Human pluripotent stem cell-derived hepatocyte-like cells for hepatitis D virus studies

  1. Huanting Chi
  2. Bingqian Qu
  3. Angga Prawira
  4. Lars Maurer
  5. Jungen Hu
  6. Rebecca M. Fu
  7. Florian A. Lempp
  8. Zhenfeng Zhang
  9. Dirk Grimm
  10. Xianfang Wu
  11. Stephan Urban
  12. Viet Loan Dao Thi

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Current culture systems available for studying hepatitis D virus (HDV) are suboptimal. In this study, we demonstrate that hepatocyte-like cells (HLCs) derived from human pluripotent stem cells (hPSCs) are fully permissive to HDV infection across various tested genotypes. When co- infected with the helper hepatitis B virus (HBV) or transduced to express the HBV envelope protein HBsAg, HLCs effectively secrete infectious progeny virions. We also show that HLCs expressing HBsAg support extracellular spread of HDV, thus providing a valuable platform for testing available anti-HDV regimens. By challenging the cells along the differentiation with HDV infection, we have identified CD63 as a potential HDV/HBV co-entry factor, which was rate-limiting HDV infection in immature hepatocytes. Given their renewable source and the potential to derive hPSCs from individual patients, we propose HLCs as a promising model for investigating HDV biology. Our findings offer new insights into HDV infection and expand the repertoire of research tools available for the development of therapeutic interventions.

