Original Article doi: 10.1111/joim.13484 Altered T-cell subset distribution in the viral reservoir in HIV-1-infected individuals with extremely low proviral DNA (LoViReTs) Cristina Gálvez1 , Víctor Urrea1 , Maria del Carmen Garcia-Guerrero1 , Sílvia Bernal1,2 , Susana Benet1,3 , Beatriz Mothe1,2,3,4 , Lucía Bailón3,5 , Judith Dalmau1 , Andrea Martinez3 , Aroa Nieto3 , Lorna Leal6 , Felipe García6 , Bonaventura Clotet1,2,3,4 , Javier Martinez-Picado1,2,4,7 & Maria Salgado1,4 From the 1 IrsiCaixa AIDS Research Institute and Institute for Health Science Research Germans Trias i Pujol (IGTP), Hospital Germans Trias i Pujol, Badalona, Spain; 2 Chair in Infectious Diseases and Immunity, University of Vic – Central University of Catalonia (UVic-UCC), Vic, Spain; 3 Lluita contra la SIDA Foundation, Infectious Diseases Department, Hospital Germans Trias i Pujol, Badalona, Spain; 4 CIBER de Enfermedades Infecciosas, Madrid, Spain; 5 Department of Medicine, Autonomous University of Barcelona, Catalonia, Spain; 6 Infectious Diseases Department Hospital Clínic, University of Barcelona, Barcelona, Spain; and 7 Catalan Institution for Research and Advanced Studies (ICREA), Barcelona, Spain Abstract. Gálvez C, Urrea V, Garcia-Guerrero MdC, lower intact HIV-DNA, cell-associated HIV-RNA, Bernal S, Benet S, Mothe B, Bailón L, et al. Altered and ultrasensitive viral load than controls. The pro- T-cell subset distribution in the viral reservoir portion of intact versus total proviruses was similar in HIV-1-infected individuals with extremely low in both groups. We found no differences in the per- proviral DNA (LoViReTs). J Intern Med. 2022;292: centage of host factors. In peripheral blood, 71% of 308–320. LoViReTs had undetectable replication-competent virus. Minimum levels of total HIV-DNA were found Background. HIV cure strategies aim to eliminate in rectal and lymph node biopsies compared with viral reservoirs that persist despite successful HIV-infected individuals receiving ART. The main antiretroviral therapy (ART). We have previously contributors to the reservoir were short-lived tran- described that 9% of HIV-infected individuals sitional memory and effector memory T cells (47% who receive ART harbor low levels of provirus and 29%, respectively), indicating an altered distri- (LoViReTs). bution of the HIV reservoir in the peripheral T-cell subpopulations of LoViReTs. Methods. We selected 22 LoViReTs matched with 22 controls ART suppressed for more than 3 years Conclusion. In conclusion, LoViReTs are charac- with fewer than 100 and more than 100 HIV-DNA terized by low levels of viral reservoir in periph- copies/106 CD4+ T cells, respectively. We mea- eral blood and secondary lymphoid tissues, which sured HIV reservoirs in blood and host genetic fac- might be explained by an altered distribution of the tors. Fourteen LoViReTs underwent leukapheresis proviral HIV-DNA towards more short-lived mem- to analyze replication-competent virus, and HIV- ory T cells. LoViReTs can be considered exceptional DNA in CD4+ T-cell subpopulations. Additionally, candidates for future interventions aimed at curing we measured HIV-DNA in rectum and/or lymph HIV. node biopsies from nine of them. Keywords: CD4+ T-cell subpopulations, HIV latency, Results. We found that LoViReTs harbored not only HIV reservoir, secondary lymphoid tissues, total lower levels of total HIV-DNA, but also significantly HIV-DNA Introduction Antiretroviral therapy (ART) effectively blocks viral Javier Martinez-Picado and Maria Salgado contributed equally to replication but does not completely eliminate HIV the study. [Correction added on 08 April 2022 after online publication: infection [1] owing to the persistence of inte- Article title has been updated.] grated proviruses, mainly in CD4+ T cells. These 308 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is noncommercial and no modifications or adaptations are made. Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. proviruses constitute the viral reservoir, which is previous reduction of the HIV reservoir. The latter the main barrier to curing HIV. Thus, most HIV- approach is especially important, since previous infected individuals experience a rapid relapse of studies in participants with a phenotype similar viremia when ART is interrupted. However, an to that of the LoViReT cohort showed that when extremely small proportion of HIV-infected indi- treatment was stopped, most individuals expe- viduals control viral replication naturally in the rienced viral rebound [15, 16]. This observation absence of ART and usually harbor low viral reser- raises the questions of whether the low reservoirs voirs. These individuals are considered examples observed in blood correlate with similar levels of of functional cure and comprise phenotypes such provirus in anatomical compartments and whether as elite controllers [2, 3], exceptional elite con- the distribution of the reservoir among different trollers [4, 5], and post-treatment controllers [6], in T-cell subpopulations could alter the dynamics of whom cumulative viral and host protective factors viral persistence. are observed [7, 8]. In exceptional elite controllers (and elite controllers), we and others have observed In this study, we comprehensively analyzed 22 enriched protective HLA alleles, potent cytotoxic individuals from the LoViReT cohort who had CD8+ T-cell activity, and a lack of viral evolu- initiated ART during the early and the chronic tion, probably owing to the presence of defective phases of HIV-1 infection in order to decipher proviruses [4, 9–11]. In contrast, post-treatment the nature and mechanism of their low latent control is associated with early initiation of ART, HIV reservoir levels. We assessed protective host enriched HLA-B*35/07 alleles, weak HIV-specific factors and the distribution and functionality of cytotoxic activity, and unusual distribution of HIV their proviruses in blood and secondary lymphoid latency in CD4+ T-cell subpopulations [6]. tissues. Understanding the mechanisms by which LoViReT individuals have a reduced latent reser- These exceptional cases have improved our under- voir will be key to designing a cure strategy for standing of the mechanisms required to control the most HIV-infected persons. virus and that could be emulated through medi- cal interventions in people with HIV who currently Methods require lifelong therapy. Therefore, various strate- Participants and study design gies are being studied with the aim of reducing the latent HIV reservoir as a first step towards Of the 42 LoViReT individuals defined in an earlier curing HIV [12]. Interestingly, in a previous study cohort [13], we selected 22 (both early and chronic by our institution, a group of treated HIV-infected treated) who met the inclusion criteria and were individuals achieved this aim naturally, harbor- willing to participate in the study. We used the ing low or even undetectable levels of HIV-DNA random forest method to match them with 22 under ART. This cohort was named LoViReT (Low controls for similarity or proximity with respect to Viral Reservoir Treated) [13], and the individuals 17 clinical characteristics (e.g., sex, age, CD4+ T included account for 9% of the treated HIV-infected cells at sampling, and zenith viral load [Fig. S1]). population. We observed that these individuals ini- Participants were receiving care at Hospital Ger- tiated ART mainly during the chronic phase of mans Trias i Pujol and Hospital Clinic, Barcelona infection, in contrast to post-treatment controllers, (Spain). The inclusion criteria for both groups were which are more frequently found in early treated to be receiving suppressive ART with undetectable individuals [14]. This new phenotype is charac- viremia (HIV-RNA <50 copies/ml) for at least terized mainly by a naturally reduced viral reser- 3 years, with fewer than 100 HIV-DNA copies/106 voir and a less impaired CD8+ T-cell compartment CD4+ T cells in the case of LoViReT individuals before the start of treatment, with a faster and more and more than 100 HIV-DNA copies/106 CD4+ marked reduction of the reservoir at initiation of T cells in the case of controls. Participants were ART and in circulating HIV-specific antibodies after classified as early or chronic treated if there were ART [13]. diagnostic criteria to document whether ART had been initiated within 6 months since the estimated The LoViReT cohort constitutes an optimal plat- acquisition of HIV, as previously defined in the form for HIV cure studies focused on both research IrsiCaixa Early-ART Cohort [17]. into new mechanisms of reservoir reduction and on the study of immune-based strategies that can All the LoViReT individuals were invited to par- facilitate complete viral remission in the context of ticipate in the second phase of the study with a © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. 309 Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. more in-depth analysis of peripheral blood and Finally, residual viremia (<50 HIV-RNA copies/ml secondary lymphoid tissues. A subgroup of 14 plasma) was measured using the ultrasensitive individuals agreed to undergo leukapheresis to viral load (usVL) assay by ultracentrifugation of increase the number of cells needed to perform 9 ml of plasma with a limit of detection of 0.56 a quantitative viral outgrowth assay (qVOA) and copies/ml, as previously described [22]. measure the distribution of virus among CD4+ T- cell subpopulations. Rectum and lymph node biop- HLA typing sies were taken only from those with an infectious High-resolution HLA class I typing for B alleles was units per million (IUPM) value smaller than 0.1 who performed using sequence-based typing methods. agreed to participate (Fig. S1). Control individuals B*27 and B*57 were considered protective alleles, did not undergo this second phase of the study. and B*35 and B*07 were considered risk alleles [9]. All participants provided their signed informed CCR5 genotyping consent. The study was approved by the ethics committees at both recruiting centers (reference #: A portion of the CCR5 gene was amplified by PCR PI-17-043). with primers that flanked the 32-bp deletion [23]. Wild-type and deleted fragments of 185 bp and Quantification of the HIV reservoir in blood 153 bp, respectively, were generated and visualized To evaluate the size of the viral reservoir, lysed in 2% agarose gels. Heterozygosity (CCR5 wt/32) extracts from purified CD4+ T cells were used to was indicated by the presence of both fragments. measure total HIV-DNA by droplet digital poly- merase chain reaction (ddPCR), as previously Quantitative viral outgrowth assay described [18]. Briefly, 5 LTR or gag regions were Leukapheresis was performed in 14 LoViReT indi- amplified, and the RPP30 housekeeping gene was viduals to quantify the size of the replication- measured in parallel to normalize sample input. competent reservoir. A total of 38 million fresh Raw ddPCR data were analyzed using the QX100TM CD4+ T cells per individual were used for a limiting Droplet Reader and the software application Quan- dilution cell culture assay in the absence of autol- taSoft v.1.6 (Bio-Rad, USA). ogous contemporaneous IgG [24]. Supernatants from day 14 were quantified with p24Gag ELISA Quantification of intact provirus was assessed (PerkinElmer, USA). IUPM was determined using using lysed extracts of CD4+ T cells. Duplex IUPMStats v.1 (https://silicianolab.johnshopkins. ddPCR was performed using the packaging signal edu/) based on the maximum likelihood method. () and nonhypermutated envelope gene (Env) The detection limit was set at 0.0185 IUPM for all primer/probe sets, per the original Intact Provirus participants. DNA Assay (IPDA) protocol using IPDA, as previ- ously described [19]. The HIV packaging signal () Quantification of the HIV reservoir in anatomical and the Rev responsive element in Env were ampli- compartments fied. For those samples that failed to detect the Env or packaging signal genomic region, we used a sec- Total HIV-DNA was measured in anatomical ondary prime/probe set to target an immediately compartments from secondary lymphoid tissues, nearby region of Env or U5’-LTR, respectively, with including rectal biopsies and lymph node biopsies the aim of rescuing provirus quantification [18, in LoViReT individuals with an IUPM lower than 20]. In parallel, two primer/probe sets targeting 0.1. In the case of rectal biopsies, 4–8 endoscopic the RPP30 gene were used to normalize cell counts specimens per individual were collected and pro- and to correct for DNA shearing. Only individuals cessed immediately to minimize loss of lamina in whom both  and Env could be amplified were propria leukocytes (LPL) and/or bacterial contam- included. ination. The specimens were processed following the previously reported LPL-vDNA protocol [18]. Viral transcription was evaluated by quantification CD45+ cells from the LPLs were sorted in a BD of cell-associated HIV-RNA (caHIV-RNA) in purified FACSAria II flow cytometer using the anti-CD45 CD4+ T cells, using one-step reverse-transcription (FITC) antibody. For comparison purposes, we ddPCR [21]. The 5’-LTR or gag genes and the used historical data from HIV-infected individuals housekeeping gene of TATA-binding protein (TBP) receiving ART, which were processed using the were measured in parallel. same protocol. 