Identification of non-A, non-B viral hepatitis-specific liver mRNAs
A large core of our activities during this period centred around probing bacterial cDNA libraries derived from NANBH-infected livers in duplicate with highly radioactive P32-labelled cDNA prepared from either NANBH-infected livers (+probe), or from control, uninfected liver RNA (−probe). Using many different chimpanzee liver samples, a number of NANBH-specific cDNA clones were identified but all these were shown to be derived from chromosomal genes rather than of viral origin. The sensitivity of this method was improved such that rare RNAs of 0.01% and lower could be detected but nonetheless, no true viral clones could be identified. cDNA libraries were also prepared from NANBH-infected liver samples in which the cDNA was depleted of host sequences by prior hybridization to an excess of uninfected, control liver RNA. Again, NANBH-specific clones were identified from these libraries but none could be proven to be of viral origin.
Cross-hybridization with known viral genomes
In view of various data in the field consistent with a NANBH agent being related to either flaviviruses, togaviruses, hepadna viruses, picornaviruses or certain other viruses, RNA and DNA extracted from NANBH-infected chimpanzee liver and plasma samples (as well as cDNA libraries derived from these materials) were hybridized to radioactive probes derived from these known viral genomes. The absence of specific signals from these experiments, even under conditions of low hybridization stringency, was interpreted as reflecting either the rarity of the NANBH agent (which was known to be present at much lower infectivity titres as compared with HAV and HBV) and/or the absence of substantial sequence identity between these known agents and the NANBH agents.
Molecular characterization of the delta hepatitis genome and its use as a hybridization probe for non-A, non-B viral hepatitis agents
The delta hepatitis antigen was discovered by Mario Rizzetto as a nuclear antigen appearing in some chronic carriers of HBV that often exhibited severe hepatitis (18). Subsequently, an uncharacterized RNA molecule was found to be associated with infectious delta hepatitis (HDV) samples that was shown to require HBV as a helper virus for infectivity (19). Chimpanzees co-infected with HBV and HDV displayed the same kind of membranous tubules within the cytoplasm of hepatocytes as observed in NANBH-infected animals (20). When combined with the knowledge that infectious HDV particles had physicochemical properties similar to the NANBH TFA, this led to the hypothesis that the two agents may be related (20). At this time, the group of John Gerin had cloned and sequenced a small part of the HDV-associated RNA but the remainder proved particularly refractory to molecular cloning (21). With a view to using the HDV genome as a potential hybridization probe for the NANBH TFA, my laboratory initiated a collaboration with John Gerin to fully characterize the delta RNA molecule. Our work (with contributions from Kang-Sheng Wang, Qui-Lim Choo and Amy Weiner in my laboratory, along with Jim Ou from UCSF) was able to show that this RNA was in fact an unusual, viroid-like, covalently closed single-stranded molecule that encoded the delta antigen from an antisense open-reading frame (22, 23). The majority of the genome formed intramolecular base pairs, thus collapsing the circle into a double-stranded RNA, rod-like structure under normal physiological conditions (Fig. 1). Unfortunately, when the complete HDV genome was used as a hybridization probe for extracts of NANBH-infected plasma and liver samples, no specific hybridization could be observed, even under conditions of low stringency (24).
Attempts to visualize a putative non-A, non-B viral hepatitis genome
As a result of my collaboration with Dan Bradley of the Centers for Disease Control and Prevention (CDC), we now had access to large samples of NANBH-infected chimpanzee livers and plasmas. Because in our continuing, frustrating molecular studies, we wanted to ensure that these samples were at least equal in infectivity titre to those available elsewhere (25), Dan had determined the infectious titre of many of these samples and shown that the best plasma and liver samples had titres of ≥106 chimpanzee infectious doses (CID)/ml and 107 CID/g respectively. Such titres were comparable to the best reported elsewhere (25). In yet another ambitious approach, my colleague Amy Weiner extracted very large quantities of these samples (following pre-digestion with nucleases to remove chromosomal DNA and RNA) and then attempted to visualize a large, putative RNA or DNA genome using agarose gel electrophoresis, followed by improved, highly sensitive silver staining methods. Unfortunately, we concluded that there was still insufficient NANBH genome to detect using these methods.
