Human monoclonal antibodies that react with the E2 glycoprotein of hepatitis C virus and possess neutralizing activity†
Article first published online: 25 OCT 2005
Copyright © 2005 American Association for the Study of Liver Diseases
Volume 42, Issue 5, pages 1055–1062, November 2005
How to Cite
Schofield, D. J., Bartosch, B., Shimizu, Y. K., Allander, T., Alter, H. J., Emerson, S. U., Cosset, F.-L. and Purcell, R. H. (2005), Human monoclonal antibodies that react with the E2 glycoprotein of hepatitis C virus and possess neutralizing activity. Hepatology, 42: 1055–1062. doi: 10.1002/hep.20906
Potential conflict of interest: Nothing to report.
- Issue published online: 25 OCT 2005
- Article first published online: 25 OCT 2005
- Manuscript Accepted: 23 AUG 2005
- Manuscript Received: 3 MAR 2005
- La Ligue Nationale Contre le Cancer, “agence Nationale pour la Recherche sur le SIDA et les Hépatites Virales” (ANRS)
- European Community. Grant Number: LSHB-CT-2004-005246
Active and/or passive immunoprophylaxis against hepatitis C virus (HCV) remain unachieved goals. Monoclonal antibodies might provide one approach to protection. We derived human monoclonal antibodies from the bone marrow of a patient with a well-controlled HCV infection of 22 years duration. Five distinct antibodies reactive with the E2 glycoprotein of the homologous 1a strain of HCV were recovered as antigen-binding fragments (FAbs). They demonstrated affinity constants as high as 2 nanomolar. “Neutralization of binding” titers paralleled the affinity constants. All five FAbs reacted with soluble E2 protein only in nonreducing gels, indicating that the relevant epitopes were conformational. The FAbs could be divided into two groups, based on competition analysis. Three of the FAbs neutralized the infectivity of pseudotyped virus particles (pp) bearing the envelope glycoproteins of the homologous HCV strain (genotype 1a). The three FAbs also neutralized genotype 1b pp and one also neutralized genotype 2a pp. In conclusion, one or more of these monoclonal antibodies may be useful in preventing infections by HCV belonging to genotype 1 or 2, the most medically important genotypes worldwide. (HEPATOLOGY 2005;42:1055–1062.)
Hepatitis C virus (HCV), a positive strand RNA virus, is a member of the family Flaviviridae. HCV is a major cause of blood-borne acute and chronic hepatitis. Chronic infection with HCV is associated with a significant risk of progression to cirrhosis and hepatocellular carcinoma. Antiviral therapy with pegylated α-interferon and ribavirin, the current optimal therapeutic regimen, is successful in only about 50% of treated patients. Therefore, development of new antiviral agents or other therapeutic modalities is a major focus of present investigations. The short-term prospects for the generation of an efficacious vaccine to prevent HCV or to control HCV postinfection are not encouraging. However, it is believed that anti-HCV immune globulin preparations, similar to those used successfully to treat hepatitis B virus infections, might be useful in preventing or controlling HCV infections.
Immune globulin prepared from unscreened donors or from selected patients with chronic HCV infection has prevented hepatitis C in recipients when administered before exposure to the virus.1–6 Such protection has been linked to the presence of antibodies that can neutralize pseudotyped virus particles (pp) bearing the envelope glycoproteins of HCV.7 Such antibodies can be broadly reactive, neutralizing pp derived from different subtypes and genotypes of HCV.8–10 Human monoclonal antibodies provide an attractive alternative to polyclonal hyperimmune globulin for immunoprophylaxis and immunotherapy, since monoclonal antibodies can be more readily standardized. Human monoclonal antibodies to the envelope glycoprotein E2 of HCV have been prepared by immortalization of blood-derived B cells or by the generation of combinatorial libraries of antibody-binding sites derived from messenger RNA extracted from bone marrow aspirates.11–13 Such antibodies have been extensively studied, but little is known about their ability to neutralize HCV.14–21
We describe here the preparation of a combinatorial library of antibody gene fragments derived from the bone marrow of patient H, an asymptomatic chronic carrier of HCV who was infected 28 years ago.22 Patient H was shown previously to possess serum neutralizing antibody to HCV as determined by in vitro and in vivo neutralization assays.5, 23 Furthermore, patient H has been the source of well-characterized HCV that has been used for numerous transmission studies in chimpanzees and was the source for the first two infectious complementary DNA (cDNA) clones of HCV.24, 25 In this first analysis of the combinatorial antibody library, five HCV E2-specific antigen-binding fragments (FAbs) were generated. Their biochemical and immunological characterization is described.
