fax: (32) 9-332-63-11.
Anti-CD81 antibodies can prevent a hepatitis C virus infection in vivo†
Article first published online: 28 JUL 2008
Copyright © 2008 American Association for the Study of Liver Diseases
Volume 48, Issue 6, pages 1761–1768, December 2008
How to Cite
Meuleman, P., Hesselgesser, J., Paulson, M., Vanwolleghem, T., Desombere, I., Reiser, H. and Leroux-Roels, G. (2008), Anti-CD81 antibodies can prevent a hepatitis C virus infection in vivo. Hepatology, 48: 1761–1768. doi: 10.1002/hep.22547
Potential conflict of interest: Drs. Paulson, Reiser, and Hesselgesser own stock in Gilead.
- Issue published online: 24 NOV 2008
- Article first published online: 28 JUL 2008
- Accepted manuscript online: 28 JUL 2008 12:00AM EST
- Manuscript Accepted: 21 JUL 2008
- Manuscript Received: 18 FEB 2008
- Ghent University via a Concerted Action Grant. Grant Number: 01G00507
- Belgian State via the Interuniversity Attraction Poles Program. Grant Number: P6/36–HEPRO
- Postdoctoral fellowship
- The Research Foundation Flanders (FWO–Vlaanderen)
The viral life cycle of the hepatitis C virus (HCV) has been studied mainly using different in vitro cell culture models. Studies using pseudoviral particles (HCVpp) and more recently cell culture–derived virus (HCVcc) suggest that at least three host cell molecules are important for HCV entry in vitro: the tetraspanin CD81, the scavenger receptor class B member I, and the tight junction protein Claudin-1. Whether these receptors are equally important for an in vivo infection remains to be demonstrated. We show that CD81 is indispensable for an authentic in vivo HCV infection. Prophylactic treatment with anti-CD81 antibodies completely protected human liver-uPA-SCID mice from a subsequent challenge with HCV consensus strains of different genotypes. Administration of anti-CD81 antibodies after viral challenge had no effect. Conclusion: Our experiments provide evidence for the critical role of CD81 in a genuine HCV infection in vivo and open new perspectives for the prevention of allograft reinfection after orthotopic liver transplantation in chronically infected HCV patients. (HEPATOLOGY 2008;48:1761–1768.)
More than 170 million people worldwide are infected with the hepatitis C virus (HCV). The lack of an effective HCV vaccine and the high cost and considerable side effects of the current standard therapies have a major impact on global public health. HCV is now a leading cause of liver cirrhosis, hepatocellular carcinoma, and liver transplantation.1
The precise mechanism by which HCV attaches to and enters host cells remains unclear. Pileri et al.2 identified the tetraspanin CD81 (TAPA-1) as a putative receptor for HCV. This 21-kDa surface molecule, composed of four transmembrane and two extracellular domains, efficiently bound recombinantly produced E2, one of the two viral envelope proteins. Experiments using pseudoviral particles (HCVpp) containing the HCV envelope proteins3–6 or the recently developed infectious HCV cell culture system7–11 have shown that CD81 expression on the target cell is essential for HCV infection of transformed hepatocytes in vitro.
Despite the evidence provided by numerous in vitro experiments, definitive proof of the role of CD81 in HCV infection in vivo is missing. It is well known that cell lines are not necessarily representative for the tissue or organ of origin. Moreover, other molecules, such as scavenger receptor class B member I,12, 13 C-type lectins L-SIGN and DC-SIGN,14–17 low-density lipoprotein receptor,18, 19 glycosaminoglycans,20, 21 and more recently, Claudin-122 and lipoprotein lipase,23 have been identified as putative HCV receptors or coreceptors. In addition, all these experiments have been performed using HCVpp or cell-culture–derived virus (HCVcc), which have different characteristics compared with natural HCV particles.24 More recently, Molina et al.25 reported that infection of primary hepatocyte cultures with serum-derived HCV could be inhibited by using anti-CD81 antibodies or via transduction of small interfering RNA that inhibited CD81 membrane expression.
