Three-Dimensional Spheroids With Primary Human Liver Cells and Differential Roles of Kupffer Cells in Drug-Induced Liver Injury

Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org

Pharmaceutical Biotechnology

Three-Dimensional Spheroids With Primary Human Liver Cells and Differential Roles of Kupffer Cells in Drug-Induced Liver Injury Feng Li*, Li Cao, Sweta Parikh, Rongjun Zuo Corning Life Sciences, Bedford, Massachusetts 01730

a r t i c l e i n f o

a b s t r a c t

Article history: Received 13 November 2019 Revised 19 February 2020 Accepted 28 February 2020 Available online 5 March 2020

Drug-induced liver injury (DILI) remains a challenge and a leading risk for drug discovery. Threedimensional liver spheroids made from primary human hepatocytes (PHHs) with, or without, other liver cell types can provide more physiological relevance. In comparison to conventional 2-dimensional monolayer culture, our tests with 100 drugs of known DILI status indicate that PHH spheroids are significantly more sensitive in detecting drug-induced hepatotoxicity. To evaluate the role of Kupffer cells (KCs) in drug-induced liver toxicity, we have established conditions for generating co-culture spheroids with PHH and KCs. Inflammatory responses as shown by interleukin 6 secretion can be recapitulated in co-culture spheroids when treated with endotoxin lipopolysaccharides. KCs potentiated the cytotoxicity induced by trovafloxacin in co-culture spheroids at 48 h, but the differences between PHH spheroids and co-culture spheroids became less obvious after a 5-day treatment. Interestingly, a protective role of KCs was shown in co-culture spheroids treated with both acetaminophen and lipopolysaccharides. Additional tests with 14 DILI compounds comparing PHH spheroids and co-culture spheroids showed differential roles of KCs that were compound dependent. In summary, these 3-dimensional liver spheroid models are useful tools to understand the complex mechanisms underlying DILI. © 2020 The Authors. Published by Elsevier Inc. on behalf of the American Pharmacists Association®. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/).

Keywords: hepatocytes Kupffer cells spheroids liver toxicity

Introduction Drug-induced liver injury (DILI) remains a leading cause of drug development failures and a serious risk to patient health. More physiologically relevant human models are being developed with support from regulatory agencies and the pharmaceutical industry to improve preclinical models in predicting and derisking liver liability during drug discovery and development.1-3 Primary human hepatocytes (PHHs)4 are considered the gold standard for studying liver biology and are broadly used in a variety of drug discovery and safety studies. However, PHHs quickly lose cell viability and function in conventional 2-dimensional (2D) monolayer culture.5-7 Much efforts have been invested to optimize 2D monolayer culture of PHH. Extracellular matrix modification by culturing PHH in a sandwich format with collagen substrate and a Matrigel overlay has shown to be beneficial for prolonging the cell viability. However, the loss of hepatic gene expression and changes of drug metabolic functions

Conflict of interest: The authors are all current employees of Corning Life Sciences and therefore declare no conflict of interest. * Correspondence to: Feng Li (Telephone: þ1-781-301-3134). E-mail address: [email protected] (F. Li).

lead to unstable hepatocyte phenotypes even in a sandwich culture and therefore limits its utilization for liver toxicity studies.8 Other more sophisticated approaches include using mouse or human stromal cells in either micropatterned or mixed co-cultures with PHH to support long-term viability and functions.9 Microfluidic systems have also been applied to improve hepatocyte culture, but their incompatibility with high throughput assays has been a recognized issue. By contrast, 3-dimensional (3D) liver spheroids simply formed by gravity-induced aggregation and self-assembly have great potential to become a robust liver model with high throughput capacity. Liver spheroids have been shown to sustain hepatic cellular phenotypes and viability in extended culture up to 4 to 5 weeks.10-12 The retention of physiologically relevant 3D architecture and the establishment of important cell-cell interactions in liver spheroids are important for hepatic cell phenotype improvement and are missing in conventional 2D culture. Furthermore, transcriptomic and proteomic characterizations indicate that these liver spheroids are very representative of native liver compared with other liver models in 2D or 3D cell culture format.13,14 Compared with other 3D cell culture systems such as 3D bioprinting, liver spheroids are more cost-effective and adaptable to a higher throughput workflow. These characteristics make liver spheroids an attractive model system for

https://doi.org/10.1016/j.xphs.2020.02.021 0022-3549/© 2020 The Authors. Published by Elsevier Inc. on behalf of the American Pharmacists Association®. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

in vitro liver toxicity screening assays and liver disease modeling such as fatty liver disease.15-17 Liver spheroids can be made with PHH or PHH together with liver nonparenchymal cells (NPCs) in a co-culture. Kupffer cells (KCs) are resident macrophages in the liver that play an important role in inflammatory responses, immune-mediated hepatotoxicity, and liver injury.18-20 As an important component of liver innate immunity, KCs are scavengers in the liver and key responders to danger signals released by stimuli such as bacterial infection, drug treatment, and liver disease.21 A plethora of cytokines secreted by KCs (including interleukin 6 [IL6] and tumor necrosis factor [TNF]a) can trigger signaling pathways that regulate cell viability, proliferation, and cell death, as well as other cellular function changes such as hepatocyte drug-metabolizing enzyme activities.19,22 However, the roles of KCs and inflammation in DILI are often overlooked and less defined. This is due, at least in part, to the lack of proper human liver models.20 Herein, we report the use of 3D PHH liver spheroids for improving in vitro human liver toxicity assays and establishing culture conditions for co-culture spheroids of PHHs and KCs. To demonstrate the usefulness of PHH spheroids and to improve the physiological relevance of in vitro liver toxicity assays, a comparative study was conducted using 2D monolayer and 3D spheroid culture of PHHs with a selection of 100 drug compounds from known DILI categories.15,23 One of our objectives in this study is to test the robustness of using PHH liver spheroids for highthroughput in vitro liver toxicity assay. A simple endpoint of spheroid viability measurement by bioluminescent ATP assay is therefore chosen for its sensitivity and fast and reliable readouts with liver spheroids that are made of only 1-2 thousand cells. For comparison, 2D monolayer culture without Matrigel overlay was prepared from the same lot of PHHs that we used for 3D liver spheroids culture because of the concern of nonspecific compound binding and no Matrigel use in liver spheroid culture. As shown in Figure 1, the 2D culture was treated with a single dosing of testing compounds 24 h after plating, while the hepatic phenotype including gene expression and drug metabolizing activity still largely remain under this condition. For 3D liver spheroid culture, 3 repeated dosing over 2 weeks for testing compounds were applied because PHH liver spheroids have been shown to sustain the viability and function of hepatocytes for this period without additional manipulations. To establish a co-culture liver spheroid condition, we evaluated the impact of dexamethasone, cell ratios, and treatment of different endotoxin concentrations. In vitro liver toxicity tests comparing PHH and PHH/KCs co-culture spheroids were performed to understand the role of KCs and inflammation in druginduced liver toxicity. This type of co-culture liver spheroids with hepatocytes and KCs can be used to study signaling pathways involved in drug toxicity or disease development and their impact on hepatocytes. For example, the cytokines secreted by activated KCs can induce major effects on hepatocyte drug metabolizing enzyme activity.1

