Viral Interference of Hepatitis C and E Virus Replication in Novel Experimental Co-Infection Systems

Background: Hepatitis C virus (HCV) constitutes a global health problem, while hepatitis E virus (HEV) is the major cause of acute viral hepatitis globally. HCV/HEV co-infections have been poorly characterized, as they are hampered by the lack of robust HEV cell culture systems. This study developed experimental models to study HCV/HEV co-infections and investigate viral interference in cells and humanized mice. Methods: We used state-of-the art human hepatocytes tissue culture models to assess HEV and HCV replication in co- or super-transfection settings. Findings were confirmed by co- and super-infection experiments in human hepatocytes and in vivo in human liver chimeric mice. Results: HEV was inhibited by concurrent HCV replication in human hepatocytes. This exclusion phenotype was linked to the protease activity of HCV. These findings were corroborated by the fact that in HEV on HCV super-infected mice, HEV viral loads were reduced in individual mice. Similarly, HCV on HEV super-infected mice showed reduced HCV viral loads. Conclusion: Direct interference of both viruses with HCV NS3/4A as the determinant was observed. In vivo, we detected reduced replication of both viruses after super-infection in individual mice. These findings provide new insights into the pathogenesis of HCV-HEV co-infections and should contribute to its clinical management in the future.


Introduction
Hepatitis E virus (HEV) is a member of the Orthohepevirus genus within the family of Hepeviridae. The positive-sense RNA virus is an understudied pathogen, which accounts for (GLuc) [28] and GFP replicon [8], the subgenomic Sar55 S17 GLuc replicon [29], the subgenomic HEV83-2-27 GLuc replicon [30] (kind gift of the laboratory of Takaji Wakita) were all previously described. The Kernow-C1 p6 GFP-Neo plasmid was cloned from a Kernow C1 p6 GLuc-Neo template (kindly provided by Viet Loan Dao Thi) by replacing the GLuc gene with a GFP.

Compounds and Reagents
Ribavirin was received from Sigma Aldrich (St. Louis, MO, USA). Sofosbuvir and telaprevir were purchased from MedChemExpress (Monmouth Junction, NJ, USA). 2 C methyladenosine (2 -CMA) was a gift from Timothy Tellinghuisen (The Scripps, Jupiter, FL, USA). All compounds were diluted in DMSO and stored at −80 • C.

In Vitro Transcription and Electroporation
All HCV containing plasmids as well as HEV Kernow-C1 p6 were linearized using MluI (New England Biolabs, Fankfurt a. M., Germany), HEV 83-2-containing plasmids were linearized with HindIII (New England Biolabs), and Sar55-containing plasmid with EcoRV (New England Biolabs). HCV-based plasmids were in vitro transcribed as indicated [31]. HEV-based plasmids were transcribed as indicated [8]. For transfection we used the electroporation technique in accordance to the previous reports [28]. In brief, 5 × 10 6 Huh7.5 cells in 400 µL cytomix containing 2 mM ATP and 5 mM glutathione were mixed with a total of 5 µg of RNA. Electroporation was carried out with a Gene Pulser system (Bio-Rad, Munich, Germany). Cells were immediately transferred to 12.1 mL of DMEM complete and the cell suspension was seeded in respective plates depending on the experiment (2 × 10 4 cells/well seeded in a 96-well plate for luciferase assays, 2 × 10 5 cells/well in 12-well plates for flow cytometry analysis and 7 × 10 4 cells/well in 24 well plates for immunofluorescence (IF) analysis).

Immunofluorescence Staining
Cells were fixed at the end of respective incubation time in 3% PFA solution for at least 10 min at room temperature before permeabilization with 0.2% Triton X-100 in PBS and blocking by 5% horse serum (ThermoFisher Scientific, 26050-088) in PBS at room temperature for at least 1 h. For co-infection and sequential infection experiments, HEV capsid was stained with rabbit anti-HEV-ORF2 serum (kind gift of Prof. Rainer G. Ulrich, Friedrich Loeffler Institute, Germany, 1/4000 in PBS supplemented with 5% horse serum), while for super-infection experiments, HEV was stained with rabbit anti-GFP (Invitrogen, A11122, 1/1000 in PBS supplemented with 5% horse serum). HCV was stained with murine anti-HCV-NS5A-9E10 antibody (kind gift of Prof. Charles M. Rice, Rockefeller University, New York, USA, 1/10,000 in PBS supplemented with 5% horse serum), all overnight at 4 • C. HEV secondary staining was performed with Alexa 488-labeled goat anti-rabbit IgG (Invitrogen A11008, 1/1000 in PBS supplemented with 5% horse serum), HCV with Alexa 555-labeled donkey anti-mouse IgG (Invitrogen, A32773, 1/1000 in PBS supplemented with 5% horse serum). Plates were imaged at an Olympus IX81 microscope (Olympus) (transfection experiments) or a Keyence BZ X810 (Keyence Deutschland GmbH, Neu-Isenburg, Germany).

