GUIDESDrug TestingAssay Methods
Practical Guide

Drug Testing Methods

Complete Compound Evaluation in Human Simulation Models

Written by J Radler | Patient Analog
Last updated: January 2025

What You'll Learn

← Back to Guides

🧪 WHY THIS MATTERS

Drug attrition due to toxicity costs the pharmaceutical industry billions annually. Human-relevant in vitro models can predict clinical toxicity earlier in development, potentially saving years of research and billions in failed trials. Proper drug testing methodology is essential for generating reliable, translatable data.

90%
Drug candidates fail in clinical trials
30%
Fail due to toxicity alone
$2.6B
Average cost per approved drug
87%
Liver-chip DILI prediction accuracy

⚠ PREREQUISITES

Required Skills

  • Established organoid or organ-chip cultures
  • Aseptic compound handling
  • Plate reader operation
  • Basic statistics and curve fitting
  • Compound stability assessment

Model Requirements

  • Validated cell model with stable baseline
  • Established functional endpoints
  • Historical data for reference compounds
  • Quality-controlled cell source
  • Reproducible culture conditions

Safety Requirements

  • Appropriate PPE for compound handling
  • Chemical fume hood access
  • MSDS review for test compounds
  • Proper waste disposal protocols
  • Spill response procedures

🧪 COMPOUND HANDLING BEST PRACTICES

Stock Preparation

  • Dissolve in DMSO at 10-100 mM
  • Verify solubility before use (no precipitate)
  • Aliquot to avoid freeze-thaw cycles
  • Store at -20C or -80C per compound stability
  • Label with compound ID, concentration, date, lot

Working Dilutions

  • Dilute in complete culture medium
  • Final DMSO: =0.1% (max 0.5%)
  • Prepare fresh dilutions day-of-use
  • Pre-warm to 37C before adding to cells
  • Always include vehicle control at same DMSO%

Stability Considerations

  • Verify stability in culture medium (37C, 5% CO2)
  • Check for precipitation after dilution
  • Some compounds require daily medium changes
  • Light-sensitive compounds: foil-wrap plates
  • Protein binding may affect free concentration

Quality Control

  • Verify compound identity (LC-MS if available)
  • Check purity (>95% for reliable results)
  • Document lot numbers in lab notebook
  • Reference standards from reputable suppliers
  • Measure actual concentrations in effluent if possible

📊 DOSE-RESPONSE EXPERIMENTAL DESIGN

Concentration Range Selection

Design 8-10 point concentration series spanning 4-5 log units using half-log (3.16x) dilutions:

100, 31.6, 10, 3.16, 1, 0.316, 0.1, 0.0316 µM

Starting concentration selection:

  • If Cmax known: Start at 10-30x Cmax
  • If unknown: Start at 100 µM (typical)
  • For high-potency compounds: Start lower
  • For biologics: Use ng/mL or pM range
Study Type Exposure Time Concentration Points Replicates Primary Use
Acute Toxicity Screen 24-48 hours 6-8 points n=3 High-throughput screening
IC50 Determination 48-72 hours 8-10 points n=3-4 Lead compound evaluation
Subchronic Toxicity 7 days 6-8 points n=4-6 Mechanism studies
Chronic Exposure 14-28 days 4-6 points n=6+ Regulatory-relevant data
Recovery Assessment 72h exposure + 7d recovery 2-3 points (at IC50) n=4-6 Reversibility studies

🔬 TOXICITY ENDPOINTS BY ORGAN

🧬 Liver (Hepatotoxicity)

Viability: ATP, LDH release

Function: Albumin secretion, urea production

Metabolism: CYP450 activity (3A4, 1A2, 2C9)

Injury markers: ALT, AST, miR-122

Morphology: Bile canaliculi, lipid accumulation

Target: Identify DILI liability early

💓 Heart (Cardiotoxicity)

Function: Beat rate, contractility, calcium transients

Electrophysiology: APD90, field potential duration

Injury markers: Troponin I/T, BNP, LDH

Structural: Sarcomere organization

Arrhythmia: EADs, beat irregularity

Target: CiPA-relevant QT/arrhythmia risk

🧬 Kidney (Nephrotoxicity)