Materials and Methods Chemicals and Reagents Unless otherwise stated, all chemicals used were purchased from MilliporeSigma (St. Louis, MO). Stock solutions were prepared in either DMSO or culture medium. Serial dilutions of drug solutions were prepared freshly before treatment. The concentration for each compound stock was at 100 Cmax or maximumly dissolvable in solvent at room temperature. Lipopolysaccharide

1913

(LPS) stock solution was prepared at 1 mg/mL in PBS and then diluted in the culture medium to make final concentrations between 1 and 10 mg/mL for spheroid treatment. For compound treatment with liver spheroids, 2 of final concentrations were prepared for each compound’s serial dilution. For spheroid culture in Corning 96-well Spheroid Plates (Corning® Cat. No. 4515 or 4520), each well contained 200 mL medium. During the dosing, 100 mL medium was first removed from one well of the spheroid culture and then 100 mL serial dilutions of 2 the final concentration of the testing compound was added to the spheroid culture to bring the total volume to 200 mL. Bioluminescent-based ATP assays were performed using a CellTiter-Glo® 3D Cell Viability Assay Kit from Promega (Catalog #G9682) with modified procedures. Briefly, the ATP assay began by removing 100 mL media from each well of the 96-well spheroid plate and adding 100 mL/well of assay reagent. The plate’s contents were mixed vigorously for 5 min on a plate shaker at 600 rpm at room temperature. After incubating the plate at room temperature for an additional 5 min to stabilize the luminescent signal, luminescence was measured with a SpectroMax plate reader for analysis. 2D PHH Monolayer Cell Culture Cryopreserved 3D spheroid-qualified PHHs (Corning Cat. No. 454552) were thawed using a high-viability recovery medium to ensure post-thaw viability at greater than 80%. William’s E medium supplemented with insulin, transferrin, selenium, and 0.1 mM dexamethasone and antibiotics was used for both 3D PHH spheroid culture and 2D PHH monolayer culture. After spinning at 100 g for 10 min, the cell pellet was resuspended in the plating medium prepared by supplementing the cell culture medium with 10% fetal bovine serum (FBS). Trypan blue cell counting was performed to determine the cell density with a hemocytometer. 2D monolayer cultures were prepared by seeding PHH suspension at 60,000 cells/well in Corning® BioCoat™ Collagen I 96-well microplates (Corning Cat. No. 354407). Cells were plated in William’s E medium with 10% FBS for approximately 4 to 5 h. The plating medium was changed to remove dead cells. A single donor lot of hepatocytes (lot 348a) was used for both 2D monolayer culture and 3D PHH spheroid cultures to compare hepatotoxicity results between the different culture formats. Compound dosing started 24 h after plating as described previously. Bioluminescent ATP assays were performed 24 h after dosing. 3D PHH Spheroid Culture and Compound Dosing After thawing as described previously, PHHs were seeded at 1000 cells/well in Corning 96-well spheroid microplates (Corning Cat. No. 4515 or 4520) to generate PHH spheroids for all the hepatotoxicity assays described in this study. On day 2, the spheroid culture was gently rotated on an orbital shaker for 5 min at room temperature. On day 3, the spheroid culture was fed by adding 100 mL/well PHH culture medium with no FBS. After spheroid formation (usually on day 6 or day 7), half medium changes were repeated to remove residue FBS from initial plating, and serum-free medium was used for DILI and control compound hepatotoxicity assays. 3D PHH spheroid cultures were treated with 3 repeated dosing on days 8, 12, and 15 (Fig. 1). Working solutions for testing compounds were prepared at 2 final concentration of serial dilutions from stock solutions. For dosing PHH liver spheroids, half the medium (100 mL/ well) was removed and 100 mL media containing 2 final concentration of serial dilutions was added to each well as described previously. Bioluminescent ATP assays were performed after 24 h of the last dosing.

1914

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Figure 1. (a) Diagram of DILI (categories 1, 2, and 3) and control compounds (categories 4 and 5) (see Garside et al 2014; Proctor et al., 2017) (b) Compound dosing regimen for 3D PHH spheroids and 2D PHH monolayer cultures. Both spheroid culture and monolayer culture were set up with the same donor lot of PHH (lot 348a). A short-term toxicity assay was applied to 2D PHH monolayer cultures because of the limited cell viability of PHHs under these conditions. 3D PHH spheroids were subjected to a 2-wk hepatotoxicity assay with three repeated dosing of DILI or control compounds. (c) Summary of margin of safety (MOS, i.e., the ratio of IC50 and Cmax) values from in vitro toxicity tests. The dotted line indicates MOS at 50, one of the thresholds used to score the positive or negative detection of liver toxicity for each tested compound. From the left to right are separate graphs for compounds within each DILI category 1, 2, and 3. Control compounds within DILI categories 4 and 5 are graphed together. For compounds with no detection (ND) of toxicity within the tested concentration ranges, IC50/Cmax ratio was plotted on top of the graph at 2000.

3D Liver Spheroid Co-Culture Cryopreserved primary human KCs were obtained from Samsara Sciences (San Diego, CA). KCs were thawed in cold PHH medium, as described previously, supplemented with 10% FBS and kept on ice before mixing with PHH cell suspension. For co-culture

spheroid culture, compound dosing was performed after spheroid formation in serum-free medium by using 2 dilution of test compounds as described previously. 3D spheroid cultures were treated with a single dosing (serial dilutions or one concentration as indicated in text) for 5 days without a medium change inbetween.

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

ELISA Assay The concentrations of IL6 and TNF-a were determined using Human IL6 high-sensitivity ELISA kit (Abcam, ab46042) and Human TNF alpha ELISA kit (Abcam, ab181421), respectively. Albumin was measured using Human Serum Albumin SimpleStep ELISA kit (Abcam, ab179887). Supernatants were collected from liver spheroid culture 48 h after LPS treatment or as indicated in the text. Standard curves were prepared accordingly to calculate the concentration of either IL6 or TNF-a in the samples. Results PHH Liver Spheroids Show Superior Sensitivity to 2D Monolayer Culture in Detecting Drug-Induced Hepatotoxicity The prolonged hepatocyte viability and sustained functions in PHH spheroid culture can be beneficial for in vitro liver toxicity assays in comparison to 2D monolayer culture where PHHs quickly dedifferentiate and die. Using a 3D PHH spheroid model, we have tested 100 drugs with known DILI status from severe DILI to no DILI in clinical observations (Table 1 and Fig. 1a). For comparison, 2D monolayer cultures were established from the same donor lot of PHH, and compound toxicity was determined using an ATP-based cell viability assay 24 h after drug treatment. Serial-dilution toxicity assays were performed using either single dosing for 2D monolayer cultures or three repeated dosing for PHH spheroids in 2-week culture (Fig. 1b). Spheroid viability was measured and dose response curves were generated when applicable to calculate the IC50 values for each drug as described in the methods and summarized in Table 1. Values of margin of safety (MOS, i.e., ratios of IC50 over Cmax) in Figure 1c were calculated for all tested compounds based on toxicity data from 2D PHH monolayer and 3D spheroid cultures. Analysis demonstrates that 3D PHH spheroids with repeated dosing were more sensitive than 2D monolayer culture of the same donor lot of PHH in detecting compounds within DILI categories 1 to 3 (severe, high clinical DILI concern and low clinical DILI concern) when using an MOS threshold at 50 times Cmax, that is, MOS 50. Using MOS 10 or MOS 30, thresholds reach similar results as summarized in Table 2. In many cases, 2D monolayer cultures failed to detect liver toxicity of these tested DILI compounds. For DILI categories 4 and 5, however, we observed increased false positive by PHH spheroids in comparison with the 2D monolayer cultures as more MOS values for 3D spheroids fell below the MOS 50 threshold (Fig. 1c). As shown in Table 2, the choice of MOS threshold impacts both sensitivity and specificity of the toxicity assays as more stringent criteria of MOS 10 and 30 reduced the false-positive calls observed in 3D toxicity test while the specificity calls are still better than that of 2D assays. Compared with 2D monolayer cultures, 3D PHH spheroids are 2-3 times more sensitive (i.e., sensitivity of 61% for 3D PHH spheroids vs 26% for 2D PHH monolayer) in detecting hepatotoxicity induced by DILI categories 1-3 compounds. A recent report by Proctor et al.15 also showed a similar improvement in assay sensitivity for 3D versus 2D PHH culture. Impact of Dexamethasone on Co-Culture Spheroids With PHH and KCs Given the important role of KCs in innate immunity and liver inflammatory responses, we assessed the impact of these macrophages on in vitro hepatotoxicity using 3D PHH/KC co-culture spheroids. Although 0.1 mM dexamethasone (Dex) is commonly used in PHH culture to maintain hepatocyte viability and phenotypes in vitro and is usually used at 0.1 mM for 3D PHH spheroid culture (Fig. 2a), this glucocorticoid may perturb KC survival and

