Main

Antiretroviral treatment (ART) efficiently blocks viral replication of HIV but cannot eliminate infected cells, which persist in people with HIV despite decades of treatment. These persistently infected cells establish viral reservoirs that initiate rapid viral rebound if ART is interrupted. A few exceptions have been reported for individuals who are able to durably control HIV-1 infection after discontinuation of ART, achieving a state of virological remission1,2. This outcome appears to be favored by early ART initiation3,4, although the mechanisms remain unknown. Notably, five individuals have seemingly achieved an HIV cure after undergoing allogeneic hematopoietic stem cell transplantation (allo-HSCT), for the treatment of different blood cancers, with cells from CCR5Δ32/Δ32 donors5,6,7,8,9. These donors’ cells lack CCR5 expression on the cell surface, thus providing natural protection against CCR5-tropic HIV-1 variants10.

Different studies have shown that allo-HSCT in people with HIV consistently provokes a dramatic decrease in the frequency of HIV-infected cells11,12,13,14. The reduction in the size of the HIV reservoir is unrelated to the presence (or absence) of the CCR5Δ32 mutation in donor cells15. Instead, it seems to result from a combination of cytotoxic effects of the conditioning regimens, donor allogeneic immunity during graft-versus-host reactions and the gradual dilution of the pool of infected cells during immune cell replacement12,16. However, even such pressure may not be sufficient to eliminate all infected cells. Cells carrying HIV DNA have been found after allo-HSCT in the blood of some people with HIV who did not achieve full donor chimerism12,17 or in tissue sanctuaries analyzed in necropsy studies18. Moreover, during the weeks following allo-HSCT, a window of vulnerability occurs when highly activated CD4+ T cells from both donor and recipient coexist19, thereby increasing the risk of reservoir reseeding if infection of donor cells is not prevented by pharmacological or genetic and host barriers. Accordingly, and in contrast to the five individuals who have achieved HIV cure, viral rebound has been reported so far in all cases of people with HIV who interrupted ART after receiving allo-HSCT from wild-type CCR5 donors11,12,13,14,15, and even in some individuals who received a transplant from CCR5Δ32/Δ32 donors20. These observations strongly supported the hypothesis that engraftment with CD4+ T cells that remain resistant to preexisting HIV-1 variants might be necessary to avoid HIV-1 relapse from the few infected cells that may persist after allo-HSCT.

Challenging this assumption, we describe here the case of a male individual living with HIV-1 for over 30 years who, 72 months after undergoing allo-HSCT with cells from a wild-type CCR5 donor and 32 months after ART interruption, has not shown evidence of HIV-1 rebound or replicating virus despite carrying CD4+ T cells that remain fully susceptible to HIV-1 infection.

Results

Case study

We conducted a longitudinal virological and immunological characterization of a 53-year-old male (IciStem number 34, IciS-34), who is alive and asymptomatic. This individual was diagnosed to be HIV-1 clade B positive in May 1990 in Switzerland and presented with a CD4+ T cell count of 589 cells per microliter (32%) at the time of diagnosis (category A1 according to US Centers for Disease Control and Prevention classification). He immediately started ART after diagnosis, including first-generation nucleoside reverse transcriptase inhibitors (Supplementary Fig. 1). However, despite antiretroviral exposure, his CD4+ T cell count decreased to 295 cells per microliter and the first available HIV-1 plasma viral load determination in October 1996 was 63,293 copies per milliliter (Fig. 1). At this time, he began protease inhibitor-based therapy, receiving sequentially boosted saquinavir and atazanavir, without achieving full viral suppression (median (interquartile range), 1,150 (102–7,745) HIV RNA copies per milliliter) during this period that lasted 9 years. A Iopinavir-based therapy was initiated in October 2005, resulting in a continuously suppressed plasma viral load despite evidence of multiresistance to components of three major classes of antiretrovirals (Supplementary Table 1). A progressive increase in CD4+ T cell counts and normalization of the CD4/CD8 ratio were also observed (Fig. 1). An integrase-inhibitor-based ART regimen with dolutegravir and darunavir/ritonavir (DRV/r) was initiated in January 2015. This individual has been followed in the Swiss HIV cohort study since April 1992.

Fig. 1: Treatment course and immuno-virological evolution.
figure 1

Timeline of clinical events, CD4+ T cell counts, CD4/CD8 ratio and plasma viral load since HIV-1 diagnosis and until last follow-up. The successive immunosuppressants and antiretroviral regimens are also shown (detailed information can be found in Supplementary Fig. 1). Undetectable plasma viral load concentrations are represented by empty symbols at the threshold of detection, which varied throughout the follow-up. Hexagons represent different episodes of GvHD.

Full size image

In January 2018, this person was diagnosed with a myeloid sarcoma with lymph node and bone marrow involvement. He initially received two cycles of induction chemotherapy based on anthracyclines, fludarabine and cytarabine. To avoid drug–drug interactions, DRV/r was switched to tenofovir alafenamide and emtricitabine (200/25 mg) in March 2018 (Fig. 1 and Supplementary Fig. 1). He experienced a short-term malignancy relapse in June 2018 and was treated with a hypomethylating agent, followed by allo-HSCT in July 2018. The donor was an unrelated nine-of-ten HLA-matched (Supplementary Table 2) male with no CCR5Δ32 mutation. IciS-34 received one cycle of a sequential conditioning regimen (clofarabine, cyclophosphamide, fludarabine and a total body irradiation of 8 Gy) before the peripheral stem cell (no T cell depletion) transplant. Graft-versus-host disease (GvHD) prophylaxis after transplant comprised cyclophosphamide at days 3 and 4, tacrolimus and mycophenolate mofetil. Full donor chimerism in granulocytes and mononuclear cells was achieved in blood and bone marrow less than a month after the transplant. The myeloid sarcoma remains in complete remission. A maintenance treatment with 5-azacytidine was provided from January 2019 to September 2020 (21 cycles of 5-azacytidine, 32.5 mg m−2 per day, days 1–5). Severe lymphopenia was initially detected after allo-HSCT. A rapid expansion of NK cells and CD8+ T cells was then observed, followed by CD4+ T cells and B cells (Fig. 1 and Supplementary Fig. 2a). Immune reconstitution was incomplete with relatively low CD4+ T cell counts and an inverse CD4/CD8 ratio, consistent with what others and we have observed for other people with HIV who underwent allo-HSCT19,21 (Fig. 1).

The individual developed hepatic acute GvHD 120 days after HSCT and was treated with corticosteroids and tacrolimus. Following immunosuppressive drug tapering in March 2019 (8 months after HSCT, M8), he presented with a hepatic GvHD relapse, which was treated with corticosteroids and cyclosporin. In July 2019 (M12), he further developed a mild chronic skin GvHD, and in August 2019, a third-line treatment with ruxolitinib 10 mg twice daily was initiated (Fig. 1 and Supplementary Fig. 1b). Immunosuppressive drugs were then tapered and stopped in early January 2021 (M30). Unfortunately, he had signs of a second hepatic GvHD relapse in late January 2021 and resumed a combined anti-GvHD treatment course including corticosteroid and ruxolitinib, with both prescriptions continued until October 2022 (M51). In November 2022 (M52), an atypical neurological chronic GvHD with neuropathy and small-fiber damage was diagnosed and ruxolitinib 10 mg twice daily was again prescribed together with low-dose prednisone (10 mg per day).