Teaser

A human stem cell-derived hepatocyte culture model for hepatitis D virus studies

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    1. General Statements We thank the Editors and the Reviewers for their time and constructive criticism, which has allowed us to improve our manuscript. All of our responses are indicated in blue font. Revision Figures for the Reviewers are included just below the response. The line numbers given here refer to those in the revised manuscript, where we have marked the changes in red.
    2. Description of the planned revisions If granted a full revision, we will experimentally address the following major points, which were raised by more than one Reviewer: ● Repeat experiment in Figure 4 C to assess statistical significance (Reviewer 1 and 3) ● Western blot analysis of HDV infected HLCs showing small and large delta antigens. We have already performed such an analysis on HLCs (see Revision Figure 2). In addition, we will perform a comparative analysis with common HDV infection models dHepaRG and Huh7-NTCP cells over time (Reviewers 2 and 3). ● Additional characterisation of the two HLC subpopulations at transcript and protein level (Reviewer 1 and 3). In addition, we planned to conduct the following experiments in response to the individual Reviewers: In response to Reviewer 1: We thank the Reviewer for their encouraging feedback on our model and for their helpful comments, allowing us to improve our manuscript. Figure 1: The observation of a denser subpopulation of hepatocytes more susceptible to HDV is interesting. Do you have more characterization of this cell subpopulation, by IFA, in term of hepatic maturation marker, known HDV host factors and particularly NTCP expression? We agree with the Reviewer that this is an interesting observation. We separated the two hepatocyte subpopulations to analyse the gene expression of the liver maturation markers NTCP and ALB by RT-qPCR (see Revision Figure 1A). Surprisingly, we found that the low-density population expressed higher levels of both ALB and NTCP, suggesting that they are more mature than the high-density population. In addition, we stained both markers by immunofluorescence and observed no apparent differences (see Revision Figures 1B & C). In contrast, the new host factor identified in our study, CD63, appeared to be more highly expressed in the high-density population compared to the low-density population (Fig. 6G). However, we cannot exclude the Revision Plan possibility that other factors play an additional role. As outlined in our response to Reviewer 3, we will separate the two populations and analyse the gene expression of other known HBV and HDV co-host factors to assess whether they play a role in addition to CD63 in conferring the higher susceptibility to HDV infection to the highly dense HLC population. Revision Figure 1: High-density HLCs population is not more mature than the low-density HLC population. (A) The low-density HLCs population was separated from the high-density HLC population by gentle dissociation. Total RNAs were isolated from both populations and Albumin and NTCP expression was analysed by RT-qPCR. (B & C) High-density HLCs (upper image) and low-density HLCs (bottom image) were stained with Albumin specific antibody. Shown are either images taken on an epifluorescence microscope (B) or single slices of confocal images acquired on a Airyscan confocal microscope (C). Fig 1B and C: Can a BLV control be included in the figure? Thank you for this suggestion, we will repeat the experiment for these panels and add BLV as control. Fig 1A-F: What is the overall level of NTCP between HLC, HepaRG, Huh7NTCP and HLCAAV- NTCP? Can NTCP and HDAg be stained simultaneously in your cells? This is an excellent question and we will compare the total NTCP levels between differentiated HepaRG, Huh7 NTCP, HLCs +/- AAV NTCP by Western blot analysis and immunofluorescence (IF) staining. Comparing NTCP expression in HLCs +/- AAV NTCP, we observed a strong upregulation of surface NTCP upon AAV transduction by IF staining (Figure 1D). Unfortunately, our initial attempts to simultaneously detect NTCP and HDAg were technically hampered. Since HDAg is mainly localised in the nuclei, we have to permeabilize the cells in a harsh manner, which interferes with the detection of membrane NTCP. The latter is further hampered by the availability of suitable anti-NCTP antibodies for IF staining. In our study, we used high doses of fluorescence-conjugated MyrB peptide to stain NTCP, but unfortunately it is very sensitive to the harsh permeabilization detergents mentioned above. However, since we have meanwhile optimised HDV infection, we will likewise try again to optimise the staining Revision Plan protocol. If we succeed, we will repeat the co-staining of NTCP and HDAg and include it in a revised manuscript. Figure 4: While the strategy is interesting, based on what has been previously shown for HCV in Wu et al., 2012, the lack of statistical data prevents the reader to really understand and see drastic difference in term of susceptibility to infection and level of expression of host genes. In panel C, is the difference between day 13 and 15 statistically significant? Same for panel D, day 17 vs 19?As a remark, day 19, the peak of susceptibility to HDV, seems to be also the peak of maturation, based on ALB RTqPCR (panel B). Thank you for this comment, and will perform another set of experiments allowing us to calculate statistical significance. The Reviewer correctly points out the correlation between HDV infection and hepatocyte maturity, which we find very intriguing. To identify potential host co- or restriction factors expressed in highly mature HLCs, we then performed the differential gene expression analysis (Figure 5). As shown in the new Figure 5A, GO analysis revealed that genes involved in pathways regulating viral entry into host cells were most significantly upregulated in mature HLCs and, as a probable consequence, they were more permissive to HDV infection. Indeed, among these factors, we identified CD63 as a novel host cofactor that renders mature HLCs susceptible to HDV infection (Figure 6). In response to Reviewer 2: We thank the Reviewer for their assessment of our study and for critically pointing out the increments over the previous study by Lange et al. We also appreciate their helpful suggestions, which allow us to improve the manuscript. The manuscript would benefit from a more detailed virological analysis, such as: •Determination of HDV genome and antigenome sequences and analysis of HDV editing. We thank the Reviewer for this comment. Accordingly, we will determine HDV genomes and antigenomes by Northern blot analysis and study HDV editing rates by sequencing in HDVinfected HLCs. •Analysis of HDV short and large antigens by western blot. We have already detected small and large HDAg in HDV-infected