The choice of ultra‐low attachment plates impacts primary human and primary canine hepatocyte spheroid formation, phenotypes, and function

Organotypic three‐dimensional liver spheroid cultures in which hepatic cells retain their molecular phenotype and functionality have emerged as powerful tools for preclinical drug development. In recent years a multitude of culture systems have been developed; however, a thorough side‐by‐side benchmarking of the different methods is lacking. Here, we compared the performance of ten different 96‐ and 384‐well microplate types to support spheroid formation and long‐term culture. Specifically, we evaluated differences in spheroid formation kinetics, viability, functionality, expression patterns, and their utility for hepatotoxicity assessments using primary human hepatocytes (PHH) and primary canine hepatocytes (PCH). All 96‐well plates enabled formation of PHH liver spheroids, albeit with differences between plates in spheroid size, geometry, and reproducibility. Performance of different 384‐wells was less consistent. Only 6/10 microplates supported the formation of PCH aggregates. Interestingly, even if PCH aggregates in these six microplates were more loosely packed than PHH spheroids, they maintained their function and were compatible with long‐term pharmacological and toxicological assays. Overall, Corning and Biofloat plates showed the best performance in the formation of both human and canine liver spheroids with highest viability, most physiologically relevant phenotypes, superior CYP activity and lowest coefficient of variation in toxicity assays. The presented data constitutes a valuable resource that demonstrates the impacts of current ultra‐low attachment plates on liver spheroid metrics and can guide evidence‐based plate selection. Combined, these results have important implications for the cross‐comparison of different studies and can facilitate the standardization and reproducibility of three‐dimensional liver culture experiments.


INTRODUCTION
The discovery of safe and effective medicines relies on the accurate clinical translation of findings based on preclinical models.The liver is among the organ systems most commonly affected by adverse drug reactions, and hepatotoxicity is a main reason for the attrition of drug candidates and the withdrawal of drugs from the market. [1]Additionally, the liver is a key determinant of the pharmacokinetics of newly developed compounds and accurate prediction of formed metabolites, drug bioavailability, and clearance constitute integral steps in preclinical drug development.Animal tests have long been an essential part of the preclinical testing pipeline with an estimated 190 million animals used in medical research globally per year [2] .Regulatory guidelines require safety and tolerability data from at least one rodent and one non-rodent species, before first-in-man trials.Of these, Beagle dogs are the most popular for preclinical safety testing and pharmacokinetic studies. [3]However, it is becoming increasingly clear that animals are overall poor predictors of human drug efficacy and safety [4][5][6] .This awareness has resulted in the increased adoption of new approach methodologies (NAMs) in preclinical pharmacology and toxicology [7] .These developments are further supported by recent changes in legislation that end the blanket mandate for animal testing [8] .
To facilitate early prediction of hepatotoxicity and pharmacokinetic properties, organotypic culture models of primary hepatocytes have emerged as versatile in vitro models [9][10][11][12] .Specifically, threedimensional liver spheroids are increasingly utilized as they allow for the maintenance of physiologically relevant phenotypes and functions for multiple weeks [13][14][15] .Furthermore, liver spheroids have been well characterized as model systems for the prediction of toxicity [16,17] and clearance [18][19][20] where they outperform previous culture systems, such as sandwich or suspension cultures.Multiple different methodologies for liver spheroid culture have been described.
Spheroid formation techniques include hanging drops, agitation-based methods, as well as aggregation in ultra-low attachment (ULA) plates [21,22] .Particularly the latter enables the rapid, inexpensive and scalable formation of thousands of spheroids of well-defined sizes in 96-or 384-well plate formats.Once spheroids are formed, they can be maintained in different media with or without the addition of serum.For media compositions, extensive careful benchmarking studies have contributed to protocol standardization and have facilitated cross-study reproducibility and comparability [23] .However, similar studies have not been conducted to compare different culture plates despite the wide range of available products on the market.
Different plates differ in geometry, which in turn can impact spheroid formation kinetics and alter evaporation [24] .[27] Surface treatments and materials can also affect the absorption of small molecules [28] , thereby changing cellular exposure levels.To optimize spheroid cultures and improve cross-study and cross-center comparability, it is thus imperative to be aware of the differences caused by plate parameters.
To provide practical guidance for plate selection, we here compared the performance of ten different 96-and 384-well ULA microplates.
We generated spheroids from primary human hepatocytes (PHH) and found considerable differences in spheroid formation kinetics, viability, functionality, gene expression profiles, and their utility in hepatotoxic assessments.To facilitate utility for translational research, we furthermore generated primary canine hepatocyte (PCH) spheroids and analyzed inter-species differences.These investigations provide a basis for evidence-based microplate selection for liver spheroid studies and aspire to promote standardization of ULA plate parameters.

