Author response: Application of human liver organoids as a patient-derived primary model for HBV infection and related hepatocellular carcinoma

Article Figures and data Abstract Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The molecular events that drive hepatitis B virus (HBV)-mediated transformation and tumorigenesis have remained largely unclear, due to the absence of a relevant primary model system. Here we propose the use of human liver organoids as a platform for modeling HBV infection and related tumorigenesis. We first describe a primary ex vivo HBV-infection model derived from healthy donor liver organoids after challenge with recombinant virus or HBV-infected patient serum. HBV-infected organoids produced covalently closed circular DNA (cccDNA) and HBV early antigen (HBeAg), expressed intracellular HBV RNA and proteins, and produced infectious HBV. This ex vivo HBV-infected primary differentiated hepatocyte organoid platform was amenable to drug screening for both anti-HBV activity and drug-induced toxicity. We also studied HBV replication in transgenically modified organoids; liver organoids exogenously overexpressing the HBV receptor sodium taurocholate co-transporting polypeptide (NTCP) after lentiviral transduction were not more susceptible to HBV, suggesting the necessity for additional host factors for efficient infection. We also generated transgenic organoids harboring integrated HBV, representing a long-term culture system also suitable for viral production and the study of HBV transcription. Finally, we generated HBV-infected patient-derived liver organoids from non-tumor cirrhotic tissue of explants from liver transplant patients. Interestingly, transcriptomic analysis of patient-derived liver organoids indicated the presence of an aberrant early cancer gene signature, which clustered with the hepatocellular carcinoma (HCC) cohort on The Cancer Genome Atlas Liver Hepatocellular Carcinoma dataset and away from healthy liver tissue, and may provide invaluable novel biomarkers for the development of HCC and surveillance in HBV-infected patients. Introduction Persistent hepatitis B virus (HBV) infection is the leading cause of chronic liver cirrhosis and hepatocellular carcinoma (HCC) world wide (MacLachlan and Cowie, 2015; Di Bisceglie, 2009; An et al., 2018). A combination of viral and host factors determines whether an individual infected with HBV will be able to clear the infection or will become a chronic carrier. Characterized by its high host species and organ specificity, HBV infection and replication is thought to orchestrate an interplay between the immune system and viral-specific factors that eventually lead to the onset of HCC. Insights into the molecular mechanisms underlying HBV-induced HCC have largely been provided by epidemiological studies (El-Serag, 2012; Fattovich et al., 2008; Jiang et al., 2012; Sagnelli et al., 2020), genome-wide analysis of viral and host characteristics (Ally et al., 2017; Fujimoto et al., 2012; Huang et al., 2012; Ji et al., 2014; Sartorius et al., 2019; Shibata and Aburatani, 2014; Ally et al., 2017), as well as by studies performed in in vitro settings using hepatoma cell lines (Thomas and Liang, 2016; Zhang et al., 2014). However, the limited availability of relevant animals or in vitro model systems to study HBV infection constitutes a major deficiency attributed to the strict viral host and cell-type tropism. Chimpanzees remain the only animal model that supports the full HBV replication cycle, while available hepatoma cell line models are unsuitable for delineating the molecular steps leading to tumorigenesis as they differ substantially from primary cells in their already tumor-derived gene expression profiles (Protzer, 2017). These systems have inherent limitations resulting in poor predictive value for clinical outcomes (Allweiss and Dandri, 2016). Primary hepatocytes present the gold standard model system for HBV research in vitro, with recent studies reporting on significant increases in the half-life of these cultures (Xiang et al., 2019). However, primary cells are difficult to obtain and, since they cannot be expanded, are not usually available in quantities sufficient to perform large-scale analyses (Hu et al., 2019). Induced pluripotent stem cell (iPSC)-derived hepatocytes, which are susceptible to HBV infection and support replication, are also a useful model for studying host-virus determinants of replication (Kaneko et al., 2016; Nie et al., 2018; Sakurai et al., 2017; Xia et al., 2017). However, iPSC-derived hepatocyte models