During the multiple episodes of GvHD, to reduce the risk of potential drug interactions, ART was further simplified to dolutegravir and lamivudine dual therapy in December 2019 (M17) and to single dolutegravir, one 50 mg tablet daily, in August 2020 (M25). Finally, on 17 November 2021 (M40), all antiretrovirals were stopped following a consensual decision between the participant and his physician to evaluate the possibility of HIV remission. At the time of this report, 32 months after interruption of ART (M72), plasma HIV viremia has remained undetectable despite frequent testing (at least monthly since ART interruption) (Fig. 1).

Decline of virologic markers after allo-HSCT

We further examined virological markers to better characterize the evolution of HIV-1 infection following allo-HSCT and ART interruption in this individual. HIV RNA could be detected with the ultrasensitive viral load assay in three plasma samples obtained before (1.33 RNA copies per ml at M3), at the time of (4.18 RNA copies per ml) and immediately after (2.22 RNA copies per ml at M1) allo-HSCT. A positive ultrasensitive viral load value (4 copies per ml) was also detected at M19 after allo-HSCT, but was undetectable (<1 copy per ml) in all the other samples analyzed, including eight samples analyzed after ART interruption (Fig. 2a). Cell-associated HIV DNA could be detected in bone marrow cells, peripheral blood mononuclear cells (PBMCs) and purified blood CD4+ T cells before allo-HSCT (1,096, 202 and 457 copies per million cells, respectively) (Fig. 3b). These frequencies rapidly decreased after allo-HSCT. Viral DNA was still detectable (316 copies) in a bone marrow sample obtained at M1 after allo-HSCT, but was undetectable in subsequent samples (Fig. 2b,c). HIV DNA was sporadically detected in PBMCs with an ultrasensitive assay22 (maximum of 5 copies per million cells at M47) and purified CD4+ T cells (maximum of 40 copies at M19, coinciding with positive ultrasensitive viral load) but was consecutively undetectable by quantitative PCR in the last 6 samples obtained after ART interruption. HIV DNA was not detected in small biopsies from the small intestine; ascending, transverse, descending and sigmoid colon; cecum; and rectum obtained at M54 (14 months after ART interruption) (<20 copies of HIV DNA per 106 cells; 1.4 million cells tested).

Fig. 2: Viral markers before and after allo-HSCT.
figure 2

a,b, Evolution of residual low-level viremia measured with an ultrasensitive viral load (VL) assay in plasma (a) and HIV DNA associated with PBMCs, purified blood CD4+ T cells or bone marrow (BM) cells (b) before and after allogeneic HSCT. The empty symbols represent undetectable levels and are shown at the threshold of the techniques, which varied depending on the amount of material analyzed. The time of allo-HSCT and of analytical treatment interruption (ATI) is indicated as a dashed vertical line. c, Summary of the outcome of IPDA analyses in blood and tissues at different timepoints before and after allo-HSCT. The number of cells per amount of material tested is indicated for each analysis. aDefective in 3′ or 5′. bDefective in 3′.

Full size image
Fig. 3: Phenotype and HIV susceptibility of CD4+ T cells.
figure 3

a,b, Percentage of CD4+ T cells from IciS-34 expressing activation markers (CD38+HLA-DR+, Ki67+) (a) and the HIV-1 coreceptors CCR5 and CXCR4 (b) at different times after allo-HSCT (samples designated as the month after allo-HSCT when they were obtained; purple). The proportions of CD4+ T cells from an independent HIV-negative (HIVneg) blood donor (gray) and one person with HIV on ART (black) are depicted for reference. c, Dynamics of viral replication in purified CD4+ T cells from IciS-34 (M59, purple) and one unrelated blood donor upon infection in vitro with HIV-1BaL. The data are shown as the mean ± s.d. (n = 3 replicates) of p24 levels in culture supernatants. d, Proportion of infected (GFP+) CD4+ T cells from IciS-34 3 days after challenge with HIVNL4.3GFPΔEnv-VSV-G particles (purple). The negative control is depicted as a dashed line. The rate of infected CD4+ T cells from one unrelated blood donor is provided as a reference (gray).

Full size image

We also investigated the presence of replication-competent virus. The intact proviral DNA assay (IPDA)23 detected potentially intact proviruses in two samples that had been obtained during ART-suppressed viremia 17 and 32 months before allo-HSCT in the context of his participation in the Swiss HIV cohort study (Fig. 2c). By contrast, potentially intact proviruses were never detected following allo-HSCT. Traces of defective proviruses were detected in PBMCs and/or bone marrow samples after allo-HSCT by IPDA, at levels over 40 times lower than those observed before allo-HSCT. Intracellular HIV RNA was also not detected in samples obtained at multiple timepoints after ART interruption (Supplementary Table 3). Finally, viral production could not be detected with an ultrasensitive p24 single molecule Simoa assay24 in the supernatants of purified CD4+ T cells from multiple samples that were cultured in the presence of a pool of activated CD4+ T cells from three different donors (Supplementary Table 3). Overall, these results indicate that the HIV-1 reservoir markedly contracted after allo-HSCT in this individual and that, although traces of viral DNA were found in some samples obtained up to 57 months after the transplant, no potentially intact proviruses or evidence of replication-competent viruses were detected after allo-HSCT and ART discontinuation.

Sustained absence of detection of antiretroviral molecules

The participant reported using on-demand pre-exposure prophylaxis during two episodes in January (M42) and November 2022 (M52), taking it for only 2–3 days during these times. To document the ART interruption period more accurately, antiretrovirals were measured retrospectively since November 2022 in all available plasma samples after ART interruption and prospectively from that point onwards. Low concentrations of emtricitabine (2.8–78 ng ml−1) and tenofovir (1–4 ng ml−1) were detected in samples obtained at M42 and M53 (Supplementary Table 4), coinciding with the self-reported use of these molecules by the participant. This is consistent with the concentrations detected being at or below the median plasma concentration levels at 24 h after single oral dose of these molecules reported in the context of the ANRS IPERGAY study25. Neither these nor other molecules were found in the other samples analyzed prospectively. These results corroborate that durable remission of HIV infection in this individual occurred in the total absence of antiretroviral molecules for extensive periods of time.

