| |
Broadly neutralizing antibodies abrogate established hepatitis C virus infection....."perhaps a vaccine"
|
| |
| |
Download the PDF here
"The largest takeaway is that HCV can rapidly be cured if the infectious cycle is interrupted," de Jong told Healthline. "This goes against the current dogma that HCV is not [damaging] to the liver cells and happily replicates in the liver without killing either its host cell or being cleared by its host cell."
Sci Transl Med 17 September 2014
Ype P. de Jong et al.
Ype P. de Jong,1,2* Marcus Dorner,2 Michiel C. Mommersteeg,2 Jing W. Xiao,2 Alejandro B. Balazs,3
Justin B. Robbins,4 Benjamin Y. Winer,5 Sherif Gerges,5 Kevin Vega,2 Rachael N. Labitt,2
Bridget M. Donovan,2 Erick Giang,4 Anuradha Krishnan,6 Luis Chiriboga,7 Michael R. Charlton,6
Dennis R. Burton,3,4 David Baltimore,8 Mansun Law,4 Charles M. Rice,2 Alexander Ploss2,5*
"Recent findings demonstrate that broadly nAbs can also efficiently suppress viremia in humanized mice (17, 44) and Rhesus monkeys infected with HIV or SHIV (45, 46), respectively. Our data extend this work to hepatotropic viral infections.
.........Our findings are the first to show that nAbs can cure liver chimeric mice (of HCV).....Seventeen to 25 days after infection, mice (n = 3 for AR pool, n = 2 for b6 controls) were treated with 500 μg of each nAb (AR3A/AR3B/AR4A) or 1.5 mg of control IgG (b6) every 3 days, which resulted in high and stable IgG levels.......nAb treatment suppressed serum infectivity to below the limit of detection within 1 day......Notably, HCV (RNA copy numbers decreased in all mice that received the AR pool and fell below the limit of detection between 5 and 11 days after starting nAb treatment.....it will be important to extend these observations to other HCV isolates and genotypes, for example, HCV genotype 3......our data imply that HCV must continuously reinfect new hepatocytes to sustain viremia......Whether these nAbs will become useful in HCV-infected patients will largely depend on the ability of upcoming direct-acting antivirals (1) to eradicate this infection in every patient.......It is therefore feasible that passive nAb transfer may be an adjunct therapeutic modality in a subset of HCV-infected individuals who cannot be cured by upcoming direct-acting antivirals."
Abstract
In most exposed individuals, hepatitis C virus (HCV) establishes a chronic infection; this long-term infection in turn contributes to the development of liver diseases such as cirrhosis and hepatocellular carcinoma. The role of antibodies directed against HCV in disease progression is poorly understood. Neutralizing antibodies (nAbs) can prevent HCV infection in vitro and in animal models. However, the effects of nAbs on an established HCV infection are unclear. We demonstrate that three broadly nAbs-AR3A, AR3B, and AR4A-delivered with adeno-associated viral vectors can confer protection against viral challenge in humanized mice. Furthermore, we provide evidence that nAbs can abrogate an ongoing HCV infection in primary hepatocyte cultures and in a human liver chimeric mouse model. These results showcase a therapeutic approach to interfere with HCV infection by exploiting a previously unappreciated need for HCV to continuously infect new hepatocytes to sustain a chronic infection.
INTRODUCTION
Hepatitis C virus (HCV) chronically infects at least 170 million worldwide, and until recently, curative therapies were poorly tolerated and ineffective in most patients (1). HCV is among the few viruses that cause human pathology that can either establish a chronic infection or be spontaneously cleared. Although an essential function for T cells is widely accepted in HCV clearance, the role of antibodies in controlling HCV infection remains elusive. Individuals almost universally seroconvert 2 to 10 months after infection (2), but it remains controversial whether early development of neutralizing antibodies (nAbs) predicts viral clearance (3-6). In addition, there are several case reports of seropositive patients who were successfully cured of their HCV and subsequently became reinfected (7). Moreover, chimpanzees that spontaneously resolved HCV infection remain susceptible to homologous rechallenge (8). These observations suggest that naturally arising immunity does not universally protect from reinfection.
Failure of the immune system to protect from rechallenge can be explained in part by HCV's remarkable genetic diversity and high proliferative rate, which readily yields mutations that allow the virus to escape from immune pressure (9). In vitro experiments in human hepatoma cell lines suggest that the effect of antibodies on ongoing infection may be further diminished by HCV's ability to spread directly from cell to cell via routes that are inaccessible to nAbs (10-12). However, clinical reports using the B cell-depleting antibody rituximab in chronically infected patients showed that HCV viremia rose between 10- and 100-fold after rituximab treatment and returned to baseline after reappearance of B cells (13, 14). Similarly, agammaglobulinemic patients have been shown to progress more rapidly to cirrhosis (15), although there are case reports that such patients retain the ability to spontaneously clear HCV (16). These clinical observations suggest that B cells and antibodies play a role in virus control but are not essential for virus clearance.