cannot be patient derived, confining studies to only ex vivo infection systems and limiting the possibility of patient-specific personalized treatment approaches (Torresi et al., 2019; Nantasanti et al., 2016). As a consequence of deficiencies in available model systems and, despite its fascinating biology, many questions regarding the life cycle of HBV and its mechanisms of persistence, including HBV-induced molecular events underlying tumorigenesis, remain largely unexplored in primary settings, and key viral and host players involved remain unknown. The dependence of hepatocytes on spatial and matrix-derived signals had until recently prevented their long-term in vitro culturing. Organoid culture technology involves the generation of cell-derived genetically stable in vitro 3D organ models of human origin. We have previously established a primary liver culture system based on isolation and expansion of primary cells that allows for the long-term expansion of liver cells as organoids (Huch et al., 2013; Huch et al., 2015). In this culture system, isolated adult hepatic cells are expanded through multiple passages in an optimized expansion medium (EM) without induction of genomic alterations (Huch et al., 2015). When switched to a differentiation medium (DM) where proliferation signals are removed and ductal (progenitor) fate is inhibited, liver organoid cultures differentiate into functional hepatocytes in vitro as exemplified by their polygonal cell shape (Figure 1A) and hepatocyte functions, including albumin production and cytochrome CYP3A4 expression and activity (Huch et al., 2015). Here we use the human liver organoid platform to model and study HBV infection and replication, as well as related tumorigenesis in patient-derived organoids generated from HBV-infected donors. This expandable model yields patient-specific organoids in quantities amenable to molecular and functional characterization and allows us to generate a living biobank of HBV-infected patient-derived cells amenable to downstream genomic, transcriptomic, and proteomic analysis as well as screening for HBV-directed therapeutics. We first described the ex vivo HBV infection of healthy donor (hD)-derived liver organoids, as a model to investigate viral infection and replication in hepatocytes. We used the HBV-infected organoid model as a platform for drug screening that can measure both drug-induced anti-HBV transcription and replication activity as well as drug-induced toxicity. We also demonstrated that transgenic modification of liver organoids provides an in vitro mechanistic platform to study the molecular determinants of HBV infection and replication. Finally, we performed transcriptomic analysis of HBV-infected patient-derived organoids and described the discovery of an early cancer gene signature, a potentially invaluable prognostic biomarker for HCC. Figure 1 with 3 supplements see all Download asset Open asset Modeling HBV infection in vitro using human liver organoids. (A) Representative images of liver organoids in expansion medium (EM) and differentiation medium (DM). (B) Experimental design of infection experiments. Arrows indicate the time points for hepatitis B virus (HBV) detection. (C) Levels of HBV DNA in supernatants of infected organoid cultures were quantified at indicated times by quantitative polymerase chain reaction (qPCR) and compared to the cultures challenged with heat-inactivated (HI) virus. (D) Schematic of the HBV genome showing open reading frames (ORFs) (arrows), HBV RNA transcripts (black circular lines) and the localization of PCR products (blue boxes). The agarose gel demonstrates expression of 3.5kb RNA transcript and total HBV RNA by nested PCR performed on complementary DNA (cDNA) obtained from two in vitro infected healthy donor (hD) organoid lines. (E) Immunofluorescent staining showing the expression of HBV core antigen (HBcAg) (green) together with sodium taurocholate co-transporting polypeptide (NTCP) (magenta), β-catenin (gray), or Hepatocyte Nuclear Factor 4 Alpha (HNF4α) (red) performed in different hD organoids 6 days after HBV infection in the differentiation medium. (F) Quantification of total intracellular HBV DNA (left Y-axis) and covalently closed circular DNA (cccDNA) (right Y-axis) purified from four hD organoid lines 6 days post infection. Quantification of total HBV DNA and cccDNA is also shown from HepG2.2.15-produced virus (inoculum, negative control for cccDNA) and from double-stranded HBV plasmid (as positive control for cccDNA) as indicated. (G) Quantification of HBV early antigen (HBeAg) was