Viability: ATP, MTT, live/dead imaging

Function: GGT activity, albumin uptake

Injury markers: KIM-1, NGAL, clusterin

Transport: OAT1/3, OCT2 activity

Morphology: Brush border integrity

Target: Proximal tubule toxicity detection

🧬 Gut (GI Toxicity)

Barrier: TEER, permeability (Lucifer Yellow)

Function: Mucus production, P-gp activity

Viability: ATP, cytokine release

Inflammation: IL-8, IL-6, TNF-alpha

Morphology: Villus structure, tight junctions

Target: Oral drug barrier disruption

🧪 RECOMMENDED ASSAY KITS

Endpoint Kit Name Supplier Detection Est. Cost
Cell Viability (ATP) CellTiter-Glo 2.0 Promega Luminescence $450/500 rxns
Cytotoxicity (LDH) CyQUANT LDH Invitrogen Absorbance $280/1000 rxns
Apoptosis (Caspase) Caspase-Glo 3/7 Promega Luminescence $380/500 rxns
Live/Dead Imaging LIVE/DEAD Kit Invitrogen Fluorescence $240/kit
Albumin (Liver) Human Albumin ELISA Bethyl Labs Colorimetric $395/96 wells
CYP3A4 Activity P450-Glo CYP3A4 Promega Luminescence $520/500 rxns
Cardiac Troponin I hcTnI ELISA Life Diagnostics Colorimetric $480/96 wells
KIM-1 (Kidney) Human KIM-1 ELISA R&D Systems Colorimetric $550/96 wells

📝 STEP-BY-STEP DRUG TESTING PROTOCOL

Day -1: Preparation

1
Verify Model Readiness

Confirm organoids/chips meet quality criteria: appropriate size/confluence, stable baseline function, no contamination. For chips, verify TEER is at target value. Image and document baseline morphology.

2
Prepare Compound Stocks

Retrieve compound from -20C storage. Allow to equilibrate to RT before opening (prevents condensation). Verify identity and concentration. Prepare intermediate dilutions in DMSO if needed.

3
Plan Plate Layout

Design plate layout with randomization to minimize edge effects. Include: vehicle controls (n=6), positive control (known toxic compound), 8-10 test concentrations (n=3-4 each). Allocate wells for baseline measurements.

Day 0: Compound Treatment

4
Collect Baseline Samples

Remove and save conditioned medium from representative wells for baseline biomarker measurements (albumin, LDH, etc.). Document baseline TEER for barrier models. Image for morphology baseline.

5
Prepare Working Dilutions

Prepare serial dilutions in pre-warmed complete medium. Work quickly to minimize compound degradation. Verify DMSO concentration is matched across all dilutions. Include vehicle-only control at same DMSO %.

6
Initiate Treatment

Remove existing medium and gently add compound-containing medium according to plate layout. For organoids, minimize disruption. For chips, ensure consistent flow rates. Document exact treatment start time.

Days 1-7: Monitoring & Sampling

7
Daily Visual Assessment

Observe cultures under microscope daily. Note any changes in morphology, detachment, or debris. Image high/low dose and vehicle wells. Document observations in lab notebook.

8
Sample Collection (Chronic Studies)

For multi-day studies: collect medium samples at defined timepoints (24h, 48h, 72h, Day 7). Store at -80C with protease inhibitors. Replenish with fresh compound-containing medium after sampling.

9
Functional Monitoring

For chips: measure TEER daily. For cardiac models: record beating activity. For liver: collect medium for albumin/urea at each medium change. Track trends over time.

Endpoint: Analysis

10
Terminal Viability Assay

At study endpoint, perform ATP viability assay (CellTiter-Glo) or equivalent. For multiplexing, perform LDH assay on spent medium first, then ATP on cells. Include lysis controls for maximum signal.

