Challenge Inoculum for Hepatitis C Virus Controlled Human Infection Model

Abstract

For any controlled human infection model (CHIM), a safe, standardized, and biologically relevant challenge inoculum is necessary. For hepatitis C virus (HCV) CHIM, we propose that human-derived high-titer inocula of several viral genotypes with extensive virologic, serologic, and molecular characterizations should be the most appropriate approach. These inocula should first be tested in human volunteers in a step-wise manner to ensure safety, reproducibility, and curability prior to using them for testing the efficacy of candidate vaccines.

A controlled human infection model could provide an accelerated pathway to evaluate the effectiveness of experimental hepatitis C virus (HCV) vaccines [1]. To establish clinical trials based on this model, the generation of appropriate challenge inocula is necessary as the first step. In this article, we focus on potential viral inocula and present an extensive overview of factors that must be considered. Potential viral inocula must be characterized fully to establish that they will reproducibly infect all controlled human infection model (CHIM) participants, comply with all safety and regulatory criteria, recapitulate the natural course of infection including progression to chronicity, address HCV genetic diversity, and be reliably cured with antiviral therapy. Here we pose a series of questions and put forth a set of recommendations for the selection and characterization of viral inocula for HCV CHIM.

SOURCE OF CHALLENGE INOCULUM

For this consideration, we need to assess all criteria that are necessary for safety and regulatory requirements, production process, and biological authenticity (reflection of natural transmission) of the viral inoculum. Humans are the only natural host for HCV. Although some animal models such as human liver chimeric mice can be used to produce infectious HCV [2], chimpanzees are the only robust immune-competent model available to study infection of the virus [3]. HCV derived from experimentally infected animals are valuable tools to study HCV replication; diverse strains have been generated and characterized in these models. However, virus derived from these models is not suitable for CHIM due to safety concerns of unknown animal-derived pathogens. Although animal-derived blood products have been used in humans, they are purified as a blood component, tested for adventitious pathogens, and undergo rigorous pathogen-inactivation processes [4]. Furthermore, international shipment of infectious chimpanzee serum currently faces import/export restrictions. Thus, animal-derived HCV is not feasible for a CHIM inoculum.

HCV can be produced in cell culture (HCVcc) [5–7] and is infectious in chimpanzees and human liver chimeric mice [6, 8]. Biochemical characterization of HCVcc revealed a higher density compared to virus produced in vivo, likely due to less association with human lipoproteins [8]. Initially, the ability of HCV to replicate efficiently in cell culture is dependent on at least a partial genome derived from the JFH-1 isolate [9]. Subsequently, other strains unrelated to JFH-1 have been produced in cell culture, but they do not replicate as well and require adaptive mutations to ensure higher levels of replication and infectivity [10–14]. Human host factors have also been identified to enhance permissiveness of HCV replication in cell culture [15]. Because a vaccine will need to work against many isolates of HCV, HCVcc derived from JFH-1, a genotype 2a strain derived from a patient with fulminant hepatitis C, is not a suitable challenge virus for evaluating the breadth of vaccine protection. Furthermore, HCVcc can only be produced with high yield using human hepatoma Huh7-derived cells, which currently would not meet regulatory criteria for human use. Re-derivation of these cells could be performed but would face a long and difficult regulatory approval process. Finally, relevant human data with these HCVcc preparations are completely lacking. Even with recombinant isolates, they may not replicate or cause persistent infection like infectious serum in chimpanzees [16, 17].

We suggest that the best source of HCV inoculum would be from a human donor, which is different from other CHIM models [18] and would have different regulatory requirements. Blood banks already have strict safety standards for blood products to prevent the spread of infectious agents. The use of human-derived blood products is widely accepted and approved for various medical indications, with excellent safety profiles. Thus, serum or plasma from HCV infected humans as an inoculum source should be feasible.

A key question is whether the inoculum should be obtained during the acute or chronic phase of the disease. Presumably, the spread of HCV within the at-risk population is from sharing contaminated materials with a chronic carrier, although acute-to-acute transmissions have been reported [19]. Therefore, using inoculum from the chronic phase would recapitulate this situation. Interestingly, the infecting viral strains analyzed during the acute phase tend to be limited in sequence diversity, suggesting only certain species is transmitted from the source [19]. More importantly, the infectivity of HCV serum usually drops dramatically in the presence of immune complexes during the chronic phase [20]. The difference in infectivity can be 1000-fold [20, 21]. Thus, a larger volume of inoculum would be needed. As the inoculum volume increases, so does the risk of introducing other adventitious agents or undesirable contaminants. HCV-specific antibody response usually requires 8–12 weeks to develop after infection [22]. During the acute phase of infection, HCV is not complexed with antibodies and serum samples obtained during this early phase of infection would have high specific infectivity. To prevent immune escape, a successful vaccine candidate should be effective against the diverse quasispecies (swarm of distinct but related variants) of HCV that typically develops during natural infection. The HCV titer of acute phase serum can often exceed 10 million genome copies/mL. Notably in chimpanzees, a dose of 100 HCV RNA molecules free of immune complex consistently established HCV infection [23, 24]. Although the exact amount of HCV RNA needed to establish consistent infection in humans has not been determined, we anticipate this value will be similar to that seen in chimpanzees. Therefore, using a conservative dose of 3000 RNA molecules, 1 mL of acute sera could provide enough doses for >1000 infections. At this scale, it would be easy to archive enough material to standardize all CHIM studies in the future. From the technical perspective, titers of antibody-free HCV can be determined by density gradient analysis [20]. In addition, human liver chimeric mice can be used to confirm infectivity and standardize infectious titer of the inoculum, although the infectivity titers in these mice may be lower than those in chimpanzees [25].