Quantification of HCV and HEV-Positive Cells in Fluorescence Microscopy Pictures
Mean fluorescence intensities of HCV immunofluorescence were obtained from a 5-pixel wide cytoplasm ring (cytoring) following segmentation of DAPI-stained nuclei using CellProfiler. HEV immunofluorescence intensities were also obtained from this cytoring. To distinguish noninfected cells from infected cells, minimum intensity thresholds were applied. Cells were classified into four categories: positive for both HCV and HEV, positive for HCV only, positive for HEV only or double negative.2.8. Western Blot.
Cells were detached with trypsin and taken up in DMEM, washed once with PBS and taken up in SDS sample buffer. After freeze-thaw, samples were digested with 1 µL benzonase (25-29 Units/µL, Millipore, Burlington, MA, USA; 70664-10KUN) until solution was easy to pipet. Samples were boiled at 98 • C for 5 min. A total of 300,000 cells each were separated on an SDS PAGE and immobilized on a polyvinylidene fluoride membrane (GE Healthcare #10600023). Membrane was blocked in PBS with 0.5% Tween20 (PBS-T), supplemented with 5% skimmed milk powder. NS3 was detected with murine anti NS3 #337 F3A6 (homemade, 1/500) for 3.5 h at room temperature, ß actin with anti-murine anti ß-actin antibody (Sigma-Aldrich, St. Louis, MO, USA; A2228, 1/1000) for 1 h at room temperature. Bound primary antibodies were detected by incubation with anti-mouse-IgG-HRP (Sigma, A4416, 1/20,000) for 1 h at room temperature.

VSV Infection
The vesicular stomatitis virus was kindly provided by Dr. Gert Zimmer [32]. Cells were infected with VSV-GFP for 16 h followed by fixation with 3% paraformaldehyde (PFA).

Flow Cytometry
After respective incubation times, cells were trypsinized, washed and fixed in PBS supplemented with 1% FCS and 0.5% PFA for 10 min at room temperature. Cells were analyzed at a BD Acurri C6 Plus FACS (BD Biosciences, Franklin Lakes, NJ, USA).

HCVcc and HEVcc Production
For infectious cell-culture derived HCV (HCVcc) production, HCV IVTs were electroporated into Huh7.5 cells. Supernatant was harvested 48 and 72 h post electroporation, pooled and centrifuged at 200× g. Virus titer was determined by TCID50 and aliquoted and frozen at −80 • C until usage. For infectious cell culture derived HEV (HEVcc) production, HEV IVTs were electroporated into HepG2 cells. Cells were trypsinized and resuspended in fresh DMEM 7 days post electroporation and lysed via three freeze thaw cycles. The lysate was cleared from cell debris by a 10,000× g centrifugation for 10 min and titrated on HepG2/C3A cells to determine viral titers. Virus was frozen at −80 • C until usage.

HEV Super-Transfection Experiments
To prepare Huh7-Lunet cells stably expressing NS3 3 _dg_ JFH-1NS5Aaa2359_RFP for super-transfection experiments, G418 selection was discontinued with simultaneous introduction of sofosbuvir, ribavirin or DMSO treatment. Cells were cultivated under these conditions for 3 days before electroporation and subsequently cultivated for 5 days before subjecting cells to IF or flow cytometry analysis. For DAA treatment of HCV selected replicons, selection antibiotic was removed, and cells were treated with indicated drugs for 5 days before super-transfection with HEV.