11
Biomarker Analysis

Run ELISA for tissue-specific biomarkers on collected medium samples. Normalize to cell number or protein content. Calculate fold-change vs. vehicle control.

12
Data Processing & IC50 Calculation

Normalize all data to vehicle control (100%). Fit 4-parameter logistic curve using GraphPad Prism or similar. Calculate IC50 with 95% CI. Compare to clinical Cmax to calculate therapeutic index.

💡 EXPERT TIPS

Reference Compounds

Always include known toxic (positive control) and non-toxic (negative control) reference compounds. For liver: acetaminophen, trovafloxacin. For heart: dofetilide, terfenadine. Build historical database.

Edge Effects

Outer wells evaporate faster. Either avoid using edge wells or fill with sterile water/PBS as evaporation barriers. Randomize plate layout to distribute edge effects across conditions.

Z-Factor Validation

Calculate Z' factor for assay quality: Z' = 1 - (3*(SD_pos + SD_neg) / |mean_pos - mean_neg|). Z' > 0.5 indicates excellent assay performance. Run before compound testing.

Compound Binding

Lipophilic compounds bind to plastic and PDMS. Pre-saturate plates with compound-containing medium. Measure actual concentrations in medium if critical. Consider glass or low-binding plates.

🔧 TROUBLESHOOTING GUIDE

Problem Possible Causes Solutions
Vehicle toxicity DMSO concentration too high; Cell type sensitive Reduce final DMSO to =0.1%. Test DMSO tolerance curve for your cell type. Use alternative vehicles (PBS, ethanol) if compatible.
Compound precipitation Poor solubility; Protein binding; Wrong vehicle Verify solubility in vehicle. Use cyclodextrin or co-solvents. Pre-warm medium. Lower concentration. Filter through 0.22×m.
Variable IC50 values Cell batch variation; Compound instability; Assay drift Standardize cell source and passage. Use fresh compound aliquots. Include reference compound on every plate. Track historical data.
Poor curve fit Insufficient concentration range; Bell-shaped response Extend concentration range. Add more points around IC50. Check for hormesis at low doses. Verify complete killing at high dose.
High background Assay interference; Autofluorescence; Compound color Include compound-only wells (no cells) as background. Choose orthogonal assay method. Use longer wavelength detection.
No toxicity detected Concentration too low; Wrong endpoint; Model insensitive Increase top concentration. Extend exposure time. Use more sensitive endpoint. Verify positive control works. Consider alternative model.
Edge well artifacts Evaporation; Temperature gradients Fill edge wells with sterile PBS. Use breathable seals. Randomize plate layout. Pre-equilibrate plates in incubator.
Contamination Non-sterile compound; Poor technique Filter-sterilize compound stocks (0.22×m). Use antibiotics during treatment. Work in BSC. Minimize plate opening.
Delayed toxicity Metabolism required; Slow mechanism; Accumulation Extend exposure time (7-14 days). Use metabolically competent cells. Add repeat dosing. Check for cumulative effects.
Discordant results Model limitations; Species differences; Wrong context Compare to clinical data. Use multiple models. Consider mechanistic relevance. Document model limitations in reports.

📈 DATA ANALYSIS & REPORTING

IC50 Calculation

Fit dose-response data to 4-parameter logistic (4PL) equation:

Y = Bottom + (Top - Bottom) / (1 + 10^((LogIC50 - X) * HillSlope))
Top: Maximum response (vehicle control = 100%)
Bottom: Minimum response (maximum toxicity)
IC50: Concentration at 50% inhibition
HillSlope: Steepness of curve (typically 0.5-2)

Therapeutic Index Calculation

The Therapeutic Index (TI) compares in vitro toxicity to clinical exposure:

TI = TC50 (or IC50) / Cmax
  • TI less than 10: High risk - toxicity likely at therapeutic doses
  • TI 10-30: Moderate risk - monitor closely in development
  • TI greater than 30: Low risk - adequate safety margin