performed by enzyme-linked immunosorbent assay (ELISA) from the supernatant of infected organoids. Challenge with HI virus and uninfected organoids was used as a control for HBeAg present in the inoculum. Positive and neg bars correspond to positive and negative controls provided by the kit manufacturer. (H) Expression of intracellular HBV RNA relative to beta-2-microglobulin in three hD organoid lines infected with organoid-produced HBV (concentrated from pooled supernatants of organoid cultures infected with HepG2.2.15-produced HBV). (I) Immunofluorescent staining showing the expression of HBV surface antigen (HBsAg) (green) in DM organoids infected with recombinant HBV and patient serum. Scale bars represent 50 µm. Bar graphs show total HBV RNA levels in the culture at the time of staining. Figure 1—source data 1 Source data for Figure 1F. https://cdn.elifesciences.org/articles/60747/elife-60747-fig1-data1-v2.xlsx Download elife-60747-fig1-data1-v2.xlsx Figure 1—source data 2 Source data for Figure 1G. https://cdn.elifesciences.org/articles/60747/elife-60747-fig1-data2-v2.xlsx Download elife-60747-fig1-data2-v2.xlsx Results Human liver organoids allow modeling of HBV infection in vitro We first used the previously characterized liver organoid platform (Huch et al., 2015) on materials generated from healthy donors to set up a novel ex vivo HBV-infection system to study HBV replication. Liver organoids from healthy donors were grown in either EM or DM (Figure 1A) for 7 days prior to infection with recombinant HBV generated from HepG2.2.15, an HepG2 cell line subclone stably expressing HBV (Figure 1B). As control for the inoculum, we also infected organoids with heat-inactivated (HI) HBV. HBV infection and replication were validated by quantifying the levels of HBV DNA in the supernatant (Figure 1C), detection of intracellular HBV RNA (Figure 1D), visualizing HBV-specific proteins by immunofluorescence microscopy (Figure 1E), and quantifying intracellular covalently closed circular DNA (cccDNA) from infected organoids (Figure 1F). HBV DNA was detected in organoid culture supernatants from 4 days post infection, but not from the HI virus-infected cells, pointing to successful HBV replication (Figure 1C). Differentiated organoids maintained in DM were more efficiently infected and produced higher viral titers than organoids maintained in EM (Figure 1C–D). The RNA intermediates necessary for protein production and viral replication (3.5 kb RNA transcript and total HBV RNA) were present in infected DM organoids and detected by nested polymerase chain reaction (PCR) analysis, but not in the HI virus-infected cells (Figure 1D). As a further measure of active HBV replication, HBV early antigen (HBeAg) was also measured in supernatants of infected organoids and quantified (Figure 1G). Immunostaining, using antibodies recognizing HBV core antigen (HBcAg), showed specific nuclear and cytoplasmic staining in multiple infected healthy donor liver organoid lines, confirming the presence of foci of HBV replication in HBV-infected cells predominantly in infected DM organoids (Figure 1E and Figure 1—figure supplement 1A–B). Furthermore, infection of DM organoids resulted in the production of cccDNA, a definitive marker of HBV replication, as detected by a quantitative polymerase chain reaction (qPCR)-based cccDNA detection method of intracellular HBV DNA after digestion with a nuclease to specifically remove non-cccDNA (Figure 1F and Figure 1—figure supplement 1C). Inoculum that lacks cccDNA was used as a negative control for the cccDNA-specific qPCR and HBV plasmid DNA was used as a positive control (Figure 1F and Figure 1—figure supplement 1C). HBV replication, infection, and spread appeared to be persistent until 8 days after infection when viral production dropped significantly, likely because of the limited half-life of differentiated organoids in culture (Figure 1—figure supplement 2A). Periodic culturing of the organoids in EM in order to stimulate the recovery and proliferation of the organoids modestly extended the half-life of the infected cultures, where viral production was maintained for approximately 1 month post infection (Figure 1—figure supplement 2B). Donor-specific differences in efficiency of HBV infection were observed consistent with variable HBV permissiveness of primary human hepatocytes (Shlomai et al., 2014), while HBV infection was observed with similar efficiency in different passages of the same donor line (Figure 1—figure supplement 2C–D). To determine whether the organoids are capable of producing