We therefore investigated whether HCV infection could be prevented in vivo by administering anti-CD81 mAbs to human liver-uPA-SCID mice. uPA-SCID mice are immunodeficient mice that suffer from a severe, transgene-induced liver disease.26 These animals can be successfully transplanted with primary human hepatocytes, which will gradually repopulate the diseased mouse liver while retaining their normal functions.27, 28 These chimeric mice are to date the only hosts other than humans and chimpanzees that can be reproducibly infected with HCV.28, 29
Our experiments provide evidence for the critical role of CD81 in a genuine HCV infection.
Materials and Methods
Monoclonal Antibodies, Immunofluorescence, and Flow Cytometry.
The following anti-human CD81 monoclonal antibodies (mAbs) were used: JS-81 (mouse immunoglobulin G1 [IgG1]; BD Pharmingen, San Jose, CA), 184.108.40.206 (mouse IgG1; Chemicon, Temecula, CA), and 1D5 (mouse IgG1; Serotec, Dusseldorf, Germany). To verify that anti-CD81 antibodies bound to native CD81 expressed on the surface of human and primate lymphocytes and hepatocytes, the respective cell populations were analyzed by immunofluorescence and flow cytometry. Peripheral blood mononuclear cells were isolated via standard density gradient centrifugation, while primary hepatocytes were obtained from Bioreclamation (Hicksville, NY). Immunofluorescence and flow cytometry were performed as decribed.30 An isotype control antibody (clone 107.3; mouse IgG1, BD Pharmingen) was used in all experiments.
Cell Culture Studies.
To evaluate whether anti-CD81 antibodies were able to block entry of HCV, we performed cell culture infections with human Huh-7-Lunet hepatoma cells (kindly provided by Dr. R. Bartenschlager). Huh-7-Lunet is a cured replicon cell line that supports efficient HCV replication.31 Briefly, Huh-7-Lunet cells were preincubated for 1 hour with anti-CD81 antibodies (JS-81, 220.127.116.11, and 1D5) or isotype control IgG1 (107.3). All three antibodies were tested at concentrations of 1 μg/mL and 10 μg/mL. After preincubation, J6/JFH-RLuc virus was added to the cell cultures for 6 hours, and subsequently the viral inoculum was aspirated and the cells were washed. After 72 hours of culture, the cells were harvested and Renilla luciferase signal was quantified as a marker for viral entry and replication.
The neutralizing capacity of the JS-81 antibody was also evaluated in the HCV cell culture system.7 One day before infection with the chimeric virus H77/JFH1, 6 × 103 Huh7.5 cells were plated on a 96-well plate precoated with poly-D-Lysine (BD Biosciences, Erembodegem, Belgium). Both the H77/JFH1 virus and Huh7.5 cells were kindly provided by Dr. Charles Rice. One day later, the cells were preincubated at 37°C with JS-81 antibody (1 μg/mL or 25 μg/mL) or a control antibody (25 μg/mL). One hour later, 4.44 × 105 IU of cell culture–derived H77/JFH1 was added to each well. After 4 hours of incubation, the cells were washed and left for 3 days before staining. HCV-infected Huh7.5 cells were visualized using pooled serum of HCV-infected patients, in combination with a horesradish-peroxidase–labeled anti-human IgG (Bethyl Laboratories, Montgomery, TX). Staining was developed using True Blue peroxidase substrate (KPL, Gaithersburg, MD). Infection was scored by quantifying the amount of foci-forming units. One foci-forming unit was defined as one or more infected cells, separated from other infected cells by at least two uninfected cells.