HLCs (see Revision Figure 2). To also satisfy Reviewer 3, we will additionally compare the S/L-HDAg ratios over time in HLCs, dHepaRGs, and Huh7-NTCP cells and include the results in a revised manuscript. Revision Figure 2: Detection of small and large delta antigen in HDV-infected HLCs. Mature HLCs were infected with HDV (MOI= 5 Int. Units/cell) and harvested 1 or 3 days post-infection. Cell lysates were analysed by Western blotting using antibodies against HDAg and b-actin. Revision Plan •Analysis of HBV-related virological parameters in monoinfected and co-infected cells. We agree with the Reviewer and we will include the characterisation of more HBV-related virological parameters in our mono- and co-infected HLCs. Accordingly, we will assess HBV cccDNA, RNA, and DNA by RT-qPCR, as well as released HBsAg and HBeAg via ELISA and add the results to the revised manuscript. In response to Reviewer 3: We thank the Reviewer for their positive evaluation, and we acknowledge their helpful comments, which will help us to improve our manuscript. Line 143: the authors describe two forms of HLCs (less and more confluent with differences regarding the susceptibility to HDV infection). The characteristics of the less and more confluent HLCs should be described in more detail-what is causative for the differences in susceptibility for HDV infection of these two forms? We thank the Reviewer for this comment. We likewise find this observation intriguing. As stated in our response to Reviewer 1, we have ruled out that NTCP and/or other mature markers such as ALB are differentially expressed between the two subpopulations. As one factor that could make a difference, we have identified CD63, which is highly expressed in the high-density HLC population and less so in the low-density HLC population (Figure 6G). Nevertheless, we will separate the two populations and analyse by RT-qPCR the expression of other known HBV and HDV host co-factors that may be additional factors governing the increased susceptibility of the highly dense HLC population. The statistical analyses should be improved: There are no p-values provided for the data presented in the supplement and a variety of figures lacks p-values We have added p-values to the Supplementary Figures (see revised Supplementary Fig. S2) and will repeat the experiments for Fig. 4 and Supplementary Fig. S1B and Fig S3 so that we can calculate the corresponding p-values. Kinetic of the infection: Here it would be interesting to see a comparative analysis by western blot investigating the ratio HBsAg/HDAg over the time in HLCs, HepaRGs and NTCP oe cells We thank the Reviewer for his comments. As stated in our response to Reviewer 2, we will perform this WB analysis to detect S/L-HDAg over time in infected HLCs, dHepaRG, and Huh7- NTCP cells. Line 157: What is the experimental evidence for the proper localization and functionality of the ectopically expressed NTCP in HLCs. Did the authors study the taurocholate transport after overexpression of NTCP? We thank the Reviewer for this comment. We analysed endogenous and ectopic NTCP expression by microscopy using a fluorescently conjugated peptide Atto-MyrB-565, which specifically binds to the ectodomain of human NTCP (Figure 2D) and found that both Revision Plan endogenously and ectopically expressed NTCP are located on the cell surface. To further confirm the correct localisation, we will perform NTCP co-staining with a cell membrane marker. We will also test the proper function of the ectopically expressed NTCP using a specific taurocholate transport assay as shown in our previous study (Ni et al, 2014, Gastroenterology). Line 169: The authors should include data comparing the number of double positive cells in HLCs, HepaRGs and Huh7NTCP o.e. expressing cells under the chosen experimental conditions We thank the Reviewer for this suggestion. We have already performed HBV/HDV co-infection of dHepaRG cells (Revision Figure 3) and we will perform the same experiment with Huh7-NTCP cells. Revision Figure 3: HBV/HDV co-infection of dHepaRG cells. Differentiated HepaRG were infected with HBV (MOI = 450 genome copies/cell) and HDV (MOI = 5). Cells were stained against HBV core (HBc), HDAg, and nuclei (DAPI) ten days p.i.. HBc- and HDAg-positive cells were counted using Cellprofiler imaging software to quantify HBV (pink) and HDV (green) single and co-infection (white) events. Images are representative of three independent differentiations. Line 291: expression analysis by RT-PCR is not sufficient. It will be important to study by CLSM if the identified factors are really present as proteins and properly localized. To satisfy this Reviewer, we will be happy to perform WB analysis of lysates from cells obtained at different stages of HLC differentiation to detect LDLR, LAMP1 and SR-B1 to further confirm our transcriptome analysis. As protein expression is easier to compare by WB analysis, we prefer this method to microscopic analysis. Regarding the role of CD63: what is the evidence for a direct role of CD63 for HDV entrycan the authors exclude that CD63 is relevant for targeting other factors to the surface? What is the impact of loss of CD63 on the functionality of the autophagosomal-MVB-EV system in HLCs? Since downregulation of CD63 before but not after impairs HDV infection, we conclude that CD63 is likely to be important for the early steps of the HDV life cycle, namely cell entry. Indeed, we speculate that CD63 may be critical for HDV trafficking to the vesicle, where fusion of the HBV glycoproteins is induced to allow capsid entry, based on the following observations: Although neither the precise site of HBV viral fusion nor the cues that induce fusion are currently fully understood, studies suggest that HBV can be co-transported with EGFR and NTCP to late endosomes for trafficking (Herrscher et al; 2020, Cells). We speculate that similar to what has been described for Lujo virus, CD63 may be involved in either HDV trafficking and/or virus fusion in the endosomal system (late endosome or lysosome) (Tominaga et al, 2014 Molecular cancer). Revision Plan CD63 is a ubiquitously expressed protein that localises to the endosomal system and, in its glycosylated form, to the cell surface. Non-glycosylated CD63 is not properly trafficked and aggregates at the nuclear periphery instead of the cell membrane (Tominaga et al., 2014, Molecular Cancer). According to the Western blot analysis in Figure 6, immature HLCs appear to express less glycosylated CD63 than mature HLCs. We will confirm the glycosylation by treating the cell lysates with PNGase F. Although AAV transduction enhanced CD63 expression of all three HLC stages tested (see new Supplementary Figure S6 in the revised manuscript), it only enhanced HDV infection of immature HLCs, in which the non-glycosylated form of CD63 appears to be the predominant form. To demonstrate that the glycosylated form of CD63 is involved in HDV entry, we will rescue WT CD63 in parallel with a glycosylation-deficient CD63 mutant (Yoshida et al., 2009, Microbiology and Immunology) in immature HLCs. We will also stain CD63 in both immature and mature HLCs to compare the subcellular localisation (plasma membrane/endosomes vs. nuclear membrane) of CD63 between the two stages.
