Abstract Body

Background:

Characterizing HIV-1 early founder viruses is essential for informing vaccine design and broadly neutralizing antibody (bnAb) design. We analyzed variable region characteristics (VC) in HIV-1 seroconverters and compared them with predicted bnAb resistance/sensitivity using machine learning techniques.

Methods:

We analyzed HIV-1 near-full-length proviral sequences from 140 adults with documented recent HIV-1 seroconversion who were enrolled in a previous population-based household study (BCPP, 2013-2018). We determined the variable loop (V1-V5) length and net charge patterns using the Variable Region Characteristics (VC) tool in the LANL HIV sequence database (hiv.lanl.gov). VC were stratified by bNAb resistance predicted using the bNAb-ReP algorithm (https://github.com/RedaRawi/bNAb-ReP). Wilcoxon ranksum test was used for assessment of variables between predicted bnAb resistance or sensitivity. We also assessed the presence of signature mutations associated with predicted resistance.

Results:

A total of 140 consensus sequences were included in this analysis. Median log viral load (VL) of participants was 3.9 (Q1, Q3: 3.2, 4.4), median age 27 years (Q1, Q3: 22,33) and 20.7% were male. Few differences were observed between resistant and sensitive variants when looking at V1-V4 loop lengths, except for VRC26.25 and NIH45-46 where V1 loop was shorter in resistant compared to sensitive strains (p<0.01 and p=0.04) respectively. The most significant differences in variable regions between resistant and sensitive strains were observed with V5 loop length; where CD4 bnAb 2G12, DH270.5 and FP interface 35O22 resistant strains had shorter V5 loop lengths; p = <0.01, <0.01 and 0.02 consecutively. V1 charge distributions were significantly different for VRC26.25 (p =0.01), PGT121(p <0.01) and VRC01 (p = 0.01) resistant strains. In terms of mutations, E164 was observed and associated with resistance to both CH01 and PGT145. We observe the mutation T234N in sequences resistant to 3BNC117 and b12. We did not record mutations N671T, W672L, WG80G and F673L among all 2F5 resistant strains. We did however observe K683R mutation in all strains resistant to 2F5.

Conclusions:

Our findings further highlight the need to evaluate VC from sequence data, to determine the optimal vaccine design and best bnAb combinations to use for neutralization of highly variable HIV-1 subtype C.