HEV Infection Experiments
For all infection experiments, 2 × 10 4 cells/well were seeded in 24-well plates and allowed to adhere overnight. Infection was performed the next day. Virus inoculum was removed after 24 h. Cells were either fixed 4 days post infection (d p.i.) or received a second virus infection for 24 h followed by 4 days incubation and subsequent fixation and preparation for IF. All infections were carried out with a multiplicity of infection of 1.

HCV Super-Infection on HEV Expressing Cells
A total of 2 × 10 4 cells/well were seeded in 24-well plates and allowed to adhere overnight without G418 selection agent. Infection was performed the next day and inoculum was replaced after 24 h with DMEM. 5 d p.i., cells were fixed with PFA and prepared for IF.

Determination of Viral Titers
To determine titers of infectious HCV, tissue culture infectious dose (TCID 50 /mL) assay was performed, as described previously [33]. To determine infectious HEV titers, focus forming units were determined as described [8].

Production and Infection of Human Liver Chimeric Mice
uPA +/+ -SCID mice were transplanted with cryopreserved primary human hepatocytes (PHH), as previously described [34]. Successful engraftment was assessed by human albumin concentration in mouse plasma, determined by ELISA (Bethyl Laboratories, Montgomery, TX, USA). Mice were either intraperitoneally injected with a fecal suspension containing 10 6 IU HEV-3f or intrasplenically with a plasma sample containing 1.5 × 10 5 IUHCV GT 1a. Fecal and blood samples were routinely collected and stored at −80 • C until analysis. The animal study protocol was approved by the Animal Ethics Committee of the Faculty of Medicine and Health Sciences of Ghent University. HEV and HCV RNA was extracted from fecal or plasma samples, respectively, using the NucliSENS easyMAG system (Biomérieux, Craponne, France). HEV RT qPCR was performed as described by Sayed et al. [35]. HCV RNA was detected using the RealStar ® HCV RT PCR kit (Altona, Hamburg, Germany) and the LightCycler 480 (Roche Diagnostics Deutschland GmbH, Mannheim, Germany).

HCV Impairs HEV Replication in Subgenomic Reporter Co-Transfection Assays
To test whether HEV and HCV influence each other's replication when introduced into the same cell population, subgenomic replicons harboring either a Gaussia (HEV) or Firefly (HCV) luciferase reporter were transfected into Huh7.5 cells by electroporation, either alone or together ( Figure 1A). Gaussia luciferase is secreted into the supernatant, while Firefly luciferase remains intracellular, enabling dual-luciferase measurements to simultaneously monitor HCV and HEV replication kinetics. Ribavirin (HEV) and 2 -Cmethyladenosine (HCV) were included as inhibitors for the respective viruses. Of note, 2 -C-methyladenosine exhibited also anti-HEV properties, which is in line with findings from the literature [36]. HEV replication was significantly reduced after 72 and 96 h post transfection when compared to single transfection ( Figure 1B, upper panels), while HCV replication was not affected ( Figure 1B, lower panels). These results indicate that HEV replication was influenced by co-replicating HCV, but not vice versa.