The historical H77 sample from patient Hutchinson obtained during his acute phase of HCV infection [26, 27] has been the gold standard for many HCV infection studies and characterized extensively in vitro and in vivo. This sample has high infectivity and can cause chronic infection in chimpanzees. Ample sample volume is available in storage. The complete genomic sequence of the virus is known, and extensive reagents have been developed for this viral strain. Thus, the H77 sample could be readily available as a challenge inoculum for HCV CHIM. However, the stock has been in storage for many years under non-clinically approved condition and/or accessed repeatedly for various studies. The sample would need to be re-characterized as a potential challenge inoculum.

An alternative approach would be to inoculate the cloned HCV full-length genome directly into the liver of a CHIM volunteer for establishment of infection. This approach, which was initially used to successfully establish HCV H77C infection in chimpanzees [28, 29] and subsequently for HCV of other genotypes [30], would alleviate the concern about blood product safety issues. Synthetic HCV genome RNA could also be delivered using modern messenger RNA (mRNA) lipid nanoparticle technologies, which have been developed for many human applications [31]. Although this approach would allow initiation of infection, it is unclear how well they would mimic natural transmission and consequently be used as a CHIM inoculum for vaccine study. This approach, however, could be considered for generation a well-characterized viral stock in a CHIM volunteer for subsequent studies.

A theoretical concern is that many vaccine candidates are based on the H77 strain, and thus a challenge study with a homologous virus may not reflect desired vaccine efficacy. An additional concern is that patient Hutchinson had never been treated with direct-acting antiviral (DAA), which could potentially be an exclusion criterion. Although H77C strain has been shown to be highly sensitive to DAA in cell culture [32] and can be tested in human liver chimeric mouse model for DAA response, it is not clear whether such a test would be acceptable to the regulatory agencies.

SAFETY BENCHMARKS FOR CHALLENGE INOCULUM

Use of challenge inoculum from an HCV-infected human would have to conform to all transfusion and blood product safety requirements. The product, in this case, serum or plasma, would need to be collected in an approved blood bank facility with appropriate equipment and procedures, and screened by standard tests, such as ABO blood type and infectious agents (other than HCV). It may be necessary to conduct sequence-independent next-generation sequencing for other adventitious infectious agents. The sample would have to be processed and stored under Good Laboratory/Manufacturing Practices. In addition to determining the infectivity and genotype, as described above, the HCV strain should be sequenced for DAA resistant-associated substitutions [33] and probably subjected to full-genome sequencing.

In selecting appropriate donors, several conditions should be considered. The donors should satisfy regulatory requirements for blood product donation and be in otherwise good physical health and be treated successfully with DAA after blood donation for potential inoculum. At this point, it is not clear whether degree of hepatitis (liver function tests) should be considered to determine donor's eligibility. For donors with acute hepatitis, it is not unusual to see substantial ALT elevation during the acute phase; however, fulminant hepatitis (ALT > 1000 IU/mL and other signs of acute liver failure), which is rare, might exclude such a sample. For donors in the chronic phase, it would be preferable to avoid those with cirrhosis.

Another potential source of inoculum could be post-transplant HCV patients, especially HCV-negative recipients receiving HCV + donor organs, who often have very high viral titers and low immune complexes [34], a desirable feature for an inoculum. On the other hand, these patients may carry occult adventitious pathogens because of intensive immunosuppression post-transplant. Although it may be possible to exhaustively rule out such a possibility, this potential risk needs to be balanced against the benefit.

Regarding blood type, it is preferable to use AB + donors who are universal donors for serum or plasma. This criterion may not be practical because AB + donors are uncommon. If donors from other blood types are used, the volume of inoculum should be minimized. Having a high-infectivity inoculum, as described above with a minuscule amount of inoculum (microliter range), would mitigate this concern.

