Resources - Application Notes

Cost-Effective Automated Hepatotoxicity Testing using in vitro 3D Bioprinting


Related Products: Cytation 5

August 17, 2017

Authors: Brad Larson, Principal Scientist, BioTek Instruments, Inc., Winooski, VT USA; Alexa Rosenberg, Kristina Klette, and Michael Mouradian, Hamilton Company, Reno, NV USA; Paul Cizdiel, REPROCELL Inc., Yokohama, Japan; Glauco R. Souza, Nano3D Biosciences, Inc., Houston, TX USA; Glauco R. Souza, University of Texas Health Science Center at Houston, Houston, TX USA

     REPROCELL     n3D


Hepatocytes were aggregated into 3D spheroids using a magnetic particle-based bioprinting technology. Aggregation in 96-well microplates was controlled by using a pin shaped magnet drive over the course of 48 hours. All cell plating and reagent additions for hepatotoxicity testing were automated using a cost-effective liquid handling system which provided suitable dose responses from known hepatotoxins.


Prescription medications, environmental toxins, and non-prescription herbal remedies together form the major causes of hepatic injury. When looking at drugs alone, induced hepatic injury is the most common reason cited for warnings or withdrawal of an approved drug, depending on the severity of the induced hepatotoxicity. Due to this reality, a paradigm shift has taken place in the way toxicology studies are being performed, including determination of hepatotoxicity. Early phase assessment of potential hepatotoxic effects from a wider range of drug candidates and existing drugs or chemicals is now the norm. Coupled with the change in the development phase where testing takes place, is the desire to use cell models that close the gap between in vitro and in vivo systems.

Hepatotoxicity studies historically were performed by repeatedly dosing hepatocytes cultured on the bottom of a microplate with multiple concentrations of a test drug or compound. Because hepatocytes rapidly dedifferentiate, lose metabolic activity, and lack the communication networks found in vivo when cultured in this manner, results may be inaccurate and yield misleading claims regarding the safety of the test agent. To combat this shortfall, three-dimensional (3D) spheroidal models, such as those using hiPSC-hepatocytes, can be incorporated that allow cells to aggregate and retain typical long-term viability, functionality and communication found in vivo; allowing for the generation of repeatable, accurate data.

To meet the demand for increased hepatotoxicity testing, automation has been incorporated to streamline the procedure and reduce the need for large scale manual manipulations. Typically, included liquid handling systems have been large, expensive, and many times required placement into clean rooms for sterile processing. While this type of solution is suitable for pharmaceutical, biotech, and even larger core facilities, the size and cost can be prohibitive to the typical academic research lab. Therefore, a smaller, less expensive instrumentation set, which can still provide accurate and repeatable results, is necessary.

Here we demonstrate the ability to combine liquid handling with a novel cell imaging multi-mode reader to perform automated 3D hepatotoxicity studies. hiPSC-hepatocytes were aggregated into spheroids easily and efficiently using magnetic bioprinting technology. Dose responses from multiple known hepatotoxicants were evaluated with this system to demonstrate utility.

Materials and Methods



ReproHepato™ hiPSC-Hepatocytes (Catalog No. RCDH001N) were donated by REPROCELL (Yokohama, Japan).

Assay and Experimental Components

The 96-Well BiO Assay™ Kit (GBO Catalog No. 655846, consisting of NanoShuttle™-PL, 6-Well Levitating Magnet Drive, 96-Well Spheroid and Holding Magnet Drives (2), 96-Well Deep Well Mixing Plate, 6-Well and 96-Well Clear Cell Repellent Surface Microplates), and additional Collagen Type I coated 6-well microplates (Catalog No. 657950) and Cell Repellent Surface 6-Well (Catalog No. 657860) microplates were donated by Nano3D Biosciences, Inc., (Houston, TX) and Greiner Bio-One (Monroe, NC). ReproHepato Culture Medium (Catalog No. RCDN101) and ReproHepato Assay Medium (Catalog No. RCDH301) were donated by ReproCELL. P450-Glo™ CYP3A4 Luc-IPA (Catalog No. V9002), CellTiter-Glo® 3D assays (Catalog No. G9683) and CellTox™ Green Dye (Catalog No. G8731) were obtained from Promega Corporation (Madison, WI).

Inhibitor Compounds

Daunorubicin (Catalog No. ALX-380-043-M010), Aflatoxin B1 (Catalog No. ALX-630-093-M001), Tamoxifen (Catalog No. ALX-550-095-G001), Troglitazone (Catalog No. BMLGR210-0005), Nicardipine (Catalog No. ALX-550-273-G001), Mitomycin C (Catalog No. BMLGR311- 0002), Simvastatin (Catalog No. BML-G244-0050), and Phenylbutazone (Catalog No. ALX-430-112-G005) were donated by Enzo Life Sciences (Farmingdale, NY).

