A Dynamic Interplay of Circulating Extracellular Vesicles and Galectin-1 Reprograms Viral Latency during HIV-1 Infection

To investigate the role of Gal-1 in HIV-1 pathogenesis, we first studied serum concentrations of this lectin during HIV infection. By evaluating 69 samples from 13 healthy HIV-seronegative donors and 51 HIV-infected individuals (see Table S1 in the supplemental material), we detected considerably larger amounts of circulating Gal-1 in HIV-1-infected than in uninfected individuals (P < 0.0001 [Fig. 1A]). To better understand the spatiotemporal and context-dependent regulation of this glycan-binding protein, we classified our study group according to the course of the disease (early versus chronic infection) and to treatment status (untreated or on cART). In addition, we analyzed a group of elite controllers, defined as individuals with undetectable viral load and preserved CD4⁺ T cell counts (median = 869.5 cells/mm³; interquartile range from 25 to 75% [IQ25-75%] = 760.75 to 942) in the absence of cART. Thus, four groups were constituted: (i) early infection, baseline (treatment naive); (ii) chronic infection, treatment naive; (iii) chronic infection, on cART; and (iv) elite controllers. Remarkably, we observed that despite highly significant differences in viral load and CD4⁺ T cell numbers (Table S1), serum concentrations of Gal-1 were similar in all four groups analyzed (Fig. 1B), suggesting that the HIV-1-associated increase of Gal-1 is independent of viral load and immunological status. This was further confirmed by a longitudinal analysis of serum Gal-1 levels over time (between 12 and 48 months) in 10 individuals. Samples were obtained before initiation of cART (baseline) and at different periods following initiation of treatment. We observed that Gal-1 serum concentrations did not change significantly over time within each individual (Fig. 1C), even though cART treatment was effective, as evidenced by a dramatic decrease in viral load (Fig. 1D) and a sustained increase in CD4⁺ T cell count (Fig. 1E). Thus, increased concentrations of Gal-1 in sera from HIV-1-infected individuals are independent of changes in viral load, CD4⁺ T cell count, and treatment status.

To expand our results further beyond this local cohort, we conducted a meta-analysis on publicly available transcriptomics data from HIV-infected patients (52). First, we narrowed all available studies down to 14 targeted studies including 15 data sets (Table S2) and analyzed LGALS1 (encoding Gal-1) expression. In 10 out of 15 data sets included, LGALS1 mRNA was increased (fold change >1) in infected individuals, even in those undertaking cART (Fig. S1). Supporting our findings (Fig. 1C to E), no significant changes were observed in LGALS1 mRNA expression in peripheral blood mononuclear cells (PBMCs) from HIV-infected individuals at baseline and after 1 year of initiation of cART (Bl versus cART condition in study 11 [Fig. S1]). Taken together, our results and meta-analysis of published data demonstrate that HIV infection itself is associated with increased transcription and secretion of Gal-1. Interestingly, as revealed by our meta-analysis, LGALS1 mRNA was particularly upregulated in myeloid cells isolated from subjects under cART, suggesting that these cells could be a major source of this lectin.

We next explored the association between serum Gal-1 concentrations and three inflammatory markers previously shown to be increased in cART-treated individuals (53–57). We observed a positive correlation between serum levels of Gal-1 and sCD14 (Fig. 1F) and IP-10 (Fig. 1G) but did not find any correlation with sCD163 (Fig. 1H).

Finally, we investigated the association between serum Gal-1 concentrations and the reservoir size in circulating CD4⁺ T cells, as estimated by the copy numbers of unspliced RNA (US-RNA) or total viral DNA (57, 58) (Table S3). Interestingly, Gal-1 serum levels positively correlated with US-RNA in CD4⁺ T cells (Fig. 1I). In contrast, we did not observe any correlation between Gal-1 concentrations and total DNA copy numbers, although a trend toward an association between Gal-1 serum levels and DNA copies was observed (Fig. 1J). Given that total DNA reflects the size of the proviral reservoir, whereas US-RNA levels mirror HIV-1 transcriptional activity during effective treatment, these results suggest that increased circulating Gal-1 could promote HIV-1 transcription in vivo.

Thus, plasma from HIV-1-infected individuals contains increased levels of circulating Gal-1, which are independent of viral load or CD4⁺ T cell counts but are associated with inflammation and with HIV-1 transcriptional activity.