To better define the role of nAbs in HCV infection in model systems that more reliably capture some aspects of human physiology, we used three different systems: primary hepatocyte cultures, mice expressing the human HCV entry factors, and human liver chimeric mice. We chose three potent nAbs and assessed their ability to prevent infection in all three systems. In addition, we tested their effects on established infection in primary hepatocyte cultures and liver chimeric mice.
RESULTS
Adeno-associated virus-delivered nAbs neutralize across HCV genotypes
We recently showed that recombinant adeno-associated viruses (AAVs) are highly efficient vectors for antibody delivery after intramuscular injection (17). We constructed AAV8 vectors expressing the three HCV nAbs: AR3A, AR3B (18), and AR4A (19). Injection of 1011 genome copies of AAV-AR3A, AAV-AR3B, and AAV-AR4A or an anti-HIV control monoclonal antibody (mAb) (B12) (20) into the gastrocnemius muscle of highly immunocompromised non-obese diabetic (NOD) Rag1-/- IL2Rγcnull (NRG) mice or immunocompetent FVB mice resulted in stable, prolonged expression of human immunoglobulin G (IgG) expression for more than 4 months (Fig. 1, A and B). It was previously shown that AR3A, AR3B, and AR4A potently inhibit HCV entry in cell lines. To test the capacity of in vivo-expressed human nAb to inhibit HCV infection, we performed in vitro neutralization assays using a broad spectrum of intergenotypic chimeras harboring the structural proteins of diverse HCV genotypes (21-23). Serum containing anti-HCV nAbs efficiently neutralized most HCV genotypes, preventing infection of Huh-7.5 hepatoma cells. Of the three nAbs, AR4A was the most potent and showed IC50s (median inhibitory concentrations) between 1 and 3 log10 lower than the previously published nAb 3/11 (12) (Fig. 1C).
Three nAbs protect genetically humanized mice from HCV infection
Having established that the AAV-delivered nAbs could efficiently neutralize HCV in vitro, we set out to test their ability to block productive viral entry in vivo. We used a genetically humanized mouse model based on adenoviral delivery of human HCV entry factors into Gt(ROSA)26Sortm1(Luc)Kaelin (Rosa26-Fluc) mice, in which expression of firefly luciferase is repressed under steady-state conditions via a loxP site-flanked transcriptional stop cassette (24).
Rosa26-Fluc mice received nAb-expressing AAV vectors (n = 12 per group).
Subsequent challenge with a bicistronic HCV genome expressing Cre recombinase (25) showed that each of the three nAbs alone or as a pool efficiently prevented HCV entry as determined by a lack of in vivo bioluminescence (Fig. 1D). We next examined the ability of the nAbs to prevent infection of human liver chimeric mice, the only small-animal model that robustly supports the entire HCV life cycle. We constructed a novel xenorecipient strain by crossing the fumaryl acetoacetate hydrolase (Fah) knockout allele (26) for 13 generations onto the NRG background. After transplantation of adult human hepatocytes into the resulting FNRG mice, mouse liver damage was induced by intermittent withdrawal of the protective drug nitisinone and engraftment levels were followed over time by measuring human albumin (hAlb) levels in the sera. In closely related Fah-/- Rag2-/- IL2Rγnull (FRG) mice, hAlb levels were previously shown to correlate with hepatocyte chimerism (27). Although engraftment varied by human donor, FNRG mice consistently engrafted to higher levels than did FRG mice (fig. S1A) (27, 28). The number of human cells could be quantified by flow cytometry for human CD81+ cells in mouse liver (fig. S1B) and confirmed by histological staining for Fah (fig. S1C). Having established this human liver chimeric FNRG model, we selected animals that were engrafted with adult human hepatocytes to hAlb levels >1 mg/ml (huFNRG), because this had previously been shown to correspond to the minimal engraftment level required for HCV permissiveness (27, 29, 30). To test whether the AAV-delivered nAbs could prevent infection, we inoculated huFNRG mice with a pool of the three nAb-expressing AAV vectors (AR3A/AR3B/AR4A) or the control AAV vector (B12). Sixteen days after AAV injection, the mice displayed serum IgG levels between 26 and 126 μg/ml and, 3 days later, were infected with a low dose (about 3000 copies) of HCV genotype 1a clone H77 (31). Whereas two of three control mice became viremic, none of the three mice that received the AR pool displayed viremia (Fig. 1E and table S1). Because AAV injection resulted in a decrease of hAlb serum levels (fig. S2A), which could affect the human graft's ability to support infection, a separate group of huFNRG mice received three injections of a pool of the purified nAbs or b6 control nAb before challenge with low-dose H77. Similar to the AAV protection experiments, both mice that received the three nAbs remained aviremic, whereas three of three control nAb recipients rapidly became viremic. These data show that nAbs AR3A, AR3B, and AR4A delivered through an AAV vector can effectively neutralize HCV across most genotypes and that these nAbs can protect human liver chimeric mice from HCV infection with a low-dose inoculum.