Cytation™ 5 Cell Imaging Multi-Mode Reader

Cytation 5 from BioTek Instruments is a modular multimode microplate reader combined with automated digital microscopy. Filter- and monochromator-based microplate reading are available, and the microscopy module provides up to 60x magnification in fluorescence, brightfield, color brightfield and phase contrast. The instrument can perform fluorescence imaging in up to four channels in a single step. With special emphasis on live-cell assays, Cytation 5 features temperature control to 65 °C, CO2/O2 gas control and dual injectors for kinetic assays, and is controlled by integrated Gen5™ Microplate Reader and Imager Software. The software was also used for dual-masking and automated analysis.

Microlab® NIMBUS® Automated Multichannel Pipetting Workstation

The compact Microlab NIMBUS automated multichannel pipetting workstation family from Hamilton Robotics includes the NIMBUS4 Enclosed. This model features four independent 1000 μL air-displacement pipetting channels, a locking cover set to minimize environmental contamination, and optional Clean Air Protection (CAP) system with HEPA filter for positive sterile airflow. It uses proprietary Compressed O-Ring Expansion (CO-RE®) technology to attach disposable tips without mechanical stress or aerosol creation. Features include anti-droplet control (ADC™), monitored air displacement (MAD) for real-time tracking of aspiration performance, and a traceable digital audit trail using Total Aspirate and Dispense Monitoring (TADM™) for robust, reliable operation. The highly accurate system was used for all reagent and cell aspirate, dispense, wash and transfer steps, and to perform plate-wide compound serial dilutions.


3D Hepatocyte Cell Preparation

ReproHepato hiPSC-Hepatocytes were thawed from cryopreservation, resuspended in ReproHepato Culture Medium, plated onto Collagen Type I coated 6-well microplates, and incubated at 37 ºC/5% CO2 for four days. After which, a 100 μL volume of NanoShuttle-PL was added to the microplate wells as part of the 96-Well BiO Assay Kit protocol (Figure 1), and the plates were returned to the incubator for one additional day. After the combined incubation period of five days, the cells were removed from the collagen-coated microplates and added to Cell Repellent Surface 6-Well microplates. A 6-well magnet drive was placed atop the plates to levitate the cells, where extracellular matrix (ECM) formation was induced during an eight-hour incubation at 37 ºC/5% CO2. After incubation, cells and ECM were broken up and resuspended. Using the NIMBUS4 Enclosed, a total of 5000 cells in 100 μL of medium were added to wells in four 96-well cell repellent microplates. A magnet was placed under each plate, and the plates were incubated at 37 ºC/5% CO2 for approximately 48 hours to allow the cells to aggregate into spheroids within each well, and to allow the hepatocytes to further mature after the combined seven-day.

Figure 1. 96-Well BiO Assay™ Kit protocol.

Figure 1. 96-Well BiO Assay™ Kit protocol.

3D Hepatocyte Cell Health Assessment

Medium was manually removed from wells and replaced with 50 μL of pre-warmed medium containing Luc-IPA substrate. The plate was then incubated at 37 ºC/5% CO2 for one (2D) or four hours (3D). Following incubation, 50 μL of supernatant was transferred to a separate white 96-well plate, and an equal volume of P450-Glo Luciferin Detection Reagent (LDR) was added to the same wells. The plate was shaken and incubated for 20 minutes at room temperature (RT) followed by luminescent signal detection. In the original cell plate, 50 μL of CellTiter-Glo 3D reagent was added to the wells, the plate was shaken at RT for 1 minute (2D) or 5 minutes (3D), followed by an additional 25 minute RT incubation. The luminescent signal was once again quantified.

Automated Long-Term Hepatotoxicity Testing

NIMBUS4 Enclosed was used to perform all steps of the automated hepatotoxicity testing procedure. Serial titrations of eight known hepatotoxins Daunorubicin, Aflatoxin B1, Tamoxifen, Troglitazone, Nicardipine, Mitomycin C, Simvastatin, and Phenylbutazone were prepared. After which, culture medium was removed and replaced with medium containing compound in four separate spheroid-containing plates (Figure 2). This process was repeated every 48 hours. Following a 1, 3, 5, and 7-day incubation, culture medium containing compound was removed from the wells of a single plate and replaced with ReproHepato Assay Medium containing Hoechst 33342 and CellTox Green Dye. The plate was incubated for five hours, then fluorescence images were captured using the Cytation 5 Cell Imaging Multi-Mode Reader.

. 96-well spheroid microplate hepatotoxicant dosing scheme.

Figure 2. 96-well spheroid microplate hepatotoxicant dosing scheme.

