a comparative analysis of 3d platforms in high throughput … · 3d results match in 384 vs 1536...

A Comparative Analysis of 3D Platforms in High Throughput Formats to Assess the Cytotoxic Effect of Drugs for Discovery

Tim SpicerSenior Scientific Director-Molecular MedicineDirector of Discovery Biology and HTS at the [email protected] Tel: (561) 228-2150

http://hts.florida.scripps.edu/

May 2018

• Evidence that multiple 3D technologies are rapid, cost effective and homogeneous solutions for physiologically relevant HTS

• A comparison of both 2D and 3D cell culture models vs large approved drug libraries testing primary pancreatic tumours including their cancer associated fibroblasts

• Direct evidence to support 3D models exhibit outcomes that are clinically relevant

• Co-culturing models of pancreatic tumours and CAFs in HTS behave differently

• 3D modelling of brain tumours (Glioblastoma) in HTS vs >10K INDs, Drugs or pre-IND molecules

• Rapid 3D technological implementation directed towards precision medicine

TSRI © 2018. All rights reserved.

• SRIMSC began in 2005• Biologists, Biochemists, Programmers, Chemists & Engineers

(currently >600 employees)• Industrial scale HTS lab with Kalypsys/GNF automated platform

• >646K Proprietary (largest in academia, ~30k unique compounds, focused sub-libraries, professionally curated)

• >360K Public Domain (NIH)• Funding is driven by NIH grants and Collaborations with Pharma and

Biotech

Where robotics, chemistry and biology join forces to help discover new drugs

Biologists, Biochemists, Programmers, Chemists & Engineers

Drug ScreeningComprehensive HTS/HIT

PlatformsVariety of formats supported

1536-well format HTS/HCSMillion samples/24 hrs

Pharma-grade IT infrastructure

Assay Development“From test tube to plate”

Assay DesignCell-based assay development

Biochemical assay developmentReagent Generation

Compound Management >1MM molecule screening libraries:~640K Scripps’ drug discovery library

~370K NIH’s MLPCN libraryLC-MS QC & Sample AuditingCheminformatics Expertise


1.7m Arm Incubators Transfer Dispense/Wash PE Suite FLIPR

• Built specifically for 1536-well plate screening (384-well plate screening also possible)

• Capable of over one million wells screened in a 24-hr period (1536-well format)• Patented lid prevents evaporation of plate contents• Long (>96 hr) plate incubation protocols possible• Plate capacity >1,500 (>2.3 million wells in 1536-well plate format)• Plate incubation from 4₀ to 50₀C, 0-100% rH, any gas (CO2, N2, Ar, etc.) • 1536-well plate washers and transfer pipettor enable heterogeneous

assays/fixing steps• Luminescence, BRET, Absorbance, Fluorescence Intensity, FP, TRF, FRET, TR-FRET,

AlphaScreen, AlphaLISA, FLIPR, High-Content…


CellInsight


• > 300 uHTS campaigns completed• >125 million data points

• 50:50 ratio NIH to SDDL• 107 peer reviewed publications• 77 molecular probes identified

• http://pubchem.ncbi.nlm.nih.gov

• 14 out the 18 licensed probes• Multiple molecules in the clinics

• Receptos (sold for $7.2BB), Cleave, Etc• Multiple NCEs

Licensed MLPCN Probes

Cell 161, June 4, 2015


• Pancreatic cancer remains a leading cause of cancer-associated death, with a 5-year survival rate less than 10%.

• Genetic KRAS (Ki-ras2 Kirsten rat sarcoma viral oncogene homolog) mutations are found in more than 90% of pancreatic cancer patients.

• Genetic alterations in tumor suppressors TP53, CDKN2A, SMAD4, ARID1A and MLL3 also accumulate in the pancreatic cancer patient.