Spheroid culture
Cryopreserved PHH were obtained from patients undergoing liver resection at the Department of Surgery at Uppsala University Hospital, Sweden (Ethical Approval no.2009/028, amended 2018/1108).PHH were isolated as previously described [29] .In brief, the liver tissue was rinsed using Hypothermosol FRS (Biolife Solutions), cannulated, and perfused with collagenase and protease buffers to release cells from the tissue matrix.For cryopreservation, the hepatocytes were resuspended in CryoStor CS10 solution (BioLife Solutions) supplemented with 10% fetal bovine serum (FBS) and stored at -150 • C until use.Cryopreserved PCHs from male Beagle dogs were commercially acquired (BioIVT).
Spheroids were seeded and cultured as previously described [30] .
After cell counting, PHH were then seeded into ten different models of 96-or 384-well ULA plates at a concentration of 1500 viable PHH or PCH per well (Table S1).In 96-well plates, cells were cultured in 100 µL while spheroids were cultured in 80 µL of medium in 384-well plates.When the spheroids were sufficiently compact, medium was exchanged to serum-free culture medium.Notably, both spheroid formation and maintenance occurs under scaffold-free conditions.Spheroid exposures were started 7 days (PHH) or 10 days (PCH) after seeding.Spheroids were maintained for up to 3 weeks with a medium change every 48 to 72 h.

Gene expression analysis
RNA was isolated from ≥30 pooled spheroids per sample using the Zymo Quick-RNA Microprep Kit (Zymo Research) and expression of selected genes was quantified using the Biomark HD system (Fluidigm) with a FLEXsix integrated fluidic circuit (IFC, Fluidigm).The utilized gene expression assays are provided in Table S2.The gene expression levels were calculated using the ΔΔCt method with two to three housekeeper genes as reference.

Immunofluorescence
PHH spheroids were fixed overnight in 4% formaldehyde at 4

Metabolic profiling
Single PHH spheroids were incubated for 4 h at 37

Statistical analyses
Statistical analyses were performed with GraphPad Prism (Version 9.5.1).No outlier elimination was performed, and all obtained data points were included into the analyses.p-values < 0.05 were considered significant.

The choice of ULA microplates impacts PHH spheroid formation kinetics and morphology
To compare different microplates for liver spheroid formation, we first monitored the aggregation of PHH into spheroids.Over the course of 7 days, PHH aggregated and compacted into spheroids with diameters of 200 to 300 µm in almost all microplates (Figure 1A,B).The establishment of three-dimensional liver spheroids as a standardized methodology in preclinical pharmacology and toxicology requires high reproducibility.To this end, it is crucial that cells assemble into a single spheroid in each well without the formation of satellites.Across 96-well plates, the cell suspensions aggregated predominantly (> 90%) into one single spheroid (Figure 1C).In 384-well plates, aggregation was overall less uniform and only Biofloat384 well plates showed uniform formation of single spheroids.In contrast, in Greiner384 plates, cells formed multiple small spheroids with diameters ranging from 70 to 100 µm, suggesting that these plates are of limited use for high-throughput pharmacological or toxicological experiments.
During aggregation, spheroid diameters decreased over time with similar kinetics across 96-well plates (Figure 1D).Among the 384-well plates, spheroids formed considerably faster in the Biofloat384 plates with fully formed aggregates being visible after 3 days compared to 5 or 6 days in the other plate models (Figure 1E).Physical proximity may play an important role in the initial assembly of loose cell suspensions which depends on cell interaction through surface integrins [31] .
The compact packing of PHHs after seeding facilitated by the specific geometry of the wells and the specific coating may thus account for the rapid organization of the cells into single spheroids in this plate model.
To compare plate performance in imaging-based high-throughput assays, we imaged entire plates with an automatic microscope on day 7 after seeding (Figure 1F).We did not observe systematic differences within plates between columns or rows.Also, no clear edge effect was observed.Overall, the morphology of spheroids agreed with the images captured by manual acquisition.Spheroids could not be automatically captured in PE384 plates since most spheroids were not located at the center of the wells, likely due to the flat-bottom design.The other plates were found to be compatible with automatic plate scans.

