that their convergence occurs at the level of eitherclathrin-coated vesicles or early endosomes. The existence ofheterogeneity at early endosomes (Lakadamyali et al. 2006)supports the idea that APP and BACE1 separately internalizefrom the plasma membrane through distinct routes and thenmerge at the early endosomal level in order for BACE1 tocleave APP.Once internalized into endosomes, what is the cellular fate

of endosomal BACE1? BACE1 can be sorted to the Golgivia a retrograde pathway or to lysosomal compartments fordegradation. Members of the GGA protein family interactwith BACE1 via its acidic cluster-dileucine binding motif toregulate its trafficking from endosomes (He et al. 2005;Tesco et al. 2007; von Arnim et al. 2006) in a process alsodependent on the phosphorylation status of BACE1. Inter-estingly both GGA1 and GGA3 negatively regulate BACE1residency in endosomes and thus in Ab generation. In ADcases, both proteins have been shown to be decreasedconfirming their involvement in Ab generation (Kang et al.2010; Walker and Tesco 2013). While the mechanismsunderlying the decreased expression of GGA1 in AD are notfully understood, GGA3 is a caspase substrate that undergoesdegradation upon apoptotic stimuli-induced caspase activa-tion. As GGA3 is involved in endosome to lysosome sortingof BACE1, GGA3 deficiency leads to increased BACE1levels including in endosomes and thus higher b-cleavageand Ab generation (Kang et al. 2010).Sorting of BACE1 to lysosomes leads to its degradation

(Koh et al. 2005). Ubiquitin-proteasomal pathways havebeen shown to be important for its degradation, although howubiquitination of membrane-bound BACE1 could act as atarget for degradation via the proteasome is not clear. Asubiquitination is a signal for sorting into intraluminal vesiclesof late endosomal compartments, BACE1 ubiquitinationmost likely plays a role in sorting of BACE1 to lateendosomes and eventually lysosomes for its degradation(Kang et al. 2010).Endosomal BACE1 is not only transported to the TGN and

lysosomes, but also to recycling endosomes (Udayar et al.2013). In a recent study, a paired Rab GTPase RNAi andRab-GAP over-expression screen lead to the identification ofnovel membrane trafficking routes and thereby to a better

understanding of the complexity of APP b-cleavage. Impor-tantly, the recycling endosome protein Rab11 was identifiedas a robust player regulating b-cleavage of APP. SilencingRab11 via RNAi reduced Ab levels both in cell line modelsas well as primary neurons from wild-type mice. Mechanisticcharacterization suggested that BACE1 from early endo-somes is recycled via Rab11-dependent recycling compart-ments to the cell surface and then re-internalized back toendosomes (Fig. 2).The recycling of APP or BACE1 is of interest to the AD

community as inhibition of BACE1 recycling to route it fordegradation could be a possibility for therapeutic interven-tion. On the other hand, APP recycling and sorting of APPfrom endosomes uses a complex machinery called theretromers. The hetero-pentameric retromer proteins associateto the cytosolic side of endosomes to mediate the retrogradetransport of transmembrane proteins to the Golgi. Usually,retromers contain three VPS (Vacuolar protein sorting)

Clathrinγ-secretase

Early endosomes

Endocytosis

Recycling endosomes (rab11)

Amyloid precursor

protein (APP)

Amyloid β BACE1

Fig. 2 Beta-site amyloid cleaving enzyme (BACE)1 trafficking requiresrecycling endosomes. BACE1 traffics from plasma membrane to early

endosomes where it cleaves most of the cellular APP. From the earlyendosomes, BACE1 is routed to Rab11-GTPase positive recyclingendosomes, through which it is sorted to plasma membrane to re-initiate another round of entry into early endosomes to cleave APP. In

the absence of functional Rab11 GTPase, much of BACE1 accumu-lates in recycling compartments and fails to be recycled to earlyendosomes, as a result of which b-cleavage of APP and Ab production

are significantly decreased.


8 R. Vassar et al.

proteins termed, VPS26, VPS29 and VPS35 and two sortingnexins, though the identity and the role of sorting nexins areless clear (Siegenthaler and Rajendran 2012). Loss ofretromer function (VPS26, VPS35) and retromer-associatedproteins (Sorting protein-related receptor containing LDLRclass A repeats) has been correlated with risk for Alzheimer’sdisease and elevated levels of Ab peptide (Andersen et al.2005; Rogaeva et al. 2007). This suggests that whileretromers and retromer-associated proteins sort uncleavedAPP from early endosomes to TGN (Andersen et al. 2005;Small and Gandy 2006; Rogaeva et al. 2007; Siegenthalerand Rajendran 2012), Rab11-dependent pathway recyclescellular BACE1 to early endosomes. Interestingly, a recentlive cell imaging study in hippocampal neurons showed thatupon internalization from the dendritic surface, BACE1undergoes exclusive retrograde transport to the soma and thatthis polarized transport in dendrites requires Eps15 homol-ogy domain (EHD) proteins 1/3, whereas axonal BACE1sorting requires Rab11 GTPase activity. Thus Rab11 andEHD proteins coordinate trafficking and axonal transport ofendocytic BACE1 in recycling endosomes in polarizedhippocampal neurons (Buggia-Prevot et al. 2013, 2014).Intriguingly, a variant of Rab11 has recently been identifiedto be associated with late-onset AD thus linking trafficking ofBACE1, Ab generation and AD (Udayar et al. 2013).

BACE1 in vivo validation

Soon after the discovery of BACE1, a related membrane-bound aspartic protease, BACE2, was identified that shares~ 64% amino acid similarity to BACE1. The high homologybetween BACE1 and BACE2 suggested BACE2 might alsoserve as a functional b-secretase enzyme in the brain.However, unlike b-secretase, BACE2 is expressed at lowlevels in neurons compared to BACE1 (Bennett et al. 2000a;Laird et al. 2005). Furthermore, although BACE2 cangenerate Ab in vitro, it prefers to cleave APP at residuesPhe+19 and Phe+20 of Ab (Farzan et al. 2000; Yan et al.2001; Fluhrer et al. 2002; Basi et al. 2003). Thus, BACE2behaves more like a-secretase in that it precludes Abformation. Taken together, the characteristics of BACE2suggest it is unlikely to be a major b-secretase in the brain.Nevertheless, it was critical to validate BACE1 as the

major cerebral b-secretase in vivo. To do so, several groupsemployed gene targeting technology to generate BACE1knockout (�/�) mice (Luo et al. 2001; Roberds et al. 2001;Cai et al. 2001; Dominguez et al. 2005). Initial reportssuggested that BACE1�/� mice lacked an overt phenotypeand were viable, fertile, and normal appearing in grossmorphology, behavior, tissue histology, and blood cell andclinical chemistry. These results implied that BACE1inhibitor drugs should be free of mechanism-based side-effects. Well-established lines of APP transgenic (Tg) micehave been generated that develop amyloid plaques with age.

Several of these APP Tg lines were crossed with BACE1�/�

mice to produce APP Tg/BACE1�/� bigenic mice that wereshown to lack Ab production, amyloid deposition, and Ab-dependent memory deficits (Ohno et al. 2004; Luo et al.2003; Laird et al. 2005; Ohno et al. 2007; McConlogueet al. 2007). These results demonstrate that BACE2 does notcompensate for BACE1 deficiency for the generation of Ab.In addition, they validate BACE1 as the major b-secretaseenzyme in the brain and suggest that BACE1 inhibitionshould be efficacious for lowering cerebral Ab levels in AD.

BACE protease substrates – overview

The function of a protease is determined by its substrates.Initially, APP was the only known BACE1 substrate and earlyreports suggested BACE1 knockout mice were normal.However, now it is evident that there are complex BACE1null phenotypes that provide insights into importantBACE1 physiological functions in vivo. Identifying allBACE1 substrates and understanding how BACE1 cleavagemodulates their functions is essential as it will be thesefunctions that will also be blocked by therapeutic BACE1inhibition. Likewise, new phenotypes and substrates are beingdiscovered using BACE2-deficient mice. We start by describ-ing our current knowledge about substrates and functions ofBACE1 in the peripheral nervous system, followed by theCNS. Other phenotypes in BACE1-deficient mice, whichappear to be independent of the nervous system, such asincreased lethality in the first month after birth and protectionagainst diet-induced obesity will not be discussed in detail(Dominguez et al. 2005; Meakin et al. 2012). Likewise,substrates which are mostly expressed outside of the nervoussystem have been listed in a recent publication (Dislich andLichtenthaler 2012). Afterward, substrates and functions ofthe homologous protease BACE2 will be described.

BACE1 substrates and functions in the peripheralnervous system

Hypomyelination in BACE1�/� animals

One of the best-understood functions of BACE1 is its role inthe proteolytic processing and activation of neuregulin 1(NRG1) type III (Willem et al. 2009; Fleck et al. 2012). InBACE1 knock-out mice, loss of cleavage of this substratecould be clearly functionally linked to a loss-of-function in asignaling pathway of pivotal importance during post-natalmyelination (Fleck et al. 2012; Willem et al. 2009). A firsthint for such a function came from a very simple experiment.A developmental western blot revealed an extremely highexpression of BACE1 during the first post-natal week(Willem et al. 2006). Subsequently, BACE1 expressiondecreases and in adulthood only very little BACE1 couldbe detected. This finding along with the observations thatBACE1 is preferentially expressed in neurons, myelination


BACE proteases 9

occurs right after birth and the major signaling pathwayinvolved in myelination (NRG signaling) appears to requireproteolytic activation (Birchmeier and Nave 2008), led to thehypothesis that BACE1 might be the protease required toregulate axonal myelination. Indeed, detailed electron micro-scopic analyses of sciatic nerves revealed two very strikingphenotypes (Willem et al. 2006; Hu et al. 2006). First, axonsin BACE1�/� animals displayed a hypomyelination pheno-type (Fig. 3a; left panel); second, sorting of small-diameteraxons by Schwann cell processes within Remak bundles wasdramatically disturbed (Fig. 3a; right panel). Axons withinRemak bundles were not separated from each other bySchwann cells, and the number of axons within Remakbundles was significantly increased (Fig. 3a; right panel).Strikingly, these myelination abnormalities within theperipheral nervous system (PNS) phenocopied a well-knownknockout phenotype of one of the central players in theregulation of myelination – namely Nrg 1 type III (Birchmeierand Nave 2008; Garratt et al. 2000; Michailov et al. 2004;Taveggia et al. 2005). The neurotrophic members of theNrg1 family signal to ErbB receptors and are involved incardiac and neuronal development. The Nrg1 type III isoformis preferentially expressed in neurons, and axonally locatedNrg1 binds to ErbB2/3 or ErbB4 receptors expressed onoligodendrocytes and Schwann cells, respectively. In linewith that, reduction of Nrg1 type III in a heterozygousknockout model (Michailov et al. 2004; Taveggia et al.2005) (the homozygous knockout is embryonically lethal)results in hypomyelination and disturbed formation ofRemak bundles as described above for the BACE1 knockout.Only BACE1 but not BACE2 contributes to myelination(van Bebber et al. 2013; Willem et al. 2006; Rochin et al.2013). Whereas the BACE1 knockout-dependent hypomye-lination phenotype within the PNS could be observed inseveral independent BACE1�/� lines (Willem et al. 2006;Hu et al. 2006), inconsistent findings regarding BACE1function in myelination of the CNS were obtained. Whereasone study failed to observe hypomyelination within the CNS(Willem et al. 2006), another study reported not onlyreduced myelination in the PNS but also within the opticnerve, hippocampus and cerebral cortex (Hu et al. 2006). Toprovide further insight into this issue, BACE1 knockoutzebrafish have recently been generated via genome editing(van Bebber et al. 2013). Crossing of BACE1 knockoutzebrafish with a reporter line, which expressed greenfluorescent protein under the myelin specific claudin Kpromotor, allowed in vivo imaging of myelination. Thismodel revealed a selective loss in myelination of the PNS-derived posterior lateral line nerves upon BACE1 deletionwhile myelination of CNS derived Mauthner axonsensheathed by oligodendrocytes was normal (Fig. 3b).Furthermore, an anti-sense gripNA-mediated knockdown ofNRG1 type III in zebrafish also revealed selective hypomye-lination of the PNS further supporting that BACE1

selectively affects PNS but not CNS myelination (vanBebber et al. 2013), at least in zebrafish.

Proteolytic processing of NRG1 type III by BACE1 and other

sheddases

Alternative splicing generates numerous isoforms of NRG1(Willem et al. 2009; Fleck et al. 2012). All isoforms containan epidermal growth factor (EGF) domain, which binds andactivates ErbB receptors (Falls 2003). However, signaling isonly possible if the EGF domain is liberated or exposed byproteolytic cleavage. While most Nrg1 isoforms contain onlyone transmembrane (TM) domain, Nrg1 type III forms ahairpin-like protein with two TM domains (Falls 2003)(Fig. 3c). Nrg1 type III undergoes regulated intramembraneproteolysis (RIP) by a rather large variety of differentsheddases and intramembrane cleaving proteases and maythus represent a “super-RIP” substrate (Willem et al. 2009;Fleck et al. 2012). The identification of Nrg1 type III as aphysiological BACE1 substrate has already predicted thatbesides BACE1 other alternative sheddases cleave andactivate Nrg1 type III, as the BACE1 knockout did notpresent with a complete amyelination phenotype but ratherwith hypomyelination (Willem et al. 2006; Hu et al. 2006).Indeed, studies over-expressing and knocking down/inhibit-ing sheddases of the ADAM (a disintegrin and metallopro-teinase) family showed that the stalk region of Nrg1 type IIIis also cleaved by ADAM10 and ADAM17 (Fleck et al.2013; Luo et al. 2011; La Marca et al. 2011) (Fig. 3c) in amanner very similar to the proteolytic processing of APP(Haass 2004). All three cleavages allow exposure of theEGF-like domain and consequently signaling to ErbBreceptors (Fleck et al. 2013; Luo et al. 2011). The remainingC-terminal stub is then cleaved within its TM domain byc-secretase, producing a secreted amyloid b-peptide likefragment (Nrg1 b-peptide; Fig. 3c) as well as an intracellularC-terminal cytoplasmic domain (Dejaegere et al. 2008)(C-ICD; Fig. 3c). The C-ICD translocates to the nucleusand may act as a transcriptional regulator in potential reversesignaling pathways (Bao et al. 2004, 2003; Chen et al.2010). More recently, it was shown that the remainingmembrane bound N-terminal fragment (NTF) produced bycleavage of either ADAM10, 17, or BACE1 undergoesfurther proteolytic cleavage to liberate a soluble (secreted)fragment containing the EGF-like domain (sEGF) (Flecket al. 2013). Strikingly, a second cleavage site for BACE1could be identified within the NTF of Nrg1 type III.Moreover, cleavage at the identified cleavage site appearsto be very efficient that exactly matches the sequence of theSwedish mutation of APP, which is known to dramaticallyincrease BACE1-mediated cleavage (Citron et al. 1992).BACE1 cleavage at this position of Nrg1 type III results insecretion of sEGF (Fig. 3c), which could be identified inconditioned media of primary neurons (Fleck et al. 2013).Similar to the multiple cleavages of Nrg1 type III within the


10 R. Vassar et al.

stalk region by several sheddases the NTF is not onlyprocessed by BACE1 but can also be processed by ADAM17(Fig. 3c; but not by ADAM10) (Fleck et al. 2013). Theobservation that dual cleavage of Nrg1 type III liberates itsEGF-like domain (sEGF) raised the question whether thisproteolytic fragment retains its signaling capacity. Indeed,monitoring the phosphorylation state of ErbB3 receptor andAKT in reporter cell lines and primary Schwann cells upontreatment with sEGF revealed that this fragment retains itssignaling capacity in vitro. Moreover, injection of in vitrotranscribed mRNAs, only encoding the sEGF-like domain,into oocytes from BACE1 knockout zebrafish (van Bebberet al. 2013) (see above) allowed rescue of the hypomyeli-nation phenotype (Fleck et al. 2013).A previous study suggested that the ADAM17 cleavage at

the C-terminus of the EGF-like domain abolishes Schwanncell myelination (La Marca et al. 2011). However, the rescueexperiments in cultured cells and in vivo using mRNAs/cDNAs encoding fragments mimicking cleavage by BACE1or ADAM17 demonstrated that both Nrg1 derivatives arefully capable of supporting Nrg1 signaling and myelination(Fleck et al. 2013). Thus, all sheddases reported to processNrg1 type III appear to exert redundant (activating) func-tions, however, BACE1 seems to be the dominant proteaseactivating Nrg1 signaling in the PNS.Finally, the membrane retained NTFs, which are generated

either by BACE1 or ADAM17 cleavage undergo furthercleavage by proteases of the signal peptide peptidase-likefamily (Voss, Fluhrer & Haass, unpublished observation).After intramembrane proteolysis a small peptide is secreted(C-peptide; Fig. 3c), and an N-terminal ICD (N-ICD) is

released from the membrane into the cytoplasm. Whether theN-ICD has a signaling function is currently not known.