    3. Description of the revisions that have already been incorporated in the transferred manuscript Based on the constructive comments by the Reviewers we already made the following changes, which are highlighted in red in the revised manuscript. In response to Reviewer 1: Fig 1B-C: the comparison with dHepaRG is very interesting, and confirms the validity of SC derived hepatocytes as a model for HDV infection. dHepaRG can be heterogeneous. Do you also see the same phenotype of enriched HDV infection within a denser subpopulations of dHepaRG We thank the Reviewer for their comment. Undifferentiated bipotent HepaRG cells are not permissive for HDV infection due to the lack of surface NTCP expression. Due to their bipotent nature and upon differentiation, two morphologically distinct populations become apparent: hepatocyte-like cells and biliary epithelial-like cells (McGill et al., 2010, Hepatology). As shown in the Figure 1 of the study by Mesnage et al. (2018, Molecular Toxicology), dense hepatocyte-like colonies are surrounded by clear epithelial cells corresponding to primitive biliary cells. In agreement with other studies, we only observe that the ALB-positive hepatocyte-like cells are permissive to HBV and HDV infection (Hantz et al., 2009, Journal of General Virology), highlighting their specific hepatic tropism and the cellular determinants required. Fig 1I is confusing. Was BLV assay also performed on the HLC infection (Day 0), or only during the titration assay in Huh7NTCP? We apologise for the confusion in this panel. BLV was only added during the titration assay on Huh7NTCP cells to confirm new and productive infections and to rule out carry-over. We have changed the order of Figures 1I - 1K to make this clearer and explain this better in the new results section (line 171-179) and figure legend (line 797-806). Revision Plan Fig 1K: x-axis is confusing... is it number of HBV, HDV and HBV/HDV positive cells? Or number of infected cells upon inoculation with HBV, HDV, or both? Please clarify. We apologise for this additional confusion caused in this panel. We infected HLCs with both HBV and HDV simultaneously and then counted the number of positive cells that were either single infected with HBV (pink cells/column), single infected with HDV (green cells/column) or double infected with both viruses (white cells/column). We have clarified this in the revised Results section (line 172-176) and in the revised Figure Legend (line 798-803). Figure 2: The AAV based vector to over express HBsAg is a very interesting tool, and the figure convincingly show production of HDV progeny viruses in HLC-AAV-HBsAg. Results shown are in agreement with previous studies based on hepatoma cell lines. We thank the Reviewer for this positive comment and we agree that AAVs represent interesting tools to genetically manipulate HLCs and other hepatocyte culture systems. Figure 2B: What is IU/ml? Infectious Unit? International Unit? Are units in Fig 1B, 2B and 2C the same? We apologise for the lack of clarity. In Figures 1B and 2C, IU corresponds to infectious units of HDV, whereas in Figure 2B, IU corresponds to international units for the assessment of secreted HBSAg levels in the supernatant. To make the difference clearer, we have changed the unit on the y-axis in Figure 2B and explicitly stated the abbreviations in the corresponding revised Figure Legends (lines 785, 786, 794, 795, 816, and 819). Figure 3: What is the overall number of transmission events observed in the co-culture setup? Can you visually observed viral spreading? Panel A shows only 1 event, making it hard to assess its efficiency. Titration assay in Fig 2C show production of up to 4-5 log of infectious HDV. But HLCs susceptibility to HDV infection may change during time... Thank you for your comment and for raising this important issue. Panel A clearly and visually demonstrates that extracellular spread of HDV had occurred in the HLCs system, as initially only WT and non-GFP positive HLCs were infected with HDV. After co-culture, the progeny of WT HLCs were able to infect GFP-HLCs (Figure 3A). The overall efficiency of HDV spread/transmission in HLC efficiency is shown in Figure 3C. If we allow spread to occur (DMSOtreated condition), the total number of HDV-positive HLCs grown in a 24-well plate is approximately 1000. When we block secondary infection of progeny with BLV and thus spread, we count only about 500 HDV-positive HLCs in a well. In general, spreading in HLCs (Figure 3C) is not as efficient as retitration to Huh7-NTCP (Figure 2C) for the following reasons: In Figure 2C, we wanted to have an estimation of the maximum amount of secreted infectious progeny from HDV-producing HLCs. To this end, we did not want the re-infection itself to be a major bottleneck and used the most susceptible model Huh7-NTCP and infected them under the best conditions, which includes the addition of 4% PEG and 2% DMSO in the culture medium. For our spread assay in HLCs, we cannot add PEG to the cells over the course of the experiment and we also wanted to be as physiological as possible. PEG significantly enhances HDV infection Revision Plan of HLCs (Supplementary Fig. S2) and Huh7-NTCP cells (Revision Figure 4), which is in agreement with previous studies (Michailidis et al., 2017, Scientific reports). In addition, as the Reviewer correctly points out, similar to other primary hepatocyte culture models, the HLC system deteriorates over time. However, we have found that HLCs can be cultured for up to 3 weeks. Nevertheless, we believe that the efficiency of HDV spread in HLCs is sufficient for drug testing (Fig. 3C & D). Revision Figure 4: PEG enhances HDV infection of Huh7-NTCP cells. Huh7- NTCP cells were infected with HDV (MOI= 5 Int. Units/cell) in the absence or presence of PEG. Cells were harvested on D5 pi and HDV genome copies were quantified by RT-qPCR. Figure 5: In panel A, GO pathways should be sorted based on significance, not Number of genes. In panel B-D, what is the scale of the heatmap on figure 5: change in CPM values, however log2, log10? Thank you for this comment, we have sorted the GO pathways based on significance (new Figure 5A). For panels B-D, we did not calculate the fold change in CPM values and they were not log transformed. Instead, we calculated the z-scores of the genes shown by comparing the expression level of a given gene (in CPM) in a given sample with the expression level of that gene across all samples. To avoid further confusion, we have added "z-score" to the new Figure