HCV Impairs HEV Replication in Subgenomic Reporter Co-Transfection Assays
To test whether HEV and HCV influence each other's replication when introduced into the same cell population, subgenomic replicons harboring either a Gaussia (HEV) or Firefly (HCV) luciferase reporter were transfected into Huh7.5 cells by electroporation, either alone or together ( Figure 1A). Gaussia luciferase is secreted into the supernatant, while Firefly luciferase remains intracellular, enabling dual-luciferase measurements to simultaneously monitor HCV and HEV replication kinetics. Ribavirin (HEV) and 2′-C-methyladenosine (HCV) were included as inhibitors for the respective viruses. Of note, 2′-C-methyladenosine exhibited also anti-HEV properties, which is in line with findings from the literature [36]. HEV replication was significantly reduced after 72 and 96 h post transfection when compared to single transfection ( Figure 1B, upper panels), while HCV replication was not affected ( Figure 1B, lower panels). These results indicate that HEV replication was influenced by co-replicating HCV, but not vice versa. Measurements of virus-encoded luciferases only provide information on the scale of a cell population and thus do not allow us to distinguish whether the inhibition of HEV replication was due to a lower fraction of HEV-transfected cells and/or lower replication levels. To address this question, we co-transfected a GFP expressing HEV replicon based on Kernow-C1 p6 strain (p6-GFP) and an RFP expressing HCV replicon based on JFH-1 (JFH-1-RFP) into Huh7.5 cells (Figure 2A). Microscopy analysis showed only a few HEV positive cells in the HCV/HEV co-transfected setting, as compared to cells only transfected with HEV ( Figure 2B). Conversely, the number of HCV positive cells was not affected ( Figure 2B). Comparison of the respective GFP and RFP fluorescence intensities on a single cell level further revealed strongly reduced levels for GFP upon co-transfection with HCV, as compared to cells only transfected with HEV ( Figure 2C), indicative of lowered HEV replication. Flow cytometry analysis was used to quantify the inhibitory effect. HEV/HCV double positive cells had approximately 70% reduced GFP levels while the RFP signal was not changed ( Figure 2D and Figure S1). Similarly, the number of GFP positive cells dropped from 54% to 12%, while the number of RFP positive cells was unaffected ( Figure 2E, upper left panel). The effect of HCV on HEV was dependent on active HCV replication, as concomitant sofosbuvir treatment reverted the percentage of GFP positive cells back to levels similar to single HEV transfection ( Figure 2E, lower left panel). This was also the case for the co-treatment of sofosbuvir with ribavirin ( Figure 2E, lower right panel). In summary, these results demonstrate that HEV replication is hindered in HCV-replicating cells, while HCV replication is not influenced by the presence of HEV subgenomic replicons.

HCV Treatment Restores HEV Replication Capacity
In a clinical context, it is likely that an active infection with both viruses does not occur simultaneously, but rather consecutively. Therefore, Huh7 Lunet cells were transfected with a subgenomic JFH-1 replicon harboring an NS5A-RFP and a neomycin resistance cassette (Huh7 Lunet sg/neo), allowing super-transfection with luciferase-encoding HEV of fluorescently labeled HCV-positive cells ( Figure 3A). The selected cells were used to test HCV's capacity to interfere with different HEV isolates, including Kernow C1 p6, 83-2-27 (HEV-3) and Sar-55 (HEV-1). Replication was reduced for all HEV strains between 70% (Sar55) and 98% (83-2-27) in Huh7 Lunet sg/neo cells compared to Huh7-Lunet naïve cells, indicating that the previously observed effects are not co-transfected as well as not HEV-3 exclusive ( Figure 3B-D). Additionally, p6 GFP was super-transfected into Huh7-Lunet naïve and Huh7-Lunet-sg/neo cells. While establishing replication in 68% of naïve Huh7-Lunets, HEV did only replicate in 0.5% of cells that had been selected for HCV ( Figure 3E,F and Figure S3), confirming the inhibitory effect of HCV on HEV-replication. Super-infection of Huh7 Lunet naïve or Huh7 Lunet sg/neo with vesicular stomatitis virus (VSV) did not result in either reduced HCV or VSV signal ( Figure S2), indicating a specific viral interference with HEV. To test if clearance of HCV would render hepatocytes susceptible to HEV again, HCV positive cells were either treated with sofosbuvir or DMSO for 5 days and subsequently super-transfected with HEV and analyzed via flow cytometry ( Figure 3G and Figure S4). DAA pretreatment significantly enhanced the number of HEV positive cells from 2.5% to 15%, while simultaneously reducing HCV positive cells from 70% to 1.4% ( Figure 3H; left panel). To exclude a potential antiviral effect of sofosbuvir on HEV [37,38], these results were confirmed with telaprevir ( Figure S5). These data suggest that the exclusion phenotype of HEV replication required active HCV replication and could be reverted by clearance of HCV with DAAs.

HCV Treatment Restores HEV Replication Capacity
In a clinical context, it is likely that an active infection with both viruses does not occur simultaneously, but rather consecutively. Therefore, Huh7 Lunet cells were transfected with a subgenomic JFH-1 replicon harboring an NS5A-RFP and a neomycin

