Spontaneous control of HIV-1 viremia in a subject with protective HLA-B plus HLA-C alleles and HLA-C associated single nucleotide polymorphisms

Biological samples were collected after receiving formal written waiver from the institutional review board (protocol MUCIM approved in January 2007 by the Ethics Committee Ospedale San Raffaele, Milan, Italy), and signed written informed consent.

HIV-1 plasma RNA was measured using the Amplicor Monitor (Roche) assay (dynamic range: 50–750,000 copies/mL), and HIV-1 DNA quantification was performed using an in-house real-time PCR assay, as described [8].

Isolation of plasma HIV RNA and PBMC-associated HIV DNA, reverse transcription, amplification, gag and env sequencing were performed according to the published methods reported in the Additional file 1 section.

Genomic DNA was extracted from CASE1’s PBMC using the PureLink Genomic DNA kit (Invitrogen, Carlsbad, CA), and a pyrosequencing approach was used to determine SNPs. Detailed methods are reported in the Additional file 1 section.

Histocultures of intestinal biopsies were performed as previously reported [9]; both cells and histoculture supernatants were collected 24 h later. Cells were dispersed by enzymatic digestion with collagenase IV (0.5 mg/ml in complete culture medium, 30 min at 37°C), passed through a 22G needle and filtered with a 70 μm cell strainer. The digestion was repeated and cells from the two rounds were pooled and debris removed by centrifugation. Two million biopsy-derived cells were cultivated either alone or with 2x10 [6] PHA-stimulated PBMC from two different donors that had been previously depleted or not of CD8⁺ T cells by magnetic immunobeads. Both cultures and co-cultures were maintained for 30 days in IL-2 enriched medium, collecting their supernatants every 3 days for measurement of virus using either Mg⁺⁺-dependent reverse transcriptase (RT) activity [10] or HIV-1 p24 Gag antigen by ELISA.

Three independent attempts were made to isolate HIV-1 from peripheral CD4⁺ T cells according to published protocols [11]. Supernatants were collected every 3–4 days for up to 4 weeks of cultivation and tested for the presence of either Mg⁺⁺-dependent reverse transcriptase (RT) activity or HIV-1 p24 Gag antigen by ELISA.

For ex-vivo infection, CD4⁺ leukocytes from both CASE1 and her partner were isolated by negative selection from peripheral blood by Ficoll-Hypaque, washed and suspended in complete medium and purified as described above. Cells were stimulated with PHA and 3 days later washed and infected with CCR5-dependent (R5) HIV-1_BaL or CXCR4-dependent (X4) HIV-1_LAI/IIIB at a multiplicity of infection of 0.2. Culture supernatants were collected every 3–4 days for up to 4 weeks and tested for the presence of RT activity.

Thawed PBMC (≥80% viable) were plated in a 96-well plate after 4 h of resting (EuroClone) at a concentration of 1x10 [6] PBMC/well in complete RPMI medium [10% FBS (Lonza-BioWhittaker) in RPMI (Lonza-BioWhittaker)] with single/pools of HIV-1 derived peptides (2 μM) in the presence of a mixture of co-stimulatory anti-CD28 and anti-CD49d Ab (1.3 μg/ml each, Becton Dickinson). Cells were then treated and stained as previously described [13]. Detailed methods are reported in the Additional file 1 section.

A previous female intravenous-drug user, (CASE1), was diagnosed with HIV-1 infection in 1993 (age: 23 years old) in the course of a pregnancy screening test; she was confirmed to be HIV-1⁺ in the following years. CASE1 was repeatedly found seropositive for anti-HIV-1 antibodies (Ab), but negative for anti-HIV-2 Ab. Her viremia levels were undetectable when assessed for the first time in May 1997, but 1,900 copies of HIV-1 RNA/ml were demonstrated in Nov 1997 that dropped to 200 copies/ml in Aug 1999. Since then, HIV-1 viremia levels, that were monitored at least twice a year, have remained <50 copies/ml although CASE1 never assumed any antiretroviral agent. She also never reported clinical episodes suggestive of immunodeficiency, while her peripheral CD4⁺ and CD8⁺ T cell counts have remained in the normal range of uninfected healthy individuals through all these years (Figure 1A).

CASE1 reported >20 years of unprotected sexual intercourse with her male partner, who remained HIV-1 seronegative. She also had a child born in 1994 by vaginal delivery who has remained uninfected.

In 1997 CASE1 was diagnosed with hepatitis C virus (HCV, genotype 3a) infection, becoming negative for HCV viremia with normal levels of transaminases in 2005 after therapy with Pegylated Interferon plus Rebetol (Additional file 1: Figure S1).

In June 2009 colon biopsies were taken for diagnostic purposes. Hematoxylin-eosin, Giemsa, PAS staining and transmission electron microscopy revealed normal histology of the intestinal mucosa, lack of inflammatory infiltrates, microbes, virions or cell injury (Additional file 1: Figure S2). The tissue was also negative for the presence of HIV-1 p24 Gag⁺ cells by immunohistochemistry, although its viral DNA content was estimated to be 881 copies/10⁶ cells. In vitro cultivation of either 6 colon biopsy histocultures or of colon-derived leukocytes (2x10⁶ cells) co-cultivated with CD8-depleted T cell blasts of two seronegative donors were negative for virus replication in terms of either RT activity or p24 Gag production (data not shown). Furthermore, neither allogeneic cultivation of 5x10⁵ PBMC-purified CD4⁺ cells from CASE1 with T cell blasts of two seronegative donors, nor direct cultivation of the same number of CASE1 cells in medium containing IL-2, induced virus replication (data not shown).