1915

function. Therefore, we first assessed the impact of dexamethasone on the formation of spheroids made with 1000 PHHs and 500 KCs. Without dexamethasone, co-culture spheroids were smaller and spheroid morphology appeared unhealthy with rough edges (Fig. 2a). Adding dexamethasone to the culture medium increased the spheroid size by ~20% (Fig. 2b). Consistent with the smaller spheroid size, ATP levels were lower in the absence of dexamethasone. Comparable co-culture spheroid ATP levels were observed among 0.01 mM, 0.05 mM, and 0.1 mM dexamethasone concentrations (Fig. 2c). When treated with LPS at 10 mg/mL for 48 h, ATP levels measured on day 9 for both PHH and PHH/KC co-culture liver spheroids decreased. The decrease of ATP levels was more obvious at lower dexamethasone concentrations (e.g., 0.01 mM) for coculture spheroids, suggesting PHH/KC co-culture spheroids are more vulnerable to LPS treatment under these conditions. Next, we examined the effect of dexamethasone concentrations (e.g., 0.01mM, 0.02mM, 0.04mM, and 0.05mM) on the viability and function of co-culture liver spheroids over a 2-week culture timeline. PHH spheroids (1000 cells/spheroid) were included as a control for LPS treatment induced inflammatory responses. ATP measurements were performed to monitor the viability of spheroid cultures and remained comparable with the tested dexamethasone concentrations from 0.02 to 0.05 mM (not shown). PHH spheroids showed minimum IL6 secretion, as expected, during the culture of 2 weeks (Fig. 2d). In co-culture spheroids, LPS treatment stimulated the activation of KCs as measured by IL6 secretion. Dose-dependent inhibitory effects of dexamethasone on IL6 secretion were observed in PHH/KC co-culture liver spheroids. In addition, IL6 secretion in response to LPS treatment was maintained in co-culture liver spheroids for 2 weeks in culture. Although IL6 secretion was higher at lower dexamethasone concentrations (Fig. 2d), 0.05 mM dexamethasone concentration was chosen for drug toxicity testing using co-culture spheroids because this glucocorticoid is necessary to maintain hepatocyte function and phenotype. This concentration of dexamethasone also allowed the reliable and quantitative detection of IL6 secretion by KCs after LPS stimulation. Impact of Cell Ratios on Co-Culture Spheroids With PHH and Kupffer Cells We examined the impact of the ratio of PHHs and KCs in coculture spheroids. As shown in Figure 3, 2 lots of primary human KCs (KC-1 and KC-2) were used with a single lot of PHH. The number of PHH cells was held constant (1000 cells) and the number of KCs was varied from 125 to 500 cells, to span cell ratios of 8:1, 4:1, and 2:1. All cultures were performed in a media supplemented with 0.05 mM dexamethasone. ATP levels and IL6 secretion were measured at 48 h after LPS treatment (Fig. 3). Spheroid formation time was comparable for all cell ratio conditions tested and spheroids formed in approximately 5 to 6 days. IL6 secretion levels increased with the increased cell number of KCs for both lots tested. There appeared to be lot-to-lot variation in terms of IL6 secretion under the same culture and treatment conditions. At 2:1 ratio, in response to 10 mg/mL LPS treatment, KC-1 lot IL6 secretion was 47.8 pg/mL and KC-2 lot was 68.5 pg/mL. This variation is expected for primary human KCs from different donors. Based on IL6 secretion results and expected lot-to-lot variation, PHH and KC ratio at 2:1 was selected for subsequent tests to allow reliable detection of IL6 secretion in PHH/KC co-culture spheroids. Titration of LPS Concentration for Treating Co-Culture Liver Spheroids As shown previously, primary human KCs showed lot-to-lot variation to endotoxin LPS treatment. We next tested 3 lots of

1916

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Table 1 In Vitro Toxicity Test IC50 and IC50/Cmax Ratio Comparison Between 2D PHH Monolayer Culture and 3D PHH Spheroid for 100 Compounds With Known DILI Categories Number

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1 2 3 4 5 6 7 8 9 10 11

Chemicals

Amiodarone Bosentan Iproniazid Isoniazid Labetalol Leflunomide Naltrexone Nefazodone Nevirapine Nomifensine Sitaxentan Sunitinib Tolcapone Troglitazone Trovafloxacin Valproic Acid Ximelagatran Amodiaquine Atorvastatin Azathioprine Celecoxib Clozapine Diclofenac Flurbiprofen Glafenine Imipramine Indomethacin Levofloxacin Nimesulide Nitrofurantoin Paroxetine Rosiglitazone Simvastatin Sulindac Tacrine Tamoxifen Tetracycline Ticlopidine Zileuton Acetaminophen Amitriptyline Astemizole Beta-Estradiol Chlorpheniramine Chlorpromazine Clomipramine Famotidine Fluorouracil Fluoxetine Furazolidone Nifedipine Pimozide Pioglitazone Propranolol Rifampicin Rosuvastatin Sulpiride Terbutaline Tolbutamide Trifluoperazine Verapamil Warfarin Amiloride Bumetanide Cycloserine Dexamethasone Dofetilide Entacapone Felodipine Fludarabine Minoxidil Mitoxantrone Nadolol

DILI Severity Category

1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 2. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4.

Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Low DILI Concern Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations Enzyme Elevations

Clinical Cmax (mM)

5.3 7.4 0.04 76.6 2.0 233.2 0.1 4.3 29.7 8.4 28.6 70.8 47.6 6.4 5.0 693.4 0.5 1.9 0.06 7.2 2.5 2.4 10.1 61.4 1.9 0.5 8.4 15.8 21.1 21 0.2 1.0 0.02 32.0 0.1 0.4 21.0 8.1 13.1 165.4 0.7 0.001 0.0006 0.04 0.9 0.8 0.6 0.6 0.05 1.5 0.3 0.1 3.0 1.2 12.2 0.01 1.8 0.1 369.9 0.02 0.6 22.7 0.1 0.08 333.0 0.2 0.1 3.3 0.03 0.2 1.2 0.4 0.8

IC50 (mM)

IC50/Cmax Ratio

2D Monolayer

3D Spheroid

2D Monolayer

3D Spheroid

209.0 ~1250.0 >10.0 >2500.0 248.7 938.7 >10.0 41.8 >1250.0 345.7 145.3 23.5 149.5 30.9 >1000.0 >25,000.0 >50.0 27.2 >10.0 177.1 279.7 149.5 529.5 3971.0 1956.0 85.3 >1000.0 >650.0 721.1 43.4 15.4 430.8 >10.0 >5000.0 >10.0 20.0 >2500.0 301.9 734.7 >20,000.0 88.7 >10.0 >2.5 >10.0 36.4 33.9 >100.0 >100.0 >10.0 >250.0 >50.0 >10.0 >500.0 183.7 >125.0 >5.0 >250.0 >50.0 >5000.0 >10.0 >50.0 >2500.0 >10.0 >10.0 >50,000.0 >1250.0 >10.0 92.9 >10.0 >50.0 >250.0 >50.0 >250.0

26.4 192.0 >10 858.6 96.9 758.5 5.0 5.2 >1250.0 85.5 170.8 6 19.3 1.0 55.0 6252.0 80.0 3.5 >10.0 <7.8 19.5 33.6 64.2 1056.0 39.1 17.1 161.2 >650.0 191.5 17.0 3.4 114.1 1.6 1481 >10.0 11.1 29.5 23.4 267.1 927.7 9.7 1.6 >2.5 >10.0 4.4 4.8 >100.0 46.9 4.8 27.2 38.7 3.1 229.6 37.4 ND >5.0 >250.0 >50.0 3169.0 4.1 31.4 1588 16.3 >10.0 3565.0 >1250 >10.0 125.1 7.9 22.5 >250.0 0.9 >250.0

39.4 ~168.2 ND ND 125.7 4.0 ND 9.7 ND 41.2 5.1 0.3 3.1 4.8 ND ND ND 14.2 ND 24.6 109.9 62.3 52.4 64.7 1041.7 159.5 ND ND 34.2 2.1 77.0 430.8 ND ND ND 49.9 ND 37.3 56.1 ND 123.0 ND ND ND 38.7 42.6 ND ND ND ND ND ND ND 158.8 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND

5.0 25.8 ND 11.2 48.9 3.3 57.1 1.2 ND 10.2 6.0 0.1 0.4 0.2 11.0 9.0 177.8 1.8 ND 1.1 7.7 14.0 6.4 17.2 20.8 31.9 19.2 ND 9.1 0.8 16.9 114.1 81.5 46.3 ND 27.8 1.4 2.9 20.4 5.6 13.4 1478.4 ND ND 4.6 6.1 ND 76.3 96.6 18.0 129.1 27.8 77.9 32.3 ND ND ND ND 8.6 167.1 57.1 69.9 187.4 ND 10.7 ND ND 38.2 254.8 132.0 ND 2.4 ND

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

1917

Table 1 (continued ) Number

12 13 14 15 16 17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Chemicals

DILI Severity Category

Nicardipine Pargyline Praziquantel Prednisolone Pyridoxine Theophylline Alendronate Cromolyn Caffeine Digoxin Flumazenil Indoramin Levocarnitine Liothyronine Mecamylamine Neostigmine Orphenadrine Phenoxybenzamine Phentolamine Practolol Procyclidine Propantheline Pyridostigmine Scopolamine Streptomycin Triprolidine Zomepirac

4. 4. 4. 4. 4. 4. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5.

Enzyme Enzyme Enzyme Enzyme Enzyme Enzyme No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI No DILI

Elevations Elevations Elevations Elevations Elevations Elevations

Clinical Cmax (mM)

IC50 (mM)

0.08 1.0 2.0 2.2 0.1 83.8 0.02 0.6 77.2 0.005 1.1 0.3 85.7 0.002 0.1 0.05 0.7 0.2 0.09 18.8 2.2 0.4 1.1 0.003 64.5 0.1 13.7

IC50/Cmax Ratio

2D Monolayer

3D Spheroid

2D Monolayer

3D Spheroid

>10.0 >250.0 >250.0 >250.0 >50.0 >500.0 >5.0 >100.0 >750.0 2.0 >1250.0 >50.0 >750.0 >10.0 >50.0 >10.0 250.0 >50.0 >10.0 >2500.0 >250.0 >250.0 >250.0 >10.0 >5000.0 >10.0 >125.0

>10.0 >250.0 35.3 127.5 >50.0 >500.0 0.8 >100.0 >750.0 0.2 ~250.0 24.9 >750.0 ND >50.0 >10.0 40.7 >50.0 >10.0 1085 47.0 >250.0 1567.0 >10.0 867.0 >10.0 >125.0

ND ND ND ND ND ND ND ND ND 405.2 ND ND ND ND ND ND 336.9 ND ND ND ND ND ND ND ND ND ND

ND ND 17.2 ND ND ND 41.5 ND ND 40.7 ~223 86.4 ND ND ND ND 54.9 ND ND 57.8 21.4 ND 1419.4 ND 13.4 ND ND

DILI categories for tested compounds are based on Garside et al., 201423; Proctor et al., 2017.15 ND, not detected based on tested concentrations.

PHH and one lot KCs to set up PHH spheroids (1000 cells/spheroid) and co-culture liver spheroids (1000 PHHþ 500 KCs). After spheroid formation, PHH spheroids or PHH/KC co-culture spheroids were treated with LPS at 0, 1, 2, 5, and 10 mg/mL 48 h after

treatment, similar IL6 secretion was observed in co-culture spheroids for PHH lot 348a and lot 397 (Fig. 4). ATP levels were largely comparable and showed only a slight decrease at the highest LPS concentration of 10 mg/mL. Interestingly, we observed some IL6

Table 2 In Vitro Toxicity Assays 3D PHH Spheroids Versus 2D PHH Monolayer 100 Compound Comparison DILI Category

MOS 10 1. Severe DILI 2. High DILI concern 3. Low DILI concern 4. Enzyme elevation 5. No DILI Total Sensitivity Specificity MOS 30 1. Severe DILI 2. High DILI concern 3. Low DILI concern 4. Enzyme elevation 5. No DILI Total Sensitivity Specificity MOS 50 1. Severe DILI 2. High DILI concern 3. Low DILI concern 4. Enzyme elevation 5. No DILI Total Sensitivity Specificity

2D PHH Monolayer

3D PHH Spheroid

# of Compound

TP

TN

FP

FN

TP

TN

FP

FN

17 22 23 17 21 100 11% 100%

6 1 0 0 0 7

0 0 0 17 21 38

0 0 0 0 0 0

11 21 23 0 0 55

8 8 4 0 0 20 32% 95%

0 0 0 15 21 36

0 0 0 2 0 2

9 14 19 0 0 42

17 22 23 17 21 100 15% 100%

6 3 0 0 0 9

0 0 0 17 21 38

0 0 0 0 0 0

11 19 23 0 0 53

12 15 7 0 0 34 55% 87%

0 0 0 14 19 33

0 0 0 3 2 5

5 7 16 0 0 28

17 22 23 17 21 100 26% 100%

8 6 2 0 0 16

0 0 0 17 20 37

0 0 0 0 0 0

9 16 21 0 0 46

13 17 8 0 0 38 61% 79%

0 0 0 13 17 30

0 0 0 4 4 8

4 5 15 0 0 24

Sensitivity% ¼ TP/(TP þ FN)*100; Specificity% ¼ TN/(FP þ TN)*100. TP, true positive; TN, true negative; FP, false positive; FN, false negative.