The HCV Protease NS3/4A Inhibits HEV Replication
To identify a potential determinant for the HCV-mediated restriction of HEV replication, we next generated Huh7.5 cells constitutively expressing the HCV protease NS3/4A as a wildtype protein or as catalytically inactive S139A mutant. The expression of the protein was determined by Western blot ( Figure 4A). The functionality of the proteins was validated by expression of a sensor consisting of a fusion protein of RFP and the NS3/4A substrate interferon-β promotor stimulator 1 (IPS-1), which are linked via a nuclear localization sequence (NLS) [39]. If a functional NS3/4A protein is expressed, IPS-1 is cleaved, exposing the NLS and resulting in the translocation of RFP into the nucleus. Representative IF pictures and the quantitative determination of the nuclear-localized RFP signal verified the successful expression of operative wildytpe NS3/4A in Huh7.5 cells, in contrast to Huh7.5 cells expressing the S139A mutant ( Figure 4B,C). Next, we analyzed the interference of the HCV protease with HEV replication. NS3/4A alone was sufficient to inhibit the replication of subgenomic luciferase replicons of both 83-2-27 ( Figure 4D) as well as Kernow-C1 p6 ( Figure 4E). This inhibitory effect was even more pronounced when infecting HCV protease expressing cells with infectious cell-culture derived HEV (HEVcc) ( Figure 4F). These data indicate that the HCV protease NS3/4A was sufficient to inhibit HEV replication and infection.

HCV Super-Infection of HEV-Replicating Cells
As we recently established improved cell culture conditions to generate HEVcc, we next investigated viral interference in an authentic HCV/HEV co-infection setup. A coinfection approach inoculating Huh7.5 cells with HCV and HEV at the same time as well as a sequential infection was conducted as depicted ( Figure 5A). In the co-infection condition, the percentage of HEV-positive cells was not significantly altered, whereas the number of HCV-infected cells was reduced from around 40% in single infections to 20% ( Figure 5B). In the sequential infection approaches, HEV infectivity was not changed, infecting approximately 5% of cells ( Figure 5C,D). However, HCV-positive cells were reduced when HCV infected first from around 40% to 25%, but remained unaffected when HEV infected first ( Figure 5C,D). As these experimental model systems might not fully reflect the potentially long timeframes between consecutive viral infections, we generated Huh7.5 cells that harbor a neomycin-selectable GFP-labeled p6 replicon to mimic a chronically HEV-infected stage before HCV super-infection ( Figure 5E). IF analysis demonstrated that approximately 65% of Huh7.5 cells selected (Huh7.5 p6-GFP-Neo) were positive for HEV ( Figure 5F). The percentage of HCV positive cells was around 40% both in naïve Huh7.5 cells as well in Huh7.5 p6 GFP-Neo cells, implying that HCV can super-infect HEV replicating cells and its infectivity was unaffected by HEV ( Figure 5F). Treatment with the NS3/4A targeting DAA paritaprevir reduced the number of HCV positive cells and had no influence on HEV replication ( Figure 5F). These results show that HCV can co-infect human hepatocytes with HEV and super-infect cells harboring a selectable HEV replicon.
NS3/4A substrate interferon-β promotor stimulator 1 (IPS-1), which are linked via a nuclear localization sequence (NLS) [39]. If a functional NS3/4A protein is expressed, IPS-1 is cleaved, exposing the NLS and resulting in the translocation of RFP into the nucleus. Representative IF pictures and the quantitative determination of the nuclear-localized RFP signal verified the successful expression of operative wildytpe NS3/4A in Huh7.5 cells, in contrast to Huh7.5 cells expressing the S139A mutant ( Figure 4B, C). Next, we analyzed the interference of the HCV protease with HEV replication. NS3/4A alone was sufficient to inhibit the replication of subgenomic luciferase replicons of both 83-2-27 (Figure 4D) as well as Kernow-C1 p6 ( Figure 4E). This inhibitory effect was even more pronounced when infecting HCV protease expressing cells with infectious cell-culture derived HEV (HEVcc) ( Figure 4F). These data indicate that the HCV protease NS3/4A was sufficient to inhibit HEV replication and infection.