CD4+ T cells remain susceptible to HIV-1 infection

Previous cases of HIV remission following allo-HSCT were associated with the reconstitution of the CD4+ T cell pool with cells that were resistant to R5 HIV-1 due to the CCR5Δ32 mutation5,6,7,8,9. We wondered whether the CD4+ T cells that expanded after allo-HSCT in this individual may possess some alternative mechanism of resistance to HIV-1 infection. As previously reported for other individuals19, CD4+ T cells from samples obtained early after allo-HSCT were characterized by high activation frequencies (15.5% of HLA-DR+CD38+ cells at M4), which decreased in later samples (2.81% of DR+CD38+ cells at M53) without reaching the basal levels observed in individuals without HIV (Fig. 3a). In agreement with the wild-type CCR5 status of the donor, CCR5 could be detected on the surface of the CD4+ T cells that expanded after allo-HSCT (Fig. 3b). These cells also expressed CXCR4, which is used by X4 HIV-1 variants. Accordingly, purified CD4+ T cells from IciS-34 obtained after allo-HSCT were highly susceptible to infection in vitro with (R5) HIV-1BaL (Fig. 3c). Moreover, we detected high levels of infected cells after their in vitro exposure to HIV-1NL4-3ΔEnv particles pseudotyped with the pantropic VSV-G envelope (Fig. 3d). These results refuted the presence of intrinsic barriers preventing HIV-1 replication in the CD4+ T cells of this individual after transplant.

Waning anti-HIV antibodies

Next, we studied whether the absence of viral rebound could be related to immune control after ART interruption. Immunoblot analyses confirmed the stable presence of anti-HIV antibodies over a period of 20 years preceding allo-HSCT. By contrast, anti-HIV antibodies began to decrease after the intervention, starting with those recognizing p17 and p31, as previously described for other PLWH who underwent allo-HSCT15 (Fig. 4a). Of note, anti-HIV antibodies continued to wane after ART discontinuation. To characterize the antibody response more thoroughly during this period, we measured the binding of purified IgG antibodies from three plasma samples after ART interruption. IgG antibodies binding to HIV-1 p24, BG505 SOSIP.664 and YU2 gp140 foldon Env trimers, gp120 and gp41 protein subunits were detected in all three samples at low levels, comparable to those found in people who are on ART since primary HIV infection (Fig. 4b). These IgGs showed very weak reactivity against consensus B Env overlapping peptides, including those from gp120 V3 loop and gp41 immunodominant regions commonly detected in other people with HIV (Fig. 4c). Accordingly, purified IgGs showed no neutralizing activity against a panel of five clade B viruses (Fig. 4d), and very weak capacity to bind to CEM.NKR-CCR5 target infected cells (Fig. 4e), and thus may have a limited potential to promote antibody-dependent cellular cytotoxicity. Overall, these results indicate that the absence of viral rebound after ART interruption was not related to an increased pressure by the antibody response.

Fig. 4: Antibody response after ART interruption.
figure 4

a, Results of immunoblots for HIV-1 antibodies in plasma samples from IciS-34 at different times before and after allo-HSCT. The arrows indicate the antigens against which reactivity was progressively lost. On the right are indicated the reactivity ratings that were automatically attributed for each antigen on the strips. b, ELISA graphs (left) comparing the reactivity of purified plasma IgG antibodies from IciS-34 (values for samples obtained at three different timepoints (M53, M55, M65) are depicted) and early- and late-treated PLWH (eART (n = 6) and lART (n = 6), respectively) against HIV-1 p24 and Env proteins. HIV-1-seronegative (Ctr−) sera (n = 5) were used as negative controls. Dot plots comparing the area under the curve (AUC) values calculated from the titration curves are shown on the right (horizontal lines indicate the median values). c, Heatmap showing the ELISA binding analysis of purified serum IgG antibodies from IciS-34 at two timepoints against consensus subtype B overlapping Env peptides. Darker colors indicate higher reactivity (absorbance values); white, no binding. Sera from eART (CO108), lART (CO107), post-treatment controller (PTC005002) and elite controller (pt3) individuals were used as positive controls. d, Heatmap comparing the in vitro neutralizing activity (as percentages) of purified serum IgG antibodies from IciS-34 (3 timepoints), CO108, CO107, PTC005002 and pt3 against selected clade B tier 1 and 2 viruses as measured in the TZM-bl assay. e, Heatmap comparing the percentage of CEM.NKR-CCR5 cells infected by laboratory-adapted (YU2 and AD8) and transmitted/founder viruses (CH058, REJO and THRO) bound by purified serum IgG antibodies from IciS-34 (2 timepoints), eART (CO42 and CO108), lART (CO105 and CO107), PTC005002 and pt3. HIV-1-seronegative (Ctr−) serum was used as negative control (c,e).

Full size image

Absence of detectable HIV-specific T cells

Allo-HSCT was performed with cells from a donor who was matched for HLA-B*27 (Supplementary Table 2), an allele that has previously been shown to favor HIV-1 control26. However, we could not detect, by intracellular cytokine staining, CD4+ or CD8+ T cells responding to 6 h of stimulation with pools of overlapping HIV-1 Gag, Nef or Pol peptides in samples obtained after allo-HSCT (M10) or after ART interruption (M45 and M64) (Fig. 5a,c). No HIV-specific cells could be amplified either after 6 days of stimulation or recall with HIV-1 peptides (Fig. 5b,c). Moreover, we did not detect CD8+ T cells binding to HLA-B*27 dextramers carrying the immunodominant KRWIILGLNK Gag epitope (Supplementary Fig. 3e). By contrast, cells responding to human cytomegalovirus (HCMV) pp65 peptides could be detected in the same samples and amplified in 6 day cultures (Fig. 5a–c). In agreement with the lack of detection of HIV-specific CD8+ T cells, purified CD8+ T cells obtained at multiple timepoints after ART interruption could not suppress ex vivo HIV-1 infection of autologous CD4+ T cells (Fig. 5d). These results argue against a role of T cells in maintaining viral control in this individual and confirm the overall lack of mobilization of the adaptive response against HIV-1 in this person despite ART discontinuation.

Fig. 5: T cell reactivity and HIV-specific T cell response.
figure 5

a, Percentage of CD8+ T cells from IciS-34 (M45) and an HIV-negative donor producing IFNγ and TNF after 6 h of stimulation with PMA/ionomycin, or with pools of HCMV pp65 peptides and HIV-1 Gag, Nef and Pol peptides. Unstimulated (Unstim) cells were used as control. Results are depicted as standard pseudocolor dot plots. b, Percentage of CD8+ T cells from IciS-34 (M45) and an HIV-negative donor that proliferated (CFSElow) after 6 days of stimulation with anti-CD3/CD28, pools of HCMV pp65 peptides, and HIV-1 Gag, Nef and Pol peptides and were able to produce IFNγ upon short polyclonal or antigen-specific restimulation. c, Heatmap comparing the percentage of CD8+ T cells from IciS-34 (two samples) and two HIV-negative donors that produced IFNγ and/or TNF and IFNγ and/or expressed CD107 after 6 h or 6 days (d) of polyclonal (positive control, Pos) or antigen-specific stimulation. Darker colors indicate higher percentages. White, undetectable; gray, not done. The percentages are indicated after subtraction of the background from the unstimulated condition. d, Comparison of the level of infection of CD4+ T cells from IciS-34 (M45) infected in vitro with HIV-1BaL cultured alone or in the presence of autologous CD8+ T cells (1:1 ratio). The results are shown as the levels of p24 in culture supernatants at day 7 after infection in vitro (mean ± s.d. of 3 experiments). Similar results were obtained at M54, M56, M57 and M59. e, Percentage of CD8+ T cells from an HIV-negative donor and IciS-34 at time of ruxolitinib discontinuation (M51), 15 days after ruxolitinib discontinuation (M52-1), 4 weeks after ruxolitinib discontinuation, time of reinitiation (M52-2) and 15 days after ruxolitinib reinitiation (M53), producing IFNγ 6 h after polyclonal stimulation with PMA/ionomycin (top). Heatmap comparing the percentages of CD4+ and CD8+ T cells from the HIV-negative donor and IciS-34 at different timepoints that produced IFNγ and/or TNF and IFNγ and/or expressed CD107 after 6 h of polyclonal or antigen-specific stimulation.