Characterization of PHH spheroid phenotype, function, and stability in different ULA microplates
To assess the impact of the different microplates on long-term culture of PHH spheroids, we first assessed cell viability for up to 3 weeks (Figure 2A,B).In Corning96 and Corning384 plates, the initial ATP levels were maintained throughout the whole investigated period.
In Biofloat384 and Biofloat96 plates, a significant drop in ATP levels could be observed at day 21 (p < 0.01).In all other plate models, ATP content dropped significantly already after 14 days.In Greiner96, PE96 and SBio384 plates, spheroid viability even dropped to levels < 25%, suggesting that these plates might be of limited use for Maintenance of phenotype and functionality of liver spheroids is critical for translational applications.We thus evaluated the expression of hepatic genes related to drug metabolism, ethanol catabolism, bile acid synthesis, and differentiation.Overall, drastic gene-and platespecific differences were observed (Figure 3A; Figure S1).Expression of the differentiation markers HNF1A, HNF4A, and ABCC2 did not markedly differ between microplates.ALB, CYP2E1, and ABCC2 showed only marginal inter-plate variability, whereas the key drug metabolizing enzymes CYP2D6 and CYP3A4, the bile acid genes CYP7A1 and ABCB11, as well as the aldehyde dehydrogenase ADH1A exhibited pronounced differences.For the best performing plates (Corning96, Corning384, Biofloat96, and Biofloat384), we moreover supplemented these results at RNA level by immunofluorescence stainings.Notably, we found that the hepatic markers CYP3A4 and albumin remained expressed at protein level throughout the spheroids without apparent differences between these plate types (Figure S2).
Particularly, spheroid culture in Nunc96 plates resulted in major transcriptional alterations after 7 days and, while these differences equilibrated for CYP2D6 and CYP3A4 at later timepoints, expression of CYP7A1 and ABCB11 remained significantly different after 14 days.
In agreement with these observations, the expression signature of spheroids cultured in Nunc96 plates formed a unique cluster after PCA at 7 days (Figure 3B).After 14 days however, two distinct clusters emerged, which separated transcriptional patterns in PE96, Nunc96, PE384, and Sbio384 from the other microplate types.
We complemented our expression analyses with measurements of metabolic activity.To this end, we incubated PHH spheroids after 7 days of culture with a cocktail of probe substrates for the metabolic enzymes (CYP1A2, CYP2C8, CYP2C9, and CYP3A4) and quantified the formed metabolites by LC-MS (Figure 3C).Striking differences between CYP activities were observed between microplates.PE96, Nunc96, PE384, and Sbio384 exhibited the lowest metabolic activities for all substrates, with CYP2C9 and CYP3A4 activity being below detection limit.In contrast, Biofloat384 and Corning384 showed the highest activities, closely trailed by Greiner96, Corning96, and Biofloat96.Among these plates, Biofloat96 and Biofloat384 plates showed the lowest variability between the different measurements.
These results indicate that the choice of microplate constitutes an important parameter that can have pronounced impacts on results and interpretation of PHH spheroid studies.

Microplates affect the outcomes of hepatotoxicity assessments in PHH spheroids
To test the impact of microplates on preclinical toxicity assays, PHH spheroids were exposed to hepatotoxins with different modes of

F I G U R E 3
The choice of culture plate affects molecular phenotypes and functions of human liver spheroids.(A) Matrix representation of expression differences across ten important hepatic genes.The bottom left quadrant and top right quadrants indicate differences after 7 days and 14 days, respectively.n = 3 per plate and time point.Grey cells indicate p ≥ 0.05 while yellow, light orange, dark orange, and red indicate p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 in a one-way ANOVA, respectively.(B) Principal component analysis (PCA) can cluster microplate types based on transcriptional differences.Note that the differences along principal component 1 (PC1) that distinguish plate types are much larger than across principal component 2 (PC2), which separate replicates.(C) Column plots showing the levels of acetaminophen, N-desethylamodiaquine, 4-hydroxy diclofenac, and hydroxymidazolam, which serve measurements of the metabolic activities of CYP1A2, CYP2C8, CYP2C9, and CYP3A4.action.Specifically, we selected APAP, TRO, and CPZ as prototypic compounds that exert toxicity via reactive metabolites, mitochondrial injury and cholestatic mechanisms, respectively.We evaluated toxicity both after a single dose (48 h exposure) and after 7 days of repeated exposures (Figure 4).
As expected, spheroid sensitivity increased with prolonged exposure for all compounds and plate types (Figure 4).Furthermore, we observed clear concentration-dependency of toxicity.This was most evident for TRO and CPZ whose higher concentrations caused complete cell death.The largest variability between plates was observed for intermediate TRO concentrations (5x c max ) where no toxicity was observed in Nunc96 and Sbio384 (viability > 80%) whereas isogenic liver spheroids in Corning96 and Corning384 were completely dead.
Overall, the highest concentration of TRO and CPZ was required to detect in vitro toxicity in Nunc96 plates, whereas PHH spheroids in Corning384 and Biofloat384 were most sensitive with toxicity being clearly evident even at the lowest concentration tested after 7 days exposure.