nAbs abrogate established HCV infection in primary hepatocyte cultures
Considering their potent HCV neutralization in vitro and their ability to protect against virus challenge in vivo, we tested the ability of these three nAbs to interfere with an established HCV infection. HCVcc infection is most reproducibly studied in the Huh7 hepatoma cell line and its derivatives, which have two major limitations: Proliferation limits the time window of any studies, and their impaired innate immunity allows for high replication to supraphysiological levels. We therefore chose to study the role of these nAbs in primary human fetal liver cultures (HFLCs), in which the hepatocytes have intact innate immunity, are nonproliferating, and can support HCVcc for several weeks (32). We first performed a neutralization assay and found that, similar to results in Huh7 cells, the three purified nAbs were able to prevent infection of a cell culture-produced HCV reporter virus that secretes Gaussia luciferase, termed HCVcc-Gluc (33) (Fig. 2A). We then used a "therapeutic" protocol in which HFLCs were infected with HCVcc-Gluc, and 3 days later, after replication had been established, purified nAbs were added and maintained in the medium for the remainder of the experiment. Longitudinal luminescence measurements (Fig. 2B) or combined endpoints from three of four livers that supported infection (Fig. 2C) showed that nAbs were able to interfere with established HCV infection in HFLCs, although not nearly as efficiently as the polymerase inhibitor 2' C-methyl adenosine (2'CMA) (34). These data suggest that extracellular spread contributes to maintenance of infection in primary hepatocytes.
HCV-infected liver chimeric mice can be cured with nAbs
These in vitro results led us to test the therapeutic effect of nAbs on established HCV infection in huFNRG mice. After intramuscular injection of FVB mice with a firefly luciferase-expressing AAV8, we observed luminescence in both liver and muscle (fig. S2B). Intramuscular injection of nAb-expressing AAVs into huFNRG (n = 8 mice total) resulted in a consistent 5- to 10-fold drop in hAlb serum levels, suggesting that AAV8 vectors were also affecting the human xenograft (fig. S2A). We therefore used a passive immunization approach to investigate the role of these nAbs on established viremia. We infected huFNRG mice with a cell culture-derived HCVcc (clone J6/JFH) (23). HCVcc was previously shown to be infectious in liver chimeric mice (35) and in vitro, which allowed us not only to determine viral RNA titers but also to quantify the number of infectious particles in limiting dilution assays. Seventeen to 25 days after infection, mice (n = 3 for AR pool, n = 2 for b6 controls) were treated with 500 μg of each nAb (AR3A/AR3B/AR4A) or 1.5 mg of control IgG (b6) every 3 days, which resulted in high and stable IgG levels (Fig. 3A and table S1) without affecting serum hAlb levels (Fig. 3B and table S1). nAb treatment suppressed serum infectivity to below the limit of detection within 1 day (Fig. 3C and table S1). Notably, HCV (RNA copy numbers decreased in all mice that received the AR pool and fell below the limit of detection between 5 and 11 days after starting nAb treatment (Fig. 3D and table S1). These data are in line with previously published observations that administration of a single dose of a nAb reduced the HCV viral load below the limit of quantification in a chronically infected chimpanzee (36). We treated HCV-infected human liver chimeric mice for 30 days with nAbs and longitudinally monitored their IgG levels in the serum until they became undetectable, 55 to 69 days after the last injection (Fig. 3A and table S1). A repeat experiment using the same conditions (n = 3 for AR pool, n = 4 for b6 controls) again showed the rapid loss of J6/JFH viremia between 4 and 8 days after starting nAb injection (Fig. 3E and table S1). In contrast to a previous study performed in experimentally HCV-1a-infected chimpanzees (36), J6/JFH viremia did not reappear after the nAb levels had fallen below the limit of detection. These data suggest that either the graft was no longer permissive to HCV or the virus was indeed cleared and the remaining HCV reservoirs may have been eliminated. To prove that the human graft could still support HCV infection, mice were challenged with a heterologous HCV-1a (clone H77) at ~2 x 104 copies (Fig. 3D and table S1). Three of three mice that still had hAlb levels >1 mg/ml became infected, illustrating that the lack of viremia was not due to graft loss or nonpermissiveness in these mice. These results show that three nAbs can efficiently abrogate serum infectivity, which leads to rapid loss of viremia that cannot be attributed to graft loss.
We next aimed to determine whether treatment with these three nAbs could abrogate viremia in mice infected with a different HCV genotype. Using a similar passive immunization strategy as before, H77 viremic huFNRG mice (n = 3 for AR pool, n = 3 for b6 controls) were treated with either the pool of three anti-HCV nAbs or an isotype control (b6). Similarly to J6/JFH-infected mice, all nAb-treated mice lost their viremia to below the limit of detection, although the difference in viremia with the control mice was less pronounced (Fig. 3F and table S1). However, and in contrast to the rapid disappearance in J6/JFH-infected animals, it took 15 to 30 days for H77-infected huFNRG mice to lose their viremia. When mice were followed, virus reappeared spontaneously, indicating that although viremia was suppressed, these mice were not cured by nAb treatment. Sequencing of H77 virus in the mice that relapsed did not reveal any escape mutations with the viral envelope associated with these nAbs (fig. S3).
DISCUSSION
Vectored immunoprophylaxis has recently been shown to protect mice efficiently against HIV or influenza A virus infection (17, 37, 38). Likewise, our results show that potent nAbs delivered by AAV vectors can prevent HCV infection, although this is generally much less efficient than in vitro neutralization assays may suggest (39, 40). Using a different nAb, others have shown that this approach may hold promise for protecting liver grafts from reinfection (41). Dosing of nAbs for this application will likely depend on their neutralization ability because AbXTL68 did not protect the graft from reinfection even at serum concentrations of 200 μg/ml (42).
Our findings are the first to show that nAbs can cure liver chimeric mice, even in the absence of an adaptive immune system. It is yet unclear if these findings are specific for J6/JFH or generalizable to all HCV isolates. Treatment of liver chimeric mice infected with HCV genotype 1a (strain H77) with the pool of antibody also lead to a decrease of viremia below the limit of detection, but HCV relapsed spontaneously, which could be explained by the shorter nAb treatment duration or by strain differences between J6/JFH and H77.
Nevertheless, our data imply that HCV must continuously reinfect new hepatocytes to sustain viremia. We speculate that the mechanism by which nAbs have the ability to cure liver chimeric mice involves the protection of uninfected hepatocytes from becoming infected and thereby allowing for clearance of HCV in already infected hepatocytes. Given the distinct kinetics in viremia decline between H77 and J6/JFH, our findings suggest differences in the survival of HCV in hepatocytes, which could be due to cytopathic mechanisms or clearance by innate immune pathways. To better dissect the relative contributions of these mechanisms, it will be important to extend these observations to other HCV isolates and genotypes, for example, HCV genotype 3. This genotype was recently shown to require prolonged suppression with a polymerase inhibitor (sofosbuvir) compared to genotype 2 isolates despite being equally interferon-sensitive (43).
Human primary hepatocyte cultures and humanized mice systems do not completely mimic the three-dimensional architecture of liver, but they are, apart from chimpanzees, arguably the most advanced systems to study HCV infection in its physiological environment. Our observations put into question whether cell-to-cell spread is a dominant route for reinfection, although it is conceivable that this process is compromised and less efficient in these experimental systems than in humans. Recent findings demonstrate that broadly nAbs can also efficiently suppress viremia in humanized mice (17, 44) and Rhesus monkeys infected with HIV or SHIV (45, 46), respectively. Our data extend this work to hepatotropic viral infections.
Whether these nAbs will become useful in HCV-infected patients will largely depend on the ability of upcoming direct-acting antivirals (1) to eradicate this infection in every patient. Although we used high doses of nAb and have not determined the minimally effective dose at which we can cure liver chimeric mice, it is currently hard to speculate what nAb levels are required to prevent infection of hepatocytes in the human liver. Passive immunization with a nAb in chimpanzees (36) was slightly above the dose typically used clinically for mAb treatments of other conditions. It is therefore feasible that passive nAb transfer may be an adjunct therapeutic modality in a subset of HCV-infected individuals who cannot be cured by upcoming direct-acting antivirals.
|
|
| |
| |
|
|
|