Identification of non-A, non-B viral hepatitis-specific antibodies
By immortalizing B cells isolated from NANBH patients and screening their supernatant immunoglobulins (Igs) for specific binding to NANBH-infected liver sections, Yohko Shimizu et al. (28) were able to identify NANBH-specific antibodies for the first time in 1985, although it remained unclear for several years afterwards whether such antibodies were binding directly to NANBH viral proteins or to host gene products upregulated by infection.
This work stimulated a large effort at Chiron to reproduce such a methodology in an attempt to isolate NANBH viral-specific antibodies. In addition, work was performed in my laboratory in collaboration with Dr Shimizu to isolate cDNA clones encoding antigens reactive with her antibodies by using them to immunoscreen expression cDNA libraries derived from NANBH-infected chimpanzee livers. Although this work was unsuccessful at isolating these genes, Yohko Shimizu and colleagues eventually proved that the antigen reacting with her antibodies was of host origin rather than viral, by purifying the antigen on antibody-affinity columns, followed by protein sequencing (29).
‘Blind’ immunoscreening of cDNA expression libraries
In 1983 and 1984, I had considered using a NANBH cloning strategy in which sera samples from clinically-diagnosed NANBH patients could be used as a presumptive source of NANBH-specific antibodies to screen cDNA expression libraries derived from infectious liver and plasma samples. However, my personal experience and that of other colleagues and consultants with practical knowledge of immunoscreening cDNA libraries informed me that such methods do not always succeed in identifying specific genes even when well-characterized, high-titre and specific antisera against particular gene products were available. In addition, there was a legitimate fear that with the NANBH agent(s) causing such a high incidence of chronic infection, perhaps the immune response elicited against such an agent(s) was defective, weak or absent. Indeed, many groups working for more than a decade in the field had failed to identify the existence of NANBH-specific antibodies and while Yohko Shimizu had identified such antibodies in 1985, it was far from clear whether they were directed to proteins of the aetiological agent(s) of NANBH or to upregulated host gene products (as they eventually turned out to be directed against). Hitherto, therefore, I had considered the implementation of such an approach to identify the aetiological agent(s) of NANBH as being too risky.
However, in 1985, during a discussion on NANBH with my colleague George Kuo, who was then working in his laboratory adjacent to mine at Chiron on a HBV vaccine and recombinant factor 8, he independently and strongly advocated the application of a blind immunoscreening cDNA cloning strategy to NANBH, arguing that unlike the potential situation in highly expressing recombinant bacteria, NANBH-specific antigen concentrations in infected livers and plasmas would be below the limits of detection using conventional immunostaining methods for identifying NANBH-specific antibodies. This rationale depended, of course, on the failure of the field in general to demonstrate the existence of NANBH-specific antibodies as being owing to limitations in target NANBH antigen concentrations in the assay format, rather than owing to the absence or only low concentrations and/or avidities of NANBH-specific antibodies. Shortly after this discussion, my long-time collaborator Dan Bradley independently raised the question of whether such a blind immunoscreening approach may work to identify the aetiological agent(s) of NANBH.
I therefore decided to initiate such an approach in parallel with many of the other activities described above. Initially, George Kuo compared the sensitivity of the available immunoscreening methods and determined that the use of radioactively labelled I125 anti-human Ig provided the best sensitivity for detecting the binding of human Ig to cDNA clones. Then, over the next 18 months, Qui-Lim Choo carefully screened many tens of millions of clones from a cDNA library that I had prepared in the efficient expression vector lambda gt11 (30) that was derived from four different NANBH-infected chimpanzee liver samples, all harvested at different relative times in the infection cycle so as to ensure the presence of replicative NANBH genome and mRNAs. Unfortunately, despite screening this library with many different chimpanzee and human sera as putative sources of NANBH-specific antibodies, no clones could be isolated that could be defined to be of NANBH aetiological origin. Many clones were identified, however, that bound circulating auto-antibodies. In addition, clones encoding bacteriophage proteins were isolated, thus demonstrating the power of this approach.
In parallel with this work, Kang-Sheng Wang, in my laboratory, had been fine-mapping B cell epitopes of the delta antigen using a lambda gt11 expression cDNA library derived from infectious HDV plasma. The great success of this approach provided strong encouragement to me to persevere with this approach for NANBH and to apply it to the best infectious NANBH chimpanzee plasma produced by Dan Bradley [derived from chimpanzee #910, which had an infectious titre of at least 106 CID/ml (31)].