8%

27%

53%

12%Localized

Regional

Distant

Unknown

Stage at diagnosis 2002-2008

23

9

2

6

0

5

10

15

20

25

Localized Regional Distant All stages

5-year relative survival rate (%)2002-2008

R. Siegel, et.al. CA CANCER J CLIN 2013, 63: 11-30J. P. Morris, et.al. Nat Rev Cancer 2010, 10 (10): 683-695


Name Tissue Mutations

hM1-CAF a resected metastatic lung lesion, same as hM1 Kraswt (fibroblast, SV40 immortalized)

hM1-2D a resected metastatic lung lesion KrasG12D, p53R175H (epithelial-like)

hF2-2D a fine-needle aspiration biopsy of a metastatic lesion KrasG12V, SMAD4loss, CDKN2Aloss (epithelial-like)

hT1-CAF resected primary tumors, same as hT1 Kraswt (fibroblast, SV40 immortalized)

hT1-2D resected primary tumors KrasG12V, P53loss, SMAD4loss, CDKN2Ahom del((epithelial-like)

• Primary human pancreatic ductal cells hM1-2D, hF2-2D, hT1-2D, and cancer-associated fibroblasts hM1-CAF and hT1-CAF, were generated from tissues of pancreatic cancer patients in the laboratory of Dr. Tuveson, M.D. at CSHL.

• 2D monolayer cells were conditioned from established pancreatic 3D organoids.Herve Triac at CHSL [S.F. Boj, et.al. Cell, 2017, 160(1-2): 324]

• Targeted sequencing analysis of human pancreatic organoids reveals their genetic makeup.2D Monolayer3D organoids in Matrigel

Method Example of Product (Manufacturer)Single

spheroid per well

Miniaturizableto 1536?

Spheroid centered

within well?

Noperipheral

device required

Homogeneous in format?

Extracellular matrices Matrigel (Corning) Qgel (Qgel)

Hanging drops GravityPLUS hanging drop system (InSphero)Perfecta3D (3D BioMatrix) Potentially

difficult Micro-patterned

surfacesAggreWell (StemCell Technologies)

3D micro pattern culture plate (Diagenode) Ultra-low attachmentsurfaces and flat-well

geometry

Cell repellent microplates (Greiner Bio-One)Corning flat ULA microplates (Corning)Nunclon Sphera microplates (Thermo)

Ultra-low attachmentsurfaces with round-

well geometry

GravityTrap ULA plates (InSphero)

ULA spheroid Microplates (Corning)

Magnetic 3D bioprinting

NanoShuttle (Nano3D) w/Greiner Bio-One cell repellent plates

TSRI © 2018. All rights reserved. Slide 9

Hanging drop cultureMatrigel-embedded culture Round-bottomed microwell culture Magnetic 3D bioprinting

Z stack Confocal Image20X objective 4X objective

Cells in 2D monolayer and in 3D spheroid/organoid format demonstrated different responses to therapeutic compounds. Drug resistance was observed in 3D model.

2D HTS 3D HTS

100µm

Slide 10

2D: hT1

-10 -9 -8 -7 -6 -5 -4-25

0

25

50

75

100DOXGEMSN-38

DOX GEM SN-38Hill Slope 1.47 0.93 0.47

IC50 3.28E-07 >9.98E-06 3.68E-08

Log[Compound], M

% In

hibi

tion

3D: hT1

-10 -9 -8 -7 -6 -5 -4-25

0

25

50

75

100DOXGEMSN-38

DOX GEM SN-38Hill Slope 1.28 / 0.58

IC50 1.13E-06 >9.98E-06 6.16E-05

Log[Compound], M

% In

hibi

tion


• Scalable

• Automation-friendly

• Single spheroid per well

• Minimize peripheral devices - i.e. adapters

• Miniaturizable to 1536-well plate format

• Spheroid to be centrally located within the well

• Well-to-well uniformity of spheroid size and morphology

• Homogenous format (no transfer of spheroids or aspiration step required)


Greiner cell repellent plates readily form spheroids using cells combined with Nanoshuttle. This can be done on thebottom side of the plates using a magnetic driver. Note that the shape of the bottom of the wells is compatible withautomated confocal microscopy.