310 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. Lymph node biopsies were processed using a non- in the case of categorical variables. The Spearman invasive scanner-guided fine needle biopsy tech- correlation coefficient was calculated to analyze nique to aspirate cells from punctures of two correlations between study variables. The analyses inguinal nodes. Cells were placed in RPMI 1640 were performed with GraphPad (v8.02). containing 1% penicillin/streptomycin and 10% fetal bovine serum (FBS). In order to increase Results the sensitivity of the technique, we concentrated Participant characteristics the infected cell population by measuring the reservoirs in memory CD4+ T cells defined as We selected 22 LoViReT individuals and 22 con- CD3+ CD4+ CD45RA– . Cells were sorted in a BD trols, all of whom had been receiving suppressive FACSAria II flow cytometer using the antibodies ART for more than 3 years. Their clinical character- CD3 (APC), CD8 (APC H7), and CD45RA (FITC). For istics are summarized in Table 1. No statistically comparison purposes, we used data on integrated significant differences were observed between the HIV-DNA from individuals receiving ART [25, 26], clinical characteristics of the groups owing to previ- which should be similar to or smaller than total ous matching. Based on clinical health records, 14 HIV-DNA in people on ART. (64%) and five (23%) LoViReT individuals were clas- sified as having initiated ART during the chronic In both cases, isolated cells were lysed, and phase or the early phase of HIV infection (i.e., total HIV-DNA was quantified using ddPCR. Unde- within 6 months from the estimated acquisition of tectable samples are expressed as the limit of HIV). The date of acquisition of HIV could not be detection, which varies between samples depend- estimated in three participants because the data ing on cell input. were missing. Cell sorting of CD4+ subpopulations The LoViReT phenotype is independent of the HLA and CD4+ T-cell subpopulations were isolated by cell CCR5 genotypes sorting from 175–450 million peripheral blood We determined whether host protective genetic mononuclear cells (PBMCs) obtained from 14 factors were enriched in LoViReT individuals. To LoViReT individuals using leukapheresis. PBMCs do so, we first measured the HLA class I alleles were incubated with live/dead (APC-Cy7), CD3 (Fig. S2a). Six LoViReTs and two controls had pro- (BV510), and CD4 (AF700) monoclonal antibod- tective alleles (two HLA-B*57 and four HLA-B*27 ies. T-cell maturation was based on the expres- vs. one HLA-B*57 and one HLA-B*27), though the sion of the surface markers CD45RA (APC), differences were not statistically significant (p = CCR7 (PE Dazzle), and CD27 (FITC) in order to 0.27). On the other hand, we did observe a ten- define TN (CD45RA+ CCR7+ CD27+ ), TCM (CD45RA– dency toward a lower frequency of risk alleles in CCR7+ CD27+ ), TTM (CD45RA– CCR7– CD27+ ), TEM the LoViReT group (one HLA-B*07 and two B*35 in (CD45RA– CCR7– CD27– ), and TTD (CD45RA+ CCR7– LoViReT vs. four HLA-B*07 and six B*35 in con- CD27– ). Subpopulations were sorted in a BD FAC- trols), though again, the differences were not sta- SAria II flow cytometer, and cells were lysed to mea- tistically significant (p = 0.07). sure total HIV-DNA using ddPCR. We also evaluated the frequencies of the CCR532 Statistical analysis mutation in the cohort (Fig. S2b) and found that 27.3% of LoViReT individuals had a CCR5wt/32 The clinical characteristics of the study popula- genotype compared with 21.4% of the controls tion were presented as percentages for categorical (p > 0.99). variables and as median and interquartile range (IQR) for continuous variables. LoViReT individu- Our results indicate that neither protective nor risk als and controls were matched using a multivari- host genetic factors are enriched in the LoViReT ate approach based on the random forest method. cohort. The proximity distance between individuals pro- vided by the random forest method was used to HIV-DNA expression and HIV-RNA expression are select the control closest to each LoViReT [27]. reduced in LoViReT individuals Differences between LoViReTs and controls were assessed using the Mann–Whitney test in the case As expected from the selection criteria, we observed of continuous variables and the Fisher exact test statistically significant differences between the © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. 311 Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. Table 1. Clinical characteristics of the subjects included in the study LoViReT group Control group (n = 22) (n = 22) p-Value Age at HIV diagnosis (years), median (IQR)a 34 (28–39) 33 (30–37) 0.7 Females, n (%)b 5 (23) 2 (9) 0.22 CCR5 -tropic virus, n (%)b,c 12 (92) 18 (78) 0.56 HIV subtype B, n (%)b,c 12 (92) 20 (95) 0.72 Mode of HIV acquisition MSM, n (%)b 10 (45) 12 (54) 0.52 Time since HIV diagnosis (years), median (IQR)a 17 (10–23) 20 (12–27) 0.52 Time receiving ART (years), median (IQR)a 13 (9–18) 17 (11–25) 0.12 Time with suppressed pVL (years), median (IQR)a 12 (9–16) 13 (11–16) 0.34 Time from ART to supressed pVL (years), median (IQR)a 0.25 (0.1–0.3) 0.4 (0.2–0.6) 0.26 Zenith pVL (log10 copies/ml plasma), median (IQR)a 4.0 (3.5–4.9) 4.6 (3.6–5.1) 0.46 CD4 T-cell count Nadir (cells/μl), median (IQR)a 300 (235–439) 297 (219–420) 0.58 CD4 T-cell count (cells/μl) at sampling, median (IQR)a 859 (527–984) 787 (573–1025) 0.86 Antiretroviral treatment Number of different regimens, median (IQR)a 5.5 (4.7–7.3) 6 (3.8–9.3) 0.55 Number of different families, median (IQR)a 3 (3–4) 3.5 (2–4) 0.59 Nucleoside reverse transcriptase inhibitors, n (%)b 22 (100) 22 (100) >0.99 Non-nucleoside reverse transcriptase inhibitors, n (%)b 16 (73) 20 (90) 0.24 Protease inhibitors, n (%)b 18 (82) 17 (77) >0.99 Integrase inhibitors, n (%)b 13 (59) 12 (55) >0.99 Entry inhibitors, n (%)b 3 (14) 0 (0) 0.23 Treatment initiation with non-HAART, n (%)b 4 (18) 8 (36) 0.31 Abbreviations: ART, antiretroviral therapy; HAART, highly active antiretroviral therapy; IQR, interquartile range; LoViReT, Low Viral Reservoir Treated; MSM, men who have sex with men; pVL, plasma viral load. a p-Value between groups: Mann–Whitney test. b p-Value between groups: Chi-square test. c Data from individuals in whom amplification of the Env gene was achieved. groups for levels of total HIV-DNA measured in reduced expression of intracellular and extracellu- CD4+ T cells (p < 0.0001) (Fig. 1a). lar HIV-RNA in the