Only few plate models support the formation of primary canine hepatocyte spheroids
Advanced organotypic cultures of primary hepatocytes of mouse, rat, minipig and cynomolgus monkey have been established to guide the selection of preclinical model species [16,[32][33][34] .However, despite the popularity of Beagle dogs for preclinical drug safety testing, similar cultures have not been established for primary canine hepatocytes (PCH).We thus tested the potential of the different ULA plates to support PCH spheroid culture.In Corning96, PE96, Corning384, Sbio384, Biofloat96, and Biofloat384 microplates, PCH formed aggregates that were stable enough to be manipulated and transferred for endpoint measurements.In contrast, in Greiner96, Nunc96, Greiner384, and PE384 microplates, PCH did not form stable aggregates and these plates were thus excluded from further analyses.Spheroids generated in Corning96, PE96, and Corning384 maintained relatively high ATP levels for 17 days with the latter showing the highest ATP levels after 24 days of culture (Figure 5A).By com-parison, ATP levels dropped more rapidly in Biofloat96, Biofloat384, and Sbio384.To further characterize the molecular phenotype of PCH aggregates, we evaluated the expression levels of the canine orthologues of eight hepatocyte-specific genes (Figure 5B).Specifically, we selected the canine transporter genes ABCB11 and ABCC2, canine albumin (ALB) and transferrin (TF) as secretory markers, the hepatocyte master regulator HNF4A and the cytochrome P450s CYP1A2, CYP2D15 (orthologue of human CYP2D6), and CYP3A12 (orthologue of human CYP3A4).Upon aggregation (day 10), expression signatures in Sbio384 plates were different from the other plates, particularly due to altered expression of ALB and CYP1A2.However, unlike in PHH spheroids, differences between plates were comparable in magnitude to differences between replicates (Figure 5C To the best of our knowledge, these results constitute the first demonstration that primary canine hepatocyte aggregates can be stably cultured for multiple weeks.Culture in chemically defined serum-free media with identical compositions as for PHH spheroids facilitates the overall maintenance of molecular phenotypes and functions and is compatible with both acute and chronic toxicity tests, thus establishing these three-dimensional aggregates as a viable model for translational applications.However, the choice of microplate is of tremendous importance to enable successful formation of PCH three-dimensional culture models.