HEV or HCV-Positive Human Liver Chimeric Mice Showed Reduced Viral Loads in Individual Mice after Super-Infection
Human liver chimeric mice have been established as an animal model to study both HEV [35,40] as well as HCV (reviewed in [41]). We used homozygous urokinase plasminogen activator (uPA)-SCID +/+ mice, of which the liver was reconstituted by primary human hepatocytes (PHH) to analyze HCV/HEV co-infections in vivo. To this end, animals were either inoculated intraperitoneally with HEV or intravenously with HCV. After several weeks, which allowed for the establishment of the infection, the animals were super-infected with the respective other virus. Viral loads were determined by RT qPCR from stool (HEV) or plasma (HCV) at regular intervals. Mice only infected with HEV ( Figure 6A, left

HEV or HCV-Positive Human Liver Chimeric Mice Showed Reduced Viral Loads in Individual Mice after Super-Infection
Human liver chimeric mice have been established as an animal model to study both HEV [35,40] as well as HCV (reviewed in [41]). We used homozygous urokinase plasminogen activator (uPA)-SCID +/+ mice, of which the liver was reconstituted by primary human hepatocytes (PHH) to analyze HCV/HEV co-infections in vivo. To this end, ani-Cells 2022, 11, 927 13 of 18 mals were either inoculated intraperitoneally with HEV or intravenously with HCV. After several weeks, which allowed for the establishment of the infection, the animals were super-infected with the respective other virus. Viral loads were determined by RT qPCR from stool (HEV) or plasma (HCV) at regular intervals. Mice only infected with HEV ( Figure 6A, left panel) showed viral loads between 10 6 and 10 8 IU/mL after the onset of infection and stabilized over the time monitored. Viral loads developed similarly in HEV positive mice super-infected with HCV ( Figure 6A, right panel). In contrast, HCV could not establish productive infection in two out of three HEV infected mice, while in the other mouse reduced viral loads were observed compared to the HCV mono-infection group visualized by the semitransparent red line underlined with a grey area ( Figure 6A, right panel). Restriction of HCV infection in HEV-positive mice was also observed when the HCV super-infection was conducted at an earlier time point ( Figure S6). stabilized over the time monitored. Viral loads developed similarly in HEV positive mice super-infected with HCV ( Figure 6A, right panel). In contrast, HCV could not establish productive infection in two out of three HEV infected mice, while in the other mouse reduced viral loads were observed compared to the HCV mono-infection group visualized by the semitransparent red line underlined with a grey area ( Figure 6A, right panel). Restriction of HCV infection in HEV-positive mice was also observed when the HCV superinfection was conducted at an earlier time point ( Figure S6). For the HEV on HCV super-infection protocol, mice only infected with HCV had viral loads between 10 7 and 10 8 IU/mL at all time points ( Figure 6B, left panel). HCV viral loads developed similarly in HCV positive mice super-infected with HEV, except for one mouse that could not be infected at all ( Figure 6B, right panel). HEV viral loads were reduced and delayed in two individual mice, while in the two other HCV super-infected mice comparable or higher HEV copy numbers could be observed ( Figure 6B, right panel). In sum, viral interference in the replication kinetics of HCV and HEV was observed in individual mice after super-infection.  For the HEV on HCV super-infection protocol, mice only infected with HCV had viral loads between 10 7 and 10 8 IU/mL at all time points ( Figure 6B, left panel). HCV viral loads developed similarly in HCV positive mice super-infected with HEV, except for one mouse that could not be infected at all ( Figure 6B, right panel). HEV viral loads were reduced and delayed in two individual mice, while in the two other HCV super-infected mice comparable or higher HEV copy numbers could be observed ( Figure 6B, right panel). In sum, viral interference in the replication kinetics of HCV and HEV was observed in individual mice after super-infection.