context of viral suppression. In order to determine whether the low levels of Low levels of intact proviruses are also associated with total HIV-DNA were also related to reduced residual LoViReT phenotype viral replication, we analyzed levels of intracellular and extracellular forms of HIV-RNA by measuring To further analyze proviruses, we measured lev- caHIV-RNA and usVL in CD4+ T cells and plasma, els of intact provirus using IPDA. We observed respectively. HIV-RNA expression was detected in lower levels of intact (p < 0.0001), 5 deleted all the LoViReT individuals, though levels were (p < 0.0001), and 3 deleted/hypermutated significantly lower than in controls (p < 0.0001) (p < 0.0001) proviruses in LoViReT individu- (Fig. 1b). Moreover, residual plasma viremia was als (Fig. 1e). However, the frequency of intact detectable in only 19% of the LoViReT individuals, proviruses with respect to the sum of intact, hyper- in contrast to 64% of controls (p = 0.0039) (Fig. 1c). mutated/3’defective, and 5’defective proviruses In addition, significant positive correlations were did not differ between the groups (p = 0.16) found between total HIV-DNA and caHIV-RNA (r = (Fig. 1f). Significant positive correlations were 0.47, p = 0.0014) (Fig. 1d). observed between intact provirus and total HIV- DNA (Fig. 1g) and caHIV-RNA (Fig. 1h). Therefore, Therefore, the LoViReT phenotype is determined our results indicate that LoViReT individuals not only by low HIV-DNA levels, but also by harbor low levels of intact proviruses, which were 312 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. (a) (b) (c) (d) (e) (f ) (g) (h) Fig. 1 Measurements of total HIV-DNA, HIV expression, and intact provirus. (a) Levels of total HIV-DNA, (b) cell-associated HIV-RNA, and (c) ultrasensitive viral load in plasma. (d) Correlation between total HIV-DNA and cell-associated HIV-RNA. (e) Levels of intact and defective provirus per 106 total CD4+ T cells measured using Intact Provirus DNA Assay (IPDA). (f) Percentage of intact provirus with respect to total values using the IPDA. Total proviruses were determined as the sum of intact, 5 defective and hypermutated/3 defective proviruses from each individual. (g) Correlation between intact proviruses and total HIV-DNA, and (h) correlation between intact provirus and cell-associated HIV-RNA. Open symbols represent values under the limit of detection; in these cases, the limit of detection varied based on sample/volume input. The Spearman correlation coefficient was calculated to analyze correlations between study variables. proportional to low levels of total proviruses, Low viral reservoirs were observed in secondary suggesting that there is no over-representation lymphoid tissues in LoViReT individuals of defective proviruses that could explain their Since the latent reservoir may be unevenly repre- phenotype. sented in lymphoid tissues and peripheral blood, we explored potential secondary lymphoid tissue A low rate of undetectable replication-competent virus by collecting rectum and lymph node biopsies in a in peripheral blood was observed in LoViReT individuals subset of eight LoViReT individuals with an IUPM <0.1 (Fig. 2a). Minimum levels of total HIV-DNA To determine whether the low levels of proviruses were found in rectal samples, with detectable observed in LoViReT individuals were replication reservoirs in six of eight individuals (median competent, we performed leukapheresis in a sub- of 56.9 HIV-DNA copies/106 CD45+ T cells [IQR: group of 14 LoViReTs (Fig. 2a). We observed that 37.4–113.7]) (Fig. 2b). These values were seven fold 71% of the individuals had an IUPM under the limit lower than historical data from our group on HIV- of detection (0.0185), despite using up to 38 mil- infected individuals receiving ART (p = 0.0003). lion CD4+ T cells to increase the sensitivity of the qVOA assay. The four LoViReT individuals with Besides, only three out of eight individuals had detectable qVOA had values ranging from 0.244 to detectable HIV-DNA in lymph nodes, with a 0.65 IUPM, which are all below the median IUPM median of 262.6 HIV-DNA copies/106 CD45RA– T of individuals on ART from previous studies [28]. cells (36.3–2112]) (Fig. 2c). In this case, the levels © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. 313 Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. Fig. 2 Measurements of the viral reservoir in peripheral blood and secondary lymphoid tissues. The figure shows data on each LoViReT (Low Viral Reservoir Treated) individual for (a) the viral outgrowth assay, (b) rectal biopsies, and (c) lymph node biopsies compared with HIV-infected individuals receiving antiretroviral therapy (ART). LoViReT individuals treated in the chronic phase of infection are represented with a circle, those treated in the early phase of infection are represented with a square, and those with no available information on their status when they initiated ART are represented with a triangle. HIV-DNA in rectal CD45+ and lymph node CD45RA– was only assessed in LoViReT individuals with infectious units per million <0.1. The values from HIV-infected individuals receiving ART are from Morón-López et al. [28] for the viral outgrowth assay, and from Banga et al. [25] and Leyre et al. [26] for the lymph node biopsies. Median values are indicated by a horizontal black line. Open symbols represent values under the limit of detection; in these cases, the limit of detection varied based on cell input. of HIV-DNA in lymph nodes were statistically sig- and initiation of ART, n = 5). We did not find sta- nificantly lower for LoViReT individuals (p = 0.007) tistically significant differences between the groups than for individuals receiving ART in previous in peripheral blood reservoirs for total HIV-DNA studies [25, 26]. (p = 0.98), caHIV-RNA (p = 0.81), usVL (p = 0.53), or replication-competent proviruses (p = 0.55) Matching of detectable/undetectable proviral (Fig. 3a–d). Similarly, with respect to lymphoid tis- HIV-DNA reservoir in rectal biopsies agreed with sues, we did not find statistically significant differ- peripheral blood findings in 75% of individuals. ences between the groups in the median levels of We also observed 57% agreement between rectal total HIV-DNA in rectum (p = 0.19) or lymph node and lymph node biopsies. specimens (p = 0.23) (Fig. 3e,f). In summary, LoViReT individuals have a lower or even undetectable reservoir compared with people Short-lived CD4+ T-cell subpopulations contributed on ART, not only in peripheral blood but also in more than expected to the HIV reservoir in LoViReT specific lymphoid tissues such