Full size image

Notably, we observed a relative lack of T cell reactivity in this individual to short polyclonal stimulation when compared with cells from different unrelated blood donors explored in these analyses. We wondered whether this observation could be related to the ruxolitinib-based immunosuppressive therapy27 that was administered for extended periods of time to treat GvHD. We therefore analyzed the T cell responses in samples taken before, during and after a brief period of ruxolitinib discontinuation that occurred during the follow-up (between M51 and M53; Fig. 5e). Poor polyclonal reactivity was again observed in the M51 sample, when compared with cells from another blood donor (EFS639). The frequency of responding cells sharply increased in the samples taken 2 weeks (M52-1) and 4 weeks (M52-2) after ruxolitinib was stopped. Ruxolitinib was reintroduced at this time owing to relapse of GvHD, and a reduction in the frequency of responding T cells was observed 2 weeks later (M53). These results support that ruxolitinib therapy may influence the reactivity of T cells to short polyclonal stimulation. Of note, despite the stronger T cell reactivity observed during ruxolitinib discontinuation, no HIV-specific T cells could be identified during this period (Fig. 5e).

High frequency of CD16+CD56– NK cells

NK cells have been proposed to play an important role in mediating the graft-versus-leukemia effect upon allo-HSCT, while their expansion and interaction with T cells may also regulate acute and chronic GvHD28,29. On the other hand, their implication in controlling HIV after ART interruption is suggested by recent reports30,31,32. Of note, IciS-34 underwent allo-HSCT with cells from a nine-of-ten HLA-matched donor. Among the matched alleles, there were three HLA class I alleles (A*24:02, B*27:05 and B*44:02) that intrinsically express the Bw4 ligand that is recognized by NK cells and whose presence has been associated with lower levels of HIV-1 viremia33. We therefore analyzed the phenotype and antiviral capacity of NK cells. While early after allo-HSCT, NK cells were characterized by a high proportion of immature CD16−CD56++ cells, a high proportion of experienced CD16+CD56− cells expressing CD57 were observed at later timepoints (Fig. 6a,b). NK cells expressed different killer-cell immunoglobulin-like receptors (KIRs; Fig. 6c), such as KIR2DL1/S1, KIR2DL23 and, notably, KIR3DL1/S1, which are reported NK cell receptors for Bw4 (refs. 34,35,36). NK cell maturation, loss of CD56 and expression of CD57, was more preponderant among cells expressing KIRs and, in particular, KIR3DL1/S1 (Fig. 6b,c), suggesting a predominant activation of KIR-expressing cells in this case. The loss of CD56 expression has been proposed to identify NK cells with adaptive traits that became exhausted owing to repeated inflammatory and activating signals37. Although CD16+CD56− NK cells are expanded during chronic HIV infection38,39, the frequency observed here was higher than that in one person with HIV on ART whose cells were analyzed in parallel for reference (Fig. 6a and Supplementary Fig. 4b). The dynamics of NK cells in this case closely recapitulate the changes occurring in people without HIV who underwent allo-HSCT and experienced HCMV reactivation during the procedure40. Indeed, IciS-34 experienced three episodes of HCMV reactivation between August 2018 and March 2019 requiring valganciclovir treatment. HCMV reactivation was also detected between June 2019 and January 2020, but at levels that did not require treatment. We did not observe significant changes in the phenotype of NK cells during the brief period of ruxolitinib discontinuation (Fig. 6b and Supplementary Fig. 4c). While CD16+CD56− NK cells have been reported to have poor cytotoxic and antiviral potential38,39, we found that NK cells from IciS-34 were able to partially inhibit HIV-1 infection in vitro of autologous CD4+ T cells (Fig. 6d). Further analyses will be needed to better understand the role that NK cells may have played in decreasing the HIV reservoir through graft-versus-HIV reservoir or direct antiviral effects.

Fig. 6: Phenotype and antiviral activity of NK cells.
figure 6

a, Expression of CD16 and CD56 on NK cells from one person with HIV on ART and IciS-34 (M52-2). b, Proportion of CD16−CD56++, CD16+CD56+ and CD16+CD56− NK cells in the sample from the person on ART and from IciS-34 at six different timepoints (left). Expression of KIR2DL2/3 and/or KIR3DL1/S1 and CD69 and/or CD57 in the three NK cell subsets from IciS-34 (M52-2) (right). c, Expression of CD57 on NK cell subsets from IciS-34 (M52-2) defined on the absence or expression of different combinations of KIRs. MFI, median fluorescence intensity. d, Comparison of the level of infection of CD4+ T cells from IciS-34 (M59) infected in vitro with HIV-1BaL cultured alone or in the presence of autologous NK cells (1:1 and 1:3 ratios). The results are shown as the levels of p24 in culture supernatants at day 3 after infection in vitro (mean ± s.d. of three experiments). Similar results were obtained at M54.

Full size image

Discussion

We describe the case of a person who underwent allo-HSCT with cells from a wild-type CCR5 donor and whose viral load remains undetectable 32 months after interruption of ART. Multiple virological and immunological readouts confirm the absence of viral exposure since ART discontinuation and support a profound and prolonged HIV-1 remission in this individual.

At the time of allo-HSCT, this individual had been living with HIV for more than 30 years and had experienced several years of uncontrolled viremia, leading to a drop in CD4+ T cell counts, before the virus was successfully controlled through an optimized protease inhibitor-based ART regimen. IPDA confirmed the presence of replication-competent virus in samples obtained during the period of suppressed viremia under ART before allo-HSCT. Cells carrying HIV DNA were readily detectable in blood and bone marrow samples just before the intervention, and residual viremia in the plasma was detected with an ultrasensitive technique at this time. A drastic drop in all these parameters was observed following allo-HSCT. However, previous cases of people with HIV who interrupted ART after wild-type CCR5 allo-HSCT resulted in viral rebound within weeks to months of treatment discontinuation11,41, confirming that the dramatic decline in the viral reservoirs associated with allo-HSCT is generally not sufficient to achieve HIV remission or cure.