Proteolytic processing of NRG1 type I by BACE1 is required

for muscle spindle formation and maintenance

Similar to NRG1 type III, NRG1 type I, which contains anIg-like domain instead of the cysteine rich domain of NRG1type III, is also proteolytically processed by BACE1. NRG1type I has a large luminal N-terminal domain but lacks thesecond transmembrane domain, which allows NRG1 type IIIto form its hairpin structure (see above). Upon fusion ofsecreted alkaline phosphatase to NRG 1 type I and co-expression of BACE1 secreted alkaline phosphatase activitywas observed in conditioned media, which could beinhibited by the addition of a BACE inhibitor (Cheret et al.2013). As previous experiments had already shown thatreduction in NRG1 in neurons or its receptor ErbB2 inmuscles reduces muscles spindle formation (Andrecheket al. 2002; Hippenmeyer et al. 2002; Leu et al. 2003),these in vitro experiments strongly suggested that BACE1might have yet another NRG1-associated function. Indeed,muscle spindle formation is gradually reduced by deletion ofone or both copies of BACE1 (Cheret et al. 2013).Consequently, BACE1 deficiency in mice results in defectsin coordinated movement. Moreover, pharmacologicalreduction in BACE1 activity in adult mice with a BACEinhibitor also reduced the number of muscle spindles (Cheretet al. 2013). Thus BACE1-mediated processing of NRG1type 1 is not only required for the formation of musclespindles during development but also for their maintenanceduring adulthood.

C

N

EGF

BACE1

C

CTF

Lumen

Cytosol

NRG1 type III

NRG1β-peptide

C

C-ICD

ADAM10/17

N

SPPL2a/b

N

N-ICD

C-peptide(c)

sEGF

(a) wt

BACE1 –/–wt; claudin K:GFP

bace1 –/–; claudin K:GFP

(b)

N

NTF

BACE1ADAM17

NRG1 signaling

EGF

EGF

Fig. 3 Beta-site amyloid cleaving enzyme(BACE)1 controls peripheral nervemyelination and muscle spindle formation

via proteolytic processing of neuregulin 1. (a)BACE1deficiency results in hypomyelinationof peripheral axons (left panel) andabnormalities in axonal-bundling (right

panel) (Willem et al. 2006). (b) Delay ofmyelination in BACE1�/� zebrafish (vanBebber et al. 2013). At 3 days post-

fertilization myelination of the posteriorlateral line nerves (PNS) is severely reducedin BACE1�/�; claudin k:GFP (red arrows)

whereas CNS derived oligodendrocytesensheathing the Mauthner axons arenormally myelinated (black arrows). (c)

Proteolytic processing of neuregulin 1(NRG1) type III by sheddases andintramembrane proteolysis.


BACE proteases 11

Therapeutic inhibition of BACE1 cleavage and

NRG signaling

Finally, although BACE1 knockouts have profound andhighly reproducible effects on myelination of the peripheralnervous system, this may not be a severe concern for BACEinhibition in human patients. Clearly, the BACE1 knockoutphenotype on myelination is a developmental phenotype.Thus hypomyelination may not be a major concern forBACE1-directed therapies. Whether the adult function ofBACE1 in spindle maintenance (Cheret et al. 2013) wouldbe affected by BACE inhibitors in humans, must, however,be carefully monitored.

BACE1 functions in the CNS

BACE1 is concentrated in pre-synaptic terminals of differentneuron types in the CNS and PNS (Kandalepas et al. 2013;Deng et al. 2013), suggesting a role for BACE1 in synapticfunction. Consistent with high neuronal expression and pre-synaptic localization of BACE1, recent studies of BACE1�/�

mice have revealed complex neurological phenotypes, manyof which involve the CNS (Table 1). BACE1 knockoutphenotypes are the result of deficient b-secretase processingof BACE1 substrates, and recent proteomic studies incultured neurons have identified a large number of novelBACE1 substrates that are involved in neuronal functions(Kuhn et al. 2012; Zhou et al. 2012).The first phenotypes identified in BACE1�/�mice involved

poor performance on spatial and temporal hippocampus-dependent memory tests (Kobayashi et al. 2008; Ohno et al.2006, 2007, 2004; Laird et al. 2005), timid behavior (Harrisonet al. 2003), and reduced serotonin and dopamine levels inhippocampus and striatum, respectively (Harrison et al.2003). These behavioral and neurochemical phenotypesstrongly suggest functions for BACE1 in specific neuronalsystems of the brain, although the culpable BACE1 substrateshave not been definitively identified. BACE1�/� mice alsoexhibit increased post-natal lethality and growth retardation

(Dominguez et al. 2005), but whether these phenotypes arerelated to a CNS function of BACE1 is unclear.Interestingly, lack of NRG1 processing in BACE1�/�

mice, which has been described in the previous paragraphsfor the peripheral nervous system, has also been associatedwith schizophrenia endophenotypes such as impairedpre-pulse inhibition, hypersensitivity to glutamatergicpsychostimulants, spine density reduction, and hyperactivity(Savonenko et al. 2008). Overall, these NRG1 studiessuggest that BACE1, in collaboration with ADAM proteases,cleaves NRG1 to release an EGF-like domain that activatesErB receptors on neighboring cells (Fleck et al. 2013; Luoet al. 2011), a signaling pathway that appears to be involvedin both myelination and schizophrenia.In addition, BACE1�/� mice exhibit spontaneous seizures

and hippocampal neuron loss that increase with age,compared with wild-type littermates (Hitt et al. 2010; Huet al. 2010; Kobayashi et al. 2008). Both generalized tonic-clonic and absence seizures are observed. Also, kainatetreatment induces more severe seizures and greater excito-toxic CA1 neuron death in BACE1�/� mice. BACE1 nullneurons display elevated sodium current and action potentialproperties (Hu et al. 2010). Notably, the density of voltage-gated sodium channels (NaV) on the BACE1�/� neuron cellsurface is increased, consistent with elevated sodium currentsin BACE1 null neurons. Previous studies have shown thatNaV b-subunits regulate the expression and cell-surfacelocalization of sodium channels (Isom 2001; Kim et al.2007). Moreover, b-subunits are BACE1 substrates (Gersb-acher et al. 2010; Wong et al. 2005), suggesting thatBACE1 processing of b-subunits controls cell-surface NaVchannel density, neuronal excitability, and seizure suscepti-bility (Kim et al. 2011).It is plausible that phenotypes displayed by BACE1�/�

mice could be related to physiological functions of BACE1-processed APP and amyloid precursor like protein (APLP)fragments. However, studies of APP, APLP1, and APLP2knockout mice largely suggest this is not the case. Although

Table 1 BACE1 null phenotypes in CNS

Phenotype Substrate References

Astrogenesis increase, neurogenesis decrease Jag1 Hu et al. (2013)Axon guidance defects CHL1 Rajapaksha et al. (2011), Cao et al. (2012), Hitt et al. (2012)

Hyperactivity NRG1 Dominguez et al. (2005), Savonenko et al. (2008)Hypomyelination NRG1 Willem et al. (2006), Hu et al. (2006, 2008)Memory deficits – Ohno et al. (2006, 2007, 2004), Laird et al. (2005)Neurochemical deficits – Harrison et al. (2003)

Neurodegeneration w/ age NaVb2 Hu et al. (2010)Post-natal lethality, growth retardation – Dominguez et al. (2005)Retinal pathology VEGFR1 Cai et al. (2012)

Schizophrenia endophenotypes NRG1 Savonenko et al. (2008)Seizures NaVb2 Kim et al. (2007), Kobayashi et al. (2008), Hu et al. (2010), Hitt et al. (2010)Spine density reduction NRG1 Savonenko et al. (2008)


12 R. Vassar et al.

APP�/� mice are viable and fertile, they display a number ofsubtle phenotypes including reduced brain and body weight,agenesis of the corpus callosum, increased susceptibility tokainite-induced seizures, reduced locomotor activity, GAB-Aergic abnormalities, increased L-type calcium channellevels, long-term potentiation defects, and spatial memoryimpairment, among others (Muller et al. 1994; Zheng et al.1995; Li et al. 1996; Steinbach et al. 1998; Dawson et al.1999; Magara et al. 1999; Phinney et al. 1999; Seabrooket al. 1999; Ring et al. 2007; Yang et al. 2009a). Althoughsome APP�/� phenotypes are shared with BACE1�/� mice(e.g., reduced weight, seizures, memory deficits), mostBACE1 null phenotypes are different. This is not surprisinggiven the large set of diverse BACE1 substrates that might befunctionally compromised in BACE1�/� mice.Reminiscent of APP�/� mice, APLP1�/� and APLP2�/�

mice display a subtle growth deficit and no phenotype,respectively (von Koch et al. 1997; Heber et al. 2000). Incontrast, double (APLP2/APLP1 and APLP2/APP) or triple(APP/APLP1/APLP2) knockout mice die shortly after birthand show malformed neuromuscular junctions with reducedsynaptic vesicle densities (von Koch et al. 1997; Heber et al.2000; Herms et al. 2004; Wang et al. 2005; Yang et al.2005). Like the single knockouts, APLP1/APP doubleknockout mice have subtle phenotypes, suggesting redun-dancy between APLP2 and the other family members (Heberet al. 2000). In addition to post-natal death and neuromus-cular junction abnormalities, APP/APLP1/APLP2 tripleknockout mice display cortical dysplasia reminiscent of typeII lissencephaly in humans with partial loss of Cajal Retziuscells (Herms et al. 2004). Together, the phenotypes of theAPP compound knockout mice suggest an important role forthe APP family in synaptic function and maintenance.The question of whether a- and b-secretase cleaved APP

extracellular fragments have physiological functions in vivohas been recently addressed by the generation of knockin micethat solely express either the sAPPa or sAPPb secretedectodomain. Remarkably, the sAPPa knockin completelyrescues all of the phenotypes displayed by APP�/�mice (Ringet al. 2007). In addition, the sAPPa knockin prevents the post-natal lethality shown by APLP2/APP double knockout mice,but it cannot rescue the neuromuscular junction abnormalities,spatial memory impairments, and long-term potentiationdeficits of these mice (Weyer et al. 2011). Interestingly, thesAPPb knockin is not able to rescue either the APP single orthe APLP2/APP double knockout phenotypes, suggesting thatsAPPa is a more physiologically relevant APP fragment thansAPPb (Li et al. 2010). These results also suggest that bothAPP and APLP2 collaborate synergistically to supportsynaptic function in the CNS and PNS.It has also been reported that BACE1�/� mice display

retinal pathology involving lipofuscin accumulation anddegeneration of retinal cell layers, phenotypes that areassociated with reduced retinal microvasculature (Cai et al.

2012). In cell culture, vascular endothelial growth factorreceptor 1 is processed by BACE1, suggesting that thisBACE1 null retinal phenotype may result from insufficientvascular endothelial growth factor receptor 1 ectodomainshedding. However, at the recent BACE meeting severalgroups reported their unpublished data, which demonstratedthat the retinal phenotype is not seen in all of the differentlygenerated BACE1�/� mice. The molecular cause of thesedifferences is presently unclear.

Defects in axon guidance and neurogenesis in the CNS of

BACE1�/� mice

Another significant phenotype of BACE1�/� mice involvesaxon guidance defects in the hippocampus and olfactorysystem (Rajapaksha et al. 2011; Cao et al. 2012; Hitt et al.2012). The mossy fiber axon pathway from dentate gyrusgranule cells to CA3 pyramidal neurons is critical forlearning and memory and displays one of the highest BACE1levels in the brain. The majority of proximal mossy fiberaxons pass along the CA3 pyramidal cell layer on the sidewith primary dendrites. However, a subset of axons calledthe infrapyramidal bundle (IPB) runs along the opposite(axonal) side of the CA3 cell layer for a given distance beforeit crosses the cell layer to join the majority of axons in thestratum lucidum. IPB length is stereotypical for a givenmouse strain and correlates with memory performance:mouse strains with long IPBs obtain superior scores onhippocampus-dependent memory tests, whereas strains withshort IPBs perform poorly (Crusio and Schwegler 2005).Intriguingly, BACE1�/� mice have IPBs that are ~ 30%shorter than wild-type littermates (Hitt et al. 2012), anobservation consistent with BACE1 null memory deficits. Inaddition to short IPB length, BACE1�/� mice also displaypre-mature IPB axon cross-overs of the CA3 cell layer.In the olfactory system, olfactory sensory neurons (OSNs)

project axons from the olfactory epithelium to the olfactorybulb (OB). A given OSN expresses only one of ~1000different odorant receptor genes in a random pattern in theolfactory epithelium, yet it is able to send its axon to atopographically fixed glomerulus in the OB that representsthe odor quality of the odorant molecule ligand (Sakano2010). In wild-type mice, axon guidance of OSN axons tocorrect glomeruli is exquisitely precise. In contrast, BACE1null OSN axons exhibit mis-targeting to incorrect glomeruliin a dorsal to ventral gradient in the OB (Rajapaksha et al.2011; Cao et al. 2012), suggesting that the dependence ofolfactory axon guidance on BACE1 varies according to theexpressed odorant receptor. Although the significance of thisfinding is not yet clear, it implies that BACE1 inhibition maynot affect the axon guidance of all OSNs.The recent identification of the neural cell adhesion

molecule close homolog of L1 (CHL1) as a BACE1substrate in cultured neurons (Kuhn et al. 2012; Zhou et al.2012) has led to important insights into the molecular basis


BACE proteases 13

of BACE1-dependent axon guidance. CHL1 is a type Imembrane protein involved in axon outgrowth and neuronalsurvival (Naus et al. 2004). ADAM8 processing of CHL1releases a soluble ectodomain fragment that may interactwith neuropilin-1 and semaphorin 3A to influence axonguidance. Importantly, CHL1�/� mice have axon guidancedefects in the hippocampus and OB (Heyden et al. 2008;Montag-Sallaz et al. 2002) that are highly similar to thoseobserved in BACE1�/� mice (Hitt et al. 2012). These resultssuggest that release of soluble CHL1 ectodomain as a resultof BACE1 processing has an important signaling role in axonguidance of hippocampal mossy fibers and OSN axons thatmay explain axon mis-targeting in BACE1 null mice.Neurogenesis is a process critical for learning and memory

that occurs both during development and in the adult.Interestingly, BACE1�/� mice exhibit a decrease in hippo-campal neurogenesis that is accompanied by a correspondingincrease in astrogenesis during post-natal development (Huet al. 2013). This BACE1 null phenotype is associated withincreased levels of Notch 1 signaling and full-length Jagged1 protein, the Notch 1 ligand. In cell culture, BACE1 cleavesJagged 1, proving that it is a BACE1 substrate. Thus,BACE1 may regulate post-natal neurogenesis and astrogen-esis by modulating Notch 1 signaling.Taken together, the complex neurological phenotypes of

BACE1�/� mice suggest that BACE1 has diverse physio-logical functions in the CNS that result from deficientb-secretase processing of multiple BACE1 substrates. Animportant question remaining is whether these BACE1 nullphenotypes derive from lack of BACE1 during developmentor in the adult. If the former, then BACE1 inhibitor treatmentof AD adults may be relatively safe. However, processeswith ongoing operation in the adult, such as axon guidanceand neurogenesis, are impaired in BACE1�/� mice. Thissuggests that therapeutic inhibition of BACE1 may havemechanism-based side-effects. However, whether theobserved BACE1�/� phenotypes might have implicationsfor BACE1 inhibition as a future pharmacotherapy or not is amatter of ongoing discussion and will be discussed in moredetail below.