infectious HBV, supernatants containing virus produced by organoids were collected, concentrated, and used for subsequent spinoculation of hD organoids. As shown in Figure 1H, infection of hD organoids with organoid-produced HBV resulted in expression of intracellular HBV RNA, indicating that organoids produce infectious viral particles. Growth of viral isolates from patient material has been limited by the lack of an adequate primary model system. However, differentiated organoids were able to support infection and replication when challenged with HBV-infected patient sera, as shown by production of viral DNA, expression of viral transcripts, and positive immunostaining for HBV surface antigen (HBsAg) (Figure 1I and Figure 1—figure supplement 3). Differentiated liver organoids therefore provide a useful ex vivo HBV-infection platform in which the role of specific host and viral factors can be investigated. Ex vivo HBV-infected liver organoids are a viable platform for anti-viral drug screening and drug-induced toxicity We next examined whether the ex vivo infected liver organoid platform would be amenable to anti-HBV drug screening to monitor antiviral activity and drug-induced toxicity of two different drugs, tenofovir and fialuridine, according to the schematic outlined in Figure 2A. Tenofovir is a nucleoside reverse transcriptase inhibitor that inhibits the reverse transcription of HBV pre-genomic RNA to DNA. Fialuridine, also a nucleoside analog that inhibits reverse transcription, was shown to cause severe hepatotoxicity in patients (McKenzie, 1995). In the organoids, HBV viral DNA production in the culture supernatant was inhibited by both tenofovir and fialuridine in three independent hD-derived organoids, whereas, as expected, RNA levels remained the same (Figure 2B). Therefore, the organoid ex vivo infection platform not only allows measurement of drug-induced antiviral activity, but also offers insight into the mechanism of drug action by allowing delineation of distinct steps of the HBV life cycle targeted and inhibited. As expected, treatment of HepG2.2.15 cells with tenofovir and fialuridine resulted in similar decreases in released HBV DNA, but no change in intracellular HBV RNA levels (Figure 2C), reaffirming the mechanism of action of these drugs in a cell-line model of HBV replication. Due to the well-established detrimental effects of fialuridine on the viability of primary human hepatocytes, we sought to evaluate fialuridine-induced toxicity on primary human liver organoids as well as in HepG2 cells. We measured the viability of organoids and HepG2 cells using the alamarBlue viability assay and by monitoring their phenotype upon fialuridine and tenofovir treatment using microscopy. HepG2 cells demonstrated no change in cell viability upon treatment with tenofovir and increasing concentrations of fialuridine as compared to the mock-treated cells (Figure 2D). The phenotype of HepG2 cells was also comparable across all treatments as observed by microscopy (Figure 2E). Strikingly, the liver organoids treated with fialuridine at as low a concentration as 1 µM demonstrated a significant reduction in viability as measured by alamarBlue assay when compared to mock-treated cells (Figure 2F and Figure 2—figure supplement 1A). The organoids treated with higher fialuridine concentrations (5–20 µM) as well as with 20 µM tenofovir also demonstrated impaired viability (Figure 2F and Figure 2—figure supplement 1A). The decreased cell viability was also apparent in the phenotype of the 1–20 µM fialuridine-treated and 20-µM tenofovir-treated organoids compared to vehicle controls as observed by microscopy (Figure 2G and Figure 2—figure supplement 1B). This highlights that fialuridine-induced toxicity is evident and quantifiable in the primary human liver organoid model but not in the HepG2.2.15 model of HBV replication. Thus, we demonstrate that ex vivo-infected differentiated liver organoids support the full replication cycle of HBV and, following further characterization, may serve as an ideal novel primary platform for drug screening as well as elucidation of the molecular events underlying HBV replication. Moreover, human liver organoids serve as an ideal platform for monitoring drug-induced toxicity in pre-clinical studies. Figure 2 with 1 supplement see all Download asset Open asset HBV-infected liver organoids as a model for HBV antiviral drug screening and toxicity. (A) Experimental design of drug treatment of hepatitis B virus (HBV)-infected liver organoids followed by assessment of antiviral activity and