Since EVs isolated from plasma of HIV-1-infected individuals stimulate monocyte-derived macrophages (MDMs) to release proinflammatory cytokines (32), and our results showed an association between elevated Gal-1 levels and inflammatory parameters in sera during HIV infection (Fig. 1), we next analyzed whether EVs isolated from plasma of HIV-1-infected individuals (EV_HIV) could trigger Gal-1 production by macrophages. EV_HIV or EVs from healthy donors (EV_HD) were isolated by size exclusion chromatography (SEC) as previously described (32). This fractionation method is based on the size of plasma components: large components such as EVs do not reach the packed matrix pores and are eluted first, whereas soluble proteins are trapped inside the matrix pores and elute later. The purity of EVs eluted in fractions 4 to 6 was assessed by immunoblot analysis; the results showed that the EV-containing fractions were positive for tetraspanins CD63 and CD9 and did not include soluble IgG, which was eluted in later fractions (Fig. 2A). Further characterization revealed that all EV preparations, from HIV⁺ and healthy individuals, were enriched in canonical EV markers, including tetraspanins CD63 and CD9 (membrane-associated proteins) and cytosolic proteins, including ALIX and HSP70, compared to the non-EV fractions (Fig. 2B). In contrast, plasma soluble IgG, readily detectable in the non-EV fractions, was mostly excluded from the EV preparations. In addition, very low concentrations of albumin (less than 0.1 mg/mL) were detected in EV fractions (Fig. 2C), indicating that the EV isolation method was highly effective in removing plasma soluble proteins. Finally, APOA1, a high-density lipoprotein (HDL) marker, was mostly present in the non-EV fraction (Fig. 2B), indicating very low levels of contaminant lipoproteins in the EV fractions. Since EVs and HIV virion particles could be coisolated by SEC, in this study, we used samples from HIV-infected individuals on cART with undetectable viral load. However, to verify that our EV preparations were free of viral particles that could complicate the interpretation of the results, the presence of HIV RNA was investigated on EV preparations by quantitative PCR (qPCR). As expected, donors with undetectable HIV-1 in plasma also showed undetectable HIV-1 RNA on isolated EVs (Table 1). In addition, transmission electron microscopy (TEM) visualization showed the characteristic double-membrane cup-shaped structures in both healthy and HIV⁺ EV samples (Fig. 2D). To further characterize our EV preparations, we analyzed EV size distribution and concentration by different particle analysis platforms, including nano-tracking analysis (NTA) (Fig. 2E and F), dynamic light scattering (DLS) (Fig. S2A), and microfluidics resistive pulse sensing (MRPS) (Fig. S2B and C). EV_HD and EV_HIV were comparable in size as determined by all methodologies applied. NTA showed mean diameters of 199.0 ± 6.6 nm and 196.9 ± 3.5 nm and mode diameters of 146.5 ± 7.0 nm and 142.1 ± 3.5 nm for EV_HD and EV_HIV, respectively, which were not significantly different. The EV concentrations measured by NTA on isolated EVs and calculated in relation to the original plasma sample were 2.4 ± 0.6 × 10⁸ and 4.6 ± 0.9 × 10⁸ EVs/mL for EV_HD and EV_HIV, respectively, showing a slight increase in circulating EVs in HIV⁺ individuals (1.9-fold). Taken together, these data indicate that plasma EVs from healthy and HIV-positive individuals have similar protein compositions, show comparable diameters, and are present at similar frequencies in the circulation.

We next analyzed the functionality of EVs isolated from HIV-infected individuals and from healthy donors by stimulating MDMs and monitoring the production of Gal-1. We observed that MDMs incubated with EV_HIV secreted larger amounts of Gal-1 than did MDMs incubated with EV_HD (Fig. 2G). Moreover, the amounts of secreted Gal-1 increased with longer incubation periods (Fig. 2H), suggesting that Gal-1 is actively produced by MDMs following stimulation with EVs. As expected (32), EV_HIV also triggered IL-6 production by MDMs (Fig. 2I). Thus, EVs circulating in the plasma of HIV-1-infected individuals stimulate the release of Gal-1 along with IL-6 from macrophages, suggesting a regulatory role for this glycan-binding protein during HIV-1 infection and persistent immune activation.