Patients and Methods
The donor, patient H, was a 60-year-old white male who developed hepatitis C following blood transfusion in 1977. Except for his initial acute hepatitis and one recurrence of serum liver enzyme elevations early in the chronic phase of infection, the patient has had essentially normal enzyme values despite continuous relatively high titers of HCV viremia as measured by reverse-transcription polymerase chain reaction (RT-PCR). Following informed consent in writing, the patient underwent voluntary bone marrow aspiration in 1999. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the Institutional Review Board of the Clinical Center, NIH.
Construction of Phage Library.
Panning and ELISA Reagents.
In all panning experiments and enzyme-linked immunosorbent assays (ELISA) CHO cell-expressed recombinant HCV E2 glycoprotein28 was diluted to 1.0 μg/mL in 50 mmol/L sodium carbonate buffer (pH 9.6), and coated onto EIA/RIA A/2 plates (Costar, Cambridge, MA) overnight at 4°C. FAbs were detected with goat anti-human IgG (FAb-specific) antibody (Pierce, Rockford, IL). This was coated on microtiter wells at a dilution of 1:1000, in 50 mmol/L sodium carbonate buffer (pH 9.6), as described.
Screening of the combinatorial library was carried out as previously described.26–29 Phage were selected by three rounds of panning on E2-coated ELISA wells. After amplification of the selected library, the phagemid DNA was extracted and modified by restriction enzyme digestion to remove the bacteriophage coat protein III-encoding region of the phage.30 The phagemid DNAs were religated and transformed into E.coli XL-1 Blue (Stratagene, La Jolla, CA) to produce soluble FAbs. Single colonies were picked, FAb production was induced as previously described,31 and the bacterial supernatants were tested by ELISA for reactivity against E2 and for the presence of FAb.
FAb Production, Purification, and Biotinylation.
Bacterial culture and FAb fragment purification were carried out as described.31 Protein concentrations were determined by both dye-binding assay (Bio-Rad, Hercules, CA) and by optical density at A280nm (using the extinction coefficient of 1.4 optical density units equivalent to 1.0 mg/mL−1). The FAb purity was determined by polyacrylamide gel electrophoresis with colloidal Coomassie blue staining (Sigma, St. Louis, MO). The purified FAbs were diluted in sodium bicarbonate buffer (pH 9.0), and biotinylated at 4°C as per the manufacturer's protocol (Pierce). After biotinylation, the FAbs were dialyzed against phosphate-buffered saline (PBS [pH 7.4]) overnight at 4°C, and concentrated in Centricon-30 concentrators (Amicon, Enschede, The Netherlands) as required.
ELISA Analysis of FAb Specificity.
Recombinant E2 was coated on ELISA microtiter plates at 1.0 μg/mL-1 and non-specific proteins (thyroglobulin, lysozyme, glyceraldehyde-3-phosphate, chicken albumen and cytochrome C [Sigma]) were coated at 10.0 μg/mL−1. ELISAs were performed as previously detailed.27
Restriction Digestion Analysis and Nucleic Acid Sequence Analysis of E2-Specific FAb Clones.
For Bst N1 (New England BioLabs, Beverly, MA) fingerprinting, 1 microgram of plasmid DNA was digested with 1 U of enzyme overnight at 60°C. The restriction digests were analyzed on a 3% agarose gel. Nucleic acid sequencing was performed with the ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction kit by using Ampli-Taq DNA Polymerase (Perkin-Elmer, Wellesley, MA) and the sequencing primers HC1 and HC4 for the γ1-heavy chain and LC1 and LC4 for the κ-chain.31 Sequences were analyzed with the Sequencher and MacVector (Oxford Molecular Group) software packages. Sequence similarity searches were performed with the V-BASE program, which is a compilation of all of the available human variable segment Ig germ line sequences.32
Competition ELISA for Affinity Determination.