    Figure 6: Do you have info about CD63 in other mature model, like dHepaRG and PHHs? Is CD63 also limiting in these models? Our data in Figure 6 suggest that CD63 may be a limiting factor for HDV infection of immature HLCs but not mature HLCs. Both dHepaRG cells and PHHs are mature hepatocyte models and therefore we speculate that CD63 is not rate limiting. However, we will investigate whether CD63 is rate-limiting in undifferentiated HepaRG cells. In response to Reviewer 2: Additional information that needs to be added, better explained, or corrected: The authors should explain why they used different MOIs depending on the genotype. In our previous study by Wang et al. 2021 J Hepatol, we found that the different HDV genotypes are heterogeneous in their ability to infect Huh7 NCTP cells. For example, as shown in Figure 4B of Wang et al. 2021 J Hepatol, GT 4 and 5 are less infectious than other genotypes. Based on the different infectious titres of the genotypes obtained on Huh7 NTCP cells, we then decided to use different MOIs for infection of our HLCs. The aim of the present study by Chi et al. was not to Revision Plan compare the different HDV genotypes, but to analyse whether they can all infect HLCs. In order to obtain similar infection efficiencies of our HLCs with the different genotypes, we used higher MOIs for those genotypes that were less infectious in Huh7-NTCP cells compared to those genotypes that were more infectious in Huh7-NCTP cells. We apologise for not making this sufficiently clear and have added this information to the results section (line 167-170) and the corresponding figure legend (line 796) of the revised manuscript. In Figure 1, it is unclear on which day the HCLs were infected by HDV and on which day they were transduced with AAV-NTCP. We apologise for the lack of clarity in the experimental design. We transduced HLCs with AAV two days before HDV infection to ensure sufficient ectopic NTCP expression on the day of HDV infection to study its effect on HDV entry. We have clarified this in the results section (line 153,

