In Vitro Assessment of Drug-Induced Liver Injury (DILI) using a High Content Cellular Imaging System Mark Wolfe1, Shannon Einhorn2, Vanessa Ott2, and Haiyan Ma1

(1)MPI Research Inc., Mattawan, MI (2)Cellular Dynamics International, Madison, WI

Society of Toxicology Annual Meeting, March 10-14, 2013

METHODOLOGY

OBJECTIVES

INTRODUCTION RESULTS

Drug-induced liver injury (DILI) is a leading cause of

drugs failing during clinical trials and being withdrawn

from the market. Idiosyncratic drug hepatotoxicity has

not been well predictive because of its low concordance

with either standard in vitro cytotoxicity screening assay

results or regulatory animal study findings. Implementing

an in vitro cell-based predictive assay early in the drug

discovery process would help improve early compound

attrition and develop safer drug candidates. The Thermo

Scientific ToxInsight® IVT platform and DILI Assay

Cartridge offer the tools to determine the hepatotoxicity

risk of a compound by measuring multiple biomarkers of

hepatotoxicity in individual cells. The high content

imaging approach increases the sensitivity and specificity

for predicting hepatotoxicity by simultaneously detecting

five multiplexed cellular targets and properties associated

with cell loss, DNA content, cellular redox stress, and

mitochondrial stress. Combining this testing system with

a proper cellular model will further help improve the

predictability of human hepatotoxicity cost effectively.

To test the potential hepatotoxicity of nine reference test articles

in three hepatocyte model systems using the ToxInsight® DILI

Assay Cartridge. The specificity and sensitivity of the assay will

be evaluated.

• Thermo Scientific ToxInsight®

• Cell Model of Choice:

• Human hepatoma Hep G2 cell line

• Induced Pluripotent stem cell (iPSC)-derived human hepatocytes

• Fresh human primary hepatocytes

• Culture cells in a 96-well plate at 37°C overnight.

• Treat the cells with test article, vehicle control, and negative control:

• 8 concentration responses in triplicate

• 100X Cmax (if possible) with 1:2 dilutions up to 8 concentrations and 0.5%

final DMSO concentration

• Incubate with test reagents for 22-24 hours.

• Stain the cells with four fluorescent probes:

• Hoechst 33342

• Cell number

• Nuclear DNA intensity

• Monochlorobimane (mBCl)

• Reduced cellular glutathione levels

• ROS Dye

• Reactive Oxygen Species

• Mitochondria Dye

• Mitochondrial membrane potential

• Cell Imaging on the ToxInsight®

• Data Analysis to quantify the multi-channel image outputs

• Hepatotoxicity prediction using the Thermo Scientific DILI Assay

Excel Template

Hep G2 Cells

COMPOUNDS TESTED SUMMARY OF RESULTS

All three human hepatocyte models were stained efficiently with

the four dyes, and allowed for quantification of all five cellular

targets with the specific imaging protocol established for each

hepatocyte model on the ToxInsight® system.

The Hep G2 cells demonstrated a 100% specificity and 80%

sensitivity with the nine reference compounds that were tested in

the in vitro DILI assay.

There was a high level of reduced cellular glutathione levels in the

iPSC-derived hepatocytes compared to the Hep G2 cells and

fresh primary hepatocytes.

The iPSC-derived hepatocytes and fresh human primary

hepatocytes demonstrated a 100% specificity and a 100%

sensitivity with the nine reference compounds that were tested in

the in vitro DILI assay.

The sensitivity for the toxicity curves is shifted for some test

articles, based on which source of hepatocytes they were tested

in.

The rank order of toxicity for the test articles differs, based on the

source of hepatocytes.

CONCLUSIONS

All three human hepatocyte models can be used in the in vitro

DILI high content imaging assay using the ToxInsight® system.

From the nine compound reference set, the three human

hepatocyte models demonstrated high specificity and sensitivity,

except for the lack of Aflatoxin B1 toxicity in the Hep G2 cells.

The lack of toxicity of Aflatoxin B1 at the concentrations tested in

the Hep G2 cells in the in vitro DILI assay is most likely due to the

lack of metabolic activity in this human carcinoma cell line. Both

the iPSC-derived hepatocytes and primary hepatocytes are

known to retain more hepatocyte functions and metabolic

activities. The benefit of iPSC-derived hepatocytes over primary

hepatocytes include their availability and data reproducibility.

The use of eight concentrations for each test article in triplicate

helps to capture the potential hepatotoxic effects of a test article.

The sensitivity for individual toxicity cellular parameters appears

to vary depending on the source of hepatocytes. This may

suggest that one source of hepatocytes may provide a better

hepatotoxicity model for a series of test articles.