The factors underlying the absence of viral rebound in the case presented here remain unclear. Sporadic (twice) pre-exposure prophylaxis use was reported by the participant and confirmed by pharmacological analyses, but given the long-term viral remission (now getting close to 3 years), we believe intermittent ART was not a major factor in the outcome of this case. Unknown host factors may hinder HIV reseeding and amplification from residual infected cells in this case. We found, however, that CD4+ T cells obtained after ART interruption were fully susceptible to HIV-1 infection. Moreover, we could not identify any evidence of immune-driven control of infection. In particular, we could not find neutralizing antibodies or CD8+ T cells able to suppress HIV infection. On the contrary, the lack of detectable HIV-specific T cells and the weak and waning antibody levels observed after ART interruption provide further evidence of the lack of viral reactivation events since allo-HSCT in this individual. Nevertheless, we cannot rule out a potential role of NK cells in mediating viral control. The combination of Bw4 ligands and KIRs present in this person after the transplant has been previously shown to favor natural viral control35,36, and NK cells have the capacity to react to the expression of stress peptides on infected cells42 before viral antigen production. Although the CD16+CD56− NK population, highly abundant in this case, has been generally considered as functionally impaired38,39, recent reports suggest that this population may be more heterogeneous than previously thought and that at least some of these cells possess diverse functionality, including cytotoxic potential43,44. A more thorough analysis of this compartment in this and other cases of people with HIV who required allo-HSCT will be needed to better understand the potential role of NK cells in controlling infection in this setting, either through graft-versus-reservoir effects or antiviral activities.

The immunosuppressive environment provided by ruxolitinib might contribute to the prevention of viral reactivation in this individual. This inhibitor of the JAK–STAT pathway was used to treat GvHD and has been administered almost continuously since ART interruption. Of note, ruxolitinib has been shown to block HIV replication, viral reactivation and reservoir reseeding in vitro and ex vivo and may favor the decay of the viral reservoir45,46. We found that the presence of ruxolitinib was indeed associated with a relative lack of reactivity of T cells from this individual to short stimulation in vitro. Ruxolitinib was briefly discontinued during the follow-up after ART interruption, and this was accompanied by an increase in T cell reactivity in vitro. The absence of ruxolitinib did not result in viral rebound or the appearance of HIV-specific cells, suggesting that no HIV antigens were produced during this period. It is possible, however, that the discontinuation of ruxolitinib (4 weeks) was too short for stochastic viral reactivation events to occur in a context in which potential remaining infected cells would be extremely rare.

Finally, we can hypothesize that allogenic immunity during repeated graft-versus-host events in this individual led to a deeper elimination of infected cells than in previous cases, achieving HIV cure through the complete purge of cells carrying replication-competent viruses. In favor of an allogenic pressure on the HIV reservoir in this case is the progressive rarefaction after the transplant of cells carrying viral DNA, which were detected at trace levels in several samples in the months that followed the transplant. The need for graft-versus-host reactions to achieve HIV cure after allo-HSCT has been the subject of debate: while its impact on the HIV reservoir is increasingly clear (graft-versus-reservoir effects)15,47, the incidence of GvH in the reported cases of HIV cure after allo-HSCT was variable5,6,7,8,9. Of note, a recent study in a model of allo-HSCT in Simian immunodeficiency virus (SIV)-infected macaques has shown that allogeneic immunity can in some cases lead to the total clearance of the viral reservoir16. Recently, a mathematical model was applied to data from IciStem participants, including data from IciS-34 before ART interruption15. The model supports the hypothesis that the main driver of the strong reservoir reduction after allo-HSCT is graft-versus-reservoir effects rather than conditioning regimens. It is tempting to assume that the repeated graft-versus-host reactions in this case may have led to an efficient elimination of reservoir cells in the absence of the barrier provided by CCR5Δ32.

Allo-HSCT is not a therapeutic option for people with HIV who do not have a cancer requiring this approach. Nevertheless, allo-HSCT is the only medical intervention that has reproducibly led to profound remission and potential cure of HIV-1 infection. The case presented here is the first to achieve such outcome after receiving cells from a wild-type CCR5 donor. It is unclear whether the status that this person has achieved will be permanent. We cannot exclude that he may harbor rare, infected cells with competent provirus or that viral rebound may occur if immunosuppressive drugs are discontinued for longer periods of time. Viral rebound can occur even after long periods of undetectable viremia without ART, as observed in the so-called Mississippi baby48. Because of the absence of an intrinsic resistance to infection, the risk of viral rebound may be considered higher than for the cases of allo-HSCT with CCR5Δ32 cells. However, the duration of undetectable viremia is unprecedented in this context. This case opens new perspectives for the development of HIV cure strategies, particularly concerning allogeneic immunity and immunosuppressive drugs.

Methods

Ethics

The described individual was enrolled in 1992 in the Swiss HIV Cohort Study (SHCS; www.shcs.ch) and in 2018 as participant number 34 in the IciStem (IciS-34) program (www.icistem.org) at the Hôpitaux Universitaires de Genève after giving signed consent. The SHCS was approved by the Cantonal Ethics Commission at Zürich (the Central Ethics Commission in Switzerland for the SHCS), and the IciStem study by the ethical committee at the Universitair Medisch Centrum Utrecht. HSCT was done in the context of the standard protocol at Hôpitaux Universitaires de Genève. The individual signed a consent form for the use of samples for research purposes according to the regulations of the Hôpitaux Universitaires de Genève.

The decision to stop ART was reached consensually between the participant and his attending physicians after a period of treatment simplification, which was implemented to diminish the risk of interactions with immunosuppressors used to treat GvHD. Analyses from unrelated HIV-negative blood donors from the Etablissement Français du Sang (collaboration agreement with Institut Pasteur) and people with HIV on ART (with undetectable viremia for >24 months) from the ANRS EP36 XII mTOR study (approved by ethics committee Ile-de-France XI) are provided as reference.

Sample processing

Peripheral blood was collected in EDTA tubes. Fresh blood samples were centrifuged at 750g for 20 min to collect the plasma. A second centrifugation was made at 2,000g for 30 min to eliminate platelets. Plasma samples were stored at −80 °C. PBMCs were obtained by density gradient centrifugation following Ficoll Plaque Plus separation (GE Healthcare) and used fresh or cryopreserved in liquid nitrogen.

Ultrasensitive plasma viremia

Ultrasensitive HIV RNA quantifications were performed on large volumes of plasma using the Generic (Biocentric) or Abbott HIV real-time PCR assay (Abbott)12,22. In brief, 3.5–17.5 ml of plasma was ultra-concentrated at 170,000g at 4 °C for 30 min, after which viral RNA was extracted. HIV RNA was quantified with a validated in-house calibration curve, set with a limit of detection of 0.56 copies per milliliter.