    1. and in the figure legend (line 788) in the revised manuscript. It is not very clear if the authors used AAV serotype 6 consistently to transduce the cells. It would be valuable to show the transduction efficiency of AAV at different time points of HLC maturation, as it might also be affected and could explain some results. For example, in Figure 6H, why does AAV-CD63 transduction increase HDV infectivity at day15 but not at day 10? It would be interesting to repeat the anti-CD63 western blot after AAV-CD63 transduction. Thank you for this comment. Yes, we have consistently used AAV 6 due to its relatively broad tissue tropism (Verdera et al., 2020, Molecular Therapy) and we have clarified this information in the revised manuscript (see line 331). We agree with the Reviewer's concerns regarding the variable transduction efficiency. We have previously tested different AAV capsids and found that AAV6 transduced mature HLCs at high levels (Zhang et al., 2022, Hepatol Commun). In this study, we also performed Western blot analysis to confirm successful CD63 overexpression by AAV transduction at different stages of hepatocyte differentiation. As shown in new Supplementare Figure 6, although there were some differences in transduction efficiency, the majority of all cells at each stage of differentiation were successfully transduced to ectopically express CD63. The authors claim that by using AAV to express HBsAg, they are mimicking the expression of HBsAg from the integrated sequence rather than cccDNA. However, it is the opposite, as AAV genomes, like cccDNA, remain as episomes in the cells. Yes, the Reviewer is conceptually correct and we apologise for the incorrect wording. In principle, we aim to trans-complement HBsAg in a setting outside of HBV infection and thus mimic the expression of antigen from integrated cells, although AAVs of course remain mostly episomal. We have clarified this in the revised manuscript (see lines 188 & 378). In response to Reviewer 3: Line 217: the complete inhibition of cell to cell spread by myrcludex suggests that there is no spread by cell-cell contact. This should be discussed. Revision Plan Yes, there is no evidence of HDV spread by cell-cell contact because, as the Reviewer correctly points out, BLV treatment almost completely blocked HDV de novo infection (Figure 2D & E). To our knowledge, cell-to-cell spread has not been demonstrated for HDV. According to our own studies by Zhang et al, 2021/2022, Journal of Hepatology, HDV spreads either extracellularly (which can be blocked by BLV) or by cell division (discussed in lines 362). Since HLC are similar to primary human hepatocytes and do not divide in vitro, we believe that extracellular spread is the predominant mode of spread in HLC (line 365). Line 210ff:Is there any evidence for syncytia formation in this system? No, we have not observed syncytia formation. Since HDV has no glycoproteins, we would not expect syncytia to form. Line 42: secrete should be replaced by release We thank the Reviewer for pointing out the inaccuracy in our terminology. We have replaced "secrete" with "release" (line 42). Line 241: proteins are not expressed, genes are expressed Thank you, we agree and changed the wording accordingly (line 246).
    1. Description of analyses that authors prefer not to carry out In response to Reviewer 1: Fig 1B: Unit is confusing, using terms usually used for titration of infectivity, from the virus input point of view, not from the cellular point of view. Can you use % infected cells instead, or "HDV infection rate" like in Supp Fig 1B? We apologise for this confusion. For other viruses, such as but not limited to HCV or HEV, the most common method is to report focus forming units per ml (FFU/ml). HLCs do not divide and, in the absence of HBV S antigen, no cell-division mediated HDV spread can occur and only single infection events can be observed (hence infectious unit = IU/ml). Since differentiated, authentic hepatocyte culture models such as PHHs, HLCs or HepaRG cells are always characterised by strong cell heterogeneity, it is difficult to directly compare the overall percentage of infection with a homogeneous cell population such as Huh7-NTCP cells. Therefore, if the Reviewer allows us, we prefer to keep this unit in our main figures. However, and hopefully to the satisfaction of this Reviewer, we have also calculated the percentage of infected cells of this exact dataset and show it in the Supplementary figures (Suppl. Fig. S1 C). The proportion of infection efficiency comparing HLCs, dHepaRGs, and Huh7-NTCP cells does not differ when presented either as IU/ml or as percentage of infected cells.
  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    In their manuscript entitled " An HBV/HDV Infection Model Using Human Pluripotent Stem 1 Cell-Derived Hepatocyte- Like Cells for Virus Host Interactions and Antiviral Evaluation" describe the use of HLCs derived from hPSCs as infection model for analysis of HDV life cycle. The ms is well written and clearly structured. It is easy to follow the concept of the study. The ms addresses a relevant topic and could help to overcome limitations in the analysis of HDV life cycle. The authors perform in many points a detailed characterization of this experimental system but there are still a variety of open points which must be addressed:

