Evidence of variation in susceptibilities of HIV-1 India clade C to bnAbs targeting gp120 having distinct target epitope specificities.
We first examined the degree of neutralization sensitivity of HIV-1 Indian clade C to four bnAbs: VRC01 (having CD4-binding site specificity), PGDM1400 and CAP256-VRC26.25 (both having V1V2 specificity) and PGT121 (target V3 supersite). A total of 71 pseudoviruses expressing complete env gene (gp160) obtained from 40 disease stage specific samples from 28 HIV-1 infected individuals from Southern and Western India during the years 1999–2014 (the complete details of the these env clones along with their genetic properties are given in Table S1) were assessed for their degree of neutralization sensitivity towards four selected bNAbs of distinct specificities in TZM-bl neutralization assay. Several of these envs have been reported earlier [32–38]. As shown in Table 1, the degree of susceptibility of Env-pseudotyped viruses to the bnAbs varied. The neutralization data demonstrated that while VRC01 exhibited most breadth (neutralized 62/71 viruses), PGDM1400 exhibited least (46/71) of all the four bnAbs assessed. The neutralization coverage was assessed by measuring the ability of each of the bnAb to demonstrate 50% virus neutralization by IgG concentration less than or equal to 5 µg/ml (IC50). In addition, CAP256-VRC26.25 was found to neutralize the virus panel with maximal potency (IC50 of 0.0033 with 76% breadth). Interestingly, while neutralization resistance of Env-pseudotyped viruses to all four bnAbs was found to be associated with known resistance signatures (Table S2), for few envs, no known resistance signatures were evident, suggesting alternate mechanism of neutralization resistance. Overall, our data showed variation in neutralization breadth and potency of VRC01, CAP256-VRC26.25, PGDM1400 and PGT121 to HIV-1 Indian clade C envelopes that were tested in this study.
Comparing neutralization diversity of Indian clade C with other globally circulating subtypes and recombinants.
We next examined whether inter and intra clade env diversity have any association with altered neutralization phenotype. For this first we retrieved full length gp160 sequences of Env-pseudotyped viruses reported in the CATNAP database (www.hiv.lanl.gov) and compared diversity by building a phylogenetic tree (Fig. 1A). As expected, our data very clearly demarcated distinctness of Indian clade C with non-Indian clade C and other subtypes (Fig. 1A; top left). This observation also corroborates with our earlier finding [39]. The neutralization scores (IC50 values) against VRC01, CAP256-VRC26.25, PGDM1400 and PGT121 of most of the same retrieved gp160 sequences of all subtypes were next obtained from the CATNAP database for the purpose of comparing with that obtained in the current study. A total of 1020 gp160 sequences were used for the analysis with the following distribution: Subtype C (Pan Africa):290, Subtype C (India): 98 that includes 71 used in the present study, Subtype A: 76, Subtype B: 255, Subtype D: 42, CRF01_AE: 70, CRF07_B/C recombinants: 49, other subtypes and recombinant forms: 140 were assessed for their phylogenetic relatedness using IQtree (HIVdb model, non-parametric 1000 fast bootstrap with aSH-LRT test). Four heatmaps based on their responses to VRC01, CAP256-VRC26.25, PGDM1400 and PGT121 were built based on the phylogenetic tree to reflect the clustering based on their IC50 values obtained against each of these Env-pseudotyped viruses. As shown in Fig. 1A, We observed that while VRC01 appeared to be most broad, it was found to be least potent amongst all. CAP256-VRC26.25 appeared to be most potent, however our data also indicates that it has been mostly assessed against HIV-1 subtype C (including one used in the present study). These qualitative observations also reflected upon statistical assessment of viruses from India as indicated in Fig. 1B.
Finally, we compared the overall neutralization sensitivity against these four bnAbs of year matched (1999–2011), randomly selected equal number of viral gp160 datasets (n = 85) from CATNAP database with the ones assessed in this current study. The IC50 value of 5ug/ml was considered as a threshold of neutralization sensitivity. As shown in Fig. 1C, the neutralization susceptibility of Indian clade C envelopes to VRC01 were found to be distinct to that observed with the African clade C as well as other subtypes. Interestingly, while the degree of CAP256-VRC26.25 susceptibility between Indian clade C and other subtypes were observed to be comparable, this was found to be significantly different from African clade C. Overall, our data highlights disparate sensitivity of Indian clade C compared to other geographically and phylogenetically distinct HIV-1 subtypes against these four bnAbs.
Trend analysis of variation in neutralization sensitivity over time at population and individual levels
We next examined the variation in neutralization sensitivity of HIV-1 env sequences sampled over time to these four bnAbs. For this, we grouped IC50 values (in three clusters) based on corresponding year of sampling (1990–2000, 2001–2010 and 2011–2016) of the viral env sequences. As shown in Fig. 2, the PGDM1400 sensitivity of Env-pseudotyped viruses obtained from both Indian and African population (Figure S1) was observed to be decreasing overtime. For PGT121 and VRC01 bnAbs, similar trend was observed in Indian population. When compared, a gradual increase in resistant neutralization phenotype of African subtypes to PGT121 and VRC01 was also observed overtime, it was, however, not found to be significant (Figure S1). Interestingly, an increasing trend in neutralization resistance of Indian clade C envelopes to CAP256-VRC26.25 was observed in contrast to that with envelopes obtained from African population (Figure S1). Overall, we observe that although number of Indian clade C envelopes tested in this study were not large enough for an absolute conclusion, our data clearly indicate differences in neutralization susceptibility trend against these four bnAbs overtime between Indian clade C and other subtypes.
We next examined the trend of neutralization susceptibility of pseudoviruses expressing primary HIV-1 clade C envs obtained from three individuals at different time points to VRC01, CAP256-VRC26.25, PGDM1400 and PGT121. Several of these envs were reported earlier by our group [35]. The primary envs (complete gp160 sequences) from the donor NARI IVC-2 were obtained at following time point: between 0–6, 12, 18, 24, 36 and 57 months; from donors NARI IVC-3 and NARI IVC-4 at 0–6, 12 and 18 months respectively. As demonstrated in Fig. 3, the trend of susceptibility of pseudoviruses expressing envs isolated from these three donors overtime to VRC01, CAP256-VRC26.25, PGDM1400 and PGT121 varied. Out of the three donors, in one (NARI IVC-11), a clear trend of increasing susceptibility of Env-pseudotyped viruses to all the four bnAbs were observed. When tested against PGT121, for NARI IVC-2, pseudoviruses expressing envs obtained from baseline to two years, were found to be resistant, while pseudoviruses expressing envs obtained from subsequent time points till 57 months were found to be highly sensitive with IC50 over 0.01ug/ml. For all the resistant envs, N332 glycan residue was found to be absent. Interestingly, we observed that the env obtained from 57-month time point was sensitive to PGT121 despite lacking the N332 glycan. While this could possibly be compensated by the presence of a glycan at the 334 position as was reported for PGT128 with similar specificity [40], four pseudoviruses expressing envs isolated from NARI IVC-2 were found to be resistant to PGT121 despite containing N334. While pseudoviruses expressing envs obtained from all time points from this donor were found to be highly sensitive to CAP256-VRC26.25 with IC50 < 0.001ug/ml, no increasing or decreasing trend in susceptibility to VRC01 and PGDM1400 of all the envelopes were observed. For the third donor, NARI IVC-3, envelopes obtained from all the time points were found to be resistant to V1V2 -specific bnAbs CAP256-VRC26.25 and PGDM1400, while they were found to be sensitive to VRC01 and PGT121, however no clear trend in neutralization sensitivity over time observed. While resistance to different bnAbs was found to be associated with absence of known target motifs, we found evidence of neutralization resistance in presence of known epitopes that are targets of some bnAbs (Table S2). Overall, we found that while increasing neutralization resistance trend was observed at the population level, when cross sectional virus isolates across years were analyzed such a trend was not evident with longitudinal viral envs obtained from three followed up Indian donors.
Predicting optimal bnAb combination using experimental data.
Finally, based on the neutralization data of all the Indian clade C envelopes assessed in this study and neutralization scores of additional Indian clade C envs retrieved from CATNAP database (www.hiv.lanl.gov), we attempted to predict the optimal combination amongst these four bnAbs that would likely confer maximal neutralization coverage with highest potency. Towards achieving this, we employed additive model implemented in CombiNAber tool (https://www.hiv.lanl.gov/content/sequence/COMBINABER/combinaber.html) to all possible combinations of these four bnAbs (VRC01, CAP256-VRC26.25, PGDM1400 and PGT121). A total of 98 Indian clade C envelopes (71 from the current study and existing neuralization data of 27 additional Indian clade C retrieved from HIV database) for preparing for coverage analysis. As shown in Fig. 4A, the cumulative coverage analysis indicated that each of VRC01, CAP256-VRC26.25, PGDM1400 and PGT121 bnAbs individually could provide 80%, 70%, 65% and 74% coverage respectively at IgG concentration of equal to or below 5 µg/ml. When assessed for the minimum combination that would provide maximal neutralization coverage, two combinations of three bnAbs each (VRC01 + CAP56-VRC26.25 + PGT121 and PGDM1400 + CAP256-VRC26.25 + PGT121) appear to provide 100% cumulative coverage at total IgG concentration of 5 µg/ml or less as predicted by CombiNAber. In parallel, we also examined whether these two bnAb combinations of three bnAbs could achieve the maximal neutralization coverage at low doses. Towards this, the median IC50 values of Indian clade C envelopes (n = 98) obtained against single and different combinations of all the four bnAbs were obtained and the coverage versus potency assessment was statistically validated by Mann-Whitney test on the dataset obtained by implementation of additive model by the CombiNAber tool with target limit of IgG concentration of 5 µg/ml. As shown in Fig. 4B, we observed that combination of CAP256-VRC26.25 + PGDM1400 and PGT121 could achieve near 100% breadth with highest potency (Fig. 4B). This was found to be strikingly different with what we observed with African subtype C, where two combinations of each of these four bnAbs appeared to provide near 100% neutralization coverage with maximal potency. For all the different combination we assessed, we compared the median IC50 obtained in our study with that reported for African clade C envelopes (n = 250) by retrieving neutralization IC50 values from CATNAP database (www.hiv.lanl.gov). As expected, we found that the same combinations of three bnAbs (VRC01 + CAP256-VRC26.25 + PGT121 and PGDM1400 + CAP256-VRC26.25 + PGT121) demonstrated maximal coverage with geometric cumulative mean IC50 of 0.0109 and 0.01137 µg/ml respectively and are observed to be comparable to what was observed with combination of all four bnAbs. Of interest, when compared with African clade C envelopes, we observed that combination of two bnAbs that commonly include CAP256-VRC26.25 (such as CAP256-VRC26.25 + PGDM1400, CAP256-VRC26.25 + VRC01 and CAP256-VRC26.25 + PGT121) appear to provide comparable coverage with significantly lower doses than what we observed with the three bnAb combinations with Indian clade C viruses (Fig. 4B). This is likely because CAP256-VRC26.25 was found to demonstrate better neutralization coverage of African clade C with increased potency compared to what we observed with Indian clade C viruses.