Cell-associated HIV DNA and RNA levels

Total DNA was isolated from frozen PBMCs or CD4+ T cells sorted from PBMCs (StemCell Technologies) using the DNeasy Kit (Qiagen). Total HIV DNA was quantified with an ultrasensitive method using the real-time PCR GENERIC HIV-DNA assay (Biocentric)22,49

Cell-associated RNA was extracted from PBMCs with an AllPrep DNA/RNA Mini Kit (Qiagen). During extraction, cell-associated HIV RNA was treated using DNase I (Qiagen). Cell-associated HIV RNA was quantified by semi-nested real-time PCR targeting the gag region with previously described primers and probes50 shown in Supplementary Table 5. Reverse transcription was performed with random hexamers and SuperScript IV (Invitrogen). The first PCR was performed with Taq ADN polymerase (Merck) for 15 cycles, then the product of the first PCR was used as a template in the second PCR. The semi-nested real-time PCR was performed with Platinum qPCR SuperMix-UDG w/ROX (Invitrogen) for 50 cycles. To normalize cell-associated HIV RNA per µg total RNA, ribosomal RNA was quantified from the same cDNA by real-time PCR using the Ribosomal RNA Control Reagents kit (Applied Biosystems).

IPDA

The presence of potentially intact DNA HIV-1 was determined in PBMCs using a duplex droplet digital PCR (QX200 ddPCR system, Bio-Rad) targeting two regions in the viral genome23: the packaging signal in the 5′ and the Rev response element in env in the 3′. Genomic DNA was extracted using the AllPrep DNA/RNA Mini Kit (Qiagen) with precautions to minimize DNA shearing. To normalize and calculate DNA shearing, a second duplex droplet digital PCR was used, targeting the human RPP30 gene. Primers and probes were previously described and are shown in Supplementary Table 6.

CD4+ T cell culture for viral amplification

CD4+ T cells were isolated from fresh PBMCs after positive selection with magnetic beads (EasySep Human CD4 Positive Selection Kit II, StemCell Technologies, 17852). Cells were stimulated with phytohemagglutinin-L (2 μg ml−1, Sigma-Aldrich, L4144) and IL-2 (200 UI ml−1, Miltenyi Biotec, 130-097-746). After 3 days of stimulation, cells from IciS-34 (1× 106–2 × 106 cells) were put in culture with a pre-activated pool of HIV-susceptible CD4+ T cells from 3 HIV-negative donors (1:3 ratio of total cells) at a final concentration of 106 ml−1 in RPMI 1640 with glutamax (Gibco, 61870-044) supplemented with 10% heat-inactivated fetal calf serum and IL-2 at 200 UI ml−1. Culture supernatants were collected every 3 to 4 days and fresh medium was added to the cultures. Supernatants were stored at −80 °C before analysis.

HIV-1 p24 was analyzed by ultrasensitive digital ELISA (Simoa Quanterix). Cell supernatants were thawed at room temperature and centrifuged at 845g for 5 min; 200 μl was transferred into a SimOa 96-well plate and inactivated with 20 μl of Triton 20%. HIV-1 Gag p24 was determined on a Simoa HD-1 analyzer using the Simoa HIV p24 kit (Quanterix, 102215) following the manufacturer’s instructions. Four-parameter logistic regression fitting was used to estimate the concentration of p24. Samples below the limit of quantification were based on the established cutoff (it was determined based on the p24 average number of enzymes per bead (AEB) signal in the standard 0 and calculated as 2.5 standard deviations from the mean of the p24 AEB signal).

CD4+ T cell susceptibility to HIV-1 infection

Productive HIV-1 infection in vitro was studied in activated CD4+ T cells (106 cells per ml in triplicate) exposed to the HIV-1BaL strain (R5; p24 10 ng ml−1). The cells were cultured in 96-U-well plates for 14 days. Every 3–4 days, the culture supernatants were removed and replaced with fresh culture medium. Viral replication was monitored in the supernatants by p24 ELISA (XpressBio). Single-round infections were performed with HIV-1 NL4.3ΔenvΔnef/GFP (ref. 51) pseudotyped with the VSV-G envelope protein by transiently cotransfecting (SuperFect; Qiagen) 293 T cells with the proviral vectors and the VSV-G expression vector pMD2.G. Activated CD4+ T cells were infected in triplicate (5 × 104 cells per well, 200 µl) with 35 ng per 1 × 106 HIV-1 NL4.3Δnef/GFP/VSV-G. Active HIV-1 infection was estimated by flow cytometry (BD Fortessa, BD Biosciences) as the percentage of GFP-expressing CD4+ T cells 72 h after infection.

Flow cytometry phenotyping

T cell phenotyping

Frozen PBMCs were thawed and incubated overnight in RPMI, 10% fetal bovine serum, 1% penicillin–streptomycin and IL-15 (0.1 ng ml−1, Miltenyi Biotec). Cells were stained with a Live/Dead Fixable Aqua Dead Cell Stain Kit (Life Technologies) followed by surface staining (CD3–FITC (SK7, 344804, dilution 1:13, BioLegend), CD4–BUV496 (OKT4, 750977, 1:65, BD Biosciences), CD8–BUV496 (RPA-T8, 612942, 1:65, BD Biosciences), CCR5–PECy7 (2D7, 557752, 1:7, BD Biosciences), CXCR4–PE (12G5, 555974, 1:7, BD Biosciences), CD45RA–APC_H7 (HI100, 560674, 1:26, BD Biosciences), CCR7–PE_Dazzle_594 (G043H7, 353236, 1:13, BioLegend), CD27–APC_R700 (M-T271, 565116, 1:26, BD Biosciences), HLA-DR–BV786 (G46-6, 564041, 1:26, BD Biosciences), CD38–BV605 (HIT2, 740401, 1:65, BD Biosciences) and Brilliant Stain Buffer Plus (563794, 1:3, BD Biosciences)). For intranuclear staining, cells were fixed and permeabilized (Cytofix/Cytoperm, BD Biosciences) and stained with anti-Ki67-eFluor450 (20Raj1, 48-5699-42, 1:26, eBioscience). All samples were acquired on an LSRFortessa flow cytometer (BD Biosciences). The differentiation into naive, central memory, transitional memory, effector memory and late effector T cells over time was analyzed via the expression of CCR7, CD27 and CD45RA (Supplementary Fig. 3a).

