High-throughput sequencing analysis of post-liver transplantation HCV E2 glycoprotein evolution in the presence and absence of neutralizing monoclonal antibody

Abstract

Chronic hepatitis C virus (HCV) infection is the most common cause of end-stage liver disease, often leading to liver transplantation, in which case circulating virions typically infect the transplanted liver within hours and viral concentrations can quickly exceed pre-transplant levels. MBL-HCV1 is a fully human monoclonal antibody recognizing a linear epitope of the HCV E2 envelope glycoprotein (amino acids 412-423). The ability of MBL-HCV1 to prevent HCV recurrence after liver transplantation was investigated in a phase 2 randomized clinical trial evaluating six MBL-HCV1-treated subjects and five placebo-treated subjects. MBL-HCV1 treatment significantly delayed time to viral rebound compared with placebo treatment. Here we report results from high-throughput sequencing on the serum of each of the eleven enrolled subjects prior to liver transplantation and after viral rebound. We further sequenced the sera of the MBL-HCV1-treated subjects at various interim time points to study the evolution of antibody-resistant viral variants. We detected mutations at one of two positions within the antibody epitope--mutations of N at position 415 to D, K or S, or mutation of N at position 417 to S. It has been previously reported that N415 is not glycosylated in the wild-type E2 protein, but N417S can lead to glycosylation at position 415. Thus N415 is a key position for antibody recognition and the only routes we identified for viral escape, within the constraints of HCV fitness in vivo, involve mutating or glycosylating this position. Evaluation of mutations along the entire E1 and E2 proteins revealed additional positions that changed moderately before and after MBL-HCV1 treatment for subsets of the six subjects, yet underscored the relative importance of position 415 in MBL-HCV1 resistance.

Figure 1. Changes in viral loads and epitope sequence distributions upon MBL-HCV1 antibody administration.

A–F: Epitope distributions for subjects A, B, C, D, E and F respectively. X axis represents time point in days, 0 denoting pre-transplant baseline sample. Note the depletion of wild-type sequence in all patients and the emergence of resistant epitope sequences (see proportions in left Y axis), in conjunction with viral rebound (right Y axis). Positions 415 and 417 are highlighted in blue in the legend. G: Summary of antibody-resistant mutations in epitope sequences, assuming maximum parsimony. Glycosylation motifs are highlighted in orange.

Figure 2. Changes in viral loads and epitope sequence distributions in placebo subjects.

A–E: Epitope distributions for subjects G, H, I, J and K respectively. X axis represents time point in days, 0 denoting pre-transplant baseline sample. Left Y axis represents proportion of reads with a particular sequence and right Y axis represents viral loads across time points (X axis). Subject J did not have viral titers ≥10,000 IU/ml by day 21; therefore analysis of post-transplant epitope sequence distributions could not be performed.

Figure 3. Positions in E1/E2 sequence that show the most significant changes in response to antibody treatment.

Left: Median-shifted chi-square statistic values of the top 15 amino acid positions across E1 and E2 that show the most difference in average chi-square statistic between MBL-HCV1- and placebo-treated subjects. Placebo subject J was unable to be analyzed due to post-rebound viral titers <10,000 IU/ml. The chi-square statistic quantifies changes in amino acid distributions between day 0 and day of rebound for each subject in each position across the genome. A higher statistic represents a larger change in the amino acid distribution (darker red in heatmap). Positions labeled in green are those within the MBL-HCV1 epitope. Positions labeled in blue are known to participate in CD81 binding. Positions in pink belong to the E1 sequence. Right: Differences in mean chi-square statistic between MBL-HCV1-treated and placebo patients. Note that change in position 415 is most significant being 10-fold higher than the second-ranked position 417.