1918

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Figure 2. (a) Representative images of PHH and PHH/KC co-culture spheroids cultured with different dexamethasone (Dex) concentrations in the medium. p values from 2-tailed unpaired student t tests are shown as ** ¼ p < 0.01, *** ¼ p < 0.001. (b) Spheroid size (diameter, mm) was measured with 16 spheroids from each culture condition on day 7. Coculture spheroids size comparison was made for different conditions. (c) Day 7 liver spheroids were cultured in a media supplemented with different concentration of Dex and then treated for 48 h with or without bacterial lipopolysaccharides (LPS) at 10 mg/mL. Spheroid viability was measured on day 9 by bioluminescent ATP assays. Comparison of spheroid ATP levels under different conditions was made to show the impact of Dex and LPS on spheroid viability. (d) IL6 secretion was measured by ELISA assays as described in Methods for week 1 and week 2 cultures of liver spheroids with or without LPS treatment. Comparisons of IL6 secretion with or without LPS were made for different Dex concentrations.

secretion in PHH spheroids (PHH lots 348a and 404), suggesting that these 2 PHH lots may have liver NPCs such as KCs or other IL6secreting cells. By contrast, lot 397 PHH spheroids showed minimum IL6 secretion. A review of the data revealed that LPS at 1mg/mL effectively induces activation of KCs in co-culture liver spheroids and does not adversely impact the viability of co-culture liver spheroids based on ATP measurement (Figs. 4a-4c). Toxicity Responses of Co-Culture Spheroids to Trovafloxacin and Acetaminophen With the established culture conditions described in the previous sections, we tested known DILI compounds to compare the in vitro cytotoxicity responses of PHH and co-culture liver spheroids. PHH lot 397 was used for these tests because this lot showed minimum background of IL6 secretion in response to LPS treatment. Two different lots of KCs were used in these tests. The ATP levels of co-culture spheroids were measured at 48 h or 5 days after treatment with trovafloxacin (Fig. 5a). When treated with trovafloxacin alone (no LPS), IC50 values at 48 h calculated from dose response curves showed >3 fold decrease for co-culture liver spheroids compared with PHH spheroids (573 mM vs 151 mM). With LPS (10 mg/mL) co-treatment, PHH and co-culture spheroids had IC50 values for trovafloxacin at 142 mM and 27 mM, respectively, representing a 3- to 5-fold decrease. These results show that including KCs in the co-culture liver spheroids potentiates the toxicity induced by trovafloxacin, which is consistent with the proposed immune-mediated toxicity mechanism for this DILI drug. At the same time, PHH spheroids also had a toxic response to trovafloxacin likely through a distinct and intrinsic hepatotoxicity mechanism (Figs. 5a and 5b). Interestingly, the difference in response to trovafloxacin-induced toxicity (no LPS) between PHH

(IC50 87mM) and PHH/KC co-culture spheroids (IC50 65 mM) became less obvious after a 5-day treatment. With LPS co-treatment, coculture liver spheroids also showed potentiated toxicity (IC50 10 mM) induced by trovafloxacin under these conditions. Contradictory roles of KCs in acetaminophen-induced liver toxicity have been reported in previous animal tests.24,25 KCs have been reported to contribute to or protect against acetaminopheninduced liver toxicity.26,27 However, whether these proposed roles of KCs in acetaminophen-induced liver injury can be translated to humans has yet to be determined. We tested acetaminophen with both PHH and co-culture spheroids using a serial dilution of acetaminophen with or without LPS co-treatment (Figs. 5c and 5d). Compared with PHH spheroids, acetaminophen treatment alone (no LPS) showed a decrease in IC50 values in coculture spheroids (IC50 of 3280 mM in PHH vs IC50 of 2247 mM in PHH/KC co-culture liver spheroids). Adding LPS did not affect the toxicity induced by acetaminophen in PHH spheroids with comparable IC50 values between no LPS and LPS co-treated spheroids. Interestingly, with LPS and acetaminophen co-treatment, IC50 values in co-culture spheroids increased 1.9-fold (2462 mM no LPS vs 4642 mM LPS), suggesting a protective role of activated KCs in acetaminophen induced toxicity in co-culture spheroids. We examined the secretion of inflammatory cytokines IL6 in coculture spheroids after treatment with serial dilutions of trovafloxacin or acetaminophen together with LPS cotreatment (Fig. 6). IL6 secretion by KCs showed different patterns in response to these 2 compounds. In acetaminophen-treated samples, IL6 secretion increased (up to 36.8 pg/mL) with significant amount of IL6 detected at 10,000 mM acetaminophen. By contrast, IL6 secretion (up to 14 pg/mL) decreased quickly with increasing trovafloxacin concentrations. IL6 generally plays a protective role against liver cell injury triggered by inflammation, disease, or drug-induced

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Figure 3. Cell ratio tests for co-culture liver spheroids with PHH and Kupffer cells. (a) Representative images from day 2 and day 5 cultures of liver spheroid with indicated cell ratio for PHH and Kupffer cells (KCs). 1000 PHHs were used with varied number of KCs at 500, 250, and 125 cells per spheroid. p values from 2-tailed unpaired student t tests are shown as * ¼ p < 0.05, ** ¼ p < 0.01, *** ¼ p < 0.001. (b) ATP measurement of day 9 liver spheroids with or without 48 h LPS treatment from day 7 culture. Two different lots of primary KCs were used (KC-1 and KC-2). (c) IL6 secretion in co-culture liver spheroids of PHH and KCs at the indicated ratio was measured by ELISA assays as described in Methods.

toxicity and promotes cell proliferation.25,28,29 On the other hand, TNF-a has been documented to play a more direct role in mediating hepatocyte cell death.30-33 The different response patterns of IL6 could potentially modulate the outcome and differential cytotoxicity responses observed in 3D liver spheroids induced by either acetaminophen or trovafloxacin. Differential Role of Kupffer Cells in In Vitro Liver Toxicity Tests We expanded the comparison of PHH and PHH/KC co-culture liver spheroids with or without LPS co-treatment with an additional 14 known DILI compounds (Table 3 and Fig. 7). In these tests, a single concentration for each compound was chosen based on the IC50 values (shown in Table 3) determined in the 2-week 3D liver spheroid toxicity test with 3 repeated dosing experiments described previously. In this single dosing test, toxicity induced by