A lambda gt11 expression library that I prepared from this material failed to isolate true NANBH clones following screening with sera from individuals convalescent from NANBH infection (that based on HAV and HBV infections might be expected to be the best source of NANBH-specific antibodies). For the next cDNA library, however, that I prepared from total nucleic acid (both RNA and DNA) derived from chimpanzee 910 plasma, I decided to try a screening serum derived from a patient diagnosed with chronic NANBH infection that exhibited unusually high serum alanine aminotransferase elevations, indicative of severe liver damage.
When screened by Qui-Lim Choo, a few clones were scored positive including one that he labelled as 5-1-1. Other positive clones were shown by Qui-Lim to be host-derived but gradually, over the course of the next several months, we were able to show that 5-1-1 really was derived from the genome of a true NANBH viral genome [although by this stage in the work, we had to regularly pinch ourselves to make sure we were not just dreaming!; see Fig. 2 and (32)].
Proof of the aetiological origin of clone 5-1-1 was provided by a sequential set of experiments demonstrating that:
Clone 5-1-1 was not a host gene derived from the chimpanzee or the human genome (using Southern blot analysis).
That clone 5-1-1 and subsequent overlapping clones derived from the same cDNA library hybridized to a large, single-stranded RNA molecule of about 10 000 nucleotides that was present only in NANBH-infected samples, not in control, uninfected samples. Furthermore, this RNA contained a large open-reading-frame that encoded a novel sequence that exhibited very distant homology with flaviviral genomes.
That these clones encoded an antigen that various NANBH-infected chimpanzees seroconverted to (i.e. elicited specific antibodies) but not chimpanzees infected with HAV or HBV. Further, this immunoreactive antigen was derived from the RNA molecule itself thus indicating that it was positive-stranded like the flaviviridae genomes.
That the majority of a small cohort of NANBH patients (provided by Dr Gary Gitnick of UCLA) had circulating antibodies specific for the gene products of clone 5-1-1 whereas control human sera did not.
At this point in the analyses of clone 5-1-1, we considered that we had clear proof for a viral aetiology of clone 5-1-1 and therefore termed this NANBH viral agent HCV (32). These data, proving the molecular identification of HCV, were presented for the first time at a public seminar that I gave at the University of California in San Francisco in 1988.
Subsequently, 5-1-1 and adjacent HCV clones were used to produce the c100-3 antigen in recombinant yeast and an enzyme immunoassay for circulating antibodies to c100-3 was developed. This assay was used to generate data showing that circulating HCV antibodies could be found in most infectious blood donors and chronically infected NANBH patients (thus enabling the development of a first-generation blood screening test) and that HCV was the major cause of parenterally-transmitted NANBH around the globe (33).
Many applications were now possible, with the most urgent priorities being the development of blood tests to protect the blood supply, along with diagnostics and new, more effective drugs for the treatment of HCV patients. The first-generation blood test was available in 1990 and prevented the majority of transfusion-associated transmissions of HCV. This test and subsequent improved versions were also valuable in establishing HCV as a major cause of hepatocellular carcinoma (34) and other HCV-associated diseases including non-Hodgkin's B cell lymphoma (35). In addition, medical practitioners now had a means to monitor responses to antiviral therapy using assays for circulating HCV RNA. These assays also showed that a subset of patients spontaneously resolve acute infection and do not progress to chronic, persistent infection (36). Over the next few years, a series of more sensitive diagnostics for HCV antibodies and RNA were developed and these effectively eradicated the incidence of post-transfusion hepatitis (37), thus proving that HCV was the major cause of parenteral NANBH worldwide.
The isolation of clone 5-1-1 enabled the whole genome to be identified and sequenced. From this, we were able to discern highly conserved sequence motifs, indicating that HCV encoded its own unique protease, replicase and helicase enzymes (38). These enzymes are now very active targets of drug discovery programmes around the world. In addition, the molecular virology of HCV could now be studied from which many important accomplishments have been made from other groups including the demonstration of infectious HCV RNA (39), the generation of cell lines replicating the HCV genome (40), the propagation of HCV in vitro (27) and the identification of many other potential drug targets (41–43).
Finally, we first showed the feasibility of developing a vaccine against HCV by protecting vaccinated chimpanzees against homologous (44) and heterologous viral challenges (45). As such, the difficult and long effort and commitment from Chiron, the CDC and the dedicated field of viral hepatitis has been rewarded with the clear prospect of being able to effectively control HCV infection and disease.