Generation of Chimeric uPA-SCID Mice.
uPA-SCID mice were generated by repeatedly backcrossing B6SJL-TgN(Alb1Plau)144Bri mice onto CBySmn. CB17-PrkdcSCID mice. Both mouse strains were initially obtained from The Jackson Laboratories (Bar Harbor, ME). Animals homozygous for the uPA-transgene32 were transplanted with primary human hepatocytes, essentially as described.28 All animals received hepatocytes from the same donor. Five weeks after transplantation, mouse plasma was analyzed for the presence of human albumin using an Elisa technique (Bethyl Lab.). Animals with a human albumin level >1 mg/mL were considered successfully engrafted and were used for infection studies. All procedures were approved by the Animal Ethics Committee of the Faculty of Medicine at the Ghent University.
Infection and Prophylactic Treatment of Chimeric Animals.
Chimeric uPA-SCID mice were infused on day −1 and day +1 with different quantities of antibodies specifically targeting the human CD81 molecule (clone JS-81, BD Pharmingen). A control group was injected with an isotype control antibody (clone 107.3; BD Pharmingen). On day 0, all animals were infected with 104 IU HCV RNA of the genotype 1a reference strain H77C or with the same dose of the genotype 4a consensus strain ED43 (both generously provided by Dr. R. Purcell). These viral inocula were previously shown to induce HCV infection in all injected chimeric animals. All injections were performed intraperitoneally.
Postexposure Treatment of Chimeric Mice.
To evaluate whether it was possible to prevent an HCV infection in a postexposure setting, chimeric mice were first infected with the H77C virus, and administration of anti-CD81 antibodies was initiated 6 hours later. Additional antibody doses were injected 2 and 5 days later. All injections contained 400 μg of antibody.
Immediately before infection (injection of virus on day 0), the quantity of anti-CD81 antibodies present in mouse plasma was determined using flow cytometry (FACSCalibur, BD Bioscences). Huh7.5 cells, known for their high expression of CD81 on the membrane, were incubated for 30 minutes at 4°C with defined amounts of anti-CD81 or with a dilution series of mouse plasma. After extensive washing, the cells were incubated with a phycoerythrin-conjugated anti-mouse IgG antibody (clone X56, BD). The concentration of anti-CD81 antibodies present in the mouse plasma (μg/mL) was then quantified by comparing the median fluorescence signal of cells incubated with the mouse plasma with the median fluorescence of Huh7.5 cells incubated with known quantities of anti-CD81. The absolute amount of circulating anti-CD81 (μg) was calculated by assuming that the total blood volume of a mouse is approximately 80 mL/kg body weight. At selected time points after infection, HCV RNA in mouse plasma was quantified using the COBAS AmpliPrep/COBAS TaqMan HCV assay (Roche Diagnostics, Mannheim, Germany). This assay has a detection limit of 15 IU/mL, but because the mouse plasma samples had to be diluted 100-fold, the limit of detection was 1,500 IU/mL. At selected time points, mouse plasma was diluted only 25-fold to improve the limit of detection to 375 IU/mL.
Validation of Anti-CD81 Antibodies In Vitro.
Three anti-human CD81 mAbs (clones JS-81, 18.104.22.168, and 1D5; all mouse IgG1) and an isotype control antibody (clone 107.3) were used. Initial analyses using immunofluorescence and flow cytometry demonstrated that all three anti-CD81 mAbs specifically bound native CD81 on the surface of human and primate lymphocytes and hepatocytes (data not shown). Next, we examined whether these anti-CD81 antibodies could block entry of HCV into human Huh-7-Lunet hepatoma cells in vitro. Huh-7-Lunet cells were incubated for 1 hour with purified anti-CD81 antibodies or isotype control IgG1 at two concentrations (1 and 10 μg/mL) before J6/JFH-Rluc virus was added. The Renilla luciferase signal was quantified as a marker for viral entry and replication 72 hours later. As shown in Fig. 1A, all anti-human CD81 mAbs, but not the isotype control mAb, inhibited HCV entry in the JFH1 infectious culture system. The JS-81 and 22.214.171.124 antibodies were the most effective inhibitors, reducing the signal to background levels at both concentrations tested. Antibody 1D5 was a less potent inhibitor with 37.5% signal at 1 μg/mL but background levels at 10 μg/mL. The JS-81 mAb was selected for further in vivo studies.