    Specific points:

    Line 143: the authors describe two forms of HLCs (less and more confluent with differences regarding the susceptibility to HDV infection). The characteristics of the less and more confluent HLCs should be described in more detail-what is causative fir the differences in susceptibitly for HDV infection of these two forms? The statistical analyses should be improved: There are no p-values provided for the data presented in the supplement and a variety of figures lacks p-values Kinetic of the infection: Here it would be interesting to see a comparative analysis by western blot investigating the ratio HBsAg/HDAg over the time in HLCs, HepaRGs and NTCP oe cells

    Line 157: What is the experimental evidence for the proper localization and functionality of the ectopically expressed NTCP in HLCs. Did the authors study the taurocholate transport after overexpression of NTCP?

    Line 169: The authors should include data comparing the number of double positive cells in HLs, HepaRGs and NTCP o.e. expressing cells under the chosen experimental conditions

    Line 217: the complete inhibition of cell to cell spread by myrcludex suggests that there is no spread by cell-cell contact. This should be discussed.

    Line 210ff:Is there any evidence for syncytia formation in this system?

    Line 291: expression analysis by RT-PCR is not sufficient. It will be important to study by CLSM if the identified factors are really present as proteins and properly localized. Regarding the role of CD63: what is the evidence for a direct role of CD63 for HDV entry-can the authors exclude that CD63 is relevant for targeting other factors to the surface? What is the impact of loss of CD63 on the functionality of the autophagosomal-MVB-EV system in HLCs?

    Minor points:

    Line 42: secrete should be replaced by release Line 241: proteins are not expressed, genes are expressed

    Significance

    The manuscript describes the use of HLCs derived from hPSCs as infection model for analysis of HDV life cycle. The ms is well written and clearly structured. It is easy to follow the concept of the study.

    The ms addresses a relevant topic and could help to overcome limitations in the analysis of HDV life cycle. The authors perform in many points a detailed characterization of this experimental system but there are still a variety of open points which must be addressed.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    In the present work, Chi et al. demonstrated that Hepatocyte-like cells (HCLs) derived from human pluripotent cells (hPCs) can be infected by HBV. The development of new HDV cellular models is of great value for understanding HDV biology and developing new treatments. However, the relevance of the present work is limited by a recent publication by Lange et al., in which they also showed that HCLs derived from hPCs can be infected by HDV, inducing the activation of the innate immune response, as previously demonstrated in cells and mice.

    The authors added new information to the work of Lange et al, including:

    • HLCs derived from human pluripotent cells can be infected by different HDV genotypes.
    • They proved that infectious HDV particles are formed.
    • They identified CD63 as a potential HDV coreceptor.

    The manuscript would benefit from a more detailed virological analysis, such as:

    • Determination of HDV genome and antigenome sequences and analysis of HDV editing.
    • Analysis of HDV short and large antigens by western blot.
    • Analysis of HBV-related virological parameters in monoinfected and co-infected cells.

    Additional information that needs to be added, better explained, or corrected:

    The authors should explain why they used different MOIs depending on the genotype.

    In Figure 1, it is unclear on which day the HCLs were infected by HDV and on which day they were transduced with AAV-NTCP.

    It is not very clear if the authors used AAV serotype 6 consistently to transduce the cells. It would be valuable to show the transduction efficiency of AAV at different time points of HLC maturation, as it might also be affected and could explain some results.

    For example, in Figure 6H, why does AAV-CD63 transduction increase HDV infectivity at day 15 but not at day 10? It would be interesting to repeat the anti-CD63 western blot after AAV-CD63 transduction.

    The authors claim that by using AAV to express HBsAg, they are mimicking the expression of HBsAg from the integrated sequence rather than cccDNA. However, it is the opposite, as AAV genomes, like cccDNA, remain as episomes in the cells.