DISCUSSION
Protocol standardization is imperative to facilitate the acceptance of three-dimensional tissue models as reliable preclinical assays.As of September 2023, more than 4500 spheroid experiments can be retrieved from the MISpheroID database, indicating that the potential of the spheroid model in biomedical research is increasingly recognized [35] .Notably, comparison of the different selected spheroid methods revealed diverse impacts on spheroid metrics, thus indicating the need for careful benchmarking.For liver spheroids, this should encompass the choice of cell model, seeding protocol, media composition as well as microplates.Most of these parameters have been benchmarked in recent studies.The phenotype and function of primary hepatocytes have been shown to be superior to hepatoma cells and stem cell-derived hepatocytes [36][37][38][39] .Furthermore, the effects of different spheroid formation protocols and media compositions have been extensively evaluated.Our study is the first to test the impacts of different microplates and demonstrates that they can have pronounced impacts on formation kinetics, morphology, gene expression patterns, compound toxicity and long-term stability.While effects on human liver spheroids were mostly quantitative, strong qualitative differences were observed for canine hepatocyte spheroids, which could only be successfully formed in six out of the ten plates tested.
To implement three-dimensional liver spheroids in preclinical phar- Cultivation of three-dimensional liver spheroids for multiple weeks is essential for chronic disease modelling and repeated drug exposures.
Our data indicate that Corning and Biofloat microplates, both in 96and 384-well formats, are best suited for long-term culture.Which factors are driving these differences however, remains unclear.Previous studies demonstrated that E-cadherin-mediated cell-cell interactions inhibit apoptosis and promote cell survival in human liver spheroids compared to isolated cells or two-dimensional cultures [40] .Further factors with potential impacts on cell viability include microstructure and leaching of compounds from the coating itself or the underlying bulk material [41][42][43] .
In addition to changes in long-term viability, we also observed that the choice of microplate affected gene expression patterns.We have previously reported that primary hepatocytes undergo dynamic transcriptomic changes during aggregation into spheroids which recapitulate molecular events observed during liver regeneration [44] .
Based on these findings, one can assume that differences in aggregation kinetics between plates could directly translate into altered expression patterns.These findings have important implications as they suggest that the use of different plates could have significant impacts on the biological interpretations of the obtained results even if otherwise identical protocols are used.
Different microplates furthermore affected the sensitivity of liver spheroids to hepatotoxins with different modes of action.APAP hepatotoxicity is caused by its metabolic byproduct, the reactive metabolite N-acetyl-p-benzoquinonimine (NAPQI).As NAPQI formation is dependent on the hepatic enzymes CYP2E1 and CYP3A4, dedifferentiated and metabolically inactive hepatic cell models are largely protected from APAP toxicity [45] .Here we find that APAP sensitivity of PHH spheroids was overall similar across plates, which is in agreement with similar expression levels of CYP2E1 and CYP3A4.In contrast, CPZ and TRO exert toxicity at least in part by inhibiting the bile salt export pump (BSEP; encoded by ABCB11) at the transcriptional and/or protein level [46,47] .At high concentrations, both compounds were clearly toxic across all microplates; however, toxicity of these compounds at lower concentrations (CPZ 1x c max and TRO 5x c max ) differed by > 10-fold, with spheroids cultivated in Nunc96 plates being the most resistant.These results align with increased BSEP expression in this plate type, which might bestow hepatocytes with an increased capacity to export bile salts and thus blunt CPZ and TRO toxicity.Differences in the sensitivity to hydrophobic compounds between different plates may be also due to differences in compound absorption.Cell culture plastics are prone to absorption of hydrophobic small molecules [48] .
TRO and CPZ are both hydrophobic with logP values of 3.

F I G U R E 1
Morphology and formation kinetics of human liver spheroids.(A) Time series of brightfield images showing PHH spheroid formation.Scale bar = 200 µm.(B) Diameter of PHH spheroids on Day 7 after seeding in different types of plates.** and **** indicate p < 0.01 and p < 0.0001 in an ANOVA test with Welch's correction.(C) The fraction of wells in which cells aggregated into one single spheroid.* and **** indicate p < 0.05 and p < 0.0001 in a Kruskal-Wallis test.Error bars indicate SEM.(D-E) Compaction kinetics of PHH over 7 days after seeding in 96-well plates (D) and 384-well plates (E).(F) High-throughput scan of spheroids in different types of plates.Only one representative row is shown per plate.

F
I G U R E 2 Cell viability of human liver spheroids during 3 weeks of culture.(A-B) Viability is shown for PHH spheroids after 7 days, 14 days, and 21 days of culture in 96-well plates (A) and 384-well plates (B).n = 8 per plate and time point.*, **, and *** corresponds to p < 0.05, p < 0.01, and p < 0.001 in a one-way ANOVA, respectively.Error bars indicate SEM.The dotted lines indicate 80% cell viability.(C-D) The coefficient of variation is shown for the different replicates after 7 days, 14 days, and 21 days of culture in 96-well plates (C) and 384-well plates (D).long-term liver spheroid experiments.The variability between individual spheroids increased with culture time and was in general lower in 96-well plates compared to 384-well plates (Figure 2C,D).Most plates allowed reproducible measurements of cell viability (coefficient of variation [CV] < 20%) except for Greiner, PE and SBio plates in which CVs exceeded 30%.