Discussion
Although HCV and HEV are both important pathogens with a considerable disease burden that targets the same organ, clinical information on HCV/HEV co-infections is limited, while molecular characterizations are absent. To address this gap in knowledge, we investigated HCV/HEV co-infections in vitro and in vivo. To address co-infection in vitro, a transfection model and an infection model were used that differed in several aspects and thus have distinct advantages. The transfection model's advantage is that it guarantees high percentages of HEV-and HCV-positive cells. This high percentage is important, because it statistically increases the number of cells, where both viruses aim to establish an infection and is hence more sensitive observing viral interference. The co-infection model with an infectious virus is more complex, as it does reflect not only interference during replication but also during entry. It also is more complex with regard to both viruses, as they also express their respective structural proteins in a co-infection setting. We were able to demonstrate the restriction of HEV replicons by concurrent HCV replication. Similarly, established HCV replication blocked HEV from replicating, which was revertible upon HCV treatment with DAAs. Of note, the ectopic expression of the HCV NS3/4A protease alone was sufficient to suppress HEV replication, suggesting this as the potential mechanism of action. It is tempting to speculate that NS3/4A might cleave the HEV ORF1 polyprotein, rendering it less functional. We hypothesized that the HEV ORF1 protein might be cleaved by NS3/4A. Because our ORF1 antibodies failed to detect ORF1 in the Western blot, we could not address this question experimentally. However, it is known which part of MAVS [27,42] or TRIF [43], is recognized and, additionally, one study tested a library of small peptides to find 613 sequences that could be cleaved by NS3/4A [44]. Using a bioinformatic approach, we checked if any of these sequences were present in the ORF1 of HEV p6, but none were (data not shown). Hence, cleavage of the HEV polyprotein might likely not be the mechanism of HEV suppression. There have been several epidemiological studies conducted to investigate a link between HCV and HEV infections. Some studies investigated HEV seropositivity in HCV-RNA positive patients with mixed results. A higher HEV seropositivity compared to the control group was reported in one study [45] with another confirming these findings in a mixed cohort of chronic HBV and HCV patients [46], while a third did not find any difference in HEV seropositivity between chronic HCV patients and an aged-matched control group [47]. HCV-IgG was enhanced in HEV-IgG positive persons from the 15,000 US NHANES III study population [48], while an Iranian study of 324 hemodialysis patients did not find a link between HCV and HEV seropositivity [49], as did a study analyzing almost 900 HIV-I infected Spanish patients [50].
To recapitulate HCV/HEV super-infections in vivo, liver humanized mice were utilized. In the HCV super-infection of HEV positive mice, only one mouse was productively infected and vice versa; reduced viral loads of HEV were noted in comparison to the mono-infected group. Two different explanations for this phenomenon seem conceivable. First, that the mutual inhibitory effect of HEV and HCV is due to the first virus eliciting an innate immune response that restricts the latter, irrespective of direct interactions. Second, that the direct interaction is responsible, but it differs between the model systems. If the former were true, then HEV's innate immune response should be stronger, as HCV is more efficiently restricted by present HEV than vice versa. However, transcriptomic profiling in chimpanzees suggest that it is reverse. Both more genes were differentially expressed and their expression was more robust [51]. Regarding the latter, there are several differences between the model systems. First, we were not able to determine the exact number of infected transplanted PHHs, which was possible in vitro. For HCV patients, it has been described that HCV-infected hepatocytes tend to be found in clusters [52]. Using a laser capture microdissection experimental setup, the fraction of HCV RNA positive cells was reported to range from 21% to 45% in human liver tissue [53], which was similar to that described with antigen staining (7-20%) [52]. The number of HEV-infected hepatocytes in biopsies has been demonstrated to be highly variable, ranging from <2% to 50% in antigen staining and 0-50% in RNA staining [54]. The number of cells positive for each virus in vivo is therefore lower than in cell culture experiments, which might reduce the effect size of inhibitory effects. Secondly, the high variability of hepatocytes positive for either virus could be one reason for the variation of co-infection courses in individual mice. Furthermore, PHHs differ substantially from hepatoma cell lines with respect to innate immunity, host factors and the fact that the former are polarized. For HCV it has been demonstrated that, besides comparatively low antigen levels and replication efficiency [52], the host transcriptional response does differ significantly in chronic HCV patients [55] or PHHs [56] compared to hepatoma cell lines, such as Huh7 [57] or Huh7.5 [58]. Regarding HEV, the infection of PHHs is efficient and transcriptional profiling has been performed recently in PHHs [8], but potential differences between the host response compared to hepatoma cell lines have not been addressed so far. Further studies addressing these differences in HEV-host interplay in different model systems would not only deepen our understanding of HEV pathogenesis, but also of HCV/HEV co-infections, especially in the more complex in vivo situation. In conclusion, co-infection of HCV and HEV was studied for the first time in different experimental model systems, including humanized mice, demonstrating viral interference of HEV and HCV replication. These findings provide new insights into the pathogenesis of HCV/HEV co-infection and may contribute to its clinical management in the future.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. For the analysis of HCV cleavage sites within the HEV p6 genome, publicly available data were analyzed. This data can be found here: https://journals.plos.org/plosone/ article?id=10.1371/journal.pone.0035759#s5, Table S1. Last accessed on 15 January 2022.