Further testing for reproducibility within a hepatocyte model

should be conducted to determine if one model lends itself to a

more consistent toxicity profiling model; i.e., passage number,

donor characteristics, and metabolic activity.

REFERENCES &

ACKNOWLEDGEMENTS

1. Xu J.J. et al. (2008) Cellular Imaging Predictions of Clinical Drug-

Induced Liver injury. Toxicol Sci. 105 (1): 97-105

2. O’Brien P.J. et.al. (2006) High concordance of drug-induced

human hepatotoxicity with in vitro cytotoxicity measured in a

novel cell-based model using high content screening. Arch.

Toxicol. 80: 580-604

3. Thermo Scientific ToxInsight Drug Induced Liver Injury (DILI)

Assay User Guide version 1.0 (2011)

4. Jubert C. et.al. (2009) Effects of Chlorophyll and Chlorophyllin

on Low-Dose Aflatoxin B1 Pharmacokinetics in Human

Volunteers. Cancer Prev. Res. 2(12): 1015-1022

I would like to acknowledge MPI Research management team for

supporting the work involved in establishing the present data set.

I would like to acknowledge Carter Cliff from CDI for arranging the

collaborative efforts between MPI Research and CDI research teams.

Human iPSC-derived Hepatocytes

Fresh Human Primary Hepatocytes

The Boolean decision table from the DILI Assay Analysis Template: For each compound, if any of the five target responses go beyond the toxicity threshold

in the concentration range, the compound is flagged as toxic (red box). Out of the five targets, the number of targets scored as toxic is also reported.

Dose response plots, from the DILI Assay Analysis Template, of all the compounds for all five cell targets. Shown are the normalized responses and the

dashed lines are the toxicity thresholds.

Fluoxetine 25X Cmax Mefenamic Acid 25X Cmax Nalidixic Acid 50X Cmax Phenylbutazone 3.1X Cmax

Representative image overlay of the Hep G2 cells post-compound treatment. DILI fluorescent probes were used to monitor the five cellular targets:

Dark blue identifies the nucleus and nuclear content of each cell; lighter blue identifies the reduced cellular glutathione levels; green identifies the reaction

oxygen species; and red identifies the mitochondria membrane potential.

Phenylbutazone 6.25X Cmax Aspirin 100X Cmax Nalidixic Acid 50X Cmax Aflatoxin B1 25X Cmax Mefenamic Acid 6.25X Cmax Melatonin 100X Cmax Nalidixic Acid 25X Cmax Aflatoxin B1 100X Cmax

Staining of nuclei (blue), cellular glutathione (light blue), ROS (green), and mitochondria membrane potential (red) in Hep G2 cells after treatment with the test

compounds. The concentrations of each compound are listed at a fold of the literature Cmax. All compounds were tested at 100X Cmax with a 1:2 serial

dilution except Phenylbutazone (12.5X) and Mefenamic Acid (50X), due to solubility. Aflatoxin B1was tested at 2500X Cmax to increase range-finding toxicity.

The Boolean decision table from the DILI Assay Analysis Template: For each compound, if any of the five target responses go beyond the toxicity threshold

in the concentration range, the compound is flagged as toxic (red box). Out of the five targets, the number of targets scored as toxic is also reported.

Dose response plots, from the DILI Assay Analysis Template, of all the compounds for all five cell targets. Shown are the normalized responses and the

dashed lines are the toxicity thresholds.

The Boolean decision table from the DILI Assay Analysis Template: For each compound, if any of the five target responses go beyond the toxicity threshold

in the concentration range, the compound is flagged as toxic (red box). Out of the five targets, the number of targets scored as toxic is also reported.

Dose response plots, from the DILI Assay Analysis Template, of all the compounds for all five cell targets. Shown are the normalized responses and the

dashed lines are the toxicity thresholds.

Staining of nuclei (blue), cellular glutathione (light blue), ROS (green), and mitochondria membrane potential (red) in iPSC-derived hepatocytes after

treatment with the test compounds. The concentrations of each compound are listed at a fold of the literature Cmax. All compounds were tested at 100X

Cmax with a 1:2 serial dilution except Phenylbutazone (12.5X) and Mefenamic Acid (50X), due to solubility. Aflatoxin B1was tested at 2500X Cmax to

increase range-finding toxicity.

Staining of nuclei (blue), cellular glutathione (light blue), ROS (green), and mitochondria membrane potential (red) in primary hepatocytes after treatment

with the test compounds. The concentrations of each compound are listed at a fold of the literature Cmax. All compounds were tested at 100X Cmax with

a 1:2 serial dilution except Phenylbutazone (12.5X) and Mefenamic Acid (50X), due to solubility. Aflatoxin B1was tested at 2500X Cmax to increase

range finding -toxicity.