toxicity. Arrows indicate time points for HBV detection or assessment of viability. Levels of HBV DNA (orange) in the supernatant and intracellular HBV RNA (blue) (normalized to beta-2-microglobulin) were quantified by quantitative polymerase chain reaction (qPCR) and reverse transcription PCR (RT-PCR), respectively, for three independent healthy donors (B) and HepG2.2.15 cells (C) upon treatment with control vehicle, fialuridine (10 μM), or tenofovir (10 μM) as indicated. Data are shown as mean ± SD of at least three replicate treatments (paired two-tailed t-test); *p<0.05; **p<0.01; ***p<0.001. (D) Relative viability of HepG2 cells was measured using the alamarBlue cell viability assay after treatment with vehicle control, fialuridine, or tenofovir for 2 or 6 days as indicated, normalized to vehicle control, and plotted as the average of percent viability ± SD of at least four replicate treatments (paired two-tailed t-test) (ns = not significant). (E) Representative bright-field images taken of HepG2 cells treated with antiviral drugs for 2 or 6 days as indicated. (F) Bar diagrams representing relative cellular viability of healthy donor (hD) liver organoids after 2 or 6 days of treatment with fialuridine or tenofovir at the different concentrations indicated using the alamarBlue cell viability assay. All values are normalized to the vehicle-treated control and plotted as the average of percent viability ± SD of at least three replicate treatments (paired two-tailed t-test); *p<0.05; **p<0.01. The dotted line represents the lower limit of quantification based on values obtained from wells free of organoids containing the basement membrane matrix (BME) only. (G) Representative bright-field images taken of liver organoids treated for 2 or 6 days with the vehicle control or increasing concentrations of the antiviral drugs tenofovir or fialuridine as indicated. Figure 2—source data 1 Source data for Figure 2B. https://cdn.elifesciences.org/articles/60747/elife-60747-fig2-data1-v2.xlsx Download elife-60747-fig2-data1-v2.xlsx Figure 2—source data 2 Source data for Figure 2C. https://cdn.elifesciences.org/articles/60747/elife-60747-fig2-data2-v2.xlsx Download elife-60747-fig2-data2-v2.xlsx Figure 2—source data 3 Source data for Figure 2D. https://cdn.elifesciences.org/articles/60747/elife-60747-fig2-data3-v2.xlsx Download elife-60747-fig2-data3-v2.xlsx Figure 2—source data 4 Source data for Figure 2F. https://cdn.elifesciences.org/articles/60747/elife-60747-fig2-data4-v2.xlsx Download elife-60747-fig2-data4-v2.xlsx HBV replication can be investigated in transgenically modified liver organoids We observed higher HBV infection efficiency in DM organoids as compared to EM organoids (Figure 1C). This correlated with higher levels of sodium taurocholate co-transporting polypeptide (NTCP) expression, a cellular receptor expressed on the surface of hepatocytes implicated in HBV entry (Yan et al., 2012), in differentiated organoids as compared to organoids in EM (Figure 3A–B). HBV infection of liver organoids was dependent on NTCP as (pre)treatment with the competitive entry inhibitor myrcludex-B decreased infection of organoids as shown by quantitation of HBV intracellular RNA (Figure 3C) as well as HBV DNA (Figure 3D) and HBeAg (Figure 3E) in the supernatant of infected organoids. Exogenous expression of NTCP in hepatoma cell lines was shown to confer susceptibility to infection (Yan et al., 2012) in line with our observations of increased HBV infection in differentiated organoids, likely because of the higher level of NTCP expression. Since differentiated organoid cultures have a limited half-life, we sought to generate transgenically modified hD organoids exogenously expressing NTCP under expansion conditions (Figure 3F–H and Figure 3—figure supplement 1A) in order to improve infection efficiency and facilitate downstream analyses and investigation of the molecular events involved in HBV replication. We used a lentiviral construct harboring the coding sequence of Flag-tagged NTCP ubiquitously expressed under a Cytomegalovirus (CMV) promoter, followed by a blasticidin selection marker (Figure 3F). Immunofluorescence experiments performed on NTCP-liver organoids in the expansion phase confirmed high levels of NTCP protein expression correctly localized to the cellular membrane (Figure 3H and Figure 3—figure supplement 1B). Cholesterol target genes were induced in response to statin treatment in NTCP transgenic organoids, confirming the functionality of exogenously expressed NTCP (Figure 3—figure supplement 1B–C). We then viral