Results and Discussion

Image-Based Spheroid Formation Validation

Brightfield, label-free live cell imaging was performed throughout the hepatocyte aggregation process to ensure spheroid formation (Figure 3). As is evident in the three brightfield images, hepatocyte spheroid formation required 48 hours.

Spheroid formation conformational imaging

Figure 3. Spheroid formation conformational imaging. Brightfield images captured of hepatocyte spheroid formation. using 2.5x objective following 5000 hepatocyte addition to 96-well cell repellent plate and (A) 24, (B) 36, and (C) 48-hour incubation.

Confirmation of Long-Term 3D Hepatocyte Cell Health

To ensure that the observed results of long-term hepatotoxicity testing are truly from the interaction of test molecules with 3D hepatocyte spheroids, it is essential to confirm that untreated aggregated cells remain healthy during the dosing period. This was carried out using the P450-Glo CYP3A4 Luc-IPA and CellTiter-Glo 3D assays to assess CYP3A4 activity and cell viability, respectively. Assays were performed immediately following spheroid formation and for 1, 3, 5 and 7 days subsequent. Media exchanges were performed every 48 hours on nonassayed wells.

Cell health assessment of long-term 3D hepatocyte cultures.

Figure 4. Cell health assessment of long-term 3D hepatocyte cultures. Raw RLU and normalized percent of initial (A) P450-Glo CYP3A4; and (B) CellTiter-Glo 3D values plotted after 0, 1, 3, 5, and 7 day incubations at 37 ºC/5% CO2. n=6 for each condition and timepoint tested.

As seen in Figure 4, the raw luminescent units (RLU) from both the CYP3A4 activity and viability assays remain consistent at each assessed incubation period. By normalizing the raw values from subsequent timepoints to those from the initial readings taken at time 0, it is evident that ReproHepato hiPSC-Hepatocytes aggregated into 3D spheroids maintain levels of CYP3A4 activity and cell health throughout the complete 7-day hepatotoxicity testing period that are equivalent to levels seen immediately following spheroid formation. This finding agrees with previously published findings regarding 3D hepatocyte cultures1.

Automated Long-Term Hepatotoxicity Testing

To aid in the discernment of the hepatotoxic effect of each test compound on the 3D aggregated hepatocytes, spheroid images were taken on multiple z-planes using the z-stacking and z-projection capabilities in the Gen5 Software (Table 1 and Figure 5A-C).

Table 1. Gen5 z-stacking and projection criteria.

Table 1. Gen5 z-stacking and projection criteria.

The imaging bottom of the 96-well plate used was optimized so that focusing took place at the most bottom location of the spheroids closest to the well bottom. This allowed for a fixed focus setting to be used which served to increase the speed of image capture. The number of slices chosen allowed for focusing on the cells within the spheroid ranging from those at the bottom middle of the spheroid to those at the widest portion of the spheroid. The step size incorporated was the depth of field for a 10x objective, ensuring that all cells within the total included span of z-heights would be imaged.

Following image capture, a final z-projected image was created using the Focus Stacking projection method (Figure 5D). This algorithm allows for only the most in focus portions of each image within the stack to be included in the final projection. The combination of the image capture and projection method enables a primary object mask to be placed around the widest portion of the spheroid, which serves to create the most accurate size and area measurements possible.

Z-stacking and projection of 3D spheroid images

Figure 5. Z-stacking and projection of 3D spheroid images. (A-C) Images captured at individual z-planes. (D) Final z-projected image of hepatocyte spheroid. DAPI: Hoechst 33342 stained nuclei, GFP: CellTox Green stained dead cells.

Analysis of 3D Captured Images

The ability of the Cytation 5 and Gen5 Software to accurately detect the increasing number of CellTox Green labeled necrotic nuclei due to induced hepatotoxicity was initially examined. As seen in Figure 6, untreated hepatocyte spheroids exhibit a low level of cellular necrosis (Figure 6A). The number of green labeled necrotic cell nuclei within each z-projected image then increases following a 24-hour exposure to increasing doses of troglitazone (Figures 6B and C), agreeing with previously published results regarding the compound’s toxic effect on 3D cultured hepatocytes2,3. This validates the ability of the NIMBUS4 to accurately perform the steps of the hepatotoxicity testing procedure, as well as the Cytation 5 to detect compound induced hepatotoxicity.

3D hepatocyte spheroid hepatotoxicity imaging

Figure 6. 3D hepatocyte spheroid hepatotoxicity imaging. Z-projected images captured with a 10x objective following 24-hour spheroid treatments with (A) 0; (B) 1.2; and (C) 100 μM troglitazone. DAPI: Hoechst 33342 stained nuclei, GFP: CellTox Green stained dead cells.