CHARACTERIZATION OF CHALLENGE INOCULUM

As discussed above, infectivity of the challenge inoculum would need to be determined first. The goal is to achieve 100% infection in CHIM participants so the efficacy of a vaccine candidate can be assessed in a clinical study that is sufficiently powered and ethically appropriate. Quantification of antibody-free HCV would provide an initial guide. The sample can then be tested in humanized liver chimeric mice, which would provide an infectivity titer as in 50%-mouse infectious dose (MID₅₀). Previous studies suggested that HCV inocula may have variable infectivity per genome-equivalent in this mouse model when compared to chimpanzees [25]. Confounding factors, such as presence of antibodies, efficiency of human hepatocyte engraftment, or lack of direct and large-sample size comparison may complicate interpretation of these data. Direct comparison of the 50%-chimpanzee infectious dose (CID₅₀) with MID₅₀ of a well-characterized historical sample, such as the H77 sample, could provide potentially useful guidance. On the other hand, these data may not completely correlate with human infectious dose. Thus, it would be necessary to conduct a limited dose-finding study in a few participants of the HCV CHIM to define more fully the appropriate infectious dose.

Additional studies of the inoculum virus may be of interest. For example, the virus can be tested for its neutralization sensitivity by various well-characterized broadly neutralizing antibodies [35] and/or antisera from previously vaccinated mice and humans [36]. The envelope glycoprotein genes (E1 and E2) of the inoculum virus can be cloned and introduced into either HCV pseudovirus or full-genome HCVcc constructs [9, 37, 38] for neutralization studies. These experiments would be for research interest and might help clarify any subsequent vaccine breakthrough infections, but should not be a precondition for a challenge inoculum.

INOCULUM VIRAL GENOTYPE

Although it would be ideal to have inocula for each major genotype (1–6), it may be impractical and unnecessary for the CHIM model, which is an intermediate step for vaccine development. On the other hand, having more than 1 inoculum to assess the cross-genotype protection of a vaccine candidate would be important and perhaps necessary for further clinical testing. Because genotypes 1 and 3 are the most common strains infecting persons who inject drugs, generation of inocula for these 2 genotypes would be of high priority. Concern has been raised about using genotype 3 as a CHIM inoculum because it may have a lower response to DAA. However, use of more potent DAA regimens in recent studies have shown that patients with genotype 3 infection without cirrhosis have similarly high response rates and treatment failure can be attributed to baseline resistance-associated substitutions [39, 40], for which we will screen. Recent studies suggest that serotyping or immunotyping based on antibody-neutralization assay may provide a better distinction in considering cross-protection of a vaccine for HCV vaccine development [38, 41], but without a clearly defined or standard means to serotype HCV samples, it is probably not practical to consider this approach. It seems prudent to begin with a genotype 1 (1a and/or 1b) challenge inoculum, with others to follow. For practical reasons, collection and preclinical characterization of multiple genotype inocula can be pursued simultaneously.

INITIAL INOCULATION STUDY

Once a suitable inoculum has been identified and characterized, the challenge study design must be optimized. To establish the safety and reproducibility of the inoculum, it would be advisable to conduct a small-scale inoculation study. First a dose-finding study would be necessary to ensure reproducible infection. Based on in vitro and in vivo characterization of infectivity, infection study with a dose that is 50-times of the estimated infectious dose could be tested in one CHIM volunteer. If infection occurs, additional volunteers (3–4) can be tested with the same dose for reproducibility of infection, disease activity and rate of chronicity. If infection does not occur, a 10-fold higher dose would then be used in another CHIM volunteer and so on until infection can be consistently initiated. The infected individuals can then be followed for a period of time before beginning treatment with DAA. The timing of the DAA treatment would depend on the final protocol for the vaccination/challenge study [42]. It is reasonable to treat the first participant at an earlier time point than the proposed wait time of the vaccination/challenge study, just to ensure successful DAA treatment. Blood samples acquired from these patients during the acute phase can be characterized as above for additional inoculum if necessary. These participants can also form part of a control arm for subsequent vaccination/challenge studies.

CONCLUSION

In this article, we systematically review many important and relevant questions for the generation of a challenge inoculum of HCV CHIM. We provide a set of provisions and recommendations (summarized in Figure 1) based on extensive discussions among the experts and thorough literature review. Collectively, we propose that a well-characterized acute phase serum sample from a suitable donor would be the most appropriate source for CHIM inoculation; however, alternative sources should be considered. We acknowledge that our recommendations could engender additional debates and alternative considerations. We believe this guidance is a rational, evidence-based starting point and will lead to the establishment of a robust panel of challenge inocula for an HCV CHIM study.

Figure 1.

A schematic illustration of the proposed steps in establishment and characterization of HCV challenge inocula for HCV CHIM. Details are provided in the text. Created with BioRender.com. Abbreviations: CHIM, controlled human infection model; DAA, direct-acting antiviral; HCV, hepatitis C virus.

Notes

Ackowledgments. The authors thank Drs Marian Major and Francis Chisari for careful review and critique of the article.

Supplement sponsorship. This article appears as part of the supplement “Controlled Human Infection Model for HCV Vaccine Development,” sponsored by Toronto General Research Institute, United States National Institutes of Health, Johns Hopkins University, the Canadian Institutes of Health Research (CIHR), and the Canadian Network on Hepatitis C (CanHepC). CanHepC is funded by a joint initiative of CIHR (HPC-178912) and the Public Health Agency of Canada.