F I G U R E 4
Toxicity of prototypic hepatotoxins in human liver spheroids.Cell viability was measured after a 48 h single exposure (A) and 7 days repeated exposures (B) for acetaminophen (APAP), troglitazone (TRO), and chlorpromazine (CPZ).n = 8 per plate, drug, and time point.Error bars indicate SEM.*, **, and *** correspond to p < 0.05, p < 0.01, and p < 0.001 in one-sample t-tests, respectively.
; 10 days).In contrast, after extended culture (day 17), Corning384 plates showed differential expression of ABCB11, ABCC2, ALB, CYP1A2, CYP2D15, CYP3A12, and TF compared to the other microplates (Figure 5C; 17 days).At the phenotypic level, Corning96 and Biofloat96 exhibited the highest levels of metabolic activity, whereas activities in Sbio384, Corning384, and Biofloat384 were around two-fold lower (Figure 5D).To evaluate hepatotoxicity in the different microplates, we exposed PCH aggregates to different concentrations of APAP, TRO, and CPZ (Figure 5E,F).After exposure to APAP, PCH aggregates generally main-F I G U R E 5 Establishment and characterization of primary dog liver spheroids.(A) Viability for PCH aggregates is shown after 10 days, 17 days, and 24 days of culture in 96-well plates and 384-well plates.Results are only shown for the six plate types that supported the formation of PCH aggregates.n = 8 per plate and time point.*, **, and *** corresponds to p < 0.05, p < 0.01, and p < 0.001 in a one-way ANOVA, respectively.Error bars indicate SEM.The dotted lines indicate 80% cell viability.(B) Matrix representation of expression differences across eight important hepatic genes.The bottom left and top right quadrants indicate differences after 10 days and 17 days, respectively.n = 3 per plate and time point.Grey cells indicate p ≥ 0.05 while yellow, light orange, dark orange, and red indicate p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 in a one-way ANOVA, respectively.(C) Principal component analysis (PCA) of expression patterns of the eight analyzed hepatic genes.(D) Column plots showing the levels of hydroxymidazolam and dextrorphan, which serve measurements of the metabolic activities of canine CYP3A12 (orthologue of human CYP3A4) and CYP2D15 (orthologue of human CYP2D6).The sensitivity of human and dog spheroids is compared after a 48h single exposure (E) and 7 days repeated exposures (F) for acetaminophen (APAP), troglitazone (TRO), and chlorpromazine (CPZ).Three concentrations (1x, 5x, and 20x c max ) are tested per compound.Error bars indicate SEM.*, **, and *** correspond to p < 0.05, p < 0.01, and p < 0.001 in two-tailed heteroscedastic t-tests, respectively.tained high viability except for spheroids generated in PE96 plates in which 20x c max concentration was clearly toxic.Across plates, spheroid sensitivity to TRO and CPZ was similar, with one exception.While TRO showed toxicity only after repeated exposure with the highest concentration in all plates, this drug was clearly hepatotoxic already at 5x Cmax for spheroids generated in Biofloat96 plate.Compared to PHH spheroids, PCH aggregates were overall less sensitive to all compounds tested, particularly after repeated exposure.
macology and toxicology, reproducibility of spheroid formation plays a crucial role.This in turn requires that all seeded cells integrate into one structure and form a single spheroid in each well.The formation of one spheroid per well could be achieved in all tested plates, except for Greiner384 plates in which cell aggregation resulted in multiple spheroids per well.In addition to reproducibility, rapid spheroid formation is important, particularly in the context of using three-dimensional liver spheroids for drug screening and toxicological analyses.Compaction of PHH seeded in Biofloat384 was the fastest among the investigated plates, resulting in spheroid formation already at 3 days after cell seeding while 5 to 6 days were required in all other tested plates.Interestingly, PCH aggregated more loosely compared to PHH across all microplates.As a result, aggregation was slower (10 days vs. 7 days) and only 6 out of 10 microplates supported the formation of PCH aggregates.
6 and 5.18, respectively, whereas APAP is hydrophilic (logP = 0.51).Different surface treatments may reduce absorption, resulting in different free concentrations, which translates into altered compound sensitivity at the same nominal exposure concentration.Minimizing drug absorption thus provides an important direction for the optimization of ULA coating strategies.In summary, by comparing morphology, phenotype and function of isogenic liver spheroids in side-by-side experiments, we demonstrated diverse impacts of the choice of ULA microplates.While most plates supported the formation of human liver spheroids, large plate differences were observed in hepatic gene expression and toxicity assays.Overall, spheroids in Corning and Biofloat plates exhibited the highest viability, metabolic activity and sensitivity to hepatotoxins.Thus, our results suggest that these plates are suitable for the generation of human liver models that can be implemented in preclinical drug development.However, liver spheroids cultivated in these plates still showed assay-specific differences and it is thus recommended that researchers evaluate the suitability of these plates for their own experiments and purposes.We furthermore provide the first characterization of three-dimensional spheroids from primary hepatocytes of Beagle dogs and demonstrate their compatibility with long-term pharmacological and toxicological studies.Combined, our results demonstrate how different microplates can affect results of spheroid assays and provide useful information for the optimization of future liver spheroid experiments.