as rectum and participants lymph nodes. Finally, we aimed to determine whether the low level of proviruses observed in LoViReT individu- als could be attributed to altered HIV persistence Low viral reservoir is independent of the time of in a specific CD4+ T-cell subpopulation. To do so, initiation of treatment we sorted five CD4+ T-cell subpopulations (TN , TCM , Previous studies have shown a prospectively lower TTM , TEM , and TTD ) from 14 LoViReT individuals viral reservoir in individuals who started treatment and quantified total HIV-DNA in each subpopula- early [29]; therefore, we questioned whether this tion (Fig. 4a). The reservoir was under the limit of effect was also observed retrospectively when we detection in TN and TTD for most of the LoViReT selected individuals with low-level reservoirs dur- participants (86% and 64%, respectively). ing ART. We compared reservoir levels and HIV expression after classifying LoViReT into chronic- We then evaluated the contribution of each subset treated LoViReT individuals (>6 months since HIV to the HIV reservoir by calculating the frequency of acquisition and initiation of ART, n = 14) and early- each subpopulation of CD4+ T cells (Fig. 4b). Both treated LoViReTs (<6 months since HIV acquisition TN and TTD made negligible contribution to the HIV 314 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. Fig. 3 Measurements of total HIV-DNA and HIV expression in LoViReT (Low Viral Reservoir Treated) individuals treated in the chronic or early phase of HIV infection. (a) Levels of total HIV-DNA, (b) cell-associated HIV-RNA, (c) ultrasensitive viral load in plasma, (d) quantitative viral outgrowth assay, (e) HIV-DNA in rectum biopsies, and (f) HIV-DNA in lymph node biopsies. Open symbols represent values under the limit of detection; in these cases, the limit of detection varied based on sample/volume input. Fig. 4 HIV latency distribution in T-cell subsets in LoViReT (Low Viral Reservoir Treated) individuals. (a) Total HIV-DNA measured using droplet digital polymerase chain reaction and (b) contribution of each subset to the HIV reservoir. LoViReT individuals treated in the chronic phase are represented by a circle, those treated in the early phase are represented by a square, and those with no available information on their status when they initiated antiretroviral therapy are represented by a triangle. Open symbols represent values that are below the limit of quantification; in these cases, the limit of detection varied based on cell input. © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. 315 Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. reservoir. However, unexpectedly, we observed that To complement the molecular measurements, we TTM contributed the most to the total HIV reservoir set up a highly sensitive viral outgrowth assay (median of 47%), closely followed by TEM and TCM based on ex vivo stimulation of 38 million periph- (median 29% and 24%, respectively). eral CD4+ T cells. We observed that 71% of LoViReT individuals did not have replication-competent Therefore, it appears that the distribution of the proviruses (i.e., IUPMs <0.0185). Recent studies HIV reservoir in the CD4+ T-cell subpopulations have shown that autologous contemporaneous IgG of LoViReT individuals is altered, thus potentially antibodies, if added to the culture, can block out- explaining the lower reservoir levels observed. growth of a substantial fraction of viruses in the latent reservoir for HIV-1 by direct neutralization of the virus [30]. Therefore, the already low fre- Discussion quency of reservoir viruses capable of outgrowth in We previously described a new phenotype of the absence of autologous IgG might represent an HIV-infected individuals on ART who harbor low overestimation of the replication-competent reser- levels of HIV-DNA in peripheral blood, namely, voir in these individuals. To our knowledge, this the LoViReT cohort [13]. These individuals had low level of replication-competent virus has only reduced reservoir levels in PBMCs, even before been associated with extreme phenotypes of con- initiation of ART, though their median viral loads trol, such as elite controllers or individuals who were close to 105 HIV-RNA copies/ml, that is sim- have undergone allogeneic stem cell transplanta- ilar to those of control individuals [13]. However, tion and remain on successful ART [3, 4, 24]. the nature, distribution, and functionality of this small HIV reservoir remain unknown. Here, we Despite the low level of replication-competent studied 22 LoViReT participants to understand proviruses, recent studies have shown that ana- the mechanism behind their low HIV reservoir in lytical treatment interruption in individuals with blood and lymphoid tissue. characteristics similar to those of LoViReT predom- inately led to a viral rebound within 2–12 weeks Our results confirm not only a low level of HIV after stopping ART [15, 16, 31]. For that reason, reservoirs in LoViReT individuals in CD4+ T cells, it is important to know whether lymphoid tissue but also that these proviruses were less transcrip- plays a role in the rapid viral rebound observed tionally active and produced fewer new virions. when ART is interrupted. HIV reservoirs have been This finding took the form of lower detectability widely reported to be 5–12 times larger in lymphoid of residual plasma viremia. Our observations are tissue, such as gut-associated lymphoid tissue in line with data from individuals with low levels [32, 33] and lymph node [34], than in peripheral of HIV-DNA, such as elite controllers [2], excep- blood. Our results showed that LoViReT individ- tional elite controllers [4], and post-treatment uals have limited proviruses in the rectum and controllers [6], in whom significantly lower tran- lymph nodes and that, in some cases, these even scriptional HIV-RNA activity and lower residual fall below the limit of detection. Median levels of plasma viremia have also been observed. proviruses in the rectum are only twofold larger than in peripheral blood. Moreover, when com- In addition, when we measured the nature of the pared with historical data from studies performed HIV reservoir in peripheral CD4+ T cells, we found in individuals on ART, we observed that reservoirs lower levels of intact proviruses in LoViReT par- in rectal biopsies were 7.2-fold higher than in ticipants. Nevertheless, those differences seemed the LoViReT individuals. Therefore, we concluded to be the consequence of the lower levels of total that the number of cells with HIV provirus in the reservoir, since the proportion to total