1536 well format

Incubation hours

Detach cells

384 well format


No Nanoshuttle Compound NanoshuttleIC50 (M) IC50 (M)

1.3 x 10-5 OXA 4.8 x 10-5

7.5 x 10-8 DOX 9.6 x 10-8

5.8 x 10-9 GEM 6.8 x 10-9

7.1 x 10-7 5’FU 6.8 x 10-7

2.4 x 10-9 SN-38 2.4 x 10-9

Without Nanoshuttle With Nanoshuttle

384-well Assay

Put the plate atop of the spheroid drive, Incubate the plate on drive for 4 hrs

Image spheroids using HCS

Incubate cells for 24 hrs | 37oC 5%CO2 95% RH

Harvest cells, seed 2500 cells in 25 μL culture medium to 384 Greiner cell repellent plate

Add nanoshuttle to cells in flask (0.6 mL per T175), incubate cells overnight | 37oC 5%CO2 95% RH

Incubate cells for additional 3 days | 37oC 5%CO2 95% RH

Pintool transfer 50 nL compounds

Dispense 25 μL CellTiter-Glo 3D Reagent, Shake and incubate for 60min at RT,

Read on ViewLux

Final volume = 25 μL

1536-well Assay

Put the plate atop of the spheroid drive, Incubate the plate on drive for 4 hrs

Incubate cells for 24 hrs | 37oC 5%CO2 95% RH

Harvest cells, seed 1250 cells in 5 μL culture medium to 1536 Greiner cell repellent plate

Add nanoshuttle to cells in flask (0.6 mL per T175), incubate cells overnight | 37oC 5%CO2 95% RH

Incubate cells for additional 3 days | 37oC 5%CO2 95% RH

Pintool transfer 10 nL compounds

Dispense 5 μL CellTiter-Glo 3D reagent, Shake for 10 mins,

Incubate for 60min at RT, Read on ViewLux

Final volume = 5 μL

Image spheroids using HCS

Cell Viability Assay


% 𝑖𝑖𝑖𝑖ℎ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 = 100 × (1 − 𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖 𝑊𝑊𝑇𝑇𝑊𝑊𝑊𝑊 − 𝑀𝑀𝑇𝑇𝑀𝑀𝑖𝑖𝑀𝑀𝑖𝑖 𝐻𝐻𝑖𝑖𝐻𝐻ℎ 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊

𝑀𝑀𝑇𝑇𝑀𝑀𝑖𝑖𝑀𝑀𝑖𝑖 𝐿𝐿𝑖𝑖𝐿𝐿 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊 − 𝑀𝑀𝑇𝑇𝑀𝑀𝑖𝑖𝑀𝑀𝑖𝑖 𝐻𝐻𝑖𝑖𝐻𝐻ℎ 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊)

𝑍𝑍′ = 1 −3𝑆𝑆𝑆𝑆 𝑖𝑖𝑜𝑜 𝐿𝐿𝑖𝑖𝐿𝐿 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊 + 3𝑆𝑆𝑆𝑆 𝑖𝑖𝑜𝑜 𝐻𝐻𝑖𝑖𝐻𝐻ℎ 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊

(𝐿𝐿𝑖𝑖𝐿𝐿 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊 − 𝐻𝐻𝑖𝑖𝐻𝐻ℎ 𝐶𝐶𝑖𝑖𝑖𝑖𝑖𝑖𝐶𝐶𝑖𝑖𝑊𝑊)

hT-29

PANC-1

4h (Immediately off drive) 24h 48h 72h Corning 72h

hM1

hM1-CAF

hT1

hT1-CAF

A

500µm

Cystic organoid

“Hybrid” organoid

Filled organoid

hM1

72h (20X)

100µm

hT1

B

C hT1-CAF, 48h



Cell Tracker-green staining

20X objective, Z-stack

Hoechst staining

HT-29 PANC-1500 µm

hT1-CAF

4X objective, wide fieldhT1 hT1-CAF

hT1

4X objective, wide field

HT-29 PANC-1


hT1

hT160X objective


A B C

Both lab adapted and primary tumor cells readily form 3D structures which we confirmed using Z-stack confocal analysis. HT-29, PANC-1 and CAFs formed compact spheroids, while the primary cancer cells grew as organoids. (confirmed by Dr.