Quantification of Observed Hepatotoxicity

Image pre-processing was then carried out by the Gen5 software using the z-projected images (Table 2).

Gen5 image pre-processing criteria.

Table 2. Gen5 image pre-processing criteria.

This was done to remove excess background signal, increasing the ability of the software to detect subtle changes in fluorescent probe signal, creating a more sensitive analysis of the hepatotoxic effect caused by each test compound.

Cellular analysis was then performed using the dual masking capabilities of the Gen5 software to quantify the level of induced hepatotoxicity by each of the test compounds over the entire incubation period.

Gen5 hepatosphere cellular analysis criteria.

Table 3. Gen5 hepatosphere cellular analysis criteria.

Primary cellular analysis criteria were optimized to accurately and consistently place object masks around the entire spheroid (Table 3). By increasing the minimum and maximum object size values, in addition to the larger rolling ball diameter, the analysis algorithm sees the spheroids as a single object versus numerous small individual nuclei (Figure 7A). To accurately determine necrotic cells within each spheroid, the fluorescence from the CellTox Green probe in untreated cells was analyzed. By determining the average fluorescence in the GFP channel plus two standard deviations, a threshold value could be set such that secondary masks would only be placed around areas within the spheroid containing cells exhibiting actual necrosis (Figure 7B-C).

The signal from the secondary masked areas was then used to compute fold induction of necrotic activity due to compound treatment, such that the total signal from treated spheroid masked areas (Figure 7C) divided by the total signal from untreated negative control spheroids (Figure 7B) determined fold response.

 Automated dual-mask analysis.

Figure 7. Automated dual-mask analysis. (A) Primary mask placed around the entire spheroid, and secondary mask placed around discontinuous areas of statistically increased GFP signal in (B) untreated; and (C) compound treated spheroids.

Finally, dose response curves were created for each of the hepatotoxins using calculated fold response values (Figure 8A). Note that hepatotoxicity was differentiated by both an EC50 type measurement or the magnitude of the fold response. Troglitazone and Nicardipine demonstrated the greatest fold response at about 10-fold. Results for individual compounds were also graphed to examine the effect of short- and long-term exposures (Figures 8B and C). It was apparent that for both Troglitazone and Nicardipine the extent of the fold response was unchanged over time, but the EC50 measurement became progressively shifted to lower concentrations.

Quantification of hepatotoxic effect on 3D hepatocyte spheroids.

Figure 8. Quantification of hepatotoxic effect on 3D hepatocyte spheroids. (A) Graph of hepatotoxic fold induction following 24-hour treatment with previously described hepatotoxins. (Daunorubicin curve not shown due to excessive toxic effect at highest tested concentrations not consistent with non-linear curve fit). Multi-day fold induction graphs also shown for 1, 3, 5, and 7 day treatments with (B) troglitazone; and (C) nicardipine.


ReproHepato hiPSC-Hepatocyte 3D spheroids provide a robust cell model for the performance of long-term hepatotoxicity studies, and the 384-Well BiO Assay Kit and NanoShuttle-PL particles, in addition to cell repellent plates, provide a simple, repeatable method to create biomimetic hepatocyte spheroids. Through incorporation of the NIMBUS4, the hepatotoxicity testing process can be fully automated to simplify and increase the repeatability of included procedures, without sacrificing sample integrity. Finally, spheroid aggregation and image-based analysis of induced hepatotoxicity can be automatically performed using brightfield and fluorescent imaging with the Cytation 5 and Gen5 cellular analysis. The combination of appropriate 3D cell models, assay methodology, and automated liquid handling, imaging and analysis create a robust method to determine the potential hepatotoxic effect of test molecules.


  1. Bell, C.C.; Hendriks, D.F.; Moro, S.M.; Ellis, E.; Walsh, J.; Renblom, A.; Fredriksson Puigvert, L.; Dankers, A.C.; Jacobs, F.; Snoeys, J.; Sison-Young, R.L.; Jenkins, R.E.; Nordling, A.; Mkrtchian, S.; Park, B.K.; Kitteringham, N.R.; Goldring, C.E.; Lauschke, V.M.; Ingelman-Sundberg, M. Characterization of primary human hepatocyte spheroids as a model system for drug-induced liver injury, liver function and disease. Sci Rep. 2016, 6:25187. doi: 10.1038/srep25187.
  2. Jaeschke, H. Troglitazone Hepatotoxicity: Are We Getting Closer to Understanding Idiosyncratic Liver Injury? Toxicol Sci. 2007, 97(1), 1-3.
  3. Letzsch, S.; Hänel, F.; Böttcher, K.; Kelm, J. Cytotoxicity Studies on 3D Primary Liver Microtissues; Application Note 011591A_01; PerkinElmer, Inc., Waltham, MA, 2014.