proviruses rectum and lymph nodes is less predominant in did not differ between the groups. In addition, this study group than in other individuals on ART. we observed a positive correlation between total and intact proviruses. Interestingly, levels of intact The factor that might explain the total HIV-DNA proviruses were undetectable in nine individuals in dynamic observed in these individuals is the dis- the LoViReT group (56%) and two individuals in the tribution of the proviruses in CD4+ T-cell subpop- control group (10.5%), the latter being consistent ulations. Short-lived TTM and TEM contributed con- with the 12% reported elsewhere [20]. Thus, our siderably to the total HIV reservoir in LoViReTs. IPDA technique is in line with the limit of detection The rapid turnover rate of these cells [35] con- validated in several laboratories. tributes to the expansion of identical genetically 316 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. intact proviruses [36]. However, memory CD4+ T risk alleles in post-treatment controllers [6]. We cells with a shorter cellular half-life are also more found that protective alleles were slightly more easily cleared through successful ART [37], sug- frequent in LoViReTs, though the differences did gesting that they are not a durable reservoir in not reach statistical significance. In contrast, individuals receiving successful ART [38] and thus other studies on individuals similar to those in the potentially explaining the more rapid decay of total LoViReT cohort showed enrichment in HLA protec- HIV-DNA in LoViReT individuals when they started tive alleles, though this was not enough to control ART [13]. viremia after discontinuation of treatment [16]. In our cohort, it seems that other factors at the Furthermore, TTM and TEM are characterized by transcriptome or protein level might be involved higher levels of reactivation upon treatment with in protecting TCM from infection and that this in some latency-reversing agents [39]; this could turn might have helped to attain some degree prove advantageous when trying new cure strate- of control in establishment of the HIV reservoir. gies aimed at reactivating the HIV reservoir. In Nevertheless, it would be interesting to explore contrast, in the general HIV-infected population, whether integration of the proviruses could be the latent reservoir is found predominantly in the undertaken at sites that are not transcriptionally long-lived cells such as TCM [40], while in LoViReT active, as recently shown for elite controllers [5]. individuals, long-lived CD4+ T cells—including TN and TCM —accounted for only 24% of the total Our study has some limitations. First, we did HIV reservoir. This limited contribution of TCM is not obtain leukapheresis or secondary lymphoid also found in the following scenarios: the classic tissues from control individuals, and therefore sooty-mangabey model of attenuated SIV infection those values were compared to historical controls. related to low CCR5 coreceptor expression [41]; In addition, given that we have a limited number HIV-2 infection, where the virus is less pathogenic of LoViReT individuals, we anticipate having a and replicative [42]; post-treatment controllers validation cohort, for which screening has already [6]; and long-term nonprogressors associated with started, and that will contribute to deciphering protective HLA-B*27 or B*57 alleles [43]. the mechanisms of reduction of HIV reservoirs. Nevertheless, it is important to highlight that Although LoViReT individuals have an interesting the screening of our current LoViReT cohort was phenotype with respect to reduction of the HIV carried out at two independent clinical sites in reservoir, this cohort differs from the previously Barcelona, and a similar proportion of LoViReT described phenotypes of control. For example, 13% individuals were obtained at both [13]. of post-treatment controllers had been treated in the early phase of the infection while only 4% In conclusion, we showed that LoViReT individ- had been treated in the chronic phase [14]. Inter- uals differ from other HIV control phenotypes. estingly, we found that the time of initiation of Factors associated with their status are related to treatment does not affect the levels of proviruses impaired establishment of the reservoir in different in LoViReTs, either in peripheral blood or in lymphoid compartments and to altered proviral secondary lymphoid tissues. This observation distribution towards CD4+ T-cell subpopulations contrasts with data from prospective studies, with shorter half-lives. Their special characteris- where the earlier ART is started, the smaller the tics might make them potential candidates for HIV reservoir becomes [44, 45]. LoViReT individuals, in cure strategies. These strategies should be aimed particular, are characterized by a lower reservoir at further boosting the immune response in order and better immune preservation before initiation to effectively control the potential residual virus of ART, which is followed by faster HIV-DNA decay that could reactivate from a minimal HIV reservoir observed under ART [13]. Thus, the LoViReT in these individuals. phenotype appears to be defined by other factors that are independent of the time of initiation of Acknowledgments treatment. We thank the IGTP Cytometry Core Facility and Control of HIV is characterized by the over- staff (M. A. Fernández and G. Requena), IGTP representation of protective HLA-class I alleles in Blood Bank—Banc de Sang i Teixits (J. Grífols, A. long-term nonprogressors [43], elite controllers Alonso), the endoscopy department from Germans [9–11], and exceptional elite controllers [4] and by Trias i Pujol Hospital (J. Boix), the Micrbiology © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. 317 Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. Department from Germans Trias i Pujol Hospital References (A. Hernández), and Nuclear Medicine department 1 Finzi D, Blankson J, Siliciano JD, Margolick JB, Chadwick from Germans Trias i Pujol Hospital (M. Tenesa, K, Pierson T, et al. Latent infection of CD4+ T cells provides a P. Puyalto, R. Rodríguez) for their contribution mechanism for lifelong persistence of HIV-1, even in patients to this publication. We also thank CERCA Pro- on effective combination therapy. Nat Med. 1999;5:512–7. gramme/Generalitat de Catalunya 2017 SGR 252. 2 May ME, Kwaa AK, Blankson JN. HIV-1 reservoirs in elite con- trollers: clues for developing a functional cure? Future Micro- This work was sponsored in part by Grifols, Merck biol. 2017;12:1019–22. 