❓ FREQUENTLY ASKED QUESTIONS

Q: How do I choose between acute and chronic exposure studies?
A: Acute studies (24-72h) are appropriate for high-throughput screening and identifying overt cytotoxicity. Chronic studies (7-28 days) are needed for compounds with slow mechanisms, cumulative toxicity, or metabolism-dependent effects. For regulatory submissions, chronic exposure data at clinically relevant concentrations is typically required. Start with acute screening, then follow up hits with chronic studies.
Q: How do I account for protein binding in my experiments?
A: Highly protein-bound compounds have lower free (unbound) concentrations. Measure free fraction in your culture medium (equilibrium dialysis or ultrafiltration). Calculate free IC50 and compare to free Cmax. For highly bound compounds (>95%), consider using serum-free medium or adjusting interpretation accordingly. Always report both total and free concentrations when possible.
Q: What is the minimum number of replicates needed?
A: For IC50 determination: minimum n=3 biological replicates per concentration. For regulatory-quality data: n=4-6 is preferred. For high-throughput screening: n=2 technical replicates may suffice for initial ranking. Always perform independent experiments (different cell batches, different days) to assess reproducibility. Report both technical and biological variability.
Q: How do organoid and organ-chip results compare to animal studies?
A: Human organoids and organ-chips often outperform animal models for human-relevant toxicity. For DILI prediction, liver-chips show ~87% sensitivity vs. ~50% for animal studies. However, in vitro models lack systemic exposure, immune components, and multi-organ interactions. They complement but don't fully replace animal studies yet. Use them for early screening and mechanistic understanding, with animal studies for systemic effects.
Q: Should I use primary cells or iPSC-derived cells for drug testing?
A: Primary human cells (especially hepatocytes) provide the most physiologically relevant response and are preferred for regulatory studies. iPSC-derived cells offer patient-specific modeling, unlimited supply, and batch consistency but may have lower metabolic activity (especially for CYP450). Consider: primary cells for final validation and regulatory; iPSC cells for screening, patient-specific modeling, and early discovery.
Q: How do I validate my assay for regulatory submission?
A: Document: (1) Assay precision (intra- and inter-assay CV, typically <20%), (2) Sensitivity and specificity with reference compounds, (3) Dynamic range and linearity, (4) Acceptance criteria (Z' factor >0.5), (5) Reference compound performance tracking. Follow IQ MPS consortium guidelines. Consider pre-submission meeting with FDA to discuss context of use and validation requirements.
Q: What controls should I include in every experiment?
A: Essential controls: (1) Vehicle control (same DMSO % as treated, n=6), (2) Positive control (known toxic compound at IC50), (3) Negative control (known non-toxic compound), (4) No-cell background wells, (5) Maximum signal control (lysis for viability assays). Include all controls on every plate. Track historical performance of controls over time.
Q: How do I multiplex multiple endpoints?
A: Plan endpoint order from least to most destructive: (1) Non-invasive imaging first, (2) Collect supernatant for secreted markers (LDH, albumin), (3) Perform non-lytic assays on live cells (calcium, mitochondrial function), (4) End with lytic assays (ATP, protein). Some vendors offer multiplexed kits (e.g., MultiTox-Fluor + CellTiter-Glo). Validate that earlier measurements don't affect later ones.
Q: What if my compound shows toxicity only in specific organ models?
A: Organ-specific toxicity is expected and valuable! It indicates the compound's target organ for adverse effects. Document the differential sensitivity across models. Compare to known drugs with similar profiles. For drug development, this helps design clinical monitoring strategies. Multi-organ-on-chip systems can help understand organ-organ interactions and systemic effects.
Q: How do I report results for publication or regulatory submission?
A: Include: (1) Full protocol details (cell source, passage, medium, exposure time), (2) IC50/EC50 with 95% CI, (3) Representative dose-response curves, (4) Comparison to clinical exposure (Cmax, therapeutic index), (5) Reference compound performance, (6) Raw data availability. Follow MIAME-style reporting for reproducibility. For regulatory: discuss with FDA early regarding acceptable formats and endpoints.