The affinity (equilibrium dissociation constant [Kd]) of the FAbs was determined by competition inhibition ELISA.21, 33 Briefly, log10 dilutions of FAb were titrated on E2-coated wells, and the dilution which substantially reduced the binding of the FAb was used in the competition ELISA. This concentration of FAb was incubated for 2 h at 37°C with decreasing log10 concentrations of soluble E2 in E2-coated wells. The plates were washed 4 times with PBS/Tween-20, and bound FAb was detected using anti-human IgG (FAb-specific) alkaline phosphatase-labeled secondary antibody (Sigma) at a dilution of 1:5000. The percent reduction in A405nm value was plotted and the 50% inhibition (I50) value was extrapolated. The concentration of E2 at the I50 values represented the affinity (Kd) of the antibody for the antigen.
CD81–E2 Blocking ELISA.
This assay was performed essentially as described in Forns et al.34 The FAbs were titrated for their ability to block the CD81-E2 interaction, with an irrelevant FAb, HEV#31,27 used as the negative control and a serum (H79) taken from patient H 2 years after onset of his hepatitis, as a positive control. The dilution of FAb or serum that blocked by 50% the binding of CD81 to E2 was calculated from titration curves.
Epitope Mapping by Indirect Competition ELISA.
Biotinylated FAbs were titrated on E2-coated wells to determine the dilution that gave an optical density reading of approximately 1.0 at A405nm, and did not saturate the antigen that was coated to the plate. For the competition assay, ten-fold dilutions of competing unlabeled FAbs were incubated in E2-coated wells for 1 hour at 37°C, and then washed 4 times with PBS-Tween 20. A single dilution of the biotinylated FAb was incubated in all wells for 1 hour at 37°C. After four washes with PBS-Tween 20, the binding of the biotinylated FAb was detected with streptavidin-alkaline phosphatase (Pierce).
Recombinant E2 protein was heated in 2x Laemmli buffer containing SDS in the presence or absence of the reducing agent dithiothreitol, and electrophoresed in two single-well 10% polyacrylamide gels (Novex, Invitrogen, Carlsbad, CA). Electrophoretic transfer of the protein to nitrocellulose membranes was carried out at 126 mA for 1 hour at 4°C. The membranes were blocked for 30 minutes with 5% skim milk in PBS (M-PBS) and then loaded into a multiscreen apparatus (Bio-Rad). Equal concentrations of purified HCV#1, 4, 7, 12 or HCV#13 or a 1:100 dilution of H79 serum antibodies in M-PBS were loaded into separate wells in the multiscreen apparatus prior to overnight incubation with gentle rocking at 4°C. After six washes of 10 minutes each, an anti-human IgG (FAb-specific, horseradish peroxidase-labeled [Pierce]) secondary antibody was added at a dilution of 1:5,000 in M-PBS. After 1 hour, the blots were washed, Supersignal ECL substrate (Pierce) was added and the blot imaged using an X-ray film developer.
Production and Neutralization of Retroviral Pseudoparticles Bearing HCV Envelope Glycoproteins.
Pseudoparticles were generated as previously described.35 DNA sequences of envelope E1 and E2 glycoproteins of HCV were isolated from patients infected with different genotypes of the virus as described.9 Briefly, viral RNA was isolated from 100 μL of serum and virus-specific cDNA was generated with genotype-specific primers. From these cDNA preparations, polymerase chain reaction products representing amino acid residues 170 – 74624 of the HCV polyprotein were cloned into the pCR3.1 vector (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Clones were sequenced as described.
HCV pp were produced as described from 293T cells cotransfected with a murine leukemia virus (MLV) Gag-Pol packaging construct, an MLV-based transfer vector encoding a luciferase marker protein and the E1/E2 expression constructs representing the different HCV genotypes.9 Infection of target cells (Huh-7) was performed as previously described.36
Isolation of E2-Specific FAbs and Sequence Analysis.
The FAb-phage library was selected against E2-coated ELISA wells and 168 clones were analyzed for FAb expression and E2-specificity by ELISA. Nineteen clones produced E2-specific FAbs.The restriction enzyme Bst NI37 was used to screen for different γ1-heavy chain sequences among the FAb clones. Six different digestion patterns were observed (data not shown). Sequence analysis identified five unique γ1-heavy chains. These were represented by clones HCV#1, HCV#4, HCV#7, HCV#12, and HCV#13 (Fig. 1).