Representative image overlay of the iPSC-derived hepatocytes post-compound treatment. DILI fluorescent probes were used to monitor the five cellular

targets: Dark blue identifies the nucleus and nuclear content of each cell; lighter blue identifies the reduced cellular glutathione levels; green identifies the

reaction oxygen species; and red identifies the mitochondria membrane potential.

Representative image overlay of the primary hepatocytes post-compound treatment. DILI fluorescent probes were used to monitor the five cellular targets:

Dark blue identifies the nucleus and nuclear content of each cell; lighter blue identifies the reduced cellular glutathione levels; green identifies the reaction

oxygen species; and red identifies the mitochondria membrane potential.

Summary of the Boolean decision table from the DILI Assay Analysis Template. The non-hepatotoxic compounds are in the table on the left above, and the

known hepatotoxic compounds that were tested are listed in the table on the right above. Multiparametric indices, based on the strength of the multiplexed

response over the concentration range, are used to rank the compounds (1 = strongest response).

Negative Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Sample Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Rank

Aspirin - - - - - 0 Aflatoxin B1 - - - - - 0 6

Fluoxetine - - - - - 0 Mefenamic Acid + + + - + 4 Caution 2

Melatonin - - - - - 0 Nalidixic Acid + + + - + 4 Caution 3

Phenylbutazone + + + - + 4 Caution 3

Ticlopidine + + + - - 3 Caution 5

Troglitazone + + + - + 4 Caution 1

Summary of the Boolean decision table from the DILI Assay Analysis Template. The non-hepatotoxic compounds are in the table on the left above, and the

known hepatotoxic compounds that were tested are listed in the table on the right above. Multiparametric indices, based on the strength of the multiplexed

response over the concentration range, are used to rank the compounds (1 = strongest response).

Negative Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Sample Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Rank

Aspirin - - - - - 0 Aflatoxin B1 - + + + + 4 Caution 3

Fluoxetine - - - - - 0 Mefenamic Acid + + + + + 5 Caution 2

Melatonin - - - - - 0 Nalidixic Acid + + + + + 5 Caution 5

Phenylbutazone + + + + + 5 Caution 1

Ticlopidine - + + + + 4 Caution 6

Troglitazone + + + + + 5 Caution 4

Summary of the Boolean decision table from the DILI Assay Analysis Template. The non-hepatotoxic compounds are in the table on the left above, and the

known hepatotoxic compounds that were tested are listed in the table on the right above. Multiparametric indices, based on the strength of the multiplexed

response over the concentration range, are used to rank the compounds (1 = strongest response).

Negative Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Sample Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Rank

Aspirin - - - - - 0 Aflatoxin B1 + + + + + 5 Caution 6

Fluoxetine - - - - - 0 Mefenamic Acid + + + + + 5 Caution 2

Melatonin - - - - - 0 Nalidixic Acid + + + + + 5 Caution 2

Phenylbutazone + + + + + 5 Caution 4

Ticlopidine + + + + + 5 Caution 1

Troglitazone - + + - + 3 Caution 4

Negative Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Sample Compounds

Cell

Loss DNA GSH ROS MMP #Pos Result Rank

Aspirin - - - - - 0 Aflatoxin B1 - - - - - 0 6

Fluoxetine - - - - - 0 Mefenamic Acid + + + - + 4 Caution 2

Melatonin - - - - - 0 Nalidixic Acid + + + - + 4 Caution 3

Phenylbutazone + + + - + 4 Caution 3

Ticlopidine + + + - - 3 Caution 5

Troglitazone + + + - + 4 Caution 1

Compound Drug’s Trade Name Cmax Dose

(g/mL)

Pharmacological Action Known to be

Hepatotoxic

Phenylbutazone Phenylbutazone 150 Nonsteroidal anti-inflammatory Yes

Mefenamic Acid Ponstel 6.5 Cyclooxygenase inhibitor. NSAID (used to treat

pain)

Yes

Nalidixic Acid Nevigramon, Neggram,

Wintomylon

30 Quinolone antibiotic, Antibacterial Yes

Ticlopidine Ticlid 2.13 Fibrinolytic agent, platelet inhibitor Yes

Troglitazone Rezulin, Resulin,

Romozin

2.82 Antineoplastic, Fibrinolytic, Hypoglycemic

agent, Platelet inhibitors, Vasodilator,

Thiazolidinedione

Yes

Aflatoxin B1 Aflatoxin B1 0.001 Carcinogenic mycotoxin Yes

Aspirin Aspirin 0.995 Fibrinolytic agent, Platelet inhibitors,

Cyclooxygenase inhibitor

No

Fluoxetine Prozac, Sarafem 0.015 Serotonin uptake inhibitor No

Melatonin Melatonin, Circadin 0.00557 Circadian rhythm hormone, antioxidant,

anticonvulsant, free radical scavenger

No