3 Mendoza D, Johnson SA, Peterson BA, Natarajan V, Salgado Sharp & Dohme (IISP 54925), and NIH/NIAID M, Dewar RL, et al. Comprehensive analysis of unique (UM1 AI164561). This research was also sup- cases with extraordinary control over HIV replication. Blood. ported by CIBER -Consorcio Centro de Inves- 2012;119:4645–55. tigación Biomédica en Red- (CB 2021), Insti- 4 Casado C, Galvez C, Pernas M, Tarancon-Diez L, Rodriguez C, tuto de Salud Carlos III, Ministerio de Ciencia Sanchez-Merino V, et al. Permanent control of HIV-1 patho- e Innovación and Unión Europea – NextGenera- genesis in exceptional elite controllers: a model of sponta- neous cure. Sci Rep. 2020;10:1902. tionEU (CB21/13/00063). C. G. was supported 5 Jiang C, Lian X, Gao C, Sun X, Einkauf KB, Chevalier JM, by the PhD fellowship of the Spanish Ministry of et al. Distinct viral reservoirs in individuals with spontaneous Education, Culture, and Sports (FPU15/03698). control of HIV-1. Nature. 2020;585:261–7. J.M.-P. and M.S. are supported by the grants 6 Sáez-Cirión A, Bacchus C, Hocqueloux L, Avettand-Fenoel V, PID2019-109870RB-I00 and PID2020-115931RA- Girault I, Lecuroux C, et al. Post-treatment HIV-1 controllers I00 respectively, from the Spanish Ministry of Sci- with a long-term virological remission after the interruption of early initiated antiretroviral therapy ANRS VISCONTI study. ence and Innovation. The funding organizations PLoS Pathog. 2013;9:e1003211. had no input in any of the following: study design; 7 Deeks SG, Walker BD. Human immunodeficiency virus con- collection, analyses, and interpretation of the data; trollers: mechanisms of durable virus control in the absence writing the manuscript; or decision to submit the of antiretroviral therapy. Immunity. 2007;27:406–16. study for publication. 8 Casado C, Colombo S, Rauch A, Martínez R, Günthard HF, Garcia S, et al. Host and viral genetic correlates of Conflicts of interests clinical definitions of HIV-1 disease progression. PLoS One. 2010;5:e11079. J. M.-P. received an institutional grant and 9 Pereyra F, Jia X, McLaren PJ, Telenti A, de Bakker PIW, educational/consultancy fees from Merck Sharp Walker BD, et al. The major genetic determinants of HIV- & Dohme. Outside the submitted work, he 1 control affect HLA class I peptide presentation. Science. has received institutional grants and educa- 2010;330:1551–7. 10 Boufassa F, Saez-Cirion A, Lechenadec J, Zucman D, tional/consultancy fees from AbiVax, AstraZeneca, Avettand-Fenoel V, Venet A, et al. CD4 dynamics over a Gilead Sciences, Grifols, Janssen, and ViiV Health- 15 year-period among HIV controllers enrolled in the ANRS care. The remaining authors declare that they have French Observatory. PLoS One. 2011;6:e18726. no competing interests. 11 Migueles SA, Sabbaghian MS, Shupert WL, Bettinotti MP, Marincola FM, Martino L, et al. HLA B*5701 is highly asso- Author contributions ciated with restriction of virus replication in a subgroup of HIV-infected long term nonprogressors. Proc Natl Acad Sci U Cristina Gálvez, Javier Martinez-Picado, and S A. 2000;97:2709–14. Maria Salgado conceived and designed the study; 12 Deeks SG, Lewin SR, Ross AL, Ananworanich J, Benkirane M, Cristina Gálvez, Maria del Carmen Garcia- Cannon P, et al. International AIDS Society global scientific Guerrero, Sílvia Bernal, and Maria Salgado strategy: towards an HIV cure 2016. Nat Med. 2016;22:839– performed the experiments; Cristina Gálvez, 50. 13 Gálvez C, Urrea V, Dalmau J, Jimenez M, Clotet B, Monceaux Víctor Urrea, and Maria Salgado analyzed the V, et al. Extremely low viral reservoir in treated chron- data; Cristina Gálvez, Javier Martinez-Picado, and ically HIV-1-infected individuals. EBioMedicine. 2020;57: Maria Salgado interpreted the results; Judith Dal- 102830. mau, Susana Benet, Beatriz Mothe, Lucía Bailón, 14 Namazi G, Fajnzylber JM, Aga E, Bosch RJ, Acosta EP, Sharaf Andrea Martinez, Aroa Nieto, Bonaventura Clotet, R, et al. The Control of HIV after Antiretroviral Medication Lorna Leal, and Felipe García contributed to the Pause (CHAMP) study: posttreatment controllers identified study design, recruitment of participants, sam- from 14 clinical studies. J Infect Dis. 2018;218:1954–63. 15 Pannus P, Rutsaert S, De Wit S, Allard SD, Vanham G, pling, and management of clinical data. Cristina Cole B, et al. Rapid viral rebound after analytical treatment Gálvez, Javier Martinez-Picado, and Maria Salgado interruption in patients with very small HIV reservoir and wrote the paper. All the authors read, reviewed, minimal on-going viral transcription. J Int AIDS Soc. 2020;23: and approved the final version of the paper. e25453. 318 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. 16 Calin R, Hamimi C, Lambert-Niclot S, Carcelain G, Bellet growth of a substantial but variable fraction of viruses in J, Assoumou L, et al. Treatment interruption in chronically the latent reservoir for HIV-1. Proc Natl Acad Sci U S A. HIV-infected patients with an ultralow HIV reservoir. AIDS. 2020;117:32006–77. 2016;30:761–9. 31 Castagna A, Muccini C, Galli L, Bigoloni A, Poli A, Spagnuolo 17 Bayón-Gil Á, Puertas MC, Urrea V, Bailón L, Morón-López S, V, et al. Analytical treatment interruption in chronic HIV- Cobarsí P, et al. HIV-1 DNA decay dynamics in early treated 1 infection: time and magnitude of viral rebound in adults individuals: practical considerations for clinical trial design. with 10 years of undetectable viral load and low HIV-DNA J Antimicrob Chemother. 2020;75:2258–63. (APACHE study). J Antimicrob Chemother. 2019;74:2039–46. 18 Morón-López S, Puertas MC, Gálvez C, Navarro J, 32 Chun T, Nickle DC, Justement JS, Meyers JH, Roby G, Carrasco A, Esteve M, et al. Sensitive quantification of the Hallahan CW, et al. Persistence of HIV in gut-associated lym- HIV-1 reservoir in gut-associated lymphoid tissue. PLoS One. phoid tissue despite long-term antiretroviral therapy. J Infect 2017;12:e0175899. Dis. 2008;197:714–20. 19 Bruner KM, Wang Z, Simonetti FR, Bender AM, Kwon KJ, 33 Yukl SA, Gianella S, Sinclair E, Epling L, Li Q, Duan L, et al. Sengupta S, et al. A quantitative approach for measuring the Differences in HIV burden and immune activation within the reservoir of latent HIV-1 proviruses. Nature. 2019;566:120–5. gut of HIV-positive patients receiving suppressive antiretrovi- 20 Kinloch NN, Ren Y, Conce Alberto WD, Dong W, Khadka P, ral therapy. J Infect Dis. 2010;202:1553–61. Huang SH, et al. HIV-1 diversity considerations in the appli- 34 Günthard HF, Havlir DV, Fiscus S, Zhang Z, Eron J, Mellors cation of the Intact Proviral DNA Assay (IPDA). Nat Commun. J, et al. Residual human immunodeficiency virus (HIV) type 1 2021;12:165. RNA and DNA in lymph nodes and HIV RNA in genital secre- 21 Puertas MC, Salgado M, Morón-López S, Ouchi D, Muñoz- tions and in cerebrospinal fluid after suppression of viremia Moreno JA, Moltó J, et al. Effect of lithium on HIV- for 2 years. J Infect Dis. 2001;183:1318–27. 