The neutralizing effect of the JS-81 antibody was confirmed using the HCV cell culture system. Huh7.5 cells were incubated with 1 μg/mL or 25 μg/mL of JS-81 antibody or with 25 μg/mL isotype control IgG before H77/JFH1 virus was added. HCV-positive foci were counted 3 days later. The infection was inhibited in a concentration-dependent manner, and the highest dose of JS-81 antibody inhibited the viral infection by more than 95% (Fig. 1B).
Prophylactic Protection of Human Liver-uPA-SCID Mice.
After successful inhibition of HCV infection in cell culture, we wanted to explore whether administration of the JS-81 antibodies would prevent HCV infection in chimeric mice. On day −1 and day +1, groups of three mice were given 200 μg of IgG1 isotype control antibody (group 1), 50 μg of anti-CD81 mAb JS-81 (group 2), 200 μg of anti-CD81 mAb JS-81 (group 3), or 400 μg of anti-CD81 mAb JS-81 (group 4). On day 0, all animals were infected with 104 IU HCV RNA of the reference strain H77C. Immediately before injection of the virus, a plasma sample was taken to quantify the amount of circulating anti-CD81 antibodies. Compared with the animals receiving 50 μg of anti-CD81, a considerably larger amount of anti-CD81 was present in the plasma of the animals receiving a dose of 200 μg and 400 μg (Table 1).
|Group||Treatment (d−1, d+1)||Week−2||Day 0||Week 1||Week 2||Week 4||Outcome|
|HuAlb (mg/mL)||a-CD81 (μg)||a-CD81 (% of Injected)||Viral Strain||HuAlb (mg/mL)||HCV RNA (log10 IU/mL)||HuAlb (mg/mL)||HCV RNA (log10 IU/mL)||HuAlb (mg/mL)||HCV RNA (log10 IU/mL)||# Protected|
|200 μg||11.1||4.6||2.3%||H77C||5.1||4.34||7.6||7.34||Sample not available||1/2 (w4)|
|9.4||14.5||7.2%||H77C||5.5||<3.18||Animal died||Animal died|
One, 2, and 4 weeks after injection of the virus, plasma was taken to monitor the presence of viral RNA. As shown in Table 1 and in agreement with our previous observations (not shown), all three animals injected with the irrelevant control antibody were HCV-positive after 1 week, and viremia increased up to week 4 when it reached levels exceeding 107 IU/mL (Fig. 2A). When the animals were treated with two doses of 50 μg of anti-CD81, two animals were HCV positive after 1 week, while one animal was protected from infection (Fig. 2B). When we treated the animals with 200 μg of anti-CD81 only one out of three chimeric mice showed signs of infection at week 1. Unfortunately, one of the HCV-negative animals died before plasma could be obtained at week 2, but the other animal remained negative throughout the 4-week observation period (Fig. 2C). Complete protection from HCV infection was accomplished when the animals were treated with two 400-μg doses of anti-CD81 (Fig. 2D). At week 4, the plasma samples were retested at a lower dilution, and no HCV RNA could be detected (limit of detection was 375 IU/mL).
However, people almost never become infected with a monoclonal virus. Six major genotypes of HCV have been described, and within a chronically infected individual a pool of closely related variants coexist as a quasispecies. To corroborate the previous findings, we investigated whether anti-CD81 treatment was able to prevent an infection with another HCV strain. Following the same treatment schedule as in group 4 of the previous experiment (400 μg anti-CD81 on day −1 and day 1), the animals were now injected with 104 IU of the HCV strain ED43 (genotype 4). As a control, two animals received the same dose of isotype-matched irrelevant control antibody and were injected on day 0 with HCV ED43. As shown in Fig. 2E, HCV RNA could be detected in the plasma of both control animals from week 1 on, and viral titers rapidly increased up to week 3. Unfortunately, one of the anti-CD81–treated animals died prematurely, but it was still HCV-negative after 2 weeks (<375 IU/mL [Table 2]). A second anti-CD81–treated animal died after 6 weeks but remained HCV-negative (<375 IU/mL [Table 2]) The two remaining treated chimeric animals could be monitored until week 7 after injection of the virus, and they remained HCV-negative throughout (Fig. 2F). Considering the dilution of the plasma samples, HCV RNA was below 375 IU/mL on week 7 (Table 2).