We attempted to determine the specific germ-line origin of the five clones by conducting a sequence similarity search of all the known human immunoglobulin genes. The findings are summarized in Table 1.
|MAb||VH Family||D Segment||JH Segment||VK Family||JK Segment|
An indirect competition assay was performed to determine the relative topology of the epitopes recognized by the five FAbs on the E2 protein (Fig. 2). Based upon these data the FAbs were divided into at least two groups: (1) FAbs HCV#4, and HCV#7, which competed with each other for binding to soluble E2; and (2) FAbs HCV#1, HCV#12, and HCV#13, which did not compete with FAbs HCV#4 or HCV#7. FAbs HCV #12 and HCV #13 were not further characterized by indirect competition.
The affinities of the FAbs for recombinant soluble E2 were determined by competition inhibition ELISA. The concentration of free E2 required to inhibit antibody binding by 50% was equivalent to the equilibrium dissociation constant (Kd). The Kd values for the FAbs are summarized in Table 2. The Kd value for HCV#12 could not be determined and is probably >10−7 mol/L, which is the cutoff for sensitivity of this assay.
|Fab||Kd (nmol/L)||NOB Titer (μg/ml)*|
CD81–E2 Blocking (Neutralization of Binding) Assay.
The FAbs were titrated for their ability to block the binding of a recombinant CD81-thioredoxin fusion protein to recombinant soluble E2 in an ELISA format. Four of the five FAbs blocked E2 binding to this recombinant form of the putative HCV cell receptor, CD81 (Table 2). HCV#12 did not block E2-CD81 interaction at the highest concentration of FAb attainable (20 μg/mL). The efficacy of the FAbs in blocking E2 binding to CD81 essentially mirrored the affinity of the FAbs for soluble recombinant E2 protein, i.e., the highest affinity FAbs were the most efficacious (Table 2).
Western Blotting With HCV-Specific FAbs.
A Western blotting assay was performed to determine the nature of the epitope recognized by each of the five FAbs (i.e., linear or conformational epitope). All of the FAbs and serum H79 failed to react with recombinant E2 protein that had been reduced prior to loading on the gel. However, all of the FAbs and serum H79 recognized nonreduced recombinant E2 protein (data not shown), suggesting that they are all directed to conformational epitopes on the E2 glycoprotein.
Neutralization of Pseudotyped Particles Bearing the E1 and E2 Glycoproteins of Different HCV Genotypes.
Because only FAbs #1, #4 and #7 appeared to have the potential for utility, only these three FAbs were tested for their ability to neutralize pseudoparticles. The three FAbs and a negative control FAb directed against hepatitis E virus27 were diluted in 5-fold increments and tested for their ability to neutralize pseudoparticles bearing the HCV E1 and E2 envelope glycoproteins of HCV genotypes 1a, 1b, 2a, 3a, 4a, and 5a. We were not able in this series of experiments to construct a viable genotype 6 pseudoparticle. Neutralization was performed as described previously with serial dilutions of the four FAbs at concentrations ranging from 50 down to 0.08 μg/mL. All four FAbs were tested in one test against the same pseudoparticle and most tests were performed at least twice. The results were analyzed against the results obtained with the negative control FAb and expressed as percent neutralization.
As seen in Fig. 3 and Table 3, the three HCV-specific FAbs all neutralized the pseudoparticles bearing the homologous genotype 1a HCV envelope glycoproteins, although they varied in potency. Similarly, all three of the HCV-specific FAbs neutralized pseudoparticles bearing the envelope glycoproteins of a genotype 1b virus. The potency of the three FAbs in neutralizing genotype 1b pseudoparticles paralleled their potency in neutralizing the genotype 1a construct (#4 > #7 > #1). Only FAb #4 neutralized genotype 2a pseudoparticles and none of the FAbs neutralized pseudoparticles bearing envelope glycoproteins of genotype 3a, 4a or 5a viruses at the FAb concentrations tested, although FAb #4, in particular, did demonstrate some neutralizing activity when tested at a concentration of 50 μg/mL, but the value was consistently below 50%. We attempted to test the FAbs for neutralizing activity at a concentration of 250 μg/mL, but the negative control FAb inhibited infection of the pseudoparticles at that concentration (data not shown). Thus, the most potent of the three HCV FAbs, #4, neutralized pseudoparticles bearing HCV glycoprotein of genotypes 1a, 1b and 2a in the low to high nanomolar range.