1919

Figure 4. IL6 secretion and ATP measurements for PHH spheroids and co-culture liver spheroids made from 3 different PHH lots lot 397 (a), lot 348a (b), and lot 404 (c) and one lot of Kupffer cells. LPS was added to the spheroid culture at indicated concentrations for 48 h before the collection of culture supernatant samples from 3 to 8 spheroid cultures and IL6 ELISA assays. Comparisons of IL6 secretion to no LPS (0 mM) were made for liver spheroids treated with different LPS concentrations. p values from 2-tailed unpaired student t tests are shown as * ¼ p < 0.05, ** ¼ p < 0.01, *** ¼ p < 0.001.

some of these DILI compounds was expected to be less because of the difference in PHH lot used, dosing time, and duration of the tests. ATP measurement, IL6, and TNF-a ELISA assays were performed after 5-day treatment of liver spheroids. Under these conditions, a single-dose treatment of several DILI compounds such as isoniazid, naltrexone, and troglitazone did not cause significant spheroid death in contrast to the repeated dosing experiments. These observations again show the advantages of using liver spheroids to improve the sensitivity of in vitro assays by conducting tests with both repeated dosing and extended treatment time. Other compounds such as bosentan, amodiaquine, indomethacin, or paroxetine induced toxicity responses in PHH or co-culture liver spheroids. In these tests, ATP measurements were made and normalized to the solvent control DMSO (up to 0.1%) to compare the PHH and coculture spheroids in response to druginduced cytotoxicity with or without LPS treatment. To quantify these observations, normalized ATP levels were summarized

1920

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Figure 5. (a) Representative dose response curves for trovafloxacin with or without LPS cotreatment for PHH or PHH/KC co-culture liver spheroids after 48 h (left panels) and 5 d (right). Ten microgram per milliliter LPS were used to treat the liver spheroids. (b) Average IC50 values (day 5) from multiple tests using PHH or PHH/KC co-culture liver spheroids with or without 1 mg/mL LPS. Student t test was used for statistical analysis, * ¼ p < 0.05. (c) Representative dose response curves for acetaminophen with or without LPS cotreatment for PHH or PHH/KC co-culture liver spheroids after 5 d of incubation. One microgram per milliliter LPS was used to treat the liver spheroids. DILI compounds and LPS were added simultaneously to day 7 liver spheroids to start the treatment. (d) Average IC50 values (day 5) from multiple tests using PHH or PHH/KC co-culture liver spheroids. p value were calculated using student t test, * ¼ p < 0.05.

(Table 4) and used to calculate the relative viability ratio between PHH and co-culture spheroids under these conditions. For example, nefazodone-induced toxicity responses were similar between PHH and co-culture liver spheroids with or without LPS co-treatment. Although tolcapone alone did not induce significant cytotoxicity in PHH or co-culture liver spheroids, with LPS co-treatment, co-culture liver spheroids had much less spheroid death compared with PHH spheroids (relative viability coculture/PHH ratio at 121.8) suggesting a protective role of activated KCs for co-culture liver spheroids. Without LPS co-treatment, cytotoxicity induced by paroxetine, indomethacin, and sulindac appeared to be potentiated in co-culture liver spheroids with coculture/PHH ratios less than 0.6. LPS co-treatment with either paroxetine or indomethacin showed less spheroid death in coculture liver spheroids (LPS/no LPS relative viability ratio at 2.7 and 1.7, respectively). Therefore, the status of KC activation appeared to have compound-dependent differential effects on drug-induced cytotoxicity. We also examined the secretion of IL6 and TNF-a in these tests with different DILI drugs using day 5 culture supernatant samples. IL6 secretion tightly correlated with LPS treatment (Fig. 7c). Without LPS treatment, only background levels of IL6 were measured (data not shown). In samples treated with indomethacin and paroxetine, relatively higher IL6 secretion was observed that could potentially explain the observed increased viability

with LPS co-treatment as described previously. On the other hand, TNF-a has been documented to play a more direct role in mediating hepatocyte cell death.30-33 With the exceptions of bosentan and paroxetine, TNF-a was detected in most drug-treated coculture spheroids with LPS co-treatment. Interestingly, higher TNF-a secretion levels were observed in co-culture spheroids treated with isoniazid and naltrexone without obvious spheroid death as shown by ATP levels. Discussion Advances have been made in recent years to use 3D liver spheroids made from primary human liver cells as a more physiologically relevant in vitro model for studying liver toxicity.10,11,15 In comparison to conventional 2D monolayer cultures of PHHs, our results with 100 selected DILI and control compounds demonstrate that 3D liver spheroids made with PHHs can significantly increase the sensitivity of in vitro liver toxicity assays. Together with 2 recent reports using 3D liver microtissues or PHH spheroids testing with more than 100 compounds from different DILI categories, our results provide additional evidence and independent characterization that these novel 3D liver models are useful tools to significantly improve the performance of in vitro liver toxicity assays.15,16 Notably, liver microtissues used in the previous report by Proctor et al., were made with PHHs and liver NPCs from different donors. Because our results

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

1921

Figure 6. (a) IL6 secretion from co-culture liver spheroids treated for 5 d with serial dilutions of trovafloxacin (a) or acetaminophen (b) and co-treatment of LPS at 1 mg/mL. The levels of IL6 at different drug concentrations were compared with solvent controls dimethyl sulfoxide (DMSO) or medium, respectively. p values from two-tailed unpaired student t tests are indicated in the graph as * ¼ p < 0.05, ** ¼ p < 0.01, *** ¼ p < 0.001.

with PHH spheroids without NPCs demonstrate similar sensitivity, it is not clear how and whether NPCs in those microtissues may contribute to increased sensitivity. Given the variation between PHH lots from different donors and differences in methods used among our and other studies, the reported improvement in assay sensitivity from these independent studies is largely comparable at 2 to 3 times better than 2D monolayer PHH cultures. Many DILI compounds may induce hepatocytotoxicity through multiple pathways and complex interactions of contributing mechanisms.34,35 It has been difficult to dissect the involvement of KCs and inflammation stress to DILI in humans because of the lack of suitable models.36 PHH/KC co-culture spheroids can be a useful tool for better understanding the role of innate immunity and inflammation in drug-induced toxicity and liver disease development. As shown in the results, the flexibility of applying different compound dosing conditions and comparison between PHH spheroid and PHH/KC co-culture spheroid models revealed the role of KCs in the early stage of 48-h treatment with trovafloxacin.

Table 3 List of DILI Compounds Used for PHH and Co-Culture Liver Spheroid Comparison Compound

DILI Category

Concentration for Dosing, mM

Bosentan Isoniazid Naltrexone Troglitazone Tolcapone Nefazodone Amodiaquine Imipramine Indomethacin Nitrofurantoin Paroxetine Tetracycline Sulindac Chlorpromazine

1. 1. 1. 1. 1. 1. 2. 2. 2. 2. 2. 2. 2. 3.