    Significance

    In the present work, Chi et al. demonstrated that Hepatocyte-like cells (HCLs) derived from human pluripotent cells (hPCs) can be infected by HBV. The development of new HDV cellular models is of great value for understanding HDV biology and developing new treatments. However, the relevance of the present work is limited by a recent publication by Lange et al., in which they also showed that HCLs derived from hPCs can be infected by HDV. The authors added new information to the work of Lange et al, including:

    • HLCs derived from human pluripotent cells can be infected by different HDV genotypes.
    • They proved that infectious HDV particles are formed.
    • They identified CD63 as a potential HDV coreceptor.
  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary: In the present manuscript, H Chi et al describe the infection of stem cell derived hepatocytes with HBV and HDV. They suggest that it could be used to validate antiviral treatment in a mature hepatocyte model. Moreover, they take advantage of the differentiation process of the cells to identify time points correlating with significant change in viral permissivity, and focus on one of these time points in an attempt to identify new host factors of HDV.

    Overall, the manuscript is well written and brings interesting information toward the establishement of an efficient HBV HDV coinfection model in stem cell derived hepatocytes. Particularly, comparison to dHepaRG, another model relying on in vitro differentiation and commonly used to study HBV and HDV infection, reveals the potential of stem cell derived hepatocytes. While the efficiency of co infection in the stem cell derived hepatocytes may seem low, the manuscript goes in the direction of helping establishing a new mature in vitro model of infection.

    Figure 1: The observation of a denser subpopulation of hepatocytes more susceptible to HDV is interesting. Do you have more characterization of this cell subpopulation, by IFA, in term of hepatic maturation marker, known HDV host factors and particularly NTCP expression?

    Fig 1B-C: the comparison with dHepaRG is very interesting, and confirms the validity of SC derived hepatocytes as a model for HDV infection. dHepaRG can be heterogeneous. Do you also see the same phenotype of enriched HDV infection within a denser subpopulations of dHepaRG?

    Fig 1B: Unit is confusing, using terms usually used for titration of infectivity, from the virus input point of view, not from the cellular point of view. Can you use % infected cells instead, or "HDV infection rate" like in Supp Fig 1B?

    Fig 1B and C: Can a BLV control be included in the figure?

    Fig 1A-F: What is the overall level of NTCP between HLC, HepaRG, Huh7NTCP and HLC-AAV-NTCP? Can NTCP and HDAg be stained simultaneously in your cells?

    Fig 1I is confusing. Was BLV assay also performed on the HLC infection (Day 0), or only during the titration assay in Huh7NTCP?

    Fig 1K: x-axis is confusing... is it number of HBV, HDV and HBV/HDV positive cells? Or number of infected cells upon inoculation with HBV, HDV, or both? Please clarify.

    Figure 2: The AAV based vector to over express HBsAg is a very interesting tool, and the figure convincingly show production of HDV progeny viruses in HLC-AAV-HBsAg. Results shown are in agreement with previous studies based on hepatoma cell lines.

    Figure 2B: What is IU/ml? Infectious Unit? International Unit? Are units in Fig 1B, 2B and 2C the same?

    Figure 3: What is the overall number of transmission events observed in the co-culture setup? Can you visually observed viral spreading? Panel A shows only 1 event, making it hard to assess its efficiency. Titration assay in Fig 2C show production of up to 4-5 log of infectious HDV. But HLCs susceptibility to HDV infection may change during time...

    Figure 4: While the strategy is interesting, based on what has been previously shown for HCV in Wu et al., 2012, the lack of statistical data prevents the reader to really understand and see drastic difference in term of susceptibility to infection and level of expression of host genes. In panel C, is the difference between day 13 and 15 statistically significant? Same for panel D, day 17 vs 19? As a remark, day 19, the peak of susceptibility to HDV, seems to be also the peak of maturation, based on ALB RTqPCR (panel B).

    Figure 5: In panel A, GO pathways should be sorted based on significance, not Number of genes. In panel B-D, what is the scale of the heatmap on figure 5: change in CPM values, however log2, log10?

    Figure 6: Do you have info about CD63 in other mature model, like dHepaRG and PHHs? Is CD63 also limiting in these models?

    Significance

    Overall, the manuscript brings interesting information toward the establishement of an efficient HBV HDV coinfection model in stem cell derived hepatocytes. Particularly, comparison to dHepaRG, another model relying on in vitro differentiation and commonly used to study HBV and HDV infection, reveals the potential of stem cell derived hepatocytes. While the efficiency of co infection in the stem cell derived hepatocytes may seem low, the manuscript goes in the direction of helping establishing a critical needed and long awaited mature in vitro model of HDV HBV infection.

  5. Version published to 10.1101/2023.10.13.561984v1 on bioRxiv