Additional BACE1 substrates in the CNS

The rapidly increasing number of phenotypes in BACE1�/�

mice suggests that more phenotypes are yet to be discovered.This assumption is in good agreement with the recentidentification of numerous additional BACE1 substrates inproteomic screens. A first study identified over 60 candidateBACE1 substrates in tumor cell lines over-expressingBACE1 (Hemming et al. 2009). Not all of them may bephysiological substrates, as over-expressed BACE1 is knownto cleave in cellular compartments, such as the ER, where itnormally is not localized (Vassar et al. 1999). Two recentstudies identified substrates in primary neurons expressing

endogenous BACE1 (Kuhn et al. 2012; Zhou et al. 2012).Zhou et al. (2012) identified 13 substrates from culturedneurons under serum- and protein-free conditions and useddifferential isotopic labeling of terminal amines and Ɛ-aminesfor quantification. In contrast, Kuhn et al. (2012) cultured theneurons in the presence of serum proteins using B27-containing media. That study developed the novel secretomeprotein enrichment with click sugars (SPECS) technologywhich utilizes metabolic glycan labeling and click chemistryto enrich glycoproteins out of B27 containing media ofprimary neurons to identify 34 BACE1 substrates.Besides APP and its homologs both studies identified the

novel BACE1 substrates L1, CHL1, contactin-2, Golgimembrane protein 1 (GLG1), and peptidyl amidating mono-oxygenase (PAM). The CHL1-dependent phenotype of theBACE1�/� mice has already been described above. FurtherBACE1 and also BACE2 substrates were identified in aproteomics study in pancreatic islets and insulinoma celllines (Stutzer et al. 2013), where both proteases are alsoexpressed. From these studies it became clear that BACE1does not only cleave a few neuronal membrane proteins, buthas a broad range of substrates within and outside of thenervous system.One of the proteomic studies quantified the extent to which

the cleavage of the substrates depends on BACE1 (Kuhn et al.2012). Some of the substrates, such as SEZ6, SEZ6L, andAPLP1 are ‘exclusive’ BACE1 substrates, because BACE1inhibition and BACE1-deficiency almost completely blockcleavage of these substrates. For other substrates totalcleavage was only partially reduced in the absence of BACE1activity, indicating that these substrates are not only cleavedby BACE1, but also by other proteases, which are expected tobe metalloproteases. Moreover, a compensatory increase inthe cleavage of some substrates by another protease mayoccur. This is the case for APP where ADAM10 compensatesfor loss of BACE1 cleavage (Colombo et al. 2012; May et al.2011). In addition, for some of the new substrates indirecteffects like trans-synaptic stabilization may potentially beresponsible for an apparently partial cleavage reduction uponBACE inhibition, such as has been shown in case of post-synaptic neuroligin 1 which binds to membrane bound pre-synaptic Neurexin-1 alpha and is increasingly cleaved uponthe addition of recombinant soluble neurexin 1 alpha whiletreatment with a BACE inhibitor leads to reduced amounts ofsecreted neurexin 1 alpha (Suzuki et al. 2012; Kuhn et al.2012; Boucard et al. 2005).Based on the current literature, the novel BACE1

substrates can be divided into two subgroups. The firstgroup comprises proteins that participate in synapse function,whereas the second group includes proteins that interact withthe surrounding extracellular matrix of astrocytes andoligodendroglia and thereby modulate axon outgrowth andpath finding as well as the formation of membrane microd-omains. Finally there are substrates with other functions.


14 R. Vassar et al.

Substrates involved in synapse function

The first group of identified BACE1 substrates containsneurexin 1 alpha, the neuroligin family and the latrophilinfamily. Neurexin-1 alpha and latrophilins are mostly pre-synaptic proteins which interact with each other in cis andwith post-synaptic neuroligins in trans, bridging the synapticcleft (Boucard et al. 2012). Although it is not yet clearexactly how BACE1 cleaves these proteins and how it alterstheir function, it is interesting to note that some of thephenotypes of neurexin 1 alpha and neuroligin-deficient miceare recapitulated in BACE1-deficient mice. This suggests thatsome of the BACE1-deficiency phenotypes may result fromthe reduced cleavage of neurexin 1 alpha or the neuroligins.Deletion of neurexin 1 alpha reduces excitatory synapticstrength (Etherton et al. 2009). Neuroligins in concert withneurexins are needed for synapse function but not for synapseformation (Xu et al. 2012; Bang and Owczarek 2013). Tripleknockout mice of neuroligin 1, 2, and 3 showed an abnormalrespiration behavior leading to post-natal death owing toreduced neurotransmission in the brainstem respiratorynetwork (Varoqueaux et al. 2006). Single knockouts ofneuroligins show more subtle phenotypes. Neuroligin 1knockout mice show reduced social interaction, impairedspatial working memory and reduced contextual- and cued-fear memory (Kim et al. 2008). Neuroligin-2 knockout miceshow increased anxiety-like behaviors, reduced pain sensi-tivity, motor coordination, and an irregular breathing pattern,whereas neuroligin 4 mice have symptoms of autismspectrum disorders (Xu et al. 2012; Jamain et al. 2008;Blundell et al. 2009). Neurexin 1 alpha knockout mice showimpaired nest building behavior and spend more time inapproaching other mice which might be because of impairedinterpretation of social cues or increased aggression behavior(Grayton et al. 2013). A lot of these behavioral phenotypeshave as well been observed in BACE1 knockout mice andmight imply that similar signaling pathways are impaired inthese mice (Laird et al. 2005; Savonenko et al. 2008).

Substrates involved in axon outgrowth, axoglial

interactions, and myelin microdomain structure

The second group of substrates is involved in axonaloutgrowth, axoglial interactions and electric transmission.This group includes CHL1, L1, contactin-2, and the SEZ6family. The role of BACE1 cleavage for CHL1 function hasalready been described above. The glycosyl phosphatidylinositol-anchored protein contactin-2 has been confirmed as aBACE1 substrate in vitro and in vivo. Contactin-2 expressedon Schwann cells and oligodendroglia interacts with thetransmembrane protein CaspR2 and Contactin-2 on theaxonal membrane to maintain the voltage-gated potassiumchannels Kv1.1 and Kv1.2 in juxtaparanodes of myelinatedaxons (Poliak et al. 2003; Traka et al. 2003). Contactin-2knockout leads to impaired learning behavior, shorterinternodes, and disrupted juxtaparanodes (Savvaki et al.

2008). Thus, it appears possible that BACE1-mediatedproteolytic processing of contactin-2 modulates the concen-tration and function of juxtaparanodal potassium channels.This possibility needs to be tested in further experiments.Sez6 family proteins are type I membrane proteins

localized to both dendritic and axonal compartments. Sez6knockout mice display increased dendritic branching,reduced spine density on apical dendrites, and impairedexcitability of layer V pyramidal neurons (Gunnersen et al.2007). The triple knockout of the whole Sez6 family leads tomotor coordination deficits reflected in electrophysiologicalalterations in Purkinje cell excitation in the cerebellum(Miyazaki et al. 2006). All three members of the Sez6 familyare BACE1 substrates but how BACE1 modulates theirfunctions is not yet clear as little is known about theirphysiological roles in vivo.PAM is one of the candidate BACE1 substrates that has

shown up in all four proteomic screens and has been shownto undergo regulated intramembrane proteolysis by anunknown sheddase and by c-secretase in tumor cells(Rajagopal et al. 2010). This copper-dependent enzymeamidates the C-terminus of a number of neuropeptides,which is necessary for their biologic activity (Bousquet-Moore et al. 2010). PAM-deficiency leads to intrauterinedeath caused by malfunction of the cardiovascular systemand cerebral edema that result from non-functional pro-aminoterminal peptide (PAM) and adrenomedullin becauseof their lacking amidation (Czyzyk et al. 2005). In contrast,BACE1-deficient mice may show the opposite phenotype asBACE1 negatively regulates PAM function by PAM cleav-age. Given that BACE1 expression is particularly high in thebrain, whereas PAM is expressed more ubiquitously, butmainly in the pituitary gland, brain, and the atrioventricularnode (Braas et al. 1989), it may be possible that BACE1-deficiency mostly affects PAM function in the brain, but notnecessarily in other organs.

Functional Roles of BACE2

Soon after the discovery of BACE1, a paralog termedBACE2 was identified (Bennett et al. 2000a; Farzan et al.2000). BACE2 was shown to function like an “a-secretase”in promoting the non-amyloidogenic processing of APP(Basi et al. 2003; Farzan et al. 2000; Fluhrer et al. 2002;Yan et al. 2001). In contrast to BACE1, little attention thushas been devoted to the study of BACE2, until recently.Initial studies in BACE2 knockout mice documented noapparent phenotype, except that the early post-natal lethalityobserved in BACE1�/� mice is enhanced in BACE1�/�;BACE2�/� double knockout mice (Dominguez et al. 2005).Using mouse and zebrafish model systems, investigatorshave made significant advances recently to identify newsubstrates of BACE2 and disclose novel physiological rolesof this aspartyl protease.


BACE proteases 15

Processing of a pigment protein by BACE2 to form the

amyloid matrix in melanosomes

The process of pigmentation begins with events that occur inmelanosomes distributed throughout melanocytes (Hearing2005). These endosomal organelles contain all the compo-nents required for melanin formation, including the enzymetyrosinase (TYR) and accessory proteins including TYR-related protein 1 and TYR-related protein 2 (Hearing 2005).Melanosomes undergo a four-stage maturation process that isdependent upon the structural protein termed pigment cell-specific melanocyte protein (PMEL). Mature PMEL issynthesized as an integral membrane protein enriched inthe lumen of melanosomes. Proteolysis of PMEL subse-quently leads to the structural rearrangement of the melano-some from an immature, non-pigmented, round organelle to amature, pigmented, ellipsoid melanosome (Watt et al. 2009).A stage I melanosome is characterized by its non-descriptround, non-pigmented shape that resembles an endosome.Proteolytic cleavage of PMEL by a putative protease(s)within the melanosome releases an N-terminal fragment,termed Ma, competent for formation of the scaffolding, oramyloid matrix, which elongates the melanosome (stage II;elliptically shaped). Stage III commences with the depositionof eumelanin, a black/brown pigment, produced by TYR,upon the PMEL scaffold. By stage IV, the internal melan-osomal structure is obscured because of the abundance ofdeposited eumelanin. Without the amyloid matrix, noeumelanin is able to accumulate and only pheomelanin, ared/yellow pigment, is observed in round melanosomes(Raposo and Marks 2007). Variations in the eumelanin/pheomelanin ratio account for the wide variety of coat colorin all organisms. Interestingly, natural mutations in Pmelfound in a variety of organisms, including mice, chicken, andhorse (Theos et al. 2006; Kerje et al. 2004; Brunberg et al.2006), shared a common phenotype: coat color dilution.While proteolytic processing of PMEL is thought to be a

critical step in the formation of the amyloid matrix, theenzyme(s) responsible remained elusive. Studies inBACE2�/� mice in which one of the BACE2 domainscontaining a catalytic aspartic residue is deleted of anotherwise full length protein revealed dilution of coat colorsuggesting that BACE2 could be a candidate (Dominguezet al. 2005) (Fig. 4a). As this pigmentation phenotyperesembling mice harboring mutations in Pmel observed inBace2�/�, but not Bace1�/� mice (Rochin et al. 2013),Rochin and coworkers addressed whether BACE2 could be asheddase that is responsible for cleaving PMEL for properformation of scaffold in melanosomes for pigment deposi-tion. These investigators provided several lines of evidencein favor of this idea. In support of the notion that BACE2resides in the same compartment as that for processing ofPMEL, co-localization studies showed that BACE2 istargeted to early stages (I and II) of PMEL-containingmelanosomes and associated with PMEL (Rochin et al.

2013). Secondly, both gain- and loss-of-function approachesin cultured cells revealed that BACE2 was able to cleavePMEL to release its ectodomain containing the amyloido-genic Ma-fragment in cultured cells using both gain- andloss-of-function approaches (Rochin et al. 2013). Finally,melanogenesis is impaired in Bace2�/� mice (Rochin et al.2013). Taken together, these findings are consistent with amodel in which PMEL is cleaved by BACE2 to release theluminal fragment into endosomal precursors for its formationinto fibrils and finally melanosome maturation. AlthoughBACE1 apparently does not participate in the processing ofPMEL as BACE1�/�;BACE2�/� double knockout miceexhibit a similar pigmentation phenotype comparable toBACE2 null mice, a caveat is that the BACE2 null micecarry a catalytically impaired BACE2 protein (rather than theabsence of BACE2 protein). Such a mutant protein could actin a dominant-negative fashion, obscuring the potential roleof BACE1 in PMEL processing. BACE2 null mice lackingthe complete BACE2 protein will be useful in future studiesto clarify this issue.

Role of BACE2 in migration of melanocytes

Recent studies in zebrafish revealed a surprising role ofBACE2 in the regulation of melanocyte migration. Thereexists one human BACE2 ortholog in the zebrafish genome,zBACE2 and its mRNA is expressed throughout earlydevelopment (van Bebber et al. 2013). Capitalizing on a

(a)

(b)

Fig. 4 Beta-site amyloid cleaving enzyme (BACE)2 regulates pigmen-tation in mice and melanophore migration in zebrafish. (a) Deletion ofBace2 in mice leads to dilution of coat color (Rochin et al. 2013). Note

the black coat normally seen in the C57Bl/6 strain as observed incontrol and Bace1 knockout mice. (b) Deletion of Bace2 in zebrafishalters the shape and migration of melanophores (van Bebber et al.

2013). Whole fish (left panels); boxed areas are enlarged to showdetails near end of yolk sac extension (middle panels) and fin (rightpanels).


16 R. Vassar et al.

collection of previously ENU (N-ethyl-N-nitrosourea)-mu-tagenized alleles, a C to A conversion in zBace2 wasidentified; this mutant led to a premature in-frame stop codonthat truncated a large portion of the protein, including bothcatalytic domains (van Bebber et al. 2013). Thus, anN-terminal 79 amino acid peptide is predicted to begenerated in such mutants (zBace2+/�). Cross-breeding ofzBACE2+/- zebrafish led to the generation of homozygousmutant alleles (zBACE2�/�). Although zBACE2�/� larvaewere indistinguishable from their zBACE2+/- or zBACE2+/+littermates before the third day post-fertilization, it becameclear later in development that melanophores were moredilated in larvae lacking Bace2. In addition, the migrationpattern of these melanophores around the tail fin and the yolksac extension was perturbed in zBace2�/� animals (Fig. 4b).As melanophores are neural crest derivatives, this phenotypeis consistent with the expression pattern of Bace2 in neuralcrest cells (Thisse et al. 2004).To assess whether compensatory mechanisms exist in

regulation of migration of melanocytes between zBACE1and zBACE2, zBACE1�/�; zBACE2�/� zebrafish weregenerated. Comparing defects in melanophores and melano-cyte migration observed in zBACE2�/�

fish, these abnor-malities were not enhanced in the zBACE1�/�;zBACE2�/�

animals, indicating that zBace2 serves distinct roles thanthose of zBace1 in term of melanocyte migration (vanBebber et al. 2013). These data support the notion thatBACE1 and BACE2 serve distinct physiological roles duringdevelopment and aging. Importantly, this zebrafish model isa valuable tool to assess in vivo the specificity of BACEinhibitors to identify BACE1-selective and BACE2-sparingcompounds in the event that such compounds are deemednecessary to limit mechanism-based toxicity occurring inclinical trials of Alzheimer’s disease.