Tuveson at CHSL)


R.-Z. Lin and H.-Y. Chang. Biotech J, 2008, 3: 1172 R.-Z. Lin and L.-F. Chou. Cell Tissue Res, 2006, 324: 411Andrea Ivascu and Manfred Kubbies. Int J Oncol, 2007, 31(6): 1403

hM1 hT1

hM1-CAF hT1-CAF

hT-29 PANC-1

Figure A. A model of the spheroid formation process

β1- Integrin expression: All three relatively high and close to equal E-cadherin expression: Hep3B > HepG2 > PLC/PRF/5 (8.5 : 4.1 : 1)

Figure B. Comparative analysis of spheroid-forming capability among three hepatoma cell lines

A

B

Slide 17

hT1-CAF

4hr (4X)

12hr (10X)

24hr (10X)

No Ab +anti-Integrin + anti-E-Cadherin

48hr (10X)


hT1No Ab +anti-Integrin + anti-E-Cadherin

Immunostaining

48hr (10X)

Antibody block

Nuclei Integrin E-Cadherin Composite Nuclei Integrin E-Cadherin Composite

Anti-Integrin has slight effect on hT1-CAF spheroid aggregation or compaction. E-Cadherin likely plays an important role in hT1-CAF spheroid compaction. Ongoing: Quantification of adhesion molecules and ECM proteins in these cultures.


3D results match in 384 vs 1536 but there is a shift in sensitivity associated with 3D formats. This shift in sensitivity in 2D vs 3D is less apparent in the CAFs.

2D/1536: hT1

-10 -9 -8 -7 -6 -5 -4-25

0255075

100 DOXGEMSN-38

DOX GEM SN-38Hill Slope 1.47 0.54 -0.32

IC50 3.72E-07 >9.98E-06 1.72E-07

Log[Compound], M

% In

hibi

tion

3D/384: hT1

-10 -9 -8 -7 -6 -5 -4-25

0255075

100 DOXGEMSN-38

DOX GEM SN-38Hill Slope 2.17 / 0.77

IC50 6.58E-07 >9.98E-06 2.61E-05

Log[Compound], M

% In

hibi

tion

3D/1536: hT1

-10 -9 -8 -7 -6 -5 -4-25

0255075

100 DOXGEMSN-38


IC50 9.71E-07 >9.98E-06 7.73E-07

Log[Compound], M

% In

hibi

tion

2D/1536: hT1-CAF

-10 -9 -8 -7 -6 -5 -4-25

0255075

100

DOXGEMSN-38

DOX GEM SN-38HillSlope 1.27 1.39 0.98

IC50 9.59E-08 2.50E-09 2.94E-09

Log[Compound], M

% In

hibi

tion

3D/384: hT1-CAF

-10 -9 -8 -7 -6 -5 -4-25

0255075

100

DOXGEMSN-38


IC50 1.77E-07 2.24E-08 7.7E-09

Log[Compound], M

% In

hibi

tion

3D/1536: hT1-CAF

-10 -9 -8 -7 -6 -5 -4-25

0255075

100

DOXGEMSN-38

DOX GEM SN-38HillSlope 1.19 0.92 0.70

IC50 1.77E-07 2.42E-08 3.83E-09

Log[Compound], M

% In

hibi

tion

Run Statistics (n = 9 plates)• 3290 compounds tested at 2 µM nominal concentration in triplicate• Z’ = 0.77±0.03• S:B = 44.76±0.91• IC50 of DOX = 513 nM• Hit Cutoff (Interval) = 34.34%• Hits = 42 (1.28%)