1 expression and proviral reservoir size in the CD4+ T 35 Macallan DC, Wallace D, Zhang Y, De Lara C, Worth AT, cells of antiretroviral therapy suppressed patients. AIDS. Ghattas H, et al. Rapid turnover of effector-memory CD4+ T 2014;28:2157–9. cells in healthy humans. J Exp Med. 2004;200:255–60. 22 Amendola A, Bloisi M, Marsella P, Sabatini R, Bibbò A, 36 Morcilla V, Bacchus-Souffan C, Fisher K, Horsburgh BA, Angeletti C, et al. Standardization and performance evalua- Hiener B, Wang XQ, et al. HIV-1 genomes are enriched tion of ‘modified’ and ‘ultrasensitive’ versions of the Abbott in memory CD4+ T-cells with short half-lives. mBio. RealTime HIV-1 assay, adapted to quantify minimal residual 2021;12:e0244721. viremia. J Clin Virol. 2011;52:17–22. 37 Trémeaux P, Lenfant T, Boufassa F, Essat A, Mélard A, 23 Vallejo A, Ruiz-Mateos E, Molina-Pinelo S, Soriano-Sarabia Gousset M, et al. Increasing contribution of integrated forms N, De Felipe B, Gutierrez S, et al. Immunovirologic charac- to total HIV DNA in blood during HIV disease progression from teristics of human immunodeficiency virus-infected patients primary infection. EBioMedicine. 2019;41:455–64. consisting mainly of injecting drug users on highly active 38 Soriano-Sarabia N, Bateson RE, Dahl NP, Crooks AM, Kuruc antiretroviral treatment with prolonged virologic failure. Viral JD, Margolis DM, et al. Quantitation of replication-competent Immunol. 2006;19:759–67. HIV-1 in populations of resting CD4+ T cells. J Virol. 24 Salgado M, Kwon M, Gálvez C, Badiola J, Nijhuis M, Bandera 2014;88:14070–7. A, et al. Mechanisms that contribute to a profound reduction 39 Kulpa DA, Talla A, Brehm JH, Pereira Ribeiro S, Yuan S, of the HIV-1 reservoir after allogeneic stem cell transplant. Bebin-Blackwell A-G, et al. Differentiation into an effec- Ann Intern Med. 2018;169:674. tor memory phenotype potentiates HIV-1 latency reversal in 25 Banga R, Procopio FA, Noto A, Pollakis G, Cavassini M, Ohmiti CD4+ T cells. J Virol. 2019;93:969–88. K, et al. PD-1+ and follicular helper T cells are responsible for 40 Chomont N, El-Far M, Ancuta P, Trautmann L, Procopio FA, persistent HIV-1 transcription in treated aviremic individuals. Yassine-Diab B, et al. HIV reservoir size and persistence are Nat Med. 2016;22:754–61. driven by T cell survival and homeostatic proliferation. Nat 26 Leyre L, Kroon E, Vandergeeten C, Sacdalan C, Colby DJ, Med. 2009;15:893–900. Buranapraditkun S, et al. Abundant HIV-infected cells in 41 Paiardini M, Cervasi B, Reyes-Aviles E, Micci L, Ortiz AM, blood and tissues are rapidly cleared upon ART initiation dur- Chahroudi A, et al. Low levels of SIV infection in sooty ing acute HIV infection. Sci Transl Med. 2020;12:eaav3491. mangabey central memory CD4+ T cells are associated with 27 Zhao P, Su X, Ge T, Fan J. Propensity score and prox- limited CCR5 expression. Nat Med. 2011;17:830–6. imity matching using random forest. Contemp Clin Trials. 42 Samri A, Charpentier C, Diallo MS, Bertine M, Even S, Morin 2016;47:85–92. V, et al. Limited HIV-2 reservoirs in central-memory CD4 T- 28 Morón-López S, Gómez-Mora E, Salgado M, Ouchi D, Puertas cells associated to CXCR6 co-receptor expression in attenu- MC, Urrea V, et al. Short-term treatment with interferon ated HIV-2 infection. PLOS Pathog. 2019;15:e1007758. alfa diminishes expression of HIV-1 and reduces CD4+ T- 43 Descours B, Avettand-Fenoel V, Blanc C, Samri A, Mé A, cell activation in patients coinfected with HIV and hepati- Supervie V, et al. Immune responses driven by protective tis C virus and receiving antiretroviral therapy. J Infect Dis. human leukocyte antigen alleles from long-term nonprogres- 2016;213:1008–12. sors are associated with low HIV reservoir in central memory 29 Ananworanich J, Chomont N, Eller LA, Kroon E, Tovanabutra CD4 T cells. Clin Infect Dis. 2012;54:1495–503. S, Bose M, et al. HIV DNA set point is rapidly established in 44 Buzon MJ, Martin-Gayo E, Pereyra F, Ouyang Z, Sun H, Li acute HIV infection and dramatically reduced by early ART. JZ, et al. Long-term antiretroviral treatment initiated at pri- EBioMedicine. 2016;11:68–72. mary HIV-1 infection affects the size, composition, and decay 30 Bertagnolli LN, Varriale J, Sweet S, Brockhurst J, Simonetti kinetics of the reservoir of HIV-1-infected CD4 T cells. J Virol. FR, White J, et al. Autologous IgG antibodies block out- 2014;88:10056–65. © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. 319 Journal of Internal Medicine, 2022, 292; 308–320 Altered HIV reservoir distribution in LoViReT / C. Gálvez et al. 45 Laanani M, Ghosn J, Essat A, Melard A, Seng R, Gousset M, at sampling, % of CD4 T cells at sampling, et al. Impact of the timing of initiation of antiretroviral therapy % of CD8 T cells at sampling, CD4/CD8 ratio during primary HIV-1 infection on the decay of cell-associated at sampling, number of blips, number of viro- HIV-DNA. Clin Infect Dis. 2015;60:1715–21. logical failures, zenith viral load, nadir CD4 T cells, time since undetectable viral load (50 HIV- Correspondence: Maria Salgado and Javier Martinez-Picado, Irsi- Caixa AIDS Research Institute and Institute for Health Science RNA copies/ml), total time with undetectable Research, Hospital Germans Trias i Pujol, Ctra. de Canyet, s/n, viral load, time from diagnosis to undetectable 08916 Badalona (Barcelona), Spain. viral load, area under the curve for viral load Email: msalgado@irsicaixa.es and jmpicado@irsicaixa.es normalized according to time undetectable, and AIDS events. IUPM = infectious units per million. Supporting Information Additional Supporting Information may be found in Supplementary Fig. S2 HLA class I profile the online version of this article: and CCR5wt/32 frequency. (a) Frequency of the protective alleles HLA-B*27 and B*57, risk Supplementary Fig. S1 Study design flow-chart. alleles HLA-B*07:02 and B*35, and genotype *Clinical factors used for matching of the groups CCR5wt/32 in LoViReT individuals and controls. included sex, age at diagnosis, age at sam- (b) Frequency of individuals with the CCR532 pling, CD4 T cells at sampling, CD8 T cells mutation in heterozygosis. 320 © 2022 The Authors. Journal of Internal Medicine published by John Wiley & Sons Ltd on behalf of Association for Publication of The Journal of Internal Medicine. Journal of Internal Medicine, 2022, 292; 308–320