|Group||Treatment (d−1, d+1)||Day −3||Day 0||Week 1||Week 2||Week 3||Week 4||Week 5||Week 6||Week 7||Outcome|
|HuAlb (mg/mL)||a-CD81 (μg)||a-CD81 (% of Injected)||Viral Strain||HCV RNA (log10 IU/mL)||HuAlb (mg/mL)||HCV RNA (log10 IU/mL)||HCV RNA (log10 IU/mL)||HuAlb (mg/mL)||HCV RNA (log10 IU/mL)||HCV RNA (log10 IU/mL)||HuAlb (mg/mL)||HCV RNA (log10 IU/mL)||HCV RNA (log10 IU/mL)||# Protected|
|400 μg||3.6||21.0||5.3%||ED43||<2.70||5.7||<2.57||Animal died||Animal died||Animal died||Animal died||Animal died||3/3 (w6)|
Postexposure Anti-HCV Therapy.
In addition to the prophylactic treatment, we also wanted to explore whether it would be possible to prevent an HCV infection in a postexposure setting. One chimeric mouse was infected with 104 IU HCV of strain H77C and was injected with 400 μg of anti-CD81 antibody 6 hours later. Additional antibody doses (400 μg) were administered 2 and 5 days later. As shown in Fig. 3, HCV RNA became detectable in the mouse plasma after only 1 week, and at week 2 the viral load reached levels above 107 IU/mL. There was no significant difference in viremia compared with the control animals of group 1 (Fig. 2A). In view of this negative result, no further postexposure experiments were performed.
Although cell culture models have been very useful to study viral entry and replication, they have some important limitations. First, the Huh7 cells that are frequently used in the in vitro systems are hepatoma cells that differ in several aspects from primary human hepatocytes. Second, the composition of HCVpp and even HCVcc differs from that of HCV isolated from serum or plasma of chronically infected humans, chimpanzees, and chimeric mice with human liver grafts.24 Molina et al. recently showed that CD81 was indispensable for infection of cultured primary human hepatocytes with plasma derived HCV of different genotypes.25 However, while a soluble form of CD81 (solCD81) was capable to prevent infection of Huh7.5 cells by HCVcc, it was not effective in preventing an infection of cultured primary hepatocytes by serum-derived virus. In addition, solCD81 is much more effective in blocking HCVcc infection in Huh7.5 cells than in primary hepatocytes. Therefore, data generated in vitro must be interpreted with caution and whenever possible validated with in vivo experiments.
We show that CD81 is a critical receptor for HCV infection in vivo. Prophylactic injection of monoclonal anti-CD81 antibodies could prevent infection of human liver-uPA-SCID mice. Our results show an association between the amount of injected antibody and the protective efficacy of this intervention. We assume that the amounts of antibody injected in animal groups 2 (50 μg) and 3 (200 μg) were not sufficient to block all CD81 molecules on the surface of the liver, leaving human hepatocytes susceptible to infection. These infected hepatocytes could then serve as a continuous source of new infectious virions. Quantification of the amount of anti-CD81 present in the plasma 24 hours after injection supports this hypothesis. Although the animals of groups 3 and 4 received only a four- to eight-fold higher dose of anti-CD81 mAb, their plasma contained a more than 100-fold higher concentration of circulating anti-CD81 compared with group 2. This may indicate saturation of the liver. Importantly, using the highest dose of anti-CD81, we were able to completely prevent HCV infections with viruses of genotype 1 (H77C) and genotype 4 (ED43). This confirms the data of Scheel et al.,11 who showed that in vitro pretreatment of Huh7.5 cells with anti-CD81 antibodies efficiently prevented infection with chimeric HCV ED43/JFH1. Our data suggest that CD81 is also an essential receptor for HCV infection in vivo.