Given the association of Gal-1 with the transcriptional activity of the HIV reservoir (Fig. 1I), we hypothesized that Gal-1 released by EV-stimulated macrophages could reprogram viral latency. To evaluate this hypothesis, conditioned medium from MDMs previously stimulated with EV_HD or EV_HIV was collected and added to J-LAT cells (clone 10.6), cells of a Jurkat-derived clone containing latent HIV-1 in single integration site engineered to express green fluorescent protein (GFP) when latency is reverted (59). Since J-LAT cells do not express Gal-1 themselves (Fig. S3), they represent an attractive model to analyze the effect of extracellular Gal-1 stimulation. Fluorescence-activated cell sorting (FACS) analysis of J-LAT cells exposed to conditioned medium from EV_HIV-stimulated macrophages revealed considerably greater HIV-1 latency reversal than in cells cultured with conditioned medium from macrophages stimulated with EV_HD (Fig. 3A and B). Interestingly, 2-fold serial dilutions showed a dose-response effect unveiling two functionally different groups, namely, strong and weak inductors (Fig. S4A and B). Interestingly, direct stimulation of J-LAT with EVs did not trigger HIV latency reversal (Fig. S4C). Since conditioned medium contains not only Gal-1 but also proinflammatory cytokines and other soluble factors that could modulate HIV-1 latency reversal, we evaluated the specific contribution of Gal-1 by silencing Gal-1 expression in MDMs. Silencing efficiency with two different short hairpin RNAs (shRNAs) was determined by qPCR and was revealed to be more than 90% (Fig. 3C). Next, control (scrambled [SCR]) or Gal-1-silenced MDMs were stimulated with EV_HIV for 24 h. Conditioned medium was then collected and added to J-LAT cells. We found that silencing Gal-1 in MDMs prior to stimulation with EV_HIV significantly reduced reactivation of latent HIV infection (Fig. 3D and E), indicating that Gal-1 secretion by EV-treated macrophages controls HIV latency reversal.

To further dissect the contribution of Gal-1 to the modulation of the inflammatory response triggered by EV_HIV, we performed a cytokine bead array on conditioned medium from SCR and Gal-1-silenced MDMs before and after EV_HIV stimulation. Analysis of basal vehicle-treated cultures revealed no significant changes between the experimental conditions in tumor necrosis factor alpha (TNF-α), IL-6, IL-10, or IL-8, while IL-1β and IL-12p70 showed a slight yet significant increase in Gal-1-silenced MDMs (Fig. S5A). Interestingly, Gal-1-silenced MDMs showed a clear reduction in TNF-α and IL-1β relative to SCR MDMs after EV_HIV stimulation (Fig. 3F and G), while other cytokines exhibited no significant changes (Fig. S5B).

To further confirm the role of Gal-1 in HIV latency reversal in J-LAT cells, we next stimulated five different clones of this reporter cell line (6.3, 8.4, 9.2, 15.4, and 10.6, each clone exhibiting a different HIV integration site) with recombinant Gal-1 (rGal-1). We found dose-dependent latency reversal in all clones exposed to rGal-1 (Fig. 4A).

Given the contribution of glycosylated ligands to the extracellular functions of Gal-1 (33, 37), we next assessed the carbohydrate-dependent binding of Gal-1 to J-LAT cells (clone 10.6) using lactose (30 mM), a galectin-specific competitive disaccharide (Fig. 4B). Interestingly, rGal-1 elicited dose-dependent HIV reactivation (Fig. 4C and D), which was prevented by addition of this saccharide. As expected, rGal-1 stimulation also influenced cell viability in a dose-dependent fashion as measured by annexin V and 7-aminoactinomycin D (7-AAD) staining (Fig. 4C and Fig. S6). Since the reduced and dimeric forms of Gal-1 are critical for most extracellular functions of this protein (60), we evaluated an oxidized form of Gal-1 (oxGal-1) (61) and a stable monomeric Gal-1 mutant (monGal-1 [V5D]) as potential regulators of HIV latency reversal. Our results showed that both monGal-1 and oxGal-1 failed to reactivate HIV-1 in J-LAT cells (Fig. 4E), thus emphasizing the importance of the redox status and dimerization balance in Gal-1-driven HIV latency reversal. Finally, pretreatment of J-LAT cells with tunicamycin, an inhibitor of the N-glycosylation pathway, markedly reduced binding of Gal-1 (Fig. 4F) and prevented Gal-1-driven latency reversal (Fig. 4G) in J-LAT cells, highlighting the importance of cell surface N-glycans in Gal-1-mediated reactivation of latent HIV-1 infection.

Based on these findings, we evaluated whether TNF-α could potentiate Gal-1-driven HIV latency reversal. As expected, stimulation of J-LAT cells (clone 10.6) with TNF-α (0.03 ng/mL) efficiently reversed HIV latency. Interestingly, simultaneous stimulation with TNF-α and different rGal-1 concentrations not only potentiated the effect of this proinflammatory cytokine but also lowered the threshold of Gal-1 stimulation, as the effect of rGal-1 at its lowest dose (2.5 μM) was similar to that at its highest concentration (7 μM) when used alone (Fig. 4H).

Thus, dimeric Gal-1, released by EV-stimulated macrophages, promotes HIV-1 latency and reversal in a glycan-dependent fashion.

