Viremic control and viral coreceptor usage in two HIV-1-infected persons homozygous for CCR5 Δ32

Abstract

Objectives: To determine viral and immune factors involved in transmission and control of HIV-1 infection in persons without functional CCR5.

Design: Understanding transmission and control of HIV-1 in persons homozygous for CCR5(Δ32) is important given efforts to develop HIV-1 curative therapies aimed at modifying or disrupting CCR5 expression.

Methods: We identified two HIV-infected CCR5(Δ32/Δ32) individuals among a cohort of patients with spontaneous control of HIV-1 infection without antiretroviral therapy and determined coreceptor usage of the infecting viruses. We assessed genetic evolution of full-length HIV-1 envelope sequences by single-genome analysis from one participant and his sexual partner, and explored HIV-1 immune responses and HIV-1 mutations following virologic escape and disease progression.

Results: Both participants experienced viremia of less than 4000 RNA copies/ml with preserved CD4(+) T-cell counts off antiretroviral therapy for at least 3.3 and 4.6 years after diagnosis, respectively. One participant had phenotypic evidence of X4 virus, had no known favorable human leukocyte antigen alleles, and appeared to be infected by minority X4 virus from a pool that predominately used CCR5 for entry. The second participant had virus that was unable to use CXCR4 for entry in phenotypic assay but was able to engage alternative viral coreceptors (e.g., CXCR6) in vitro.

Conclusion: Our study demonstrates that individuals may be infected by minority X4 viruses from a population that predominately uses CCR5 for entry, and that viruses may bypass traditional HIV-1 coreceptors (CCR5 and CXCR4) completely by engaging alternative coreceptors to establish and propagate HIV-1 infection.

Fig. 1

HIV-1 viral load and CD4+ T cell counts for participants A and B (top and bottom panels, respectively). Both participants experienced slow disease progression with persistent, but low-level viremia for 3–4 years prior to rising viral loads (virologic escape) and declining CD4+ T cell counts. Participant B initiated ART in January of 2012.

Fig. 2

Coreceptor usage determined by phenotypic assay for participant A and his suspected infecting sexual partner. Pseudoviruses incorporating full-length env amplicons from population viral RNA from participant A’s partner used CCR5-expressing U87 cells for entry and led to a faint, but reproducible signal in CXCR4-expressing cells which was inhibited with the addition of the small-molecule CXCR4 inhibitor AMD3100 (A). Pseudoviruses incorporating envelope amplicons from participants A used only CXCR4 for entry from pre- and post-virologic escape time-points (panels B and C, respectively). Error bars represent standard error measurement from triplicate wells of each phenotypic assay. Panel D shows unrooted phylogenetic tree of near full-length HIV-1 env sequences from plasma RNA single genome amplification from participant A, participant A’s suspected infecting partner and an outgroup of compendium of subtype B sequences. The maximum likelihood tree was generated using the best fit selected General Time Reversible model. Sequences from participant A and his partner form distinct clusters as shown in the dashed circles, but share a common branch point from other subtype B compendium sequences. One sequence from participant A’s partner (denoted by an asterisk) had a V3 loop with closer homology to V3 sequences from participant A despite an overall higher homology to his own full-length env sequences.

Fig. 3

Conservation alignment of HIV-1 env V3 loop amino acid sequences from single genome PCR amplification and sequencing from participant A before and after virologic escape (2009 and 2012 respectively) and his sexual partner in 2005. One V3 loop sequence from participant A’s partner was more similar to V3 loop sequences from participant A, with a GPGR V3 crown motif in lieu of GPGQ. A positively charged amino acid in V3 loop position 11 (arginine) in all of participant A’s sequences and the related V3 loop sequence from his partner (lysine) was observed, and the geno2pheno (G2P) algorithm predict X4-D/M virus (FPR = false positive rate of classifying an R5-virus as X4) from these sequences. Results from phenotypic coreceptor usage assays using pseudoviruses incorporating single genome envelope sequences from participants A’s partner are shown in the bottom frame. Pseudovirus incorporating a representative sequence from participantA partner’s majority V3 loop sequences [A] used only CCR5 for entry. Virus incorporating the partner envelope sequence predicted to be X4-D/M [B] used only CXCR4. Error bars represent standard error measurement from triplicate wells of each phenotypic assay.

Fig. 4

Participant B population HIV-1 env V3 loop amino acid sequences and results of the phenotypic coreceptor usage assay. Envelope sequences were obtained from plasma RNA prior to virologic escape (2008) and from cell-associated DNA from 2011 and 2012, after rising viral loads. The geno2pheno (G2P) algorithm predicted that all virus or proviruses would use CXCR4 given low FPRs and positively charged amino acids in V3 position 11 (arginine). However, pseudoviruses incorporating population envelope sequences from 2008 RNA used only CCR5 for entry. HIV-1 RNA was unable to be recovered or amplified from post-virologic escape time points but pseudoviruses incorporating cell-associated DNA envelope sequences from 2012 were able to use CXCR4 and CCR5 (dual/mixed tropic), but at lower levels than viruses derived from RNA. AMD3100 either incompletely suppressed or potentiated entry in CXCR4-expressing U87 cells. Pseudovirsues derived from DNA in 2008 and 2011 were unable to enter U87 cells expressing either CXCR4 or CCR5.

Fig. 5

Phenotypic coreceptor usage experiments for participant B using GHOST cells expressing a variety of alternative HIV-1 coreceptors. Cells were infected with pseudoviruses incorporating 2008 plasma RNA full-length population envelope sequences. Low but reproducible levels of entry as determined by luciferase activity were observed in cell lines expressing CXCR6, and to a lesser extent GPR15 and CCR1. No entry was observed in cells expressing CXCR4 as demonstrated in the prior coreceptor usage experiments, but a strong signal was observed in cells expressing CCR5. Pseudoviruses incorporating envelope from a known X4 viral clone (pNL4-3) was used as a control in parallel experiments, and was able to use CXCR4 for entry as shown in the right panel. Error bars represent standard error measurement from duplicate experiments. Luciferase activity from X4 control pseudoviruses in the CD4 parental cell line was intermittently detectable at low levels, suggesting the possibility of endogenous coreceptor expression in the parental cell line.