Hepatitis D virus infection of stem cell-derived hepatocytes triggers an IFN- and NFκB-based innate immune response unable to clear infection

Human pluripotent stem cell-derived hepatocyte-like cells (HLCs) are a valuable model to investigate host-pathogen interactions of hepatitis viruses in a mature and authentic environment. Here, we investigated the susceptibility of HLCs to the Hepatitis D Virus (HDV), a virus that in co-infection with HBV is responsible for the most severe form of viral hepatitis. Cells undergoing hepatic differentiation became susceptible to HDV infection after acquiring expression of the Na+-taurocholate cotransporting polypeptide (NTCP), the receptor mediating HBV and HDV entry. Inoculation of mature HLCs with HDV lead to increasing amounts of intracellular HDV RNA and accumulation of the HDV antigen in the cells. The infection was abrogated when using known entry inhibitors targeting NTCP or by disrupting genome replication using the nucleoside analogue Ribavirin. Upon infection, the HLCs mounted an innate immune response based on induction of the interferons IFNB and L, but not IFNA, and were associated with an upregulation of interferon-stimulated genes. The intensity of this immune response positively correlated with the level of viral replication and was dependant on both the JAK/STAT and NFκB pathway activation. Importantly, neither this innate immune response nor an exogenous treatment of IFNα2b inhibited HDV replication. However, pre-treatment of the HLCs with IFNα2b reduced viral infection, suggesting that ISGs may limit early stages of infection. This novel HDV in vitro infection model represents a valuable tool for studying HDV replication and investigating candidate antiviral drugs in cells displaying mature hepatic functions. Lay summary HDV can infect stem cell-derived hepatocytes through an NTCP-mediated entry process. Infection triggers an IFN and NFκB dependent innate immune response. However, viral replication seems unaffected by this innate response or by exogenous IFN treatment. Graphical abstract

intensity of this immune response positively correlated with the level of viral replication and 48 was dependant on both the JAK/STAT and NFB pathway activation. Importantly, neither this 49 innate immune response nor an exogenous treatment of IFN2b inhibited HDV replication. 50 However, pre-treatment of the HLCs with IFN2b reduced viral infection, suggesting that ISGs 51 may limit early stages of infection. 52 This novel HDV in vitro infection model represents a valuable tool for studying HDV replication 53 and investigating candidate antiviral drugs in cells displaying mature hepatic functions. 54 Lay summary: HDV can infect stem cell-derived hepatocytes through an NTCP-mediated entry 55 process. Infection triggers an IFN and NFB dependent innate immune response. However, 56 viral replication seems unaffected by this innate response or by exogenous IFN treatment. (IFN) based immune response that is able to clear the infection. 10  HepaRG. 17 PHHs have previously been demonstrated to produce a strong induction of 99 interferons (IFNs) and interferon stimulated genes (ISGs) when infected with HDV, but only

126
Detailed material and methods can be found in the Supplementary materials.

128
Hepatic differentiation and stage dependent susceptibility to HDV. 129 HLCs were differentiated as previously described. 24 Briefly, hPSCs (D0) are subjected to 130 sequential growth factors and small drugs treatment to successively differentiate into 131 definitive Endoderm (D4), Hepatoblasts (D8, D12), and mature HLCs (D15) (Fig 1A). RNAseq  Fig 1D). HLCs expressed higher level 144 of NTCP than PHHs and Huh7 NTCP (Fig 1D), and NTCP protein was detected in almost every 145 HLCs ( Fig 1E). Altogether, our data show that cells undergoing in vitro hepatic differentiation 146 become susceptible to HDV infection when inducting the expression of the entry factor NTCP. 148 We inoculated HLCs with infectious HDV at a multiplicity of infection (MOI) of 0.5, for 6 hours 149 in presence of 4% Polyethylene glycol. In the three following days, we could detect increasing 150 amounts of intracellular HDV RNA by RTqPCR (Fig 2A), consistent with an efficient intracellular Strategies to improve HLCs susceptibility to HDV 166 Overall, inoculation of HLCs at MOI 0.5 led to a low percentage of detectable infected cells 167 (estimated around 2%). HLCs are less permissive to infection than Huh7 NTCP , as visualized by 168 HDAg staining (Sup Fig 2B) and RTqPCR for intracellular HDV RNA (Sup Fig 2C). The majority of 169 HLCs express NTCP (Fig 1E), but the receptor may not be accessible to the virus due to cell  Fig 3A). 31 However, the pre-treatment did not significantly enhance 174 intracellular HDV RNA amounts (Sup Fig 3B) and HDAg positive cells were only increased by a 175 slight percentage (Sup Fig 3C). 176 The limited number of HDAg positive cells may also be due to inhibition of viral replication 177 within the cells. Farnesylation of the large isoform of the HDAg has been described to inhibit 178 viral genome replication. 32,33 Inhibiting HDAg farnesylation using Lonafarnib improves HDV 179 genome replication in hepatoma cell lines. 34 In our HLCs, Lonafarnib treatment increased both 180 intracellular HDV RNA amounts (Sup Fig 4A) and percentage of HDAg positive cells detected 181 by IFA significantly (Sup Fig 4B). Remarkably, a Lonafarnib treatment led to a higher HDAg was significantly upregulated as opposed to IFNA mRNA (Fig 2D). Notably, type III IFNL mRNA 193 was strongly over expressed in the first dpi, and then decreased over days. Canonical ISGs MX1 194 (MXA), ISG15 and IFIH1 (MDA5) were induced throughout the experimental time line (Fig 2D). 195 Levels of induction of IFNs and ISGs correlated with the level of intracellular HDV RNA, 196 particularly at d3 and d5pi (Sup Fig 5). Myrcludex, CsA or Ribavirin treatment prevented the 197 induction of these immune genes (Fig 2E), confirming that the triggering of the HLCs' immune 198 response is dependent upon authentic HDV infection and viral genome replication. 199 Interestingly, Lonafarnib treatment of HDV infected HLCs, which allowed higher levels of viral 200 genome replication (Sup Fig 4A), led to increased levels of innate immune activation (Sup Fig   201   4C), further confirming that the innate immune response observed is triggered by the viral 202 replication. 203 Our results show that HDV infection triggers an innate immune response dependent on and 204 correlating with the level of viral replication.  Fig 6C). In order to 214 assess the role of both IFN and NFB pathways upon HDV infection, we treated the HLCs with 215 specific inhibitors. Ruxolitinib (Fig 3, green), a specific JAK1 and 2 inhibitor blocking the IFN 216 pathway, 10 and ML120B (Fig 3, blue), a specific IKK inhibitor blocking the NFB pathway, 23 217 both abrogated the induction of IFN and canonical ISG in HDV infected HLCs (Fig 3A). However, 218 none of these treatments showed a positive effect on intracellular HDV RNA (Fig 3B) and HDAg 219 expression (Fig 3C). Induction of innate immune response could also be inhibited by 220 transfecting the cells with siRNAs targeting STAT1 and STAT2 (Fig 4A, B), but had no significant 221 effect on viral replication (Fig 4C). HDV infection of the hepatoma cell line Huh7 NTCP did not 222 induce a significant innate immune response; and treatment with inhibitors or Lonafarnib did 223 not change the level of viral genome replication in this model (Sup Fig 7A, B). It has been 224 hypothesized that the HDV ribozyme may activate PKR. 36 In order to assess the role of PKR in 225 infected HLCs, we treated the cells with the oxindole/imidazole derivative C16 PKR inhibitor. 37 226 C16 treatment had no effect on the induction of the innate immune response (Sup Fig 8A) and 227 did not modify the level of viral replication (Sup Fig 8B). 228 Altogether, our results show that upon HDV infection, immune competent HLCs mount an IFN 229 and NFB-dependent, PKR-independent innate immune response. However, this antiviral 230 immune response is not able to restrict viral replication.