NK cell phenotyping

Frozen PBMCs were thawed and incubated overnight in RPMI, 10% fetal bovine serum, 1% penicillin–streptomycin and IL-15 (0.1 ng ml−1, Miltenyi Biotec). Cells were stained with a Live/Dead Fixable Aqua Dead Cell Stain Kit (L34957, 1:2,000, Life Technologies) followed by surface staining (KIR2DL2/L3–BUV395 (clone CH-L, 743456, 1:40), KIR2DL1/S1–BUV496 (clone HP-MA4, 752510, 1:40), CD25–BUV661 (clone M-A251, 741608, 1:40), CD56–BUV737 (clone NCAM16.2, 612766, 1:20), CD14–V450 (clone M5E2, 558121, 1:250), CD19–V450 (clone HIB19, 560353, 1:20), NKG2A–BV605 (clone 131411, 747921, 1:80), CD69–BV650 (clone FN50, number 563835, 1:20), DNAM1–BV711 (clone DX11, 564796, 1:20), NKG2C–BV786 (clone 134591, 748170, 1:25), CD57–FITC (clone NK-1, 555619, 1:5), NKp46–PECy7 (clone 9E2, 562101, 1:20), CD3–AF700 (clone UCHT1, 557943, 1:50), CD16–APC Cy7 (clone 3G8, 560195, 1:20) (all from BD Biosciences); CD85j/LILRB1–PE (clone REA998, 130-116-615, 1:50), NKG2D–PE Vio615 (clone REA1228, 130-124-352, 1:50), KIR3DL1/S1–PerCPVio700 (clone REA168, 130-124-077, 1:50), NKp30–APC (clone REA823, 130-112-431, 1:50) (from Miltenyi)). The gating schemes applied to identify NK cells are shown in Supplementary Fig. 4. Boolean gating was performed with FlowJo (v10.9) and the following markers: KIR3DL1/S1, KIR2DL1/S1 and KIRDL2/3. Data were acquired using an LSRFortessa X20 flow cytometer (BD Biosciences).

T cell stimulation

Cryopreserved PBMCs were thawed in RPMI 1640 Medium, GlutaMAX Supplement, complemented with 20% fetal bovine serum. Cells were split and partly stained with carboxyfluorescein succinimidyl ester (CFSE) at 1 µM (Invitrogen, C34554) for 6 days of stimulation experiments. All cells were then kept overnight at 37 °C and 5% CO2.

6 h of stimulation

PBMCs were resuspended in RPMI 1640 Medium, GlutaMAX with 10% fetal bovine serum in the presence of anti-CD107a–BUV396 (clone H4A3, 565113,1:200) and BD FastImmune Co-stimulatory Antibodies CD28/CD49d (1 µg ml−1; BD, 347690) and left unstimulated or stimulated with either hCMV pp65 peptide pool (2 µg ml−1), HIV Gag peptides (2 µg ml−1), HIV Nef peptides (2 µg ml−1) (all of them obtained through the NIH HIV reagents program) or soluble anti-CD3 (clone OKT3, 1 µg ml−1, eBioscience, 16-0037-85) and anti-CD28 (clone CD28.2, 1 µg ml−1, eBioscience, 16-0289-85). After 30 min of incubation, brefeldin A (10 µg ml−1; Invitrogen, 00-4506-51) and BD GolgiStop Protein Transport Inhibitor (containing monensin) (1 µg ml−1, BD, 554724) were added and cells were cultured for 5 h 30 min before flow cytometry staining.

6 days of stimulation

CFSE-labeled PBMCs were resuspended in RPMI 1640 Medium, GlutaMAX Supplement, complemented with 10% fetal bovine serum and left unstimulated or stimulated in the same conditions as described above. Following 6 days of culture, cells were resuspended with anti-CD107a_BUV395 (clone H4A3, BD Biosciences, 565113, 1:200), brefeldin A and BD GolgiStop Protein Transport Inhibitor (containing monensin) and were left unstimulated or restimulated overnight with hCMV pp65 peptide pool (2 µg ml−1), HIV Gag peptides (2 µg ml−1), HIV Nef peptides (2 µg ml−1), or phorbol 12-myristate 13-acetate (PMA) (80 ng ml−1, Sigma-Aldrich, P8139-5MG) and ionomycin (500 ng ml−1, Sigma-Aldrich, I0634-5MG).

In all conditions, samples were stained using the Live/Dead Fixable Aqua Dead Cell Stain Kit (Invitrogen; L34957), then extracellular staining was performed using CD3–APCe780 (clone UCHT1, 47-0038-42, 1:9, Biolegend), CD4–BUV737 (clone OKT4, 750977, 1:36, BD Biosciences), CD8–BUV496 (clone RPA-T8, 612942, 1:36, BD Biosciences), CCR7–PEDazzle594 (clone G043H7, 353236, 1:7, Biolegend), CD45RA PECy7 (clone 5H9, 561216, 1:14, BD Biosciences) and CD27 APCR700 (clone M-T271, 565116, 1:14, BD Biosciences) antibodies. The cells were fixed and permeabilized with the BD Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences) and stained for IFNγ BV605 (clone B27, 560679, 1:6, BD Biosciences), and TNF PerCP Cy5.5 (clone Mab11, 560679, 1:6, BD Biosciences) before analysis with an LSRFortessa X20 flow cytometer (BD Biosciences).

Viral suppression assays

HIV-1 suppression was evaluated with fresh blood samples52. After PBMC isolation from peripheral blood, CD4+ T cells were separated by positive magnetic bead isolation (EasySep Human CD4 Positive Selection Kit II, 17852) and the remaining cell fraction was split for subsequent CD8+ T cell and NK cell negative selection (EasySep Human CD8+ Cell Enrichment Kit, 19053; EasySep Human NK Enrichment Kit, 19055) using a Robosep instrument (StemCell Technology). Purified cells were cultured in RPMI 1640 medium containing GlutaMAX, 10% fetal bovine serum, penicillin (10 UI ml−1) and streptomycin (10 µg ml−1). After purification, CD4+ T cells were activated for 3 days with 2 µg ml−1 of phytohemagglutinin-L (Sigma, L4144) and 200 IU ml−1 of IL-2 (human IL-2 IS, premium grade, Miltenyi Biotec, 130-097-745). In parallel, CD8+ T cells and NK cells were cultured in complete RPMI medium in the absence of cytokines (CD8+ T cells) or in the presence of IL-15 at 0.1 ng ml−1 (NK cells). Activated CD4+ T cells were infected with HIV-1BaL by spinoculation alone or with autologous CD8+ T cells (1:1 ratio) or NK cells (1:1 and 1:3 ratio). Cells were then cultured for 14 days in interleukin-2 (100 IU ml−1)-supplemented complete RPMI. Supernatants were collected and fresh medium replenished every 3–4 days. Viral replication was measured in terms of p24 production in the culture supernatants by means of ELISA (HIV-1 p24 ELISA kit, XpressBio, XB-1000). The viral inhibitory capacity of NK cells was calculated comparing p24 levels at day 3 after infection in the NK:CD4 co-cultures to CD4+ T cells cultured alone. The viral inhibitory capacity of CD8+ T cells was calculated at day 7 after infection as the log drop in p24 production when CD4+ T cells were cultured in the presence of CD8+ T cells.

Analysis of anti-HIV antibodies

Initial screening for HIV antibodies in plasma samples was done using INNO-LIA HIV Score immunoblot (Fujirebio). For deeper characterization, IgG antibodies were purified from plasma samples by affinity chromatography using Protein G Sepharose 4 Fast Flow (Cytvia, 17061805) according to the manufacturer’s instructions. Purified plasma antibodies were dialyzed against PBS using Slide-A-Lyzer Cassettes (10 K molecular weight cutoff, Thermo Fisher Scientific). Final IgG concentrations were measured using a NanoDro One instrument (Thermo Fisher Scientific). Previously purified plasma IgG antibodies from early treated (eART), late treated (lART), elite controller (Pt3), and post-treatment controller (PTC005002) donors53,54,55 were used as controls in the following experiments.