Ex vivo infection of CASE1 CD4⁺ T cell blasts with HIV-1 R5_BaL or X4_LAI/IIIB strains resulted in virus replication, although at lower levels compared to those observed after infection of CD4⁺ T cell blasts from her partner (Figure 1B and C).

Quantitative PCR from CASE1 plasma samples were repeatedly <50 copies HIV-1 RNA/ml (Figure 1A), and only 2 viral genomes were detected in a total of 9 mL of plasma from June and Nov 2009 by multiplex PCR to amplify both full length gag and gp120 env coding sequences (see Additional file 1). Her PBMC-associated HIV-1 DNA load in June, and Nov 2009 and in June 2011 was 28, <13 and 10 copies/10⁶ cells, respectively. In the same dates, 23 gag- and 20 gp120 env- sequences (all with open reading frames) were obtained from a total of 21x10⁶ PBMC.

All 15 PBMC-derived single genome HIV-1 gag coding sequences from June (N = 6) and Nov (N = 9) 2009 were identical, as were 5 out of 8 sequences from June 2011 (Figure 2A). In the case of env (Figure 2B), all 4 PBMC sequences from June 2009 were identical; all 10 sequences from Nov 2009 encoded identical proteins (4 sequences had the same synonymous site mutation) and differed from the June 2009 population at only one amino acid. Three out of 6 env sequences from June 2011 were identical to the major Nov 2009 population and 2 out of 6 were identical to the minor Nov 2009 population. One env sequence and gag sequences from 3 June 2011 were subtle outliers, with an additional 5 or 5–17 changes, respectively, relative to the major virus populations (Figure 2). Despite this exceptional overall homogeneity in cell-associated viral DNA (Mean sequence diversity across all 3 time points =0.17% in gag and 0.08% in env), the single gag and env sequences obtained from the Nov 2009 plasma sample were significantly divergent from the PBMC-associated DNA sequences (Mean =1.1% in gag and 2.5% in env), although they each shared some of the mutations that distinguished the PBMC outlier sequences from Nov2011. The changes in gag shared with each of the outlier PBMC led to the reversion of the B*57-associated AA substitution I147L in the immunodominant [14] B*57 epitope IW9, restoring the optimal CTL epitope sequence, and to the introduction of the T242N substitution in the TW10 epitope (Figure 2A), normally associated with CTL escape. No mutations were observed in the B*57-restricted epitopes KF11 or the B*14 epitope DA9 (Additional file 1: Figure S3A). Finally, 2 of the 3 shared non-synonymous site changes in env corresponded to acquisition of potential N-linked glycosylation sites (Figure 2B).

Although not carrying the protective alleles CCR5-∆32 and CCR2-64I, CASE1 does express HLA B and C alleles (B*14.02 and B*57.08; C*06 and C*08.02), previously associated, either alone [5],[15] or in combination [16],[17], with slow/non-progression in LTNP and low levels of viremia in the absence of cART assumption.

In addition, CASE1 possesses the C/C genotype in the rs9264942 (C/T) -35 SNP and the del/del polymorphism (rs67384697 G/deletion) mapping upstream to the 5′ and within the 3′ untranslated region of the HLA-C gene, respectively. These 2 SNPs have been reported to be in strong linkage dysequilibrium being associated with control of HIV-1 viremia [5],[6].

CASE1 and her partner were found to have a distribution of B lymphocyte subpopulations comparable to those of uninfected individuals (Additional file 1: Table S1). Two serum samples (tested as total and purified immunoglobulins) and cervico-vaginal derived Ab collected in June 2009 and Nov 2009 did not show neutralizing activity against a panel of different HIV clades and phenotypes (Additional file 1: Table S2). CASE1 and her partner also had a normal distribution of T cell subpopulations (Additional file 1: Table S3), and both expressed comparable levels of CXCR4 and CCR5 on CD4⁺ lymphocytes (Additional file 1: Figure S4).

The breadth and magnitude of T-cell responses were assessed using CASE1 PBMC collected in Nov 2009 and June 2011. T cell responses were observed against Gag and Nef but not against Tat peptides (Figure 3A). Recognition of Gag peptide pools was mainly directed against regions containing well-defined HLA-B*57-restricted Gag epitopes (pools D, E, G, H and I, Figure 3B), without appreciable differences between the two dates.

Among the peptides present in Gag pools D and E, those expressing HLA-B*57 restricted epitopes (I/LW9 peptides 25 and 26 from pool D; KF11 peptide 29 and 30 pool E) mounted the strongest T-cell response (Figure 3C) as reported for LTNP [18]. Interestingly, recognition of the B*14-restricted epitope DA9, previously associated with virological control [19], was also observed (peptides 52 and 53 from pool H), whereas recognition of the B57* restricted epitope QW9 (peptide 54 and 55) was absent (Figure 3C). Moreover, recognition of pool I was not directed against the B*57-restricted epitope TW10 (peptides 42 and 43), but toward the C*06 restricted epitope KL9 (peptide 59, Figure 3C).

Among CD8⁺ T cells specific for HLA-B*57 (I/L)W9 restricted Gag epitopes (peptides 25 and 26), 34.6% expressed CCL4/MIP-1β and 50.5% were polyfunctional (Figure 3D). Recognition of Nef peptide pools was also mainly directed against regions containing well-defined HLA-B*57 and C*06 restricted Nef epitopes (Figure 3E).