Regulation of pancreatic b cell mass and function by BACE2

Hyperglycemia occurring in type 2 diabetes results from thefailure of the pancreas to secrete appropriate amounts ofinsulin to match metabolic requirements associated withinsulin resistance (Kahn et al. 2006). Two major abnormal-ities observed in this illness are progressive degeneration ofbeta cell function and decrease in pancreatic beta cell mass(Kahn et al. 2006). However, the molecular mechanismunderlying beta cell loss is not clearly understood. Proteo-lytic processing of growth factors and their receptors of bcells are thought to be important for proper autocrine andparacrine signaling. One example is the type I transmem-brane protein Tmem27 (pro-proliferative plasma membraneprotein), which is a target gene of the transcription factorTcf1 (or hepatocyte nuclear factor 1a), that is linked to themost common form of maturity onset diabetes of the young(MODY3)(Shih et al. 2001). Other than the kidney (Zhanget al. 2001), expression of Tmem27 has only been docu-mented in the pancreas (Akpinar et al. 2005). Thought to

stabilize apical amino acid transporters and critical for re-absorption of amino acids (Danilczyk et al.2006;Malakauskaset al. 2007), Tmem27when abnormally accumulated inb cellscauses an increase in pancreatic cell mass (Akpinar et al.2005). Importantly, ectodomain shedding of Tmem27 regu-lates its abundance and activity in b cells as the holoproteinretains mitogenic activity (Akpinar et al. 2005). Thus, iden-tification of the protease responsible for the shedding ofTmem27 may provide an opportunity to elevate Tmem27levels as a strategy to modify b cell mass and function.To identify the responsible protease, Esterhazy and

coworkers employed a siRNA approach to screen all majorclasses of proteases in a mouse insulinoma cell line. In theirinitial screen, BACE2 was identified as a sheddase forTmem27 in both mouse and human beta cells (Esterhazyet al. 2011). Importantly, when insulin-resistant mice weretreated with a BACE2 inhibitor, these coworkers observedincreased beta cell mass and improved control of glucosehomeostasis as a result of increased insulin levels (Esterhazyet al. 2011). To complement this observation, these inves-tigators took advantage of the availability of BACE2knockout mice with inactive BACE2 (Dominguez et al.2005) to examine Tmem27 processing and physiologicalconsequences in b cells of these mutant mice. Remarkably,Esterhazy and coworkers observed that these Bace2 inactivemice have significantly lower levels of blood glucose,improved glucose tolerance and elevated b cell mass,implicating BACE2 as the sheddase that is important forthe regulation of b cell maintenance (Esterhazy et al. 2011).These data thus are consistent with the notion that Bace2controls beta cell mass and function and provide a strategy toinhibit BACE2 as a potential therapy in efforts to improvehuman b cell function.Finally, Stoffel and coworkers used a proteomic approach

in a pancreatic cell line to identify Seizure 6 protein familymembers, SEZ6L and SEZ6L2, which are islet-enriched cellsurface proteins as BACE2-specific substrates (Stutzer et al.2013). However, these substrates are cleaved by BACE1 inneurons (Kuhn et al. 2012). Why SEZ6L and SEZ6L2 areselectively cleaved by BACE2 over BACE1 in b cells is notcompletely clear. Likely possibilities include organ-depen-dent abundance of enzyme, substrate isoform or cleavage sitepreferences, and subcellular compartmentalization of eachsheddase. These investigators also identified IGF2R andSORT1 as substrates of BACE2 in islet cells (Stutzer et al.2013).In summary, over the past few years, significant progress

has been made in the identification of novel substrates ofBACE2 and revealed that some of its physiological rolesappear to be unique and distinct from those of BACE1. Thiswill help the development of tools for evaluating BACE1inhibitors that are selective over BACE2, thus potentiallyproviding improved compounds that are safe and effectivefor the treatment of AD.


BACE proteases 17

BACE inhibition as a treatment strategy for AD

Drug discovery scientists often invest years of research efforton hypotheses despite significant uncertainties owing toincomplete understanding of the fundamental biology ofdisease and the potential for entirely unanticipated and ofteninsurmountable non-mechanism-based side-effects thatemerge late in the development of a novel therapeutic.Therefore, it is critical that efforts are focused on hypothesesthat are supported by the strongest possible data which isoften derived from human genetics (Plenge et al. 2013) as isthe case with the amyloid hypothesis of AD as reviewedabove (Loy et al. 2013). The recently described rareprotective b-secretase-cleavage-sparing variant of APP(APP-A673T) further supports the rationale for BACE1inhibition in AD (Jonsson et al. 2012). Our understanding ofthe pathophysiological sequence of AD is incomplete butfindings to date from ongoing natural progression studiesconducted by the Alzheimer’s Disease Neuroimaging Initia-tive and the Dominantly Inherited Alzheimer Network(DIAN) support that Ab plaque deposition is likely an earlyand potentially initiating event of AD that drives thesubsequent development and/or expansion of tau pathologyand inflammation that together contribute to the neurotoxicenvironment that causes synapse dysfunction, cell loss, anddementia (Bateman et al. 2012; Jack et al. 2012). While thisinterpretation of existing literature supports that earlyintervention with CNS Ab lowering therapeutics will havethe best chances for success, multiple studies have shown thedirect neurotoxic effects of Ab oligomers and plaques whichcould be operant at all stages of AD such that substantiallyreducing the influx of newly synthesized Ab peptides toreduce Ab oligomer formation and plaque growth could bebeneficial even in later-stage symptomatic patients (Walshand Selkoe 2007).After over a decade of AD clinical research on amyloid-

directed agents, has the amyloid hypothesis been adequatelytested? Most clinical AD investigators will answer this withan unqualified, “No”. The primary reason given is that theamyloid-directed agents tested to date (e.g., c-secretaseinhibitors, c-secretase modulators, plaque disruptors, andanti-Ab immunotherapy) have failed to convincingly dem-onstrate an unambiguous and substantial reduction in CSFAb levels; CSF Ab is the primary biomarker of soluble brainAb status in humans (Wan et al. 2009; Blennow 2010). Forexample, the strategy of c-secretase inhibition, while stronglysupported by human genetics and discovery of compoundscapable of significantly reducing CNS Ab levels in pre-clinical studies (Best et al. 2007) and humans such assemagacestat (Bateman et al. 2009) and avagacestat (Tonget al. 2012), ultimately failed in Phase 3 efficacy studiesowing to the emergence of serious dose-limiting and thusCNS Ab-lowering limiting mechanism-based side-effectsincluding cancerous skin lesions, severe gastrointestinal

toxicity, and ultimately cognitive worsening that were likelyrelated to inhibition of Notch processing or other non-APPc-secretase substrates (Doody et al. 2013; Coric et al. 2012).Indeed the dose limiting toxicities associated with c-secretaseinhibitions did not support a robust test of the amyloidhypothesis as only a transient ~ 25% lowering of CSF Abfrom baseline was achieved in Phase 3 efficacy trials. Theconfounds arising from the combination of serious mecha-nism-based side-effects and the likely minimal impact ofsemagacestat or avagecestat on CSF Ab levels at the dosestested unfortunately limits the impact of these findings on ourunderstanding of the validity of the amyloid hypothesis(Pomara 2013) or importantly the minimum levels of Ablowering that will have a benefit. Genetic studies in mice andhumans suggest that moderate reductions in CNS Ab levelscan result in long-term benefits. CNS Ab levels in BACE1wild-type, heterozygous and knockout mice unexpectedlyrevealed that 50% knockdown of CNS BACE1 protein inBACE+/� mice resulted in only ~ 12–20% reduction in CNSAb level versus wild-type animals (McConlogue et al.2007). Despite this modest effect on Ab, aged PDAPP/BACE1+/� mice developed about 75% less amyloid plaqueburden than similarly aged PDAPP/BACE1+/+ mice(McConlogue et al. 2007). From in vitro cell studies, therare APP-A673T protective variant was associated with a40% reduction in Ab production compared to APP-WT cellsand an 8-fold reduced risk for AD (Jonsson et al. 2012)again suggesting that a relatively moderate reduction in Abproduction afforded by reduced BACE1 cleavage of APPmay confer a long-term benefit. As PDAPP/BACE+/- miceand APP-A673T human carriers experience a lifelongreduction in steady-state CNS Ab, it is difficult to inferwhether similar levels of Ab lowering achieved with BACEinhibition would benefit AD patients. Importantly the degreeof CNS Ab reduction required for a meaningful clinicalbenefit at any given point of the AD continuum remains anunknown and will require mechanisms and molecules thatare capable of safely achieving the dose–response range ofAb lowering in AD patients.

BACE inhibition – therapeutic potential of smallmolecules and biologics

The molecular identification of BACE1 and the subsequentreports of BACE1 knockout mice revealed an atypicalmember of the aspartic protease family that combined astrong therapeutic rationale for inhibitors to treat AD basedon requirement of BACE1 for Ab synthesis and the lack ofovert negative phenotypes in multiple reports of BACE1knockout mouse lines (Roberds et al. 2001; Cai et al. 2001;Luo et al. 2001). Aspartic proteases are a therapeuticallyimportant class of enzymes that includes the HIV proteaseand the mammalian renin enzymes that possess structuralhomology with the BACE1 active site (Hong et al. 2000).


18 R. Vassar et al.

Renin presented major medicinal chemistry challenges andonly after decades of effort was a clinically effective renininhibitor approved for blood pressure control (Jensen et al.2008) whereas HIV protease inhibitors are a transformationalcore component of HIV therapy (Ray et al. 2010). Despitethese successes and significant active site structural homol-ogy among aspartic proteases, the peptidomimetic nature ofclassical aspartic protease inhibitor scaffolds have proven tobe incompatible with the chemical and physical propertiesrequired for a clinically developable inhibitor of the CNS-localized BACE1 protease (Ghosh et al. 2012). Earlyexamples of peptidomimetic BACE inhibitors possessedhigh affinity and selectivity for BACE enzymes andproduced significant reductions in plasma Ab but achievedmodest reductions in brain Ab in mice (Chang et al. 2004;Hussain et al. 2007; Sankaranarayanan et al. 2008) and non-human primates (Sankaranarayanan et al. 2009) oftenrequiring exposure-boosting strategies. CTS-21166 is apeptidomimetic BACE inhibitor, derived from the earliestchemistry efforts on BACE inhibitors, that advanced intoPhase 1 clinical trials in healthy volunteers and producedsignificant and sustained reductions of plasma Ab (80% peakinhibition) following intravenous delivery however effects onCSF Ab have not yet been reported and its developmentstatus remains unknown (Albert 2009). A novel biologicsbased approach to achieving highly selective BACE1inhibition relies on a bi-functional inhibitory antibody thatbinds both the transferrin-receptor to improve blood-brainbarrier penetration and the BACE1 enzyme domain. Thisantibody potently and selectively inhibits purified BACE1and BACE1-meditated processing of APP in cells and hasachieved impressive proof of concept in vivo efficacy resultsin transgenic APP mice (Atwal et al. 2011). The affinityoptimized transferrin-receptor binding function improvedantibody uptake into the brain by ~5-fold which was criticalfor the efficacy of this approach (Yu et al. 2011). Results todate are encouraging that this approach could serve asalternative to small molecule inhibitors of BACE1 or otherCNS-localized intracellular enzymes. Importantly, thisapproach achieves exquisite selectivity for the BACE1enzyme over other aspartyl proteases including BACE2. Ina broader context, the blood–brain barrier enhancing strat-egies utilized for anti-BACE1 immunotherapy may also beharnessed to improve the performance of other CNS passiveimmunotherapies such as anti-Ab and anti-tau approaches ifcarrier-dependent mechanism-based side-effects can beavoided (Couch et al. 2013).Breakthrough discoveries in small molecule inhibitors of

BACE came with the identification of novel acylguanidine,aminopyridine, and isothiourea pharmacophores that boundwith modest affinity to the catalytic aspartic acid residues ofBACE1 (Cole et al. 2006; Congreve et al. 2007; Yang et al.2009b; Wang et al. 2010). These molecules emerged fromdiverse hit finding strategies using high-throughput X-ray

crystallography (Congreve et al. 2007), functional screens(Cole et al. 2006), or a protein NMR-based fragment screenusing an unlabeled soluble BACE1 construct (Geschwindneret al. 2007) or an 15N-labeled human BACE1 solubleenzyme domain (Wang et al. 2010). For the latter screen,the initial hit Merck D, a weak isothiourea structure thatpossessed a BACE1 binding Kd of ~ 15 lM and a functionalBACE1 IC50 of ~ 200 lM (Fig. 5c), (Wang et al. 2010)launched a structure-based design effort that led to thediscovery of the novel iminoheterocyclic class of asparticprotease inhibitors with more drug-like pharmacophore asexemplified by the iminohydantoin Merck E (Fig. 5c).Iminohydantoin BACE inhibitors possessed an unprece-dented combination of low molecular weight and highaffinity binding compared to prior peptidomimetic inhibitorsin a structure that at the time of its discovery was unknown inthe aspartic protease inhibitor literature, (Zhu et al. 2010).These efforts relied heavily on BACE1 X-ray crystallographyand computational modeling to aid rational chemical designdecisions that achieved high affinity binding coupled withrobust oral pharmacokinetic and CSF Ab lowering efficacyin rodents as exemplified by Merck F (Cumming et al.2012). Subsequent medicinal chemistry efforts pursued ringexpansion of the iminohydantoin warhead to generate thenovel 6-membered C5-substituted iminopyrimidinone core,Merck J in Fig. 5c, which again maintained a low molecularweight with high affinity binding and a minimal shift incellular potency (Stamford et al. 2012). The iminopyrimid-inone SAR optimization focused on defining the appropriatecombination of in vitro pharmacological properties withpotent and efficacious CSF Ab lowering in rats (Stamfordet al. 2012; Stamford and Strickland 2013). The novel cyclicisothiourea core BACE inhibitor, LY2811376 Fig. 5b (Mayet al. 2011), that displayed modest intrinsic affinity forpurified BACE1 (IC50~250 nM), more potent cellular activ-ity and robust CNS Ab lowering in rodents and dogs and wasadvanced into Phase 1 studies where it supported the firstpublished report of BACE inhibitor-mediated lowering ofCSF Ab in healthy volunteers (May et al. 2011). Develop-ment of this compound was terminated owing to observa-tions of retinal and neuronal and glial cell degenerationassociated with intracellular accumulations of autofluores-cent granules following 3 month dosing in rodent toxicologystudies (May et al. 2011). These effects manifested in theeye as retinal pigment epithelium cell hypertrophy anddegeneration. The histopathological findings resulting fromchronic LY2811376 treatment were retained in BACE1knockout mice implying that the effects were not likelyBACE1 dependent (May et al. 2011). The histopathologicalfindings described by May et al. are reminiscent of findingsreported for cathepsin D knockout mice (Saftig et al. 1995;Shacka et al. 2007) and suggest that the moderate selectivityof LY2811376 for BACE1 over cathepsin D of ~ 60-fold(May et al. 2011) is not sufficient to support chronic dosing.