Titration of antibody levels by ELISAs

High-binding 96-well ELISA plates (Costar, Corning) were coated overnight with purified Env proteins (His-tagged clade B YU2 trimeric gp140 and monomeric gp120 (ref. 56), BG505 SOSIP.664 (ref. 57), gp41 (group O HIV-1 and 2, 227-20101, RayBiotech) and HxB2 p24 (produced from the expression plasmid number ARP-13137, NIH AIDS reagent program; 125 ng per well in PBS). After washing with 0.05% Tween 20-PBS (PBST), plates were blocked for 2 h with 2% bovine serum albumin and 1 mM EDTA–PBST (blocking solution), washed and incubated with 1:3 serially diluted purified IgG antibodies in PBS (maximum concentration of 50 µg ml−1). After washing, plates were revealed by the addition of goat-HRP-conjugated anti-human IgG (1:2,000, 109-035-098, Jackson ImmunoResearch) and HRP chromogenic substrate (ABTS solution; Euromedex)58,59. Overlapping linear HIV-1 Env peptides (n = 211, consensus Subtype B Env peptide set, 9480, BEI Resources) were coated on high-binding 96-well ELISA plates (Costar, Corning) at 10 μg ml−1 in PBS overnight. After washing with 0.1% Tween 20–PBS, plates were blocked for 2 h with 1% Tween 20, 5% sucrose and 3% milk–PBS (blocking solution); washed with 0.1% Tween 20–PBS; and incubated with purified IgG antibodies at 10 μg ml−1 in 1% BSA and 0.1% Tween 20–PBS. Plates were revealed by the addition of secondary antibody and substrate as described. Experiments were performed using a HydroSpeed microplate washer and Sunrise microplate absorbance reader (Tecan), with absorbance measured at 450 nm (A450 nm). All antibodies were tested in duplicate in at least two independent experiments.

HIV-1 neutralization assay

Pseudoviruses (BaL.26 (11446), 6535.3 (11017), YU2.DG (12133), SC422661.8 (11058) and PVO.4 (11022); Env plasmids obtained from the NIH AIDS reagent program) were prepared by co-transfection of HEK-293T cells (CRL-3216, ATCC) with pSG3Δenv vector (11051, NIH AIDS Reagent Program) using FUGENE-6 transfection reagent (Promega)60,61. Neutralization experiments were performed by incubating in triplicate IgG antibodies at a final concentration of 250 μg ml−1 with pseudoviruses for 1 h at 37 °C. The virus–IgG mixtures were then used to infect 10,000 TZM-bl cells (8129, NIH AIDS Reagent Program) in the presence of 10 μg ml−1 of diethylaminoethyl (DEAE)–dextran. Infection levels were determined after 48 h by measuring the luciferase activity of cell lysates.

Antibody binding to infected cells

The capacity of purified antibodies to bind to HIV-1-infected cells was evaluated using laboratory-adapted (AD8 (11346) and YU2 (1350)) and transmitted/founder (CH058 (11856), REJO (11746) and THRO (11745)) viruses produced from infectious molecular clones (NIH HIV Reagent Program). CEM.NKR-CCR5 cells (4376, NIH HIV Reagent Program) were infected with inocula of selected viruses and adjusted to achieve 10–40% of Gag+ cells at 48 h after infection. Infected cells were incubated with purified IgG antibodies (50 µg ml−1 final concentration) in staining buffer (0.5% BSA, 2 mM EDTA–PBS) for 30 min at 37 °C, washed and incubated with AF647-conjugated anti-human IgG antibodies (1:400; A-21445, Life Technologies) for 30 min at 4 °C. Cells were then fixed with 4% paraformaldehyde and stained for intracellular Gag using FITC-conjugated anti-HIV-1 core FITC KC57 (1:500, 6604665, Beckman Coulter)62. Data were acquired using an Attune Nxt instrument (Life Technologies) and analyzed using FlowJo software (v10.7.1; FlowJo LLC).

Screening of antiretrovirals

The screening of antiretrovirals in plasma samples was performed using three distinct multiplex liquid chromatography coupled to tandem mass spectrometry (LC–MS/MS) methods. Bictegravir, cabotegravir, cobicistat, darunavir, dolutegravir, doravirine, elvitegravir, raltegravir, rilpivirine and ritonavir (pool A) were analyzed using a Vanquish system hyphenated to a TSQ Quantiva triple quadrupole MS. The chromatographic column was a Waters Xselect HSS T3 3.5 µm, 2.1 × 75 mm, kept at 35 °C in the LC oven. The mobile phase was made of water and acetonitrile (ACN) with 0.1% formic acid in each. The gradient program ranged from 10% to 95% ACN plus formic acid in 3.6 min, and the total method duration (including equilibration for the next injection) was 5.5 min. The flow rate and injection volume were 0.5 ml min−1 and 5 µl, respectively. For the analysis of atazanavir, efavirenz, etravirine, lopinavir, maraviroc, nevirapine and saquinavir (pool B), the gradient program ranged from 2% to 95% ACN plus formic acid in 2.81 min and the total method duration (including equilibration for the next injection) was 4.5 min. The analysis of abacavir, emtricitabine, lamivudine, tenofovir and zidovudine (pool C) was performed using a Vanquish system hyphenated to a TSQ Altis triple quadrupole MS. The chromatographic column was a Waters Xselect HSS T3 3.5 µm, 2.1 × 75 mm, kept at room temperature. The gradient program ranged from 0 to 70% ACN in 3 min, and the total method duration (including equilibration for the next injection) was 5 min. The flow rate and injection volume were 0.4 ml min−1 and 3 µl, respectively.

For the sample preparation, 150 µl of the precipitation solution containing the isotopically labeled internal standards was added to an aliquot of 50 µl of plasma for protein precipitation. For pools A and B, the mixture was then centrifugated for 10 min at 14,000g (5 °C) and the supernatant was directly injected. For pool C, the mixture was centrifugated for 10 min at 12,700g (5 °C) and the supernatant was diluted 1:1 with fresh Milli-Q water before injection.

Statistics and reproducibility

Graphs were generated using Prism version 10 (GraphPad Software). Flow cytometry data were analyzed using FlowJo cytometry analysis software v10.7 or v10.9 (Tree Star).

As this study was focused on one specific male individual, several limitations need to be noted: influence of sex or gender could not be considered; no statistical method was used to predetermine sample size; the experiments were not randomized; the investigators were not blinded to allocation during experiments and outcome assessment.

Samples at different timepoints (biological replicates) were measured in all experiments except HIV DNA determinations in gut biopsies. Technical triplicates were measured for viral suppression assays, neutralization assays and CD4+ T cell susceptibility to HIV-1 infection, and duplicates for antibody titers. All replication attempts produced consistent results. No data were excluded from the analyses.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.