🔗 RELATED CONTENT

SCIENCE
Liver Toxicity Testing
DILI prediction methods
SCIENCE
Cardiac Safety Testing
CiPA guidelines and QT risk
GUIDE
Biomarker Selection
Choosing right endpoints
GUIDE
Organ-Chip Protocols
MPS technical methods
GUIDE
Data Analysis
Statistical methods for MPS
REGULATORY
FDA Modernization Act
NAMs in drug development

🎯 NEXT STEPS

  1. Establish Baseline: Characterize your model's baseline function before testing compounds
  2. Validate with References: Test known toxic/non-toxic compounds to establish assay performance
  3. Optimize Protocol: Titrate DMSO tolerance, concentration range, and exposure time for your model
  4. Build Historical Data: Track reference compound performance over time for quality control
  5. Document Everything: Maintain detailed records for regulatory compliance and reproducibility
RELATED
Chip Protocols →
NEXT
Biomarker Selection →
← Back to All Guides

Implementation Pathway

PhaseActivitiesTimeline
PlanningDefine objectives, select platform1-2 months
SetupInstallation, training, protocols2-3 months
ValidationTesting, regulatory engagement6-12 months

Next Steps

🎯

MPS Technology

Platform deep dive

🎯

Personalized Medicine

Patient approaches

🎯

FDA ISTAND

Submission pathways

Frequently Asked Questions

What drug testing protocols work best on organ chips?

Protocols include acute toxicity testing (24-72 hour single dose), repeated dosing (daily treatments over 7-14 days), dose-response curves (6-10 concentrations), time-course studies (sampling multiple timepoints), and mechanistic investigations (combining drug with pathway inhibitors).

How do you determine drug concentrations for chips?

Use free plasma concentrations from clinical PK studies (unbound drug reaching tissues), typically 0.1-100x therapeutic Cmax. Include concentrations triggering toxicity in patients. Test 6-10 concentrations spanning 3-4 log range to capture full dose-response relationship.

What are positive and negative control compounds?

Positive controls are drugs with known toxicity in humans (acetaminophen for liver, dofetilide for heart). Negative controls are drugs proven safe in clinic. Testing 10-20 reference compounds validates that chip correctly identifies toxic versus safe drugs before testing novel compounds.

How long should drug exposures last?

Acute toxicity: 24-72 hours. Subacute: 7-14 days repeated dosing. Chronic requires extended culture platforms (21+ days). Exposure duration should match clinical use—single dose antibiotics need short exposure while daily cardiovascular drugs need repeated dosing.

Can chips test drug combinations?

Yes. Chips test synergistic combinations for cancer, HIV, and tuberculosis. Checkerboard designs test matrix of dose combinations. High-throughput arrays screen thousands of combinations. Chips reveal whether drugs interact through metabolism, transporters, or combined toxicity.

What is organ chip equivalence to animal dosing?

No direct equivalence exists. Use human therapeutic free plasma concentrations as starting point. Chips avoid first-pass metabolism and protein binding complications of whole animals. Focus on human-relevant concentrations rather than trying to match animal mg/kg doses.

How do you test prodrugs on organ chips?

Prodrugs require metabolic activation. Multi-organ chips link gut and liver enabling intestinal absorption and hepatic activation before target organ exposure. Alternatively, pre-activate prodrugs in liver chips then transfer metabolites to target tissue chips.

What are QC criteria for drug testing experiments?

Criteria include cell viability above 90 percent in vehicle controls, functional markers within validated ranges (albumin production, electrical activity), barrier integrity maintained, and appropriate response to positive controls. Failed QC invalidates experiments.

How many replicates are needed?

Biological replicates: 3-6 chips per condition (different cell preparations or batches). Technical replicates: 2-3 measurements per chip. For regulatory submissions, increase to 6-10 biological replicates demonstrating reproducibility across chip lots.

What is the future of chip drug testing?

Future includes automated platforms testing hundreds of drugs simultaneously, AI predicting toxicity from chip data, integration with patient genomics for precision medicine, and regulatory acceptance replacing animal studies for specific endpoints like hepatotoxicity and cardiotoxicity.