hT1-2D

0 3000 6000 9000 12000-50-25

0255075

100125

Well

% In

hibi

tion

Hit Cutoff


Run Statistics (n = 11 plates)• 3290 compounds tested at 2 µM nominal concentration in singlicate• Z’ = 0.80±0.06• S:B = 207.37±16.33• IC50 of DOX = 658 nM• Hit Cutoff (Interval) = 40.21%• Hits = 26 (0.88%)

hT1-3D

0 1000 2000 3000 4000-50-25

0255075

100125

Well

% In

hibi

tion

High Control

Sample Field

Low Control


hT1CAF-3D

0 1000 2000 3000 4000-50-25

0255075

100125

Well

% In

hibi

tion

Run Statistics (n = 9 plates)• 3290 compounds tested at 2 µM nominal concentration in triplicate• Z’ = 0.70±0.07• S:B = 34.52±2.39• IC50 of DOX = 95.9 nM• Hit Cutoff (Interval) = 47.42%• Hits = 82 (2.49%)

hT1CAF-2D

0 3000 6000 9000 12000-50-25

0255075

100125

Well

% In

hibi

tion

Run Statistics (n = 11 plates)• 3290 compounds tested at 2 µM nominal concentration in singlicate• Z’ = 0.75±0.04• S:B = 176.27±5.15• IC50 of DOX = 275 nM• Hit Cutoff (Interval) = 39.41%• Hits = 53 (1.64%)

Hit Cutoff

High Control

Low Control

Sample Field


Activity of 3305 approved drugs against both 2D and 3D models

hT1 vs. Approved Drugs

-100 -50 0 50 100-100

-50

0

50

100

% Inhibition (hT1-2D)

% In

hibi

tion

(hT1

-3D

)

I

II

III

hT1CAF vs. Approved Drugs

-100 -50 0 50 100-100

-50

0

50

100

% Inhibition (hT1CAF-2D)

% In

hibi

tion

(hT1

CAF

-3D

)

I

II

III

hM1 vs. Approved Drugs

-50 0 50 100-50

0

50

100

% Inhibition (hM1-2D)

% In

hibi

tion

(hM

1-3D

)

I

II

III

hM1CAF vs. Approved Drugs

-100 -50 0 50 100-100

-50

0

50

100

% Inhibition (hM1CAF-2D)

% In

hibi

tion

(hM

1CA

F-3D

)

I

II

III

Disulfiram

Submitted to SLAS Discovery


Activity of 114 NCI Oncology drugs against both 2D and 3D models

1 - Trametinib2 - Romidepsin3 - Bortezomib4 - Carfilzomib5 - Homoharringtonine8 - Doxorubicin97 - Gemcitabine

Dru

gs

Drug resistance in 3D model is cell and drug dependent. Ultimate goal towards Personalized Medicine using 3D cell models.


DrugResistance Factor

(the ratio of IC50 between 2D and 3D assay)

hT1 hT1-CAF hM1 hM1-CAF

Bortezomib 4.0 5.9 7.9 4.4

Carfilzomib 3.3 8.1 3.2 7.3

Trametinib <0.001 <0.3 6.1 <0.6

Romidepsin 24.0 3.3 28.8 1.5

Disufliram <0.03 / 0.3 >32.9


E


-10 -9 -8 -7 -6 -5-25

0

25

50

75

100

125

hT1-3DhT1-CAF-3DhM1-3DhM1-CAF-3DhT1-2DhT1-CAF-2DhM1-2DhM1-CAF-2D

Log[Disulfiram]

% In

hibi

tion

DhT1-3D

hT1-2D

-10 -9 -8 -7 -6 -5-25

0

25

50

75

100

125


Log[Carfilzomib]

% In

hibi

tion

A

2D

3D-10 -9 -8 -7 -6 -5

-25

0

25

50

75

100

125


Log[Trametinib]

% In

hibi

tion

B

-10 -9 -8 -7 -6 -5-25

0

25

50

75

100

125


Log[Romidepsin]

% In

hibi

tion

C

RedBull


Magnetic 3D bioprinting technology has been successfully integrated into our GNF/Kalypsysautomation platforms.