192.0 858.6 5.0 1.0 19.3 5.2 3.5 17.1 161.2 17.0 3.4 29.5 1481.0 4.4

Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI Severe DILI High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern High DILI Concern Low DILI Concern

Figure 7. Comparison of liver toxicity induced by 14 DILI compounds for PHH (a) or coculture liver spheroids made from PHH/KC (b). Normalized ATP levels (to DMSO solvent control) were compared between drug treatment with or without LPS cotreatment at 1 mg/mL with p values from 2-tailed unpaired student t tests shown (c). Secretion of IL6 and TNF-a was measured by ELISA assays using the medium supernatant from co-culture liver spheroid cultures after treatment with 1 mg/mL LPS for 5 d. The levels of IL6 were compared between drug-treated samples and DMSO solvent control with p values from two-tailed unpaired student t test shown in the bar graph as * ¼ p < 0.05, ** ¼ p < 0.01.

The differences in the secretion of IL6 and TNF-a in co-culture spheroids with LPS treatment can only partially explain the observations of the protective role of activated KCs against acetaminophen-induced hepatotoxicity. In a previous report by Bonzo et al.,37 trovafloxacin has been shown to differentially regulate the secretion of IL6 and TNF-a in a rat hepatocyte/KC co-culture. In acetaminophen-treated mice or overdosed patients, inflammatory cytokines including IL6 and TNF-a can be detected in blood. However, their roles or contribution to liver toxicity remains to be clarified.38 There is an ongoing debate on the role of innate immune cells

1922

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923

Table 4 Fourteen DILI Compound Comparison Using PHH and PHH/KC Co-Culture Liver Spheroids DILI Compound

Bosentan Isoniazid Naltrexone Troglitazone Amodiaquine Imipramine Indomethacin Nitrofurantoin Paroxetine Tetracycline Chlorpromazine Sulindac Tolcapone Nefazodone

No LPS

LPS

LPS:No LPS Ratio

PHH

PHH/KC

PHH/KC:PHH Ratio

PHH

PHH/KC

PHH/KC:PHH Ratio

PHH

PHH/KC

100.1 102.0 78.3 119.1 66.4 88.1 62.2 103.6 93.8 121.8 101.8 77.5 85.3 87.6

79.8 100.3 77.3 103.4 47.3 102.9 33.0 93.9 20.9 102.4 97.4 38.1 102.2 111.1

0.8 1.0 1.0 0.9 0.7 1.2 0.5 0.9 0.2 0.8 1.0 0.5 1.2 1.3

27.8 111.6 94.4 112.2 61.2 79.3 58.3 102.2 65.9 99.5 73.5 32.8 0.5 26.1

57.3 118.5 94.9 111.8 73.9 96.6 55.9 107.9 55.4 103.9 105.6 33.8 57.8 33.4

2.1 1.1 1.0 1.0 1.2 1.2 1.0 1.1 0.8 1.0 1.4 1.0 121.8 1.3

0.3 1.1 1.2 0.9 0.9 0.9 0.9 1.0 0.7 0.8 0.7 0.4 0.01 0.3

0.7 1.2 1.2 1.1 1.6 0.9 1.7 1.1 2.7 1.0 1.1 0.9 0.6 0.3

including neutrophils and KCs in acetaminophen-induced liver injury and acute liver failure in mice and patients.38,39 Ju et al.,27 previously reported that acetaminophen-induced liver injury is more severe in mice after depletion of KCs using liposomeentrapped clodronate. Our results using co-culture human liver spheroids, for the first time to our knowledge, demonstrate that activated KCs can protect against acetaminophen-induced cytotoxicity in a physiologically relevant in vitro human system. Therefore, the co-culture human liver spheroids described in the present study could be a useful in vitro human model to translate some of the findings in rodent models about the role the KCs and inflammation in acetaminophen-induced liver injury. From the 14-compound test comparing the PHH and PHH/KC coculture liver spheroids, the role of KCs and inflammation simulated with LPS co-treatment appeared to be compound-dependent. Using these defined conditions, further detailed analysis with additional omics approaches will certainly shed light to more humanrelevant mechanisms underlying DILI by these DILI compounds. Inflammation has been postulated as a contributing factor for idiosyncratic drugeinduced injury.40 Concurrent inflammationrelated stress due to infection or diseases could have a negative impact on liver homeostasis by inhibiting tissue repair and recovery and therefore exacerbate hepatotoxicity by certain drugs.41 Inflammation can inhibit the expression and functions of key drug-metabolizing enzymes and transporters. This could lead to an accumulation of drug compounds in the liver because of impaired drug metabolism and clearance and therefore increase the risk of exposure to the drugs.42,43 Conclusions In summary, 3D spheroid culture of PHHs significantly improves cell viability to allow long-term in vitro toxicity tests with repeated dosing. As a result, PHH spheroids are a more sensitive tool to detect drug-induced liver toxicity in comparison to conventional 2D monolayer culture of PHHs where the cells quickly lose their phenotype and viability within a few days. With optimization of culture medium supplements, cell ratio, and LPS treatment, PHH/ KC co-culture liver spheroids can recapitulate inflammatory responses as measured by IL6 secretion in a 3D culture system that is more physiologically relevant. These co-culture liver spheroids can be used to explore the roles of KCs, a critical component of innate immune system, and the role of inflammatory responses in DILI. Our results using model DILI compounds such as trovafloxacin, acetaminophen, and others demonstrate that the effect of KCs and simulated inflammation after LPS co-treatment on hepatotoxicity

are compound dependent. Further studies with genomic and other investigative tools could shed insights into the underlying mechanisms with more human relevance. Acknowledgments The authors thank Joanne Bourgea and Kirsten Cooper for technical support and Ronald Faris and Anthony Frutos for reviewing the manuscript. References 1. LeCluyse EL, Witek RP, Andersen ME, Powers MJ. Organotypic liver culture models: meeting current challenges in toxicity testing. Crit Rev Toxicol. 2012;42:501-548. 2. Godoy P, Hewitt N, Albrecht U, et al. Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative hepatocyte sources and nonparenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME. Arch Toxicol. 2013;87:1315-1530. 3. Lauschke VM, Hendriks DF, Bell CC, Andersson TB, Ingelman-Sundberg M. Novel 3D culture systems for studies of human liver function and assessments of the hepatotoxicity of drugs and drug candidates. Chem Res Toxicol. 2016;29: 1936-1955. 4. LeCluyse EL, Alexandre E, Hamilton GA, et al. Isolation and culture of primary human hepatocytes. 290. Totowa, NJ: Humana Press; 2005. 5. Kono Y, Yang S, Roberts EA. Extended primary culture of human hepatocytes in a collagen gel sandwich system. In Vitro Cell Dev Biol Anim. 1997;33:467472. 6. Elaut G, Henkens T, Papeleu P, et al. Molecular mechanisms underlying the dedifferentiation process of isolated hepatocytes and their cultures. Curr Drug Metab. 2006;7:629-660. 7. Gomez-Lechon MJ, Tolosa L, Conde I, Donato MT. Competency of different cell models to predict human hepatotoxic drugs. Expert Opin Drug Metab Toxicol. 2014;10:1553-1568. 8. Bell CC, Dankers ACA, Lauschke VM, et al. Comparison of hepatic 2D sandwich cultures and 3D spheroids for long-term toxicity applications: a multicenter study. Toxicol Sci. 2018;162:655-666. 9. Khetani SR, Bhatia SN. Microscale culture of human liver cells for drug development. Nat Biotechnol. 2008;26:120-126. 10. Messner S, Agarkova I, Moritz W, Kelm JM. Multi-cell type human liver microtissues for hepatotoxicity testing. Arch Toxicol. 2013;87:209-213. 11. Bell CC, Hendriks DF, Moro SM, et al. Characterization of primary human hepatocyte spheroids as a model system for drug-induced liver injury, liver function and disease. Sci Rep. 2016;6:25187. 12. Hendriks DF, Fredriksson Puigvert L, Messner S, Mortiz W, IngelmanSundberg M. Hepatic 3D spheroid models for the detection and study of compounds with cholestatic liability. Sci Rep. 2016;6:35434. 13. Bell CC, Lauschke VM, Vorrink SU, et al. Transcriptional, functional, and mechanistic comparisons of stem cell-derived hepatocytes, HepaRG cells, and three-dimensional human hepatocyte spheroids as predictive in vitro systems for drug-induced liver injury. Drug Metab Dispos. 2017;45:419-429. 14. Messner S, Agarkova I, Moritz W, Kelm JM. Transcriptomic, proteomic, and functional long-term characterization of multicellular three-dimensional human liver microtissues. Appl In Vitro Toxicol. 2018;4:1-12. 15. Proctor WR, Foster A , Vogt J, et al. Utility of spherical human liver microtissues for prediction of clinical drug-induced liver injury. Arch Toxicol. 2017;91:28492863.