HIV-1 transcription is dependent on the activation of host transcription factors, with NF-κB playing a prominent role in this process (62, 63). Given the relevance of NF-κB in Gal-1 function (35, 64), we hypothesized that Gal-1 action on HIV-1 latency reversal is mediated by NF-κB activation. To test this hypothesis, we analyzed expression of IκB-α, a regulatory protein that interrupts NF-κB activation by trapping this transcription factor within the cytoplasmic compartment. Exposure to rGal-1 reduced IκB-α expression in J-LAT cells (Fig. 5A), suggesting its ability to trigger NF-κB activation in HIV latently infected cells. This effect was further confirmed by analyzing nuclear translocation of p65 by immunofluorescence (Fig. 5B and C). Furthermore, treatment with an NF-κB inhibitor (BAY 117082) (Fig. 5D, light gray bars) inhibited latency reversal driven both by Gal-1 and l-phytohemagglutinin (PHA-L), a positive control, thus substantiating the relevance of the NF-κB signaling pathway in Gal-1 effects. Thus, Gal-1 promotes glycan-dependent HIV-1 latency reactivation via activation of the NF-κB pathway.

To study the effects of Gal-1 in reversing HIV-1 latency in a more physiologic setting, we first confirmed binding of this lectin to the surface of human primary CD4⁺ T cells (Fig. 6A) and analyzed its function in an in vitro model of HIV latency (65, 66). Briefly, resting CD4⁺ T cells isolated from the blood of healthy donors were pretreated with CCL19, a facilitator of the establishment of latent HIV infection for 24 h. Cells were then infected with a high multiplicity of infection of HIV-1-GFP. At day 5 postinfection, cells were sorted by FACS to remove GFP-positive cells (productively infected) (Fig. 6B). GFP-negative cells (including uninfected and latently infected cells) were exposed to different stimuli to reactivate latent HIV-1. As expected (65, 66), stimulation with anti-CD3/CD28 monoclonal antibody (MAb) and IL-7 (50 ng/mL) reversed HIV latency, as revealed by the increase in GFP expression (Fig. S7). Although purified rGal-1 was not sufficient to trigger HIV latency reversal when added alone to primary cultures, this lectin significantly potentiated reactivation of cells when they were cotreated with suboptimal doses of the activating stimulus PHA (Fig. 6C).

To further understand the physiopathological relevance of Gal-1-driven latency reversal, we performed an ex vivo assay on CD4⁺ T cells obtained from peripheral blood samples of cART-treated HIV-infected individuals (Fig. 6D and E). Consistent with the results obtained in the in vitro model, while suboptimal activation with phorbol myristate acetate (PMA) was not sufficient to induce viral transcription activity, as revealed by US-RNA quantification, costimulation with PMA and rGal-1 resulted in increased US-RNA detection (Fig. 6E). Together, in vitro and ex vivo experiments on primary CD4⁺ T cells support a role for Gal-1 in HIV-1 latency reactivation.

Collectively, our results indicate that Gal-1 secretion by macrophages triggered by EVs promotes latency reversal of infected CD4⁺ T cells in an N-glycan-dependent fashion (Fig. 6F).

A total of 64 subjects participated in this study: 13 healthy HIV-seronegative donors (HD) and 51 HIV-infected individuals, of whom 15 were enrolled during primary HIV infection (baseline), 14 were chronically infected and treatment naive (chronic treatment naive), 22 were chronically infected and under treatment (chronic under treatment), and 9 were elite controllers (EC) (Table S1). Subjects were enrolled under the following inclusion criteria: (i) detection of HIV RNA and (ii) treatment-naive status. The chronic treatment naive cohort was defined as subjects with established HIV infection for over 3 years, detectable viral load (VL; >0.50 HIV RNA copies/mL of plasma) and combined antiretroviral therapy (cART) naive, while subjects chronic under treatment were defined as subjects with established HIV infection, undetectable viral load (<0.50 HIV RNA copies/mL of plasma), and under cART for at least 1 year. EC were defined as subjects infected for more than 5 years with undetectable VL (<50 HIV RNA copies/mL of plasma) and CD4⁺ T cell counts of 0.450 cells/mL of blood and who were cART naive and had no record of opportunistic infections and/or AIDS-related diseases.

To isolate plasma EVs, HIV-positive individuals were recruited at the Fundación Huésped Medical Center (Buenos Aires, Argentina). Blood samples were collected in citrate anticoagulant tubes and processed within 1 h of venipuncture to isolate plasma for EV purification. Healthy donors were voluntary blood donors at Blood Center of the Julio Mendez Hospital (Buenos Aires, Argentina). All healthy donors were individuals older than 18 years who had completed and passed a survey on blood donation and were screened for serological markers before being accepted as donors.

The following antibodies were used: Alexa 594-anti-rabbit, Alexa 647-anti-rabbit, streptavidin Alexa 674 and Alexa 594, and anti-human IgG (Jackson ImmunoResearch Laboratories, Inc.); rabbit polyclonal anti-mouse IