P-215 Optimization of HIV-1 next generation sequencing genotyping drug resistance testing methods from both RNA and DNA on Ion GeneStudio S5 prime system

Le résumé fourni par la source

Background Genotyping drug resistance testing methods for HIV-1 are continually evolving, as exemplified by Next Generation Sequencing (NGS) technology, which is gradually replacing Sanger sequencing (SS) in clinical diagnostics. A comparison between SS and NGS was performed, considering both DNA and RNA samples from naïve or experienced people living with HIV-1 (PWH). Material and Methods 71 samples from PWH were selected based on viral loads (VL) [median (IQR) 4.7(3.5–6.1) log10 cp/mL] for NGS sequencing with the Ion GeneStudio S5 prime System by AmpliSeq (AmS) primers pool (Thermo Fisher), generating a total of 17 overlapping amplicons (from PR to INT). 20 additional samples were sequenced using the Ion Plus Fragment Library Kit (Thermo Fisher) starting from PR/RT (N=20) and INT (N=5) in-house amplicons. 71 samples were plasma RNA, and 20 were proviral DNA (prvDNA). Mutations were compared with those obtained with SS for samples with positive NGS PR/RT/INT sequencing. Sequences were interpreted by Stanford HIV-db. NGS minority variants (MV) were classified with a 5–20% frequency. Phylogeny was performed to determine viral subtypes and evaluate the proper clustering of the SS and NGS sequences from the same subject. Results By phylogeny, non-B subtypes were 51.1%. Overall, 86/91 samples (94.5%) had coverage ≥100x for PR/RT, 74/76 (97.4%) for INT, and 71/76 (93.4%) for the entire pol gene in the resistance-associated mutations (RAMs)/natural polymorphisms (NPs) positions. AmS failed PR/RT/INT sequencing in 2 RNA samples and only PR in 3 samples (1 prvDNA). SS was available for 60 samples with complete PR/RT/INT sequencing; at the 20% threshold, there were 257 RAMs/NPs. As expected, the majority (93.0%) were detected in both SS and NGS. A small proportion (2.3%) of mutations were present only in SS, whereas 4.7% were present only in NGS. Decreasing the threshold to ≥5%, 287 RAMs/NPs were identified (table 1). At this threshold, no MV to PI was detected, whereas additional RAMs (RNA) to NRTI (M41I, 6.5%, D67E, 5.3% of frequency) and INSTI (E138K, 11.0%, G140A, 5.6%, 7.0%, 11.0%) in different naïve individuals were identified. Other RAMs to NRTI/NNRTI (RNA: V75M, 19.0%, L100I, 11.0%; prvDNA: K101P, 15.0%, K103N, 11.0%) and INSTI (RNA: G140A, 6.6%, 16.0%, G163R, 17.0%; pvDNA: G140S, 16.0%) were also detected in experienced individuals. Of note, some samples with failed SS were successfully sequenced with NGS. NGS failed for samples with low VL, except for 2 samples, where SS also failed. Conclusions Overall, our results show that NGS by Ion Torrent S5 assays performance was comparable to SS, both using RNA and prvDNA at several VL. As expected, NGS detected MV setting the threshold to ≥5% which could not be detected by SS. That can improve treatment selection and clinical outcomes. However, the real weight of these MVs has yet to be determined, as is the case with implementing studies with a larger sample size.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.