F. Li et al. / Journal of Pharmaceutical Sciences 109 (2020) 1912-1923 16. Vorrink SU, Zhou Y, Ingelman-Sundberg M, Lauschke VM. Prediction of drug-induced hepatotoxicity using long-term stable primary hepatic 3D spheroid cultures in chemically defined conditions. Toxicol Sci. 2018;163: 655-665. 17. Kozyra M, Johansson I, Nordling A, Ullah S, Lauschke VM, IngelmanSundberg M. Human hepatic 3D spheroids as a model for steatosis and insulin resistance. Sci Rep. 2018;8:14297. 18. Kolios G, Valatas V, Kouroumalis E. Role of Kupffer cells in the pathogenesis of liver disease. World J Gastroenterol. 2006;12:7413-7420. 19. Dixon LJ, Barnes M, Tang H, Pritchard MT, Nagy LE. Kupffer cells in the liver. Compr Physiol. 2013;3:785-797. 20. Rose KA, Holman NS, Green AM, Andersen ME, LeCluyse EL. Co-culture of hepatocytes and Kupffer cells as an in vitro model of inflammation and druginduced hepatotoxicity. J Pharm Sci. 2016;105:950-964. 21. Bilzer M, Roggel F, Gerbes AL. Role of Kupffer cells in host defense and liver disease. Liver Int. 2006;26:1175-1186. 22. Nguyen-Lefebvre AT, Horuzsko A. Kupffer cell metabolism and function. J Enzymol Metab. 2015;1:101. 23. Garside H, Marcoe KF, Chesnut-Speelman J, et al. Evaluation of the use of imaging parameters for the detection of compound-induced hepatotoxicity in 384-well cultures of HepG2 cells and cryopreserved primary human hepatocytes. Toxicol In Vitro. 2014;28:171-181. 24. Jaeschke H, Williams CD, Ramachandran A, Bajt ML. Acetaminophen hepatotoxicity and repair: the role of sterile inflammation and innate immunity. Liver Int. 2012;32:8-20. 25. Krenkel O, Mossanen JC, Tacke F. Immune mechanisms in acetaminopheninduced acute liver failure. Hepatobiliary Surg Nutr. 2014;3:331-343. 26. Roberts RA, Ganey PE, Ju C, Kamendulis LM, Rusyn I, Klaunig JE. Role of the Kupffer cell in mediating hepatic toxicity and carcinogenesis. Toxicol Sci. 2007;96:2-15. 27. Ju C, Reilly TP, Bourdi M, et al. Protective role of Kupffer cells in acetaminophen-induced hepatic injury in mice. Chem Res Toxicol. 2002;15:1504-1513. 28. Streetz KL, Tacke F, Leifeld L, et al. Interleukin 6/gp130-dependent pathways are protective during chronic liver diseases. Hepatology. 2003;38:218-229. 29. Klein C, Wustefeld T, Assmus U, et al. The IL-6-gp130-STAT3 pathway in hepatocytes triggers liver protection in T cell-mediated liver injury. J Clin Invest. 2005;115:860-869.

1923

30. Trautwein C, Rakemann T, Brenner DA, et al. Concanavalin A-induced liver cell damage: activation of intracellular pathways triggered by tumor necrosis factor in mice. Gastroenterology. 1998;114:1035-1045. 31. Shuh M, Bohorquez H, Loss Jr GE, Cohen AJ. Tumor necrosis factor-alpha: life and death of hepatocytes during liver Ischemia/Reperfusion injury. Ochsner J. 2013;13:119-130. 32. Jones BE, Lo CR, Liu H, et al. Hepatocytes sensitized to tumor necrosis factoralpha cytotoxicity undergo apoptosis through caspase-dependent and caspaseindependent pathways. J Biol Chem. 2000;275:705-712. 33. Hatano E. Tumor necrosis factor signaling in hepatocyte apoptosis. J Gastroenterol Hepatol. 2007;22(Suppl 1):S43-S44. 34. Almario EE, Borlak J, Suzuki A, Chen M. Drug-induced liver injury. Biomed Res Int. 2017;2017:2461694. 35. Kuna L, Bozic I, Kizivat T, et al. Models of drug induced liver injury (DILI) current issues and future perspectives. Curr Drug Metab. 2018;19:830-838. 36. Shaw PJ, Ganey PE, Roth RA. Idiosyncratic drug-induced liver injury and the role of inflammatory stress with an emphasis on an animal model of trovafloxacin hepatotoxicity. Toxicol Sci. 2010;118:7-18. 37. Bonzo JA, Rose K, Freeman K, et al. Differential effects of trovafloxacin on TNF-a and IL-6 profiles in a rat hepatocyteeKupffer cell coculture system. Appl In Vitro Toxicol. 2015;1:45-54. 38. Woolbright BL, Jaeschke H. Role of the inflammasome in acetaminopheninduced liver injury and acute liver failure. J Hepatol. 2017;66:836-848. 39. James LP, Lamps LW, McCullough S, Hinson JA. Interleukin 6 and hepatocyte regeneration in acetaminophen toxicity in the mouse. Biochem Biophys Res Commun. 2003;309:857-863. 40. Jiang J, Mathijs K, Timmermans L, et al. Omics-based identification of the combined effects of idiosyncratic drugs and inflammatory cytokines on the development of drug-induced liver injury. Toxicol Appl Pharmacol. 2017;332:100-108. 41. Iorga A, Dara L, Kaplowitz N. Drug-induced liver injury: cascade of events leading to cell death, apoptosis or necrosis. Int J Mol Sci. 2017;18:1018. 42. Morgan ET. Impact of infectious and inflammatory disease on cytochrome P450-mediated drug metabolism and pharmacokinetics. Clin Pharmacol Ther. 2009;85:434-438. 43. Deng X, Luyendyk JP, Ganey PE, Roth RA. Inflammatory stress and idiosyncratic hepatotoxicity: hints from animal models. Pharmacol Rev. 2009;61: 262-282.