In this first examination of a combinatorial library from patient H we have generated five monoclonal antibodies to the E2 glycoprotein. The equilibrium dissociation constant for recombinant soluble E2 was determined for four of the five FAbs. They ranged from the highest, HCV#4 at 1.9 nmol/L, to the lowest, HCV# 13 at 40.5 nmol/L. The constant for HCV#12 could not be determined. In Western blot experiments all five FAbs reacted with E2 that had been electrophoresed on nonreducing polyacrylamide gels, but not E2 that had been electrophoresed under reducing conditions, indicating that they were directed to conformational epitopes.
At the present time in vitro propagation of HCV is extremely limited. Therefore direct in vitro neutralizing activity could not readily be determined for these monoclonal antibodies. In the absence of such a cell culture system, surrogate neutralization assays have been devised. Rosa et al.38 demonstrated that blocking of recombinant E2 binding to an HCV susceptible human T-cell line, MOLT-4, correlated with protection from infection in HCV-vaccinated chimpanzees. This model for HCV binding was modified following the discovery of a putative cellular receptor for HCV, the tetraspanin molecule, CD81.39 Both recombinant E2 and HCV virions could be inhibited from binding to recombinant CD81 by the same chimpanzee immune sera that blocked in the Rosa et al. assay. We have performed a similar assay using an ELISA format to measure the blocking of the CD81-E2 interaction (NOB-ELISA). This assay was previously shown to give an equivalent neutralization of binding (NOB) titer to that determined using the Rosa et al. assay for the patient serum sample, H79.34 The H79 serum sample also had in vitro neutralizing activity when tested in a chimpanzee.5
Using the NOB-ELISA format, four of the five FAbs inhibited recombinant E2 binding to recombinant CD81. Generally speaking, the ranking of the NOB-ELISA titers could be predicted by the antibody affinity for the E2 protein. Only those FAbs that had affinities in the nmol/L range blocked the binding of CD81 to E2. In this NOB-ELISA, the affinity of the interaction between CD81 and E2 was 3.2 μmol/L.40 Therefore, the data obtained with our FAbs are in agreement with the hypothesis that NOB-positive antibodies must have an affinity value for E2 that is greater than the affinity of CD81 for E2.41 Our data agree with those for other published FAbs where antibody affinity and NOB activity have been examined.12, 13, 18, 19 However, it should also be pointed out that not all E2-specific monoclonal antibodies have NOB activity.11, 18, 19
The most recently described surrogate test for neutralization of HCV is the neutralization of pp bearing HCV envelope glycoproteins.35, 42 We tested the three FAbs with the greatest potential for utility, #1, #4, and #7. All three FAbs neutralized pseudoparticles bearing the homologous HCV envelope glycoproteins.
Little is known about how the different genotypes of HCV relate to each other in terms of serotypes for the envelope glycoproteins, again due to the lack of a robust in vitro neutralization test. However, a previous report showed that greater than 90% of serum from a panel of HCV-infected individuals had antibodies to E2 proteins from genotypes 1a, 1b, 2a, and 2b.14 and these results have been confirmed and extended recently.8–10 In terms of therapeutic antibodies for preventing HCV infection or treating hepatitis C, antibodies with the broadest specificity for E2 from different HCV genotypes are obviously desirable. That patients can produce broadly reactive antibodies in response to monotypic infection has been shown recently for patient H.8, 10 Thus, convalescent sera taken from patient H reacted with all six major genotypes as measured by neutralization of pseudoparticles, although such neutralization was much less potent against genotype 2 and 3 pseudoparticles than against genotypes 1, 4, 5, and 6.10 The significance of this broad reactivity and the identity of the epitopes involved remains to be determined, but these data suggest that the epitopes detected by monoclonal and polyclonal antibodies derived from chronically infected individuals late in their disease may be highly conserved and this bodes well for the development of passive and active immunoprophylaxis against hepatitis C.