BACE proteases 19

The importance of cathepsin D selectivity is further under-scored by rare cathepsin D homozygous loss of functionmutations in humans that manifests as a subtype of thelysosomal storage disorder, Batten’s disease, characterizedby seizures, severe developmental defects, and neurodegen-eration (Siintola et al. 2006). Most genetic forms oflysosome dysfunction manifest with neurohistopathologicalfindings of aberrant intracellular accumulation of ceroid-lipofuscin deposits (Dawson and Cho 2000). Importantly, themost advanced BACE inhibitors currently in clinical trialshave been described at recent scientific meetings to bevirtually inactive at cathepsin D.The recent proliferation of novel substrates and proposed

biological functions of BACE1 and BACE2 that are

reviewed herein have tempered the early expectations thatchronic BACE inhibitor therapy would be devoid of anypotential for mechanism-based safety concerns. Insights intothe impact of reduced versus complete loss of BACE1activity during development or adulthood can be gleanedfrom comparing and contrasting the impact of varying levelsof chronic pharmacological inhibition of BACE1 activity inadult animals to BACE1 knockout mouse phenotypes(Willem et al. 2006; Hu et al. 2006; Savonenko et al.2008; Laird et al. 2005; Hitt et al. 2012; Cao et al. 2012;Cai et al. 2012; Cheret et al. 2013). Phenotypes reported forBACE1 knockout mice to date reflect the impact of a lifelong50% reduction or complete loss of BACE1 function whilepharmacological inhibition of BACE activity is targeted for

NNH

NH

OCl

Merck E (MW:293)IminohydantoinNMR Kd = 120 μMBACE1 IC50~ 300 μM(Zhu-2010)

Merck F (MW: 344)IminohydantoinBACE1 Ki = 6 nMCell IC50: 80 nMRat CSF Aβ40 ED50 ~ 30 mg/kgb/p = 0.5(Cumming-2012)

HN N

NH

OSN

Cl

NHN

HN

O

N

Merck D (MW:272)BACE1 NMR Fragment Screen HitKd (NMR) ~ 15 μMBACE1 IC50 > 200 μM(Wang-2010)

NH2S

NH

ClO

Novel, more drug-like warhead

Second novel warhead

GSK188909 (MW~600) BACE1 IC50: 5 nM; Cell IC50: 30 nM APPswe

• ~20% brain Aβ40 lowering at 250 mg/kg in TASTPM mice

(Hussain-2007)

(a) Optimized peptidomimetic inhibitor (b) First generation cyclic isothiourea inhibitor

LY2811376 (MW: 320) BACE1 IC50: 250 nMCell IC50: 300 nM

β(May-2011)

(c) Design and evolution of iminheterocyclc BACE inhibitor series

Merck J (MW 372)IminopyrimidinoneBACE1 Ki: 1.7 nMCell Aβ40 IC50: 10 nMRat CSF Aβ40 ED50 ~6 mg/kgb/p = 3(Stamford-2012)

HN S

NH

F

N

N

F

NH

HN

O OHF

NH

NSO2

CF3

Fig. 5 (a) Example of an optimized peptidomimetic b-site amyloidcleaving enzyme (BACE) inhibitor with high molecular weight and

marginal in vivo Ab lowering activity at very high doses of 250 mg/kg(Hussain et al. 2007). (b). Optimized cyclic isothiourea BACEinhibitor derived from a fragment-based screening approach (Mayet al. 2011). LY2811376 displays modest BACE1 potency but with

robust in vivo Ab lowering activity in animals and normal healthyvolunteers in Phase 1 studies. LY2811376 development wasterminated during Phase 1 owing to off target toxicities in eye and

brain. (c) Merck iminoheterocyclic BACE inhibitor series evolved from

a low affinity isothiourea hit (Merck D) identified in a protein NMR-based screen. Replacement off the isothiourea warhead with the

more drug-like iminohydantoin core (Merck E) led to further optimi-zation (Merck G) and ring expansion of the warhead to producemolecules like Merck J with several orders of magnitude improve-ments in intrinsic affinity with relatively small increase in molecular

weight. MK-8931 builds upon the favorable properties described herefor the iminoheterocyclic BACE inhibitor series with significantimprovements in intrinsic affinity, selectivity, in vivo potency, and

other key properties.


20 R. Vassar et al.

adults and at doses that do not achieve complete sustainedinhibition. Ongoing Phase 3 clinical trials of the BACEinhibitor MK-8931 are evaluating doses that achieve a rangeof 50 to 75% sustained reduction in CSF Ab. Several groupshave shown that the majority of phenotypes reported forBACE1 knockout mice are not observed in BACE1 hetero-zygous mice suggesting that a 50% loss of BACE1 functionwould be tolerated. In the course of the BACE inhibitordiscovery program at Merck Research Labs a number ofchronic BACE inhibitor studies have been conducted(M. Kennedy et al., manuscript in preparation). Thesestudies include: (i) chronic treatment of transgenic APPmice to examine the impact of Ab lowering on the course ofamyloid pathology and (ii) toxicology studies in multiplespecies at supra-therapeutic doses to assess short- and long-term safety and tolerability. These studies have evaluatedstructurally diverse inhibitors including MK-8931. Theperipheral nerve hypomyelination phenotype reported inBACE1 knockout mice was not observed in Tg-CRND8-APPSwedish/Indiana mice following 4 months of chronic phar-macological inhibition at doses of a BACE inhibitor thatproduced a sustained ~ 80% brain Ab lowering1. Otherphenotypes examined in these same studies such as reducedpre-pulse inhibition were also not impacted by 4 monthchronic BACE inhibitor treatment1. As described above,retinal degeneration was observed with chronic LY2811376treatment in rodents, however, studies by this same group inBACE1 knockout mice showed this finding to be indepen-dent of BACE1 (May et al. 2011). Subsequently, a non-progressing retinal degeneration phenotype was described inyoung BACE1 and BACE2 knockout mice suggestingBACE activity may be important for retinal homeostasisearly in life (Cai et al. 2012). However, retinal degenerationphenotypes have not been reported for all BACE1 andBACE2 knockout lines indicating that this phenotype may bestrain dependent. Taken together the emerging data suggestthat some functions of BACE1 may be more importantduring development than in adulthood which is consistentwith the dramatic drop off in CNS expression of BACE1 inmice during the first few weeks of life (Willem et al. 2006)Exploring the relationship between genetic deletion ofBACE1 versus pharmacological BACE inhibition in youngversus adult animals remains an area of intense effort andwill require collaborations that will benefit from the sharingof potent and selective BACE inhibitors. The lack ofinhibitors with substantial selectivity between BACE1 andBACE2 complicates the comparison of pharmacologicalstudies versus knockout studies. However, to date themajority of phenotypes reported for BACE1 and BACE2knockout mice appear to be non-overlapping (Dominguezet al. 2005). Beyond the role of BACE1 in AD, recentstudies using knockouts and small molecule inhibitorssuggest a potential positive impact of reduced BACE activityin traumatic brain injury (Chami and Checler 2012) and

peripheral nerve damage (Farah et al. 2011) (BACE1)whereas type 2 diabetes may benefit from inhibition ofBACE2 (Esterhazy et al. 2011).

BACE inhibitors in clinical trials

Several BACE inhibitors have progressed into Phase 1studies. The BACE inhibitors that are known to haveadvanced the furthest in clinical development areLY-2886721 (Eli Lilly), E-2069 (Eisai), and MK-8931(Merck & Co.). Structures and detailed in vitro and in vivoprofiles of these compounds have not been published in thepeer-reviewed literature as this requires the release ofproprietary chemical structures, however, these compoundshave been extensively described at scientific meetings. Allthree of these molecules are high affinity and selectiveinhibitors of the human BACE1 enzyme and potently reduceperipheral and CNS Ab1-40, Ab1-42, and sAPPb levels inrodents and large animal species such as non-human primates(E-2069 and MK-8931) or dogs (LY-2886721). While thesecompounds are highly selective for BACE1 over othermammalian aspartyl proteases such as cathepsins D and Eand renin they are equipotent inhibitors of the humanBACE2 enzyme. BACE2 does not contribute to b-secretasecleavage of APP in vivo (see above) and possesses highactive site identity with BACE1 such that achievingsubstantial (≥ 100-fold) selectivity over BACE2 whilemaintaining optimal CNS drug properties has proven chal-lenging. These pre-clinical pharmacodynamic profiles havetranslated well in Phase 1 studies where significant andsustained CSF Ab lowering of ≥ 80% following single andmultiple doses to healthy volunteers (LY-2886721, E2069and MK-8931) and AD patients (MK-8931 and E-2069) hasbeen reported. The development of LY-2886721 wasrecently terminated owing to observations of abnormal livertests indicative of liver toxicity in a chronic Phase 2 study ofAD patients. These findings were noted as unexpected basedon the lack of any liver abnormalities in pre-clinical safetytesting of LY-2886721 and were not considered related toBACE1 or BACE2 inhibition. The remaining BACE inhib-itors, E-2069 and MK-8931 are progressing into long-termclinical efficacy studies in patients. Eisai has noted itscommitment to explore the efficacy of E-2069 in early ADpatients but has not disclosed the design or status of this trial,whereas Merck started a registration supporting Phase 2/3safety and efficacy study of MK-8931 in mild-moderate ADpatients in November of 2012. This study contains a lead-insafety cohort of 200 patients that received once daily doses ofplacebo, 12, 40, or 60 mg of MK8931 for 3 months. Themain Phase 3 efficacy cohort will examine the impact of twodose levels of MK-8931 that will test the impact of 50% or75% CSF Ab lowering respectively on Alzheimer’s DiseaseAssesment Scale-cog and Alzheimer’s Disease Co-operativeStudy - Activities of Daily Living Inventory end points


BACE proteases 21

following 18-months treatment (for study details seeNCT0173934; clinicaltrials.gov). A second parallel Phase 3study of MK-8931 in prodromal AD patients commencedlate in 2013 and will test the impact of MK-8931 on CDR-SBfollowing 24 months of treatment at the same dose levelsused in mild-moderate patients. Inclusion criteria for theprodromal AD trial require that a patient has a positive18F-Flumetamol (VizamylTM, GE Healthcare, ArlingtonHeights, IL, USA) Positron emission tomography scan orhas a positive CSF tau/Ab42 biomarker profile (for studydetails see NCT01953601; clincaltrials.gov).Given the robust levels of CNS Ab lowering achieved with

BACE inhibitors such as MK-8931 it appears that a robusttest of the amyloid hypothesis is underway. While a positiveoutcome is sorely needed for patients, even a negativeoutcome will provide critically needed clinical evidence thatwill allow AD researchers to focus more effort on alternativetherapeutic mechanisms as well as novel clinical strategiesaimed at studying the earliest pre-symptomatic stages of whatwe recognize as a decades long disease process.

Outlook

Since the identification of BACE1 and BACE2 in 1999 wehave seen tremendous progress in BACE protease research,which has further intensified over the past few years as it ishighlighted in this review article. Basic research demonstratesa continuously increasing spectrum of physiological functionsof BACE1 and BACE2 both within and outside of the nervoussystem. Over the next few years we are likely to see moreBACE protease functions being discovered. The recentproteomic studies have identified a large number of additionalBACE protease substrates. Their validation in vitro and in vivowill help to uncover additional BACE protease functions. Onechallenge for the future is to link the novel substrates to thephenotypes in the BACE-deficient mice and thereby under-stand novel BACE protease functions at the molecular level.It turns out that – similar to APP - many BACE substrates

are not only cleaved by BACE1 or BACE2, but also by otherproteases, in particular metalloproteases. Thus, anotherquestion to answer is whether the different proteases haveredundant functions for an individual substrate as it seems tobe the case for neuregulin. Alternatively, substrate cleavageby either BACE or a metalloprotease may lead to a differentfunctional outcome, as it appears to happen for APP.BACE1 has taken center stage in AD drug development

programs and around a dozen companies have BACEinhibitors in clinical trials. While severe mechanism-basedside-effects of BACE inhibitors have not yet been observedin clinical trials, it remains to be seen which of the BACE1functions may be compromised upon therapeutic BACEinhibition and thereby cause side-effects. Compared to aknock-out, BACE inhibitors do not completely blockBACE1 activity and may still allow residual BACE1 activity

sufficient for essential signaling functions. Moreover, it islikely that many of the developmental BACE1 functions,such as myelination, are not affected in adult AD patients.However, there are developmental processes which alsooccur in the adult brain, such as neurogenesis and axonguidance in the hippocampus. They may be equally affectedin adults upon BACE inhibition. Likewise, muscle spindlemaintenance is an adult BACE1 function and is compro-mised in adult mice treated with a BACE inhibitor (Cheretet al. 2013). We clearly need a better understanding of adultBACE1 functions as a way to predict potential side-effects ofBACE inhibitors. This will be eased by analyzing conditionalBACE1-deficient mice, which are now available in differentlaboratories.The growing list of BACE substrates is not only helpful

for understanding BACE functions, but the substrates mayhave the potential to be used as companion diagnostics forindividualized dosing of patients and for the potentialdevelopment of BACE substrate-specific inhibitors.Since its discovery BACE2 has been the neglected cousin of

BACE1. This has changed as BACE2 substrates and functionsare also being unveiled and as BACE2 is considered apotential target for diabetes. These BACE2 functions are evenrelevant for BACE inhibitor development programs, becausemost BACE inhibitors are equally or even more potent atinhibiting BACE2. In this regard, it may be necessary togenerate conditional BACE1/BACE2 double knock-out miceto analyze phenotypes arising through loss of function of bothproteases specifically in adulthood. Despite their similarities,BACE1 and BACE2 also have differences, for example, insubstrate spectrum, expression pattern and subcellular local-ization. While significant work has been done on theregulation of BACE1 expression and activity, it will beinteresting to see whether BACE2 is regulated in the same or adifferent manner and whether this may allow developing newtherapeutic approaches selectively targeting BACE1. Anotherinteresting approach would be to achieve selective inhibitionof BACE1 cleavage of APP while sparing the cleavages ofother physiologically relevant substrates.Taken together, the recent discoveries demonstrate that we

are in an exciting time for BACE protease research, which isno longer predominantly focusing on AD, but reaching intomany other research fields. And we can expect major newinsights in the next years, both for basic BACE research andfor the use of BACE proteases as drug targets.

Acknowledgments and conflict of interestdisclosure

The authors thank Dr Patty Kandalepas for generating Fig. 1 andGuillaume van Niel for contribution of data for Fig. 4. This workwas supported by the European Research Council under theEuropean Union’s Seventh Framework Programme (FP7/2007–2013)/ ERC Grant Agreement No. 321366-Amyloid (to CH), the


22 R. Vassar et al.

general legacy of Mrs Ammer (to the Ludwig-Maximilians Univer-sity/the chair of CH), by the BMBF (KNDD to SFL and CH), theJPND RiMod (to SFL), the DAAD (to SFL), NIH R01AG022560(RV), NIH R01AG030142 (RV), Cure Alzheimer’s Fund (RV),BrightFocus Foundation, (RV), the Alzheimer’s Association (RV),the Swiss National Science Foundation, the Velux Foundation (bothto LR).

Matthew E. Kennedy is an employee of Merck Research Labs.

References

Akpinar P., Kuwajima S., Krutzfeldt J. and Stoffel M. (2005) Tmem27: acleaved and shed plasma membrane protein that stimulatespancreatic beta cell proliferation. Cell Metab. 2, 385–397.

Albert J. S. (2009) Progress in the development of beta-secretaseinhibitors for Alzheimer’s disease. Prog. Med. Chem. 48,133–161.

Andersen O. M., Reiche J., Schmidt V., et al. (2005) Neuronal sortingprotein-related receptor sorLA/LR11 regulates processing of theamyloid precursor protein. Proc. Natl Acad. Sci. USA 102,13461–13466.

Andrechek E. R., Hardy W. R., Girgis-Gabardo A. A., Perry R. L.,Butler R., Graham F. L., Kahn R. C., Rudnicki M. A. and MullerW. J. (2002) ErbB2 is required for muscle spindle and myoblastcell survival. Mol. Cell. Biol. 22, 4714–4722.

von Arnim C. A., Spoelgen R., Peltan I. D., et al. (2006) GGA1 acts as aspatial switch altering amyloid precursor protein trafficking andprocessing. J. Neurosci. 26, 9913–9922.

Atwal J. K., Chen Y., Chiu C., et al. (2011) A therapeutic antibodytargeting BACE1 inhibits amyloid-beta production in vivo. Sci.Transl. Med., 3, 84ra43.

Bang M. L. and Owczarek S. (2013) A matter of balance: role ofneurexin and neuroligin at the synapse. Neurochem. Res. 38,1174–1189.

Bao J., Wolpowitz D., Role L. W. and Talmage D. A. (2003) Backsignaling by the Nrg-1 intracellular domain. J. Cell Biol. 161,1133–1141.

Bao J., Lin H., Ouyang Y., et al. (2004) Activity-dependent transcriptionregulation of PSD-95 by neuregulin-1 and Eos. Nat. Neurosci. 7,1250–1258.

Basi G., Frigon N., Barbour R., Doan T., Gordon G., McConlogue L.,Sinha S. and Zeller M. (2003) Antagonistic effects of beta-siteamyloid precursor protein-cleaving enzymes 1 and 2 on beta-amyloid peptide production in cells. J. Biol. Chem. 278,31512–31520.

Bateman R. J., Siemers E. R., Mawuenyega K. G., et al. (2009) Agamma-secretase inhibitor decreases amyloid-beta production inthe central nervous system. Ann. Neurol. 66, 48–54.

Bateman R. J., Xiong C., Benzinger T. L., et al. (2012) Clinical andbiomarker changes in dominantly inherited Alzheimer’s disease. N.Engl. J. Med. 367, 795–804.

van Bebber F., Hruscha A., Willem M., Schmid B. and Haass C. (2013)Loss of Bace2 in zebrafish affects melanocyte migration and isdistinct from Bace1 knock out phenotypes. J. Neurochem. 127,471–481.

Benjannet S., Elagoz A., Wickham L., et al. (2001) Post-translationalprocessing of beta-secretase (beta-amyloid-converting enzyme) andits ectodomain shedding. The pro- and transmembrane/cytosolicdomains affect its cellular activity and amyloid-beta production.J. Biol. Chem. 276, 10879–10887.

Bennett B. D., Babu-Khan S., Loeloff R., Louis J. C., Curran E., CitronM. and Vassar R. (2000a) Expression analysis of BACE2 in brainand peripheral tissues. J. Biol. Chem. 275, 20647–20651.

Bennett B. D., Denis P., Haniu M., Teplow D. B., Kahn S., Louis J. C.,Citron M. and Vassar R. (2000b) A furin-like convertase mediatespropeptide cleavage of BACE, the Alzheimer’s beta -secretase.J. Biol. Chem. 275, 37712–37717.

Best J. D., Smith D. W., Reilly M. A., et al. (2007) The novel gammasecretase inhibitor N-[cis-4-[(4-chlorophenyl)sulfonyl]-4-(2,5-difluorophenyl)cyclohexyl]-1,1,1-trifl uoromethanesulfonamide(MRK-560) reduces amyloid plaque deposition without evidenceof notch-related pathology in the Tg2576 mouse. J. Pharmacol.Exp. Ther. 320, 552–558.

Birchmeier C. and Nave K. A. (2008) Neuregulin-1, a key axonal signalthat drives Schwann cell growth and differentiation. Glia 56,1491–1497.

Blennow K. (2010) Biomarkers in Alzheimer’s disease drugdevelopment. Nat. Med. 16, 1218–1222.

Blundell J., Tabuchi K., Bolliger M. F., Blaiss C. A., Brose N., Liu X.,Sudhof T. C. and Powell C. M. (2009) Increased anxiety-likebehavior in mice lacking the inhibitory synapse cell adhesionmolecule neuroligin 2. Genes Brain Behav. 8, 114–126.

Boucard A. A., Chubykin A. A., Comoletti D., Taylor P. and Sudhof T.C. (2005) A splice code for trans-synaptic cell adhesion mediatedby binding of neuroligin 1 to alpha- and beta-neurexins. Neuron48, 229–236.

Boucard A. A., Ko J. and Sudhof T. C. (2012) High affinity neurexinbinding to cell adhesion G-protein-coupled receptor CIRL1/latrophilin-1 produces an intercellular adhesion complex. J. Biol.Chem. 287, 9399–9413.

Bousquet-Moore D., Mains R. E. and Eipper B. A. (2010)Peptidylgycine alpha-amidating monooxygenase and copper: agene-nutrient interaction critical to nervous system function.J. Neurosci. Res. 88, 2535–2545.

Braas K.M., Stoffers D. A., Eipper B.A. andMayV. (1989) Tissue specificexpression of rat peptidylglycine alpha-amidating monooxygenaseactivity and mRNA.Mol. Endocrinol. 3, 1387–1398.

Brunberg E., Andersson L., Cothran G., Sandberg K., Mikko S. andLindgren G. (2006) A missense mutation in PMEL17 is associatedwith the Silver coat color in the horse. BMC Genet. 7, 46.

Buggia-Prevot V., Fernandez C. G., Udayar V., et al. (2013) A functionfor EHD family proteins in unidirectional retrograde dendritictransport of BACE1 and Alzheimer’s disease Ab production. CellRep. 5, 1552–1563.

Buggia-Prevot V., Fernandez C. G., Riordan S., Vetrivel K. S., RosemanJ., Waters J., Bindokas V. P., Vassar R. and Thinakaran G. (2014)Axonal BACE1 dynamics and targeting in hippocampal neurons: arole for Rab11 GTPase. Mol. Neurodegener. 9, 1.

Cai H., Wang Y., McCarthy D., Wen H., Borchelt D. R., Price D. L. andWong P. C. (2001) BACE1 is themajor beta-secretase for generationof Abeta peptides by neurons. Nat. Neurosci. 4, 233–234.

Cai J., Qi X., Kociok N., et al. (2012) beta-Secretase (BACE1)inhibition causes retinal pathology by vascular dysregulation andaccumulation of age pigment. EMBO Mol. Med. 4, 980–991.

Cao L., Rickenbacher G. T., Rodriguez S., Moulia T. W. and Albers M.W. (2012) The precision of axon targeting of mouse olfactorysensory neurons requires the BACE1 protease. Sci. Rep. 2, 231.

Capell A., Steiner H., Willem M., Kaiser H., Meyer C., Walter J.,Lammich S., Multhaup G. and Haass C. (2000) Maturation andpro-peptide cleavage of beta-secretase. J. Biol. Chem. 275, 30849–30854.

Chami L. and Checler F. (2012) BACE1 is at the crossroad of a toxicvicious cycle involving cellular stress and beta-amyloid productionin Alzheimer’s disease. Mol. Neurodegener. 7, 52.

Chang W. P., Koelsch G., Wong S., et al. (2004) In vivo inhibition ofAbeta production by memapsin 2 (beta-secretase) inhibitors.J. Neurochem. 89, 1409–1416.


BACE proteases 23

Chen Y., Hancock M. L., Role L. W. and Talmage D. A. (2010)Intramembranous valine linked to schizophrenia is required forneuregulin 1 regulation of the morphological development ofcortical neurons. J. Neurosci. 30, 9199–9208.

Cheret C., Willem M., Fricker F. R., et al. (2013) Bace1 andNeuregulin-1 cooperate to control formation and maintenance ofmuscle spindles. EMBO J. 32, 2015–2028.

Citron M., Oltersdorf T., Haass C., McConlogue L., Hung A. Y.,Seubert P., Vigo-Pelfrey C., Lieberburg I. and Selkoe D. J. (1992)Mutation of the beta-amyloid precursor protein in familialAlzheimer’s disease increases beta-protein production. Nature360, 672–674.

Cole D. C., Manas E. S., Stock J. R., et al. (2006) Acylguanidines assmall-molecule beta-secretase inhibitors. J. Med. Chem. 49,6158–6161.

Colombo A., Wang H., Kuhn P. H., Page R., Kremmer E., Dempsey P.J., Crawford H. C. and Lichtenthaler S. F. (2012) Constitutivealpha- and beta-secretase cleavages of the amyloid precursorprotein are partially coupled in neurons, but not in frequently usedcell lines. Neurobiol. Dis. 49C, 137–147.

Congreve M., Aharony D., Albert J., et al. (2007) Application offragment screening by X-ray crystallography to the discovery ofaminopyridines as inhibitors of beta-secretase. J. Med. Chem. 50,1124–1132.

Coric V., van Dyck C. H., Salloway S., et al. (2012) Safety andtolerability of the gamma-secretase inhibitor avagacestat in a phase2 study of mild to moderate Alzheimer disease. Arch. Neurol. 69,1430–1440.

Couch J. A., Yu Y. J. and Zhang Y. et al. (2013) Addressing safetyliabilities of TfR bispecific antibodies that cross the blood-brainbarrier. Sci. Transl. Med., 5, 183ra157, 181–112.

Crusio W. E. and Schwegler H. (2005) Learning spatial orientation tasksin the radial-maze and structural variation in the hippocampus ininbred mice. Behav. Brain Funct. 1, 3.

Cumming J. N., Smith E. M., Wang L., et al. (2012) Structure baseddesign of iminohydantoin BACE1 inhibitors: identification of anorally available, centrally active BACE1 inhibitor. Bioorg. Med.Chem. Lett. 22, 2444–2449.

Czyzyk T. A., Ning Y., Hsu M. S., Peng B., Mains R. E., Eipper B. A.and Pintar J. E. (2005) Deletion of peptide amidation enzymaticactivity leads to edema and embryonic lethality in the mouse. Dev.Biol. 287, 301–313.

Danilczyk U., Sarao R., Remy C., et al. (2006) Essential role forcollectrin in renal amino acid transport. Nature 444, 1088–1091.

Dawson G. and Cho S. (2000) Batten’s disease: clues to neuronal proteincatabolism in lysosomes. J. Neurosci. Res. 60, 133–140.

Dawson G. R., Seabrook G. R., Zheng H., et al. (1999) Age-relatedcognitive deficits, impaired long-term potentiation and reduction insynaptic marker density in mice lacking the beta-amyloid precursorprotein. Neuroscience 90, 1–13.

De Strooper B., Vassar R. and Golde T. (2010) The secretases: enzymeswith therapeutic potential in Alzheimer disease. Nat. Rev. Neurol.6, 99–107.

Dejaegere T., Serneels L., Schafer M. K., et al. (2008) Deficiency ofAph1B/C-gamma-secretase disturbs Nrg1 cleavage andsensorimotor gating that can be reversed with antipsychotictreatment. Proc. Natl. Acad Sci. USA 105, 9775–9780.

Deng M., He W., Tan Y., Han H., Hu X., Xia K., Zhang Z. and Yan R.(2013) Increased expression of reticulon 3 in neurons leads toreduced axonal transport of beta site amyloid precursor protein-cleaving enzyme 1. J. Biol. Chem. 288, 30236–30245.

Di Fede G., Catania M., Morbin M., et al. (2009) A recessive mutationin the APP gene with dominant-negative effect onamyloidogenesis. Science 323, 1473–1477.

Dislich B. and Lichtenthaler S. F. (2012) The membrane-bound aspartylprotease BACE1: molecular and functional properties inAlzheimer’s disease and beyond. Front. Physiol. 3, 8.

Dominguez D., Tournoy J., Hartmann D., et al. (2005) Phenotypic andbiochemical analyses of BACE1- and BACE2-deficient mice.J. Biol. Chem. 280, 30797–30806.

Doody R. S., Raman R., Farlow M., et al. (2013) A phase 3 trial ofsemagacestat for treatment of Alzheimer’s disease. N. Engl.J. Med. 369, 341–350.

Ehehalt R., Keller P., Haass C., Thiele C. and Simons K. (2003)Amyloidogenic processing of the Alzheimer beta-amyloidprecursor protein depends on lipid rafts. J. Cell Biol. 160,113–123.

Esterhazy D., Stutzer I., Wang H., et al. (2011) Bace2 is a beta cell-enriched protease that regulates pancreatic beta cell function andmass. Cell Metab. 14, 365–377.

Etherton M. R., Blaiss C. A., Powell C. M. and Sudhof T. C. (2009)Mouse neurexin-1alpha deletion causes correlatedelectrophysiological and behavioral changes consistent withcognitive impairments. Proc. Natl Acad. Sci. USA 106,17998–18003.

Falls D. L. (2003) Neuregulins and the neuromuscular system: 10 yearsof answers and questions. J. Neurocytol. 32, 619–647.

Farah M. H., Pan B. H., Hoffman P. N., et al. (2011) Reduced BACE1activity enhances clearance of myelin debris and regeneration ofaxons in the injured peripheral nervous system. J. Neurosci. 31,5744–5754.

Farzan M., Schnitzler C. E., Vasilieva N., Leung D. and Choe H. (2000)BACE2, a b-secretase homolog, cleaves at the b site and within theamyloid-b region of the amyloid-b precursor protein. Proc. NatlAcad. Sci. USA 97, 9712–9717.

Fleck D., Garratt A. N., Haass C. and Willem M. (2012) BACE1dependent neuregulin proteolysis. Curr. Alzheimer Res. 9,178–183.

Fleck D., van Bebber F., Colombo A., et al. (2013) Dual cleavage ofneuregulin 1 type III by BACE1 and ADAM17 liberates its EGF-like domain and allows paracrine signaling. J. Neurosci. 33,7856–7869.

Fluhrer R., Capell A., Westmeyer G., Willem M., Hartung B., CondronM. M., Teplow D. B., Haass C. and Walter J. (2002) A non-amyloidogenic function of BACE-2 in the secretory pathway.J. Neurochem. 81, 1011–1020.

Garratt A. N., Voiculescu O., Topilko P., Charnay P. and Birchmeier C.(2000) A dual role of erbB2 in myelination and in expansion of theschwann cell precursor pool. J. Cell Biol. 148, 1035–1046.

Gersbacher M. T., Kim D. Y., Bhattacharyya R. and Kovacs D. M.(2010) Identification of BACE1 cleavage sites in human voltage-gated sodium channel beta 2 subunit. Mol. Neurodegener. 5, 61.

Geschwindner S., Olsson L. L., Albert J. S., Deinum J., Edwards P. D.,de Beer T. and Folmer R. H. (2007) Discovery of a novel warheadagainst beta-secretase through fragment-based lead generation.J. Med. Chem. 50, 5903–5911.

Ghosh A. K., Brindisi M. and Tang J. (2012) Developing beta-secretaseinhibitors for treatment of Alzheimer’s disease. J. Neurochem. 120(Suppl 1), 71–83.

Goldstein J. L. and Brown M. S. (2009) The LDL receptor. Arterioscler.Thromb. Vasc. Biol. 29, 431–438.

Grayton H. M., Missler M., Collier D. A. and Fernandes C. (2013)Altered social behaviours in neurexin 1alpha knockout miceresemble core symptoms in neurodevelopmental disorders. PLoSONE 8, e67114.

Gunnersen J. M., Kim M. H., Fuller S. J., et al. (2007) Sez-6 proteinsaffect dendritic arborization patterns and excitability of corticalpyramidal neurons. Neuron 56, 621–639.


24 R. Vassar et al.

Haass C. (2004) Take five–BACE and the gamma-secretase quartetconduct Alzheimer’s amyloid beta-peptide generation. EMBO J.23, 483–488.

Haass C., Lemere C. A., Capell A., Citron M., Seubert P., Schenk D.,Lannfelt L. and Selkoe D. J. (1995) The Swedish mutation causesearly-onset Alzheimer’s disease by beta-secretase cleavage withinthe secretory pathway. Nat. Med. 1, 1291–1296.

Haniu M., Denis P., Young Y., et al. (2000) Characterization ofAlzheimer’s beta -secretase protein BACE. A pepsin familymember with unusual properties. J. Biol. Chem. 275,21099–21106.

Hardy J. and Selkoe D. J. (2002) The amyloid hypothesis of Alzheimer’sdisease: progress and problems on the road to therapeutics. Science297, 353–356.

Harrison S. M., Harper A. J., Hawkins J., et al. (2003) BACE1 (beta-secretase) transgenic and knockout mice: identification ofneurochemical deficits and behavioral changes. Mol. Cell.Neurosci. 24, 646–655.

He X., Li F., Chang W. P. and Tang J. (2005) GGA proteins mediate therecycling pathway of memapsin 2 (BACE). J. Biol. Chem. 280,11696–11703.

Hearing V. J. (2005) Biogenesis of pigment granules: a sensitive way toregulate melanocyte function. J. Dermatol. Sci. 37, 3–14.

Heber S., Herms J., Gajic V., et al. (2000) Mice with combined geneknock-outs reveal essential and partially redundant functions ofamyloid precursor protein family members. J. Neurosci. 20,7951–7963.

Hemming M. L., Elias J. E., Gygi S. P. and Selkoe D. J. (2009)Identification of beta-secretase (BACE1) substrates usingquantitative proteomics. PLoS ONE 4, e8477.

Herms J., Anliker B., Heber S., Ring S., Fuhrmann M., Kretzschmar H.,Sisodia S. and Muller U. (2004) Cortical dysplasia resemblinghuman type 2 lissencephaly in mice lacking all three APP familymembers. EMBO J. 23, 4106–4115.

Heyden A., Angenstein F., Sallaz M., Seidenbecher C. and Montag D.(2008) Abnormal axonal guidance and brain anatomy in mousemutants for the cell recognition molecules close homolog of L1 andNgCAM-related cell adhesion molecule. Neuroscience 155,221–233.

Hippenmeyer S., Shneider N. A., Birchmeier C., Burden S. J., Jessell T.M. and Arber S. (2002) A role for neuregulin1 signaling in musclespindle differentiation. Neuron 36, 1035–1049.

Hitt B. D., Jaramillo T. C., Chetkovich D. M. and Vassar R. (2010)BACE1-/- mice exhibit seizure activity that does not correlate withsodium channel level or axonal localization.Mol. Neurodegener. 5,31.

Hitt B., Riordan S., Kukreja L., Eimer W., Rajapaksha T. and Vassar R.(2012) beta-site amyloid precursor protein (APP) cleaving enzyme1 (BACE1) deficient mice exhibit a close homolog of L1 (CHL1)loss-of-function phenotype involving axon guidance defects.J. Biol. Chem. 287, 38408–38425.

Hong L., Koelsch G., Lin X., Wu S., Terzyan S., Ghosh A. K., Zhang X.C. and Tang J. (2000) Structure of the protease domain ofmemapsin 2 (beta-secretase) complexed with inhibitor. Science290, 150–153.

Hu X., Hicks C. W., He W., Wong P., Macklin W. B., Trapp B. D. andYan R. (2006) Bace1 modulates myelination in the central andperipheral nervous system. Nat. Neurosci. 9, 1520–1525.

Hu X., He W., Diaconu C., Tang X., Kidd G. J., Macklin W. B., TrappB. D. and Yan R. (2008) Genetic deletion of BACE1 in miceaffects remyelination of sciatic nerves. FASEB J. 22, 2970–2980.

Hu X., Zhou X., He W., Yang J., Xiong W., Wong P., Wilson C. G. andYan R. (2010) BACE1 deficiency causes altered neuronal activityand neurodegeneration. J. Neurosci. 30, 8819–8829.

Hu X., He W., Luo X., Tsubota K. E. and Yan R. (2013) BACE1regulates hippocampal astrogenesis via the Jagged1-Notchpathway. Cell Rep. 4, 40–49.

Hussain I., Powell D., Howlett D. R., et al. (1999) Identification of anovel aspartic protease (Asp 2) as beta-secretase. Mol. Cell.Neurosci. 14, 419–427.

Hussain I., Hawkins J., Harrison D., et al. (2007) Oral administration ofa potent and selective non-peptidic BACE-1 inhibitor decreasesbeta-cleavage of amyloid precursor protein and amyloid-betaproduction in vivo. J. Neurochem. 100, 802–809.

Isom L. L. (2001) Sodium channel beta subunits: anything but auxiliary.Neuroscientist 7, 42–54.

Jack C. R., Jr, Vemuri P., Wiste H. J., et al. (2012) Shapes of thetrajectories of 5 major biomarkers of Alzheimer disease. Arch.Neurol. 69, 856–867.

Jamain S., Radyushkin K., Hammerschmidt K., et al. (2008) Reducedsocial interaction and ultrasonic communication in a mouse modelof monogenic heritable autism. Proc. Natl Acad. Sci. USA 105,1710–1715.

Jensen C., Herold P. and Brunner H. R. (2008) Aliskiren: the first renininhibitor for clinical treatment. Nat. Rev. Drug Discovery 7,399–410.

Jonsson T., Atwal J. K., Steinberg S., et al. (2012) A mutation in APPprotects against Alzheimer’s disease and age-related cognitivedecline. Nature 488, 96–99.

Jorissen E., Prox J., Bernreuther C., et al. (2010) The disintegrin/metalloproteinase ADAM10 is essential for the establishment ofthe brain cortex. J. Neurosci. 30, 4833–4844.

Kahn S. E., Hull R. L. and Utzschneider K. M. (2006) Mechanismslinking obesity to insulin resistance and type 2 diabetes. Nature444, 840–846.

Kalvodova L., Kahya N., Schwille P., Ehehalt R., Verkade P., DrechselD. and Simons K. (2005) Lipids as modulators of proteolyticactivity of BACE: involvement of cholesterol, glycosphingolipids,and anionic phospholipids in vitro. J. Biol. Chem. 280,36815–36823.

Kandalepas P. C., Sadleir K. R., Eimer W. A., Zhao J., Nicholson D. A.and Vassar R. (2013) The Alzheimer’s beta-secretase BACE1localizes to normal presynaptic terminals and to dystrophicpresynaptic terminals surrounding amyloid plaques. ActaNeuropathol. 126, 329–352.

Kang E. L., Cameron A. N., Piazza F., Walker K. R. and Tesco G.(2010) Ubiquitin regulates GGA3-mediated degradation ofBACE1. J. Biol. Chem. 285, 24108–24119.

Kang E. L., Biscaro B., Piazza F. and Tesco G. (2012) BACE1 proteinendocytosis and trafficking are differentially regulated byubiquitination at lysine 501 and the Di-leucine motif in thecarboxyl terminus. J. Biol. Chem. 287, 42867–42880.

Kerje S., Sharma P., Gunnarsson U., et al. (2004) The Dominant white,Dun and Smoky color variants in chicken are associated withinsertion/deletion polymorphisms in the PMEL17 gene. Genetics168, 1507–1518.

Kim D. Y., Carey B. W., Wang H., et al. (2007) BACE1 regulatesvoltage-gated sodium channels and neuronal activity. Nat. CellBiol. 9, 755–764.

Kim J., Jung S. Y., Lee Y. K., et al. (2008) Neuroligin-1 is required fornormal expression of LTP and associative fear memory in theamygdala of adult animals. Proc. Natl Acad. Sci. USA 105,9087–9092.

Kim D. Y., Gersbacher M. T., Inquimbert P. and Kovacs D. M. (2011)Reduced sodium channel Na(v)1.1 levels in BACE1-null mice.J. Biol. Chem. 286, 8106–8116.

Kinoshita A., Fukumoto H., Shah T., Whelan C. M., Irizarry M. C. andHyman B. T. (2003) Demonstration by FRET of BACE interaction


BACE proteases 25

with the amyloid precursor protein at the cell surface and in earlyendosomes. J. Cell Sci. 116, 3339–3346.

Ko M. H. and Puglielli L. (2009) Two endoplasmic reticulum (ER)/ERGolgi intermediate compartment-based lysine acetyltransferasespost-translationally regulate BACE1 levels. J. Biol. Chem. 284,2482–2492.

Kobayashi D., Zeller M., Cole T., Buttini M., McConlogue L., Sinha S.,Freedman S., Morris R. G. and Chen K. S. (2008) BACE1 genedeletion: impact on behavioral function in a model of Alzheimer’sdisease. Neurobiol. Aging 29, 861–873.

von Koch C. S., Zheng H., Chen H., Trumbauer M., Thinakaran G., vander Ploeg L. H., Price D. L. and Sisodia S. S. (1997) Generation ofAPLP2 KO mice and early postnatal lethality in APLP2/APPdouble KO mice. Neurobiol. Aging 18, 661–669.

Koh Y. H., von Arnim C. A., Hyman B. T., Tanzi R. E. and Tesco G.(2005) BACE is degraded via the lysosomal pathway. J. Biol.Chem. 280, 32499–32504.

Kuhn P. H., Wang H., Dislich B., Colombo A., Zeitschel U., Ellwart J.W., Kremmer E., Rossner S. and Lichtenthaler S. F. (2010)ADAM10 is the physiologically relevant, constitutive alpha-secretase of the amyloid precursor protein in primary neurons.EMBO J. 29, 3020–3032.

Kuhn P. H., Koroniak K., Hogl S., et al. (2012) Secretome proteinenrichment identifies physiological BACE1 protease substrates inneurons. EMBO J. 31, 3157–3168.

La Marca R., Cerri F., Horiuchi K., et al. (2011) TACE (ADAM17)inhibits Schwann cell myelination. Nat. Neurosci. 14, 857–865.

Laird F. M., Cai H., Savonenko A. V., et al. (2005) BACE1, a majordeterminant of selective vulnerability of the brain to amyloid-betaamyloidogenesis, is essential for cognitive, emotional, and synapticfunctions. J. Neurosci. 25, 11693–11709.

Lakadamyali M., Rust M. J. and Zhuang X. (2006) Ligands for clathrin-mediated endocytosis are differentially sorted into distinctpopulations of early endosomes. Cell 124, 997–1009.

Leu M., Bellmunt E., Schwander M., Farinas I., Brenner H. R. andMuller U. (2003) Erbb2 regulates neuromuscular synapseformation and is essential for muscle spindle development.Development 130, 2291–2301.

Li Z. W., Stark G., Gotz J., Rulicke T., Gschwind M., Huber G., MullerU. and Weissmann C. (1996) Generation of mice with a 200-kbamyloid precursor protein gene deletion by Cre recombinase-mediated site-specific recombination in embryonic stem cells.Proc. Natl Acad. Sci. USA 93, 6158–6162.

Li H., Wang B., Wang Z., Guo Q., Tabuchi K., Hammer R. E., SudhofT. C. and Zheng H. (2010) Soluble amyloid precursor protein(APP) regulates transthyretin and Klotho gene expression withoutrescuing the essential function of APP. Proc. Natl Acad. Sci. USA107, 17362–17367.

Lin X., Koelsch G., Wu S., Downs D., Dashti A. and Tang J. (2000)Human aspartic protease memapsin 2 cleaves the beta-secretasesite of beta-amyloid precursor protein. Proc. Natl Acad Sci. USA97, 1456–1460.

Loy C. T., Schofield P. R., Turner A. M. and Kwok J. B. (2013) Geneticsof dementia. Lancet 383, 828–840.

Luo Y., Bolon B., Kahn S., et al. (2001) Mice deficient in BACE1, theAlzheimer’s beta-secretase, have normal phenotype and abolishedbeta-amyloid generation. Nat. Neurosci. 4, 231–232.

LuoY., Bolon B., DamoreM. A., Fitzpatrick D., Liu H., Zhang J., YanQ.,Vassar R. and Citron M. (2003) BACE1 (beta-secretase) knockoutmice do not acquire compensatory gene expression changes ordevelop neural lesions over time. Neurobiol. Dis. 14, 81–88.

Luo X., Prior M., He W., et al. (2011) Cleavage of neuregulin-1 byBACE1 or ADAM10 protein produces differential effects onmyelination. J. Biol. Chem. 286, 23967–23974.

Magara F., Muller U., Li Z. W., Lipp H. P., Weissmann C., Stagljar M.and Wolfer D. P. (1999) Genetic background changes the patternof forebrain commissure defects in transgenic miceunderexpressing the beta-amyloid-precursor protein. Proc. NatlAcad. Sci. USA 96, 4656–4661.

Malakauskas S. M., Quan H., Fields T. A., McCall S. J., Yu M. J.,Kourany W. M., Frey C. W. and Le T. H. (2007) Aminoaciduriaand altered renal expression of luminal amino acid transporters inmice lacking novel gene collectrin. Am. J. Physiol. Renal Physiol.292, F533–F544.

May P. C., Dean R. A., Lowe S. L., et al. (2011) Robust centralreduction of amyloid-beta in humans with an orally available,non-peptidic beta-secretase inhibitor. J. Neurosci. 31,16507–16516.

McConlogue L., Buttini M., Anderson J. P., et al. (2007) Partialreduction of BACE1 has dramatic effects on Alzheimer plaque andsynaptic pathology in APP Transgenic Mice. J. Biol. Chem. 282,26326–26334.

Meakin P. J., Harper A. J., Hamilton D. L., et al. (2012) Reduction inBACE1 decreases body weight, protects against diet-inducedobesity and enhances insulin sensitivity in mice. Biochem. J. 441,285–296.

Michailov G. V., Sereda M. W., Brinkmann B. G., et al. (2004) Axonalneuregulin-1 regulates myelin sheath thickness. Science 304,700–703.

Mitterreiter S., Page R. M., Kamp F., et al. (2010) Bepridil andamiodarone simultaneously target the Alzheimer’s disease beta-and gamma-secretase via distinct mechanisms. J. Neurosci. 30,8974–8983.

Miyazaki T., Hashimoto K., Uda A., et al. (2006) Disturbance ofcerebellar synaptic maturation in mutant mice lacking BSRPs, anovel brain-specific receptor-like protein family. FEBS Lett. 580,4057–4064.

Montag-Sallaz M., Schachner M. and Montag D. (2002) Misguidedaxonal projections, neural cell adhesion molecule 180 mRNAupregulation, and altered behavior in mice deficient for the closehomolog of L1. Mol. Cell. Biol. 22, 7967–7981.

Mullan M., Crawford F., Houlden H., Axelman K., Lilius L., Winblad B.and Lannfelt L. (1992a) A pathogenic mutation for probableAlzheimer’s disease in the APP gene at the N-terminus of beta-amyloid. Nat. Genet. 1, 345–347.

Muller U., Cristina N., Li Z. W., Wolfer D. P., Lipp H. P., Rulicke T.,Brandner S., Aguzzi A. and Weissmann C. (1994) Behavioral andanatomical deficits in mice homozygous for a modified beta-amyloid precursor protein gene. Cell 79, 755–765.

Naus S., Richter M., Wildeboer D., Moss M., Schachner M. and BartschJ. W. (2004) Ectodomain shedding of the neural recognitionmolecule CHL1 by the metalloprotease-disintegrin ADAM8promotes neurite outgrowth and suppresses neuronal cell death.J. Biol. Chem. 279, 16083–16090.

Ohno M., Sametsky E. A., Younkin L. H., Oakley H., Younkin S. G.,Citron M., Vassar R. and Disterhoft J. F. (2004) BACE1 deficiencyrescues memory deficits and cholinergic dysfunction in a mousemodel of Alzheimer’s disease. Neuron 41, 27–33.

Ohno M., Chang L., Tseng W., Oakley H., Citron M., Klein W. L.,Vassar R. and Disterhoft J. F. (2006) Temporal memory deficits inAlzheimer’s mouse models: rescue by genetic deletion of BACE1.Eur. J. Neurosci. 23, 251–260.

Ohno M., Cole S. L., Yasvoina M., Zhao J., Citron M., Berry R.,Disterhoft J. F. and Vassar R. (2007) BACE1 gene deletionprevents neuron loss and memory deficits in 5XFAD APP/PS1transgenic mice. Neurobiol. Dis. 26, 134–145.

Pastorino L., Ikin A. F., Nairn A. C., Pursnani A. and Buxbaum J. D.(2002) The carboxyl-terminus of BACE contains a sorting signal


26 R. Vassar et al.

that regulates BACE trafficking but not the formation of total A(beta). Mol. Cell. Neurosci. 19, 175–185.

Perez R. G., Soriano S., Hayes J. D., Ostaszewski B., Xia W., Selkoe D.J., Chen X., Stokin G. B. and Koo E. H. (1999) Mutagenesisidentifies new signals for beta-amyloid precursor proteinendocytosis, turnover, and the generation of secreted fragments,including Abeta42. J. Biol. Chem. 274, 18851–18856.

Phinney A. L., Calhoun M. E., Wolfer D. P., Lipp H. P., Zheng H. andJucker M. (1999) No hippocampal neuron or synaptic bouton lossin learning-impaired aged beta-amyloid precursor protein-nullmice. Neuroscience 90, 1207–1216.

Plenge R. M., Scolnick E. M. and Altshuler D. (2013) Validatingtherapeutic targets through human genetics. Nat. Rev. DrugDiscovery 12, 581–594.

Poliak S., Salomon D., Elhanany H., et al. (2003) Juxtaparanodalclustering of Shaker-like K+ channels in myelinated axons dependson Caspr2 and TAG-1. J. Cell Biol. 162, 1149–1160.

Pomara N. (2013) Semagacestat for treatment of Alzheimer’s disease. N.Engl. J. Med. 369, 1661.

Prabhu Y., Burgos P. V., Schindler C., Farias G. G., Magadan J. G. andBonifacino J. S. (2012) Adaptor protein 2-mediated endocytosis ofthe beta-secretase BACE1 is dispensable for amyloid precursorprotein processing. Mol. Biol. Cell 23, 2339–2351.

Rajagopal C., Stone K. L., Mains R. E. and Eipper B. A. (2010)Secretion stimulates intramembrane proteolysis of a secretorygranule membrane enzyme. J. Biol. Chem. 285, 34632–34642.

Rajapaksha T. W., Eimer W. A., Bozza T. C. and Vassar R. (2011) TheAlzheimer’s beta-secretase enzyme BACE1 is required for accurateaxon guidance of olfactory sensory neurons and normal glomerulusformation in the olfactory bulb. Mol. Neurodegener. 6, 88.

Rajendran L. and Simons K. (2008) Membrane Trafficking andTargeting in Alzheimer’s disease. Foundation Ipsen Series,Springer, Heidelberg, Berlin.

Rajendran L., Honsho M., Zahn T. R., Keller P., Geiger K. D., VerkadeP. and Simons K. (2006) Alzheimer’s disease beta-amyloidpeptides are released in association with exosomes. Proc. NatlAcad. Sci. USA 103, 11172–11177.

Rajendran L., Schneider A., Schlechtingen G., et al. (2008) Efficientinhibition of the Alzheimer’s disease beta-secretase by membranetargeting. Science 320, 520–523.

Raposo G. and Marks M. S. (2007) Melanosomes–dark organellesenlighten endosomal membrane transport. Nat. Rev. Mol. Cell Biol.8, 786–797.

Ray S., Fatima Z. and Saxena A. (2010) Drugs for AIDS.Mini Rev. Med.Chem. 10, 147–161.

Ring S., Weyer S. W., Kilian S. B., et al. (2007) The secreted beta-amyloid precursor protein ectodomain APPs alpha is sufficient torescue the anatomical, behavioral, and electrophysiologicalabnormalities of APP-deficient mice. J. Neurosci. 27, 7817–7826.

Roberds S. L., Anderson J., Basi G., et al. (2001) BACE knockout miceare healthy despite lacking the primary beta-secretase activity inbrain: implications for Alzheimer’s disease therapeutics. Hum.Mol. Genet. 10, 1317–1324.

Rochin L., Hurbain I., Serneels L., et al. (2013) BACE2 processesPMEL to form the melanosome amyloid matrix in pigment cells.Proc. Natl Acad Sci. USA 110, 10658–10663.

Rogaeva E., Meng Y., Lee J. H., et al. (2007) The neuronal sortilin-related receptor SORL1 is genetically associated with Alzheimerdisease. Nat. Genet. 39, 168–177.

Rossner S., Sastre M., Bourne K. and Lichtenthaler S. F. (2006)Transcriptional and translational regulation of BACE1 expression–implications for Alzheimer’s disease. Prog. Neurobiol. 79, 95–111.

Saftig P., Hetman M., Schmahl W., et al. (1995) Mice deficient for thelysosomal proteinase cathepsin D exhibit progressive atrophy of

the intestinal mucosa and profound destruction of lymphoid cells.EMBO J. 14, 3599–3608.

Sakano H. (2010) Neural map formation in the mouse olfactory system.Neuron 67, 530–542.

Sankaranarayanan S., Price E. A., Wu G., et al. (2008) In vivo beta-secretase 1 inhibition leads to brain Abeta lowering and increasedalpha-secretase processing of amyloid precursor protein withouteffect on neuregulin-1. J. Pharmacol. Exp. Ther. 324, 957–969.

Sankaranarayanan S., Holahan M. A., Colussi D., et al. (2009) Firstdemonstration of cerebrospinal fluid and plasma A beta loweringwith oral administration of a beta-site amyloid precursor protein-cleaving enzyme 1 inhibitor in nonhuman primates. J. Pharmacol.Exp. Ther. 328, 131–140.

Sannerud R., Declerck I., Peric A., et al. (2011) ADP ribosylation factor6 (ARF6) controls amyloid precursor protein (APP) processing bymediating the endosomal sorting of BACE1. Proc. Natl Acad. Sci.USA 108, E559–E568.

Savonenko A. V., Melnikova T., Laird F. M., Stewart K. A., Price D. L.and Wong P. C. (2008) Alteration of BACE1-dependent NRG1/ErbB4 signaling and schizophrenia-like phenotypes in BACE1-nullmice. Proc. Natl Acad. Sci. USA 105, 5585–5590.

Savvaki M., Panagiotaropoulos T., Stamatakis A., Sargiannidou I.,Karatzioula P., Watanabe K., Stylianopoulou F., Karagogeos D.and Kleopa K. A. (2008) Impairment of learning and memory inTAG-1 deficient mice associated with shorter CNS internodes anddisrupted juxtaparanodes. Mol. Cell. Neurosci. 39, 478–490.

Schneider A., Rajendran L., Honsho M., Gralle M., Donnert G., WoutersF., Hell S. W. and Simons M. (2008) Flotillin-dependent clusteringof the amyloid precursor protein regulates its endocytosis andamyloidogenic processing in neurons. J. Neurosci. 28, 2874–2882.

Seabrook G. R., Smith D. W., Bowery B. J., et al. (1999) Mechanismscontributing to the deficits in hippocampal synaptic plasticity inmice lacking amyloid precursor protein. Neuropharmacology 38,349–359.

Shacka J. J., Klocke B. J., Young C., Shibata M., Olney J. W., UchiyamaY., Saftig P. and Roth K. A. (2007) Cathepsin D deficiency inducespersistent neurodegeneration in the absence of Bax-dependentapoptosis. J. Neurosci. 27, 2081–2090.

Shih D. Q., Screenan S., Munoz K. N., Philipson L., Pontoglio M.,Yaniv M., Polonsky K. S. and Stoffel M. (2001) Loss of HNF-1alpha function in mice leads to abnormal expression of genesinvolved in pancreatic islet development and metabolism. Diabetes50, 2472–2480.

Siegenthaler B. M. and Rajendran L. (2012) Retromers in Alzheimer’sdisease. Neurodegener. Dis. 10, 116–121.

Siintola E., Partanen S., Stromme P., Haapanen A., Haltia M., MaehlenJ., Lehesjoki A. E. and Tyynela J. (2006) Cathepsin D deficiencyunderlies congenital human neuronal ceroid-lipofuscinosis. Brain129, 1438–1445.

Sinha S., Anderson J. P., Barbour R., et al. (1999) Purification andcloning of amyloid precursor protein beta-secretase from humanbrain. Nature 402, 537–540.

Sisodia S. S. and St George-Hyslop P. H. (2002) gamma-Secretase,Notch, Abeta and Alzheimer’s disease: where do the presenilins fitin? Nat. Rev. Neurosci. 3, 281–290.

Small S. A. and Gandy S. (2006) Sorting through the cell biology ofAlzheimer’s disease: intracellular pathways to pathogenesis.Neuron 52, 15–31.

Stamford A. and Strickland C. (2013) Inhibitors of BACE for treatingAlzheimer’s disease: a fragment-based drug discovery story. Curr.Opin. Chem. Biol. 17, 320–328.

Stamford A. W., Scott J. D., Li S. W., et al. (2012) Discovery of anorally available, brain penetrant BACE1 inhibitor that affordsrobust CNS abeta reduction. ACS Med. Chem. Lett. 3, 897–902.


BACE proteases 27

Steinbach J. P., Muller U., Leist M., Li Z. W., Nicotera P. and Aguzzi A.(1998) Hypersensitivity to seizures in beta-amyloid precursorprotein deficient mice. Cell Death Differ. 5, 858–866.

Stutzer I., Selevsek N., Esterhazy D., Schmidt A., Aebersold R. andStoffel M. (2013) Systematic proteomic analysis identifies beta-siteamyloid precursor protein cleaving enzyme 2 and 1 (BACE2 andBACE1) substrates in pancreatic beta-cells. J. Biol. Chem. 288,10536–10547.

Suh J., Choi S. H., Romano D. M., Gannon M. A., Lesinski A. N., KimD. Y. and Tanzi R. E. (2013) ADAM10 missense mutationspotentiate beta-amyloid accumulation by impairing prodomainchaperone function. Neuron 80, 385–401.

Suzuki K., Hayashi Y., Nakahara S., et al. (2012) Activity-dependentproteolytic cleavage of neuroligin-1. Neuron 76, 410–422.

Tanzi R. E. (2012) The genetics of Alzheimer disease. Cold SpringHarb. Perspect. Med. 2.

Taveggia C., Zanazzi G., Petrylak A., et al. (2005) Neuregulin-1 type IIIdetermines the ensheathment fate of axons. Neuron 47, 681–694.

Tesco G., Koh Y. H., Kang E. L., et al. (2007) Depletion of GGA3stabilizes BACE and enhances beta-secretase activity. Neuron 54,721–737.

Theos A. C., Berson J. F., Theos S. C., et al. (2006) Dual loss of ERexport and endocytic signals with altered melanosome morphologyin the silver mutation of Pmel17. Mol. Biol. Cell 17, 3598–3612.

Thisse B., Heyer V., Lux A., Alunni V., Degrave A., Seiliez I., KirchnerJ., Parkhill J. P. and Thisse C. (2004) Spatial and temporalexpression of the zebrafish genome by large-scale in situhybridization screening. Methods Cell Biol. 77, 505–519.

Tong G., Wang J. S., Sverdlov O., et al. (2012) Multicenter,randomized, double-blind, placebo-controlled, single-ascendingdose study of the oral gamma-secretase inhibitor BMS-708163(Avagacestat): tolerability profile, pharmacokinetic parameters, andpharmacodynamic markers. Clin. Ther. 34, 654–667.

Traka M., Goutebroze L., Denisenko N., et al. (2003) Association ofTAG-1 with Caspr2 is essential for the molecular organization ofjuxtaparanodal regions of myelinated fibers. J. Cell Biol. 162,1161–1172.

Udayar V., Buggia-Prevot V., Guerreiro R. L., et al. (2013) A PairedRNAi and RabGAP overexpression screen identifies Rab11 as aregulator of beta-amyloid production. Cell Rep. 5, 1536–1551.

Varoqueaux F., Aramuni G., Rawson R. L., Mohrmann R., Missler M.,Gottmann K., Zhang W., Sudhof T. C. and Brose N. (2006)Neuroligins determine synapse maturation and function. Neuron51, 741–754.

Vassar R., Bennett B. D., Babu-Khan S., et al. (1999) Beta-secretasecleavage of Alzheimer’s amyloid precursor protein by thetransmembrane aspartic protease BACE. Science 286, 735–741.

Vassar R., Kovacs D. M., Yan R. and Wong P. C. (2009) The beta-secretase enzyme BACE in health and Alzheimer’s disease:regulation, cell biology, function, and therapeutic potential.J. Neurosci. 29, 12787–12794.

Vetrivel K. S., Meckler X., Chen Y., et al. (2009) Alzheimer diseaseAbeta production in the absence of S-palmitoylation-dependenttargeting of BACE1 to lipid rafts. J. Biol. Chem. 284, 3793–3803.

Walker K. R. and Tesco G. (2013) Molecular mechanisms of cognitivedysfunction following traumatic brain injury. Front. AgingNeurosci. 5, 29.

Walsh D. M. and Selkoe D. J. (2007) A beta oligomers - a decade ofdiscovery. J. Neurochem. 101, 1172–1184.

Wan H. I., Jacobsen J. S., Rutkowski J. L. and Feuerstein G. Z. (2009)Translational medicine lessons from flurizan’s failure in

Alzheimer’s disease (AD) trial: implication for future drugdiscovery and development for AD. Clin. Transl. Sci. 2, 242–247.

Wang P., Yang G., Mosier D. R., et al. (2005) Defective neuromuscularsynapses in mice lacking amyloid precursor protein (APP) andAPP-Like protein 2. J. Neurosci. 25, 1219–1225.

Wang Y. S., Strickland C., Voigt J. H., et al. (2010) Application offragment-based NMR screening, X-ray crystallography, structure-based design, and focused chemical library design to identify novelmicroM leads for the development of nM BACE-1 (beta-site APPcleaving enzyme 1) inhibitors. J. Med. Chem. 53, 942–950.

Watt B., van Niel G., Fowler D. M., et al. (2009) N-terminal domainselicit formation of functional Pmel17 amyloid fibrils. J. Biol.Chem. 284, 35543–35555.

Weyer S. W., Klevanski M., Delekate A., et al. (2011) APP and APLP2are essential at PNS and CNS synapses for transmission, spatiallearning and LTP. EMBO J. 30, 2266–2280.

Willem M., Garratt A. N., Novak B., et al. (2006) Control of peripheralnerve myelination by the beta-secretase BACE1. Science 314,664–666.

Willem M., Lammich S. and Haass C. (2009) Function, regulation andtherapeutic properties of beta-secretase (BACE1). Semin. Cell Dev.Biol. 20, 175–182.

Wong H. K., Sakurai T., Oyama F., et al. (2005) beta Subunits ofvoltage-gated sodium channels are novel substrates of beta-siteamyloid precursor protein-cleaving enzyme (BACE1) and gamma-secretase. J. Biol. Chem. 280, 23009–23017.

Xu J. Y., Xia Q. Q. and Xia J. (2012) A review on the current neuroliginmouse models. Sheng Li Xue Bao 64, 550–562.

Yan R., Bienkowski M. J., Shuck M. E., et al. (1999) Membrane-anchored aspartyl protease with Alzheimer’s disease beta-secretaseactivity. Nature 402, 533–537.

Yan R., Munzner J. B., Shuck M. E. and Bienkowski M. J. (2001)BACE2 functions as an alternative alpha-secretase in cells. J. Biol.Chem. 276, 34019–34027.

Yang G., Gong Y. D., Gong K., et al. (2005) Reduced synaptic vesicledensity and active zone size in mice lacking amyloid precursorprotein (APP) and APP-like protein 2. Neurosci. Lett. 384, 66–71.

Yang L., Wang Z., Wang B., Justice N. J. and Zheng H. (2009a)Amyloid precursor protein regulates Cav1.2 L-type calciumchannel levels and function to influence GABAergic short-termplasticity. J. Neurosci. 29, 15660–15668.

Yang W., Fucini R. V., Fahr B. T., et al. (2009b) Fragment-baseddiscovery of nonpeptidic BACE-1 inhibitors using tethering.Biochemistry 48, 4488–4496.

Yu Y. J., Zhang Y. and Kenrick M. et al. (2011) Boosting brain uptakeof a therapeutic antibody by reducing its affinity for a transcytosistarget. Sci. Transl. Med. 3, 84ra44.

Zhang H., Wada J., Hida K., et al. (2001) Collectrin, a collecting duct-specific transmembrane glycoprotein, is a novel homolog of ACE2and is developmentally regulated in embryonic kidneys. J. Biol.Chem. 276, 17132–17139.

Zheng H., Jiang M., Trumbauer M. E., et al. (1995) beta-Amyloidprecursor protein-deficient mice show reactive gliosis anddecreased locomotor activity. Cell 81, 525–531.

Zhou L., Barao S., Laga M., et al. (2012) The neural cell adhesionmolecules L1 and CHL1 are cleaved by BACE1 protease in vivo.J. Biol. Chem. 287, 25927–25940.

Zhu Z., Sun Z. Y., Ye Y., et al. (2010) Discovery of cyclicacylguanidines as highly potent and selective beta-site amyloidcleaving enzyme (BACE) inhibitors: part I–inhibitor design andvalidation. J. Med. Chem. 53, 951–965.


28 R. Vassar et al.

function, therapeutic potential and cell biology of bace proteases: current status and future...

Documents