Once an infection occurred, no significant difference in viremia was observed between anti-CD81 treated and control animals (irrelevant antibody). This suggests that the infection blockade induced by the anti-CD81 antibodies remains intact for only a short period of time. Probably newly produced CD81 molecules quickly reappear on the surface of the hepatocytes, rendering them again susceptible to HCV infection. In addition, this can also be explained by an alternative transmission route. It was recently described that HCV can move from cell to cell via a CD81-independent mechanism.33 The ineffectiveness of a postexposure anti-CD81 therapy also corroborates this hypothesis. On the other hand, when a sufficient amount of antibody is used, complete in vivo protection is possible. This is interesting because the CD81 molecule is thought to be involved only late in the sequence of viral attachment.21, 34
Our results strongly support the use of CD81 as a clinical target for HCV prevention, especially in the context of orthotopic liver transplantation. Unfortunately, allograft reinfection is almost universal in HCV-infected patients, occurs within hours after reperfusion, and is followed by an accelerated disease progression.35, 36 This may be prevented by saturating the donor liver with anti-CD81 antibodies before and in the first days following graft reperfusion. If this blockade could be sustained until all circulating virus has been cleared, reinfection may be prevented.
The binding between CD81 and HCV E2 occurs through interaction of the major extracellular loop of CD812 with a conformational structure within the receptor-binding domain of E2, spanning amino acids 384-661 of the HCV polyprotein.37–39 Whereas residues 162, 182, 184, and 186 probably form an E2-binding site on the LEL of CD81,40 the two hypervariable regions, HVR1 and HVR2, and the intergenotypic variable region of E2 are completely dispensable for CD81 binding.41 Small molecules could be developed that specifically inhibit the interaction between CD81 and the viral envelope proteins.42–45 Compounds with this capacity may be a more economical alternative for monoclonal anti-CD81 antibodies. Preferentially, these small molecules should be specific in their action and only inhibit the interaction between the virus and the receptor without interfering with the numerous other functions of CD81. This would reduce the risk of unwanted side effects that such agents may cause.
This report demonstrates for the first time that the tetraspanin CD81 is an indispensable component of the HCV receptor complex for a primary in vivo infection. Our observations may lead to new strategies to prevent HCV reinfection after liver transplantation. Although high concentrations of antibody are needed to achieve full prophylactic protection, small molecules that specifically inhibit the interaction between the virus and CD81 may provide an alternative. Targeting the receptor instead of the surface proteins of the virus, as with neutralizing antibodies, can be a more effective strategy considering the high variability of HCV and versus the conserved nature of CD81.
In addition to CD81, other molecules that constitute the HCV receptor complex may also represent a target to prevent infection. Recently, antibodies targeting scavenger receptor class B member I have been described to prevent HCV infection in vitro.46 It will be most interesting to evaluate whether these antibodies are also effective in vivo. Apart from its role in HCV entry, CD81 is also an essential receptor for infection of hepatocytes by Plasmodium sporozoites.47, 48 Targeting CD81 may therefore also be a useful strategy to prevent or treat malaria.
We wish to thank L. Verhoye and P. Premereur for excellent technical assistance. We are also grateful to Dr. C. Rice for providing Huh7.5 cells and H77/JFH1 virus, Dr. R. Purcell and Dr. J Bukh for providing infectious H77C and ED43 virus and Dr. R. Bartenschlager for kindly providing Huh7-Lunet cells and J6/JFH-RLuc virus.
- 1Hepatitis C and liver transplantation. 2005; 436: 973–978..