232
To explore the antiviral activity of exogenous IFNs, we treated HLCs, which were inoculated 3 233 days earlier with HDV, with 1000 IU/ml of IFN2B (Fig 5A, B). 2 days post IFN treatment, we 234 confirmed that IFN2B could induce the expression of ISGs in naïve HLCs (Fig 5B, blue) and 235 could slightly increase the induction of ISGs in HDV infected HLCs (Fig 5B, purple vs red). 236 However, IFN2b treatment showed no antiviral effect (Fig 5A, red vs purple). 237 Pre-treating HLCs with 1000 IU/ml of IFN2b 16 hours before inoculation with HDV allowed a 238 mild viral restriction effect (Fig 5C, red vs purple). A pre-treatment with IFN showed a higher 239 level of induction of ISGs and IFNs than an HDV infection (Fig 5D, blue vs red). Moreover, the 240 IFN2b-induction and the HDV-mediated immune activation were not additive (Fig 5C, blue   241 vs purple). This observed viral restriction effect was not associated with a down regulation of 242 NTCP, as assessed by IFA and RTqPCR (Fig 5E). Interestingly, IFN2b exhibited a dose-243 dependent effect on the induction of canonical ISG MX1 but the observed antiviral effect could 244 not be increased when using a higher dosage (Fig 5F). Collectively, our data show that IFN2b 245 is not able to inhibit viral replication in immunocompetent infected HLCs. However, IFN2b 246 treatment renders naïve cells more resistant to viral infection.  Fig 9A, B), but no significant effect was seen on induction of IFNs and 274 ISGs (Sup Fig 9C). Here, we could decipher some aspects of the innate immune response after 275 HDV inoculation. We describe that upon HDV infection, the induction of innate immune genes 276 in HLCs is not only under the control of the JAK/STAT pathway, similar to after infection by 277 other hepatotropic RNA viruses like HCV; 10 but also relies upon the activation of the NFB 278 signalling cascade (Fig 3). In HepaRG cells, HDV has been shown to induce NFB driven genes. 23

279
Similarly, an HDV infection of HLCs transiently elevates the expression of NFB driven genes 280 (Sup Fig 6C), and abrogation of the NFB signalling cascade critically hampers the induction of 281 the IFN-based innate immune response (Fig 3). 282 Moreover, we show that viral replication is not affected by the triggered innate immune 283 response: Abrogating either JAK/STAT or NFB pathway efficiently blocks ISGs induction but 284 did not increase viral replication (Fig 3). Recently, inducers of the NFB pathway have been 285 shown to reduce HDV replication in immune competent PHHs and HepaRG. 41 The HLCs 286 monoinfection model offers the ideal platform to confirm if increased NFB activation would 287 lead to viral inhibition and whether it is dependent on ISGs induction. 288 Importantly, we show that exogenous IFN2b treatment had no effect on established HDV 289 infection of immunocompetent HLCs (Fig 5A). These findings are consistent with clinical 290 observations that IFN treatment alone rarely leads to viral clearance (reviewed in 17 ). 291 Interestingly, when pre-treating with IFN2b, the HDV infection of HLCs was partially 292 restricted ( Fig 5C). This suggests that IFN2b treatment may render naïve hepatocytes less