Although a number of human monoclonal antibodies to HCV have been described, evidence for their ability to neutralize HCV is very limited. Some have been shown to neutralize pseudotyped vesicular stomatitis viruses bearing HCV envelope proteins,20 but others have raised serious questions about the validity of such tests.43 Other monoclonal antibodies have been shown to react with virus-like particles consisting of the HCV core protein and the E1 and E2 envelope proteins expressed from insect cells.15 However, there is no direct evidence that reaction with these virus-like particles correlates with neutralization of HCV. One group of well-studied monoclonal antibodies11, 14–16 has been tested for their ability to neutralize the same pseudoparticles as employed herein, but those monoclonal antibodies were much less potent for neutralization of the pseudoparticles than the FAbs we describe herein: the former complete monoclonal antibodies either failed to neutralize at saturating concentrations of 20 μg/mL or, at best, marginally neutralized, with maximum neutralization values of only ≈60% at that concentration of antibody.44 In contrast, FAbs #4, #7, and #1, described herein, neutralized virtually 100% of the pseudoparticles at comparable concentrations of antibody and displayed 50% neutralization titers at concentrations of 0.3, 1 and 4 μg/mL, respectively. It is anticipated that complete monoclonal antibodies would have even greater neutralizing capacity. Thus, failure of the complete monoclonal antibodies to neutralize pseudoparticles fully was not the result of heterogeneity of E2 glycosylation as suggested,44 because our two most potent FAbs neutralized virtually 100% of the pseudoparticles (Fig. 3).
The reasons for the marked discrepancy in potency and breadth of reactivity of human monoclonal antibodies generated by similar techniques is not known but may be related to the duration of chronicity in the patients who have yielded such antibodies. Although the interval from infection to obtaining the monoclonal antibodies is not described, indeed is not known in many instances, the time of infection of patient H has been very well documented, as has the strength and breadth of his antibody response to this well-controlled HCV infection. Thus, a strong and broad antibody response was not detected until more than 10 years after onset of the infection and the titer and breadth of response to other genotypes has continued to increase until the present.8, 10 The bone marrow aspirate was taken from patient H 22 years after onset of his disease and well after a strong, broadly reactive antibody response was in place. It will be interesting to see if monoclonal and polyclonal antibodies in other patients with comparably long and well-controlled HCV infection are similarly potent in neutralization assays.
Based on analyses of the polyclonal antibody response by patient H to all six major genotypes, it is likely that potent monoclonal antibodies directed to each of the major genotypes could be recovered from this patient if appropriately expressed E2 envelope glycoproteins were available from each of the HCV genotypes.
There are two primary areas where a strong, broadly reactive monoclonal antibody to HCV would have clinical relevance. The first is in postexposure prophylaxis for accidental needlestick or other percutaneous or mucosal exposure. The second and most relevant application would be in the prevention or containment of recurrent HCV infection in the liver transplant setting. In essence, this would simulate the established efficacy of hepatitis B immune globulin in the prevention of posttransplant hepatitis B. In HCV infection, reinfection of the transplanted liver is universal and 20% to 30% of such recurrences result in accelerated progression of fibrosis that can lead to cirrhosis in as little as 5 years. Through a cohort effect, as the large population infected with HCV 25 to 40 years ago now reaches the age when the untoward sequelae of HCV infection are more common, the demands for liver transplantation will increase considerably. Thus, new therapeutic modalities that enhance transplant and patient survival are sorely needed. Passively administered neutralizing antibody, probably with concomitant antiviral therapy, offers a viable and promising option to suppress/eradicate HCV before it infects the naïve transplanted liver. Given this goal, there is need now to produce these monoclonal antibodies in sufficient quantity that their efficacy can be tested in vivo using the chimpanzee model or small-scale human trials.
The laboratory of François-Loïc Cosset receives support from La Ligue Nationale Contre le Cancer, “agence Nationale pour la Recherche sur le SIDA et les Hépatites Virales” (ANRS) and the European Community (LSHB-CT-2004-005246).”
The authors thank Kristin Morrow for her excellent technical assistance and Dr. Isa Mushahwar (Abbott) for the gift of the recombinant E2 glycoprotein.
- 3A trial for prevention of serum hepatitis with intravenous human γ-globulin (venoglobulin). J Jpn Soc Blood Transfus 1980; 24: 2–8., .