production following HBV infection in transgenically modified NTCP organoid lines as compared to lines (Figure Interestingly, comparable levels of HBV DNA and were observed in the supernatants of both and organoid lines (Figure suggesting that expression of NTCP is not sufficient to improve HBV infection in liver organoids grown in EM (Figure and Figure 3—figure supplement 1D). Figure 3 with 2 supplements see all Download asset Open asset transgenic liver organoids in the study of HBV. (A) expression levels of the hepatitis B virus (HBV) receptor sodium taurocholate co-transporting polypeptide (NTCP) in (EM) and differentiated (DM) organoids = levels were according to the method using as the (B) Immunofluorescent staining showing the expression of NTCP in EM and DM organoids. were with Scale bars represent 50 µm. of HBV infection of liver organoids by (pre)treatment with myrcludex-B (10 or differentiated liver organoids were infected with HBV, and intracellular HBV RNA of HBV DNA genome in the culture supernatant and HBV early antigen (HBeAg) (E) produced in the culture supernatant were days after bars represent mean ± SD from three independent donors two-tailed t-test); (F) Schematic of the for the transduction experiments. infection with a lentiviral expressing organoids were with blasticidin for days in order to obtain lines expressing NTCP in the expansion (G) Levels of expression of NTCP were by reverse transcription polymerase chain reaction in the and the (NTCP) lines. Expression of NTCP was according to the method using the gene as the gene and confirmed by immunofluorescence staining NTCP or (red) (H) HBV surface antigen (HBsAg) released in the supernatant of and NTCP organoid lines grown in EM or DM days after HBV infection was detected by enzyme-linked immunosorbent assay Challenge with heat-inactivated virus was used to control for present in the inoculum. and neg bars correspond to positive and negative controls provided by the kit manufacturer. for was as the average of negative HBV DNA in the supernatant of organoid cultures was quantified days after infection and compared to DNA detected in the supernatant of HBV-infected organoids = 3). represent the in HBV DNA detected in the and HBV-infected organoids were used as the Relative of HBV DNA and produced by healthy donor (hD) organoid lines. Figure data 1 Source data for Figure Download Figure data 2 Source data for Figure Download Figure data 3 Source data for Figure Download To further the of liver organoids to be transgenically we produced a primary liver organoid model system that can be used to study HBV transcription We a lentiviral construct to produce transgenic organoid lines containing an integrated of HBV (Figure 3—figure supplement (Figure 3—figure supplement for the generation of hD transgenic organoid lines, in which transcription from the HBV in the production of viral proteins and (Figure Figure 3—figure supplement 2D). This model system provides a primary platform to for of HBV transcription and is a to the HepG2.2.15 cell lines that produce low levels of cccDNA et al., for studies into HBV Thus, the of human liver organoids to transgenic modification investigation of HBV replication and characterization of the molecular events of non-tumor HBV-infected patient-derived liver organoids The to generate a patient-derived primary model is a key of using the liver organoid We the previously characterized method to generate liver organoids from healthy donors (Huch et al., 2015) to generate novel patient-derived organoids from HBV-infected liver (Figure The used for patient-derived organoids was cirrhotic liver tissue obtained from from infected with HBV (Figure and We generated and expanded organoid cultures from and tissue from all donors with similar efficiency not Figure 4 with 2 supplements see all Download asset Open asset of organoid cultures from liver explants of HBV-infected patients. (A) Representative showing the to generate organoid cultures or from liver (B) of liver tissue and showing the of liver organoids derived from hepatitis B virus (HBV)-infected (C) Expression of the and in EM organoids derived from liver of healthy donors (hD) = and HBV-infected = Levels of expression were according to the method using as the (D) of organoid cultures derived from liver of = and = represent the in the expression of genes cytochrome and sodium taurocholate co-transporting polypeptide and the gene in DM cultures compared to EM organoids using the (E) Immunofluorescent staining albumin (green) and (red) was performed in EM and DM organoids.