Run Statistics (n = 127 plates)• 151,977 compounds tested at 2 µM nominal concentration in singlicate• Z’ = 0.72±0.06• S:B = 188.13±23.07• IC50 of DOX = 393±58 nM• Hit Cutoff (Interval) = 34.98%• Hits = 735 (0.48%)

hT1-3D

-10 -9 -8 -7 -6 -5 -4-25

0255075

100DOXGEMSN-38


IC50 4.08E-07 >9.98E-06 1.72E-07

Log[Compound], M

% In

hibi

tion

Hit Cutoff

Well

% In

hibi

tion

This assay showed excellent Z’ values. CRC of control compounds are consistent from batch to batch. 150K library primary screen vs. hT1-3D was complete. Identified 735 hits will be directly proceeded to dose response test.


Co-culture of cancer cells and fibroblasts (hT1 + hT1-CAF)

D. Ahrens, et. al. Journal of Hematology & Oncology, 2017, 10: 76

Cancer-associated fibroblasts (CAFs) are cellular components of the desmoplastic stroma characteristic to the tumor that contributes to this treatment resistance.

2500 hT1-CAF

2500 hT1

4hrs (4X) 24hrs (4X) 48hrs (4X)

Co-culture(2500:2500)

96hrs (4X) 96hrs (10X)


10X

CellTrackergreen

OctadecylRhodamine B Chloride (R18)

hT1-3D monoculture

-10 -9 -8 -7 -6 -5 -4-25

0255075

100 DOXGEMSN-38


IC50 5.06E-07 >9.98E-06 7.76E-06

Log[Compound], M

% In

hibi

tion

Coculture

-10 -9 -8 -7 -6 -5 -4-25

0255075

100 DOXGEMSN-38


IC50 5.84E-07 >9.98E-06 >4.99E-05

Log[Compound], M

% In

hibi

tion

hT1CAF-3D monoculture

-10 -9 -8 -7 -6 -5 -4-25

0255075

100

DOXGEMSN-38


IC50 1.85E-07 3.06E-08 7.43E-09

Log[Compound], M

% In

hibi

tion

Co-culture of hT1-CAF and hT1 formed a more compact structure but not as tight as CAFs alone. Co-culture format more closely approximates the hT1 cancer cell outcome alone…

3D Co-culture 3D hT1 3D hT1-CAF

Venn diagram of compounds with IC50 < 1 µM


NCI Oncology Drugs (114) Co-culture hT1

monoculturehT1-CAF

monoculture

Replicates triplicate triplicate triplicate

S/B 211.3±10.0 193.7±13.5 192.51±8.08

Z’ 0.70±0.07 0.82±0.08 0.77±0.03

IC50 of Dox 584 nM 439 nM 262 nM

# of cmpds(IC50 < 1 µM) 7 6 16


Adapted from Bradshaw A et al Frontiers (2016) & Lathia JD et al Dev Cell (2015)

• GSCs targeting presents an attractive approach to improve treatment outcome in GBM (Classical stablished tumor cell lines HTS campaign haven’t historically shown a good correlation with clinical outcome)


0 20 40 60 800

20

40

60

80

100

Size (µm)

Freq

uenc

y

Day 2

Day 5

Day 025µm

25µm

25µm

100µm

100µm

100µm

• GSCs-spheroids presents an homogenous distribution in size and number after GSC seeding in 1536-wells

1 µm 49


Seed 1000 cells in 5 µLDay 0

Transfer compounds using 10 nL pintoolDay 2

Add 5 µL CellTiterGloDay 5

Read plates using ViewLux

Allow the spheres to form for 2 days in incubator (37oC, 5%CO2, 99% RH)

Maintain in incubator for 3 days (37oC, 5%CO2, 99% RH)

Incubate 10 min at RT


Run Statistics (n = 3 plates)• 3290 compounds tested at 2 µM nominal concentration, 1X• Ave Z’ = 0.77±0.02 • Ave S:B = 163.52±7.49• Hit cutoff (Ave+3SD) = 32.41%• Hits = 66 • Hit Rate = 2.0%

Primary Screen

-9 -8 -7 -6 -50

500

1000

1500

2000 DOX30294835

LogIC50HillSlope

IC50

DOX-7.014-1.9369.674e-008

3029-6.565-1.8572.722e-007

4835~ -7.362~ -13.21~ 4.343e-008

Log[Compound], MR

LU

GBM Approved Drug Screen

0 1000 2000 3000 4000-50

-25

0

25

50

75

100

Well

% R

espo

nse

This assay is sensitive to DMSO. Negative scatterplot is reflected in plate 3 because of empty wells

Hit cutoff


EC100: Medium +DMSO

EC0: Cells + DMSO

%CV S/B Z Z' Ave+3SD4.98 167.91 0.85 0.81 16.9

Scatterplot of 114 NCI Dose Response

0 1000 2000 3000 4000-25

0

25

50

75

100 High Control

DataLow Control

Well #

% R

espo

nse

DMSO cutoff 16.9%

Correlation between two replicates

-25 0 25 50 75 100-25

0

25

50

75

100

r² 0.9147

%Response (Replicate 1)

% R

espo

nse

(Rep

licat

e 2)

(A) Scatterplot of a DMSO plate(B) Scatterplot of 114 NCI dose response data(C) Correlation between replicates of 114 NCI dose

response data

(A) (B)

(C)


Top 5 CRC

-10 -9 -8 -7 -6 -50

25

50

75

100 RomidepsinBortezomibDocetaxelVinblastineCarfilzomib

LogIC50HillSlope

IC50

Romidepsin-8.5361.9792.910e-009

Bortezomib-8.4701.8353.388e-009

Docetaxel-8.1161.5897.664e-009

Vinblastine-8.0452.0369.015e-009

Carfilzomib-7.2920.94475.108e-008

Log[Compound], M

% C

ytot

oxic

ity


Hepatocells: derived from primary human hepatocytes (liver)hT29: human colon cancer cellshT1: patient-derived pancreatic cancer cells GBM: Glioblastoma primary brain cancer cells


Hospital

3D cell models (spheroids and organoids) are more physiologicallyrelevant than 2D monolayer cells for anti-cancer therapeutic screening.Relevance: 3D co-culture > 3D monoculture > 2D culture

Magnetic 3D bioprinting technology for HTS• Confirmed the formation of spheroids/organoids using magnetic 3D

technology.• Assay statistics and response to drugs are robust and consistent in

both 384 and 1536 well format.• Successfully completed approved drug screening in monoculture

and co-culture 3D models.• Larger chemical libraries (~150K) has been screened in 1536-well

automated platform.

Drug sensitivity shift from 2D to 3D model is cell and drug dependent,which necessitates the development of personalized medicine using 3Dcell models.



National Cancer Institute of the National Institutes of Health under IMAT Award Number

R33CA206949

1. Hou S, Tiriac H, Sridharan B, Scampavia L, Madoux F, Seldin J, Souza G, Watson D, Tuveson D, Spicer TAdvanced Development of Primary Pancreatic Organoid Tumor Models for High-Throughput Phenotypic Drug Screening. SLAS Discovery. 2018 Apr. PMID:296732792. Madoux F, Tanner A, Vessels M, Willetts L, Hou S, Scampavia L, Spicer T.A 1536-Well 3D Viability Assay to Assess the Cytotoxic Effect of Drugs on Spheroids.SLAS Discovery. 2017 Jan. PMID:28346088

a comparative analysis of 3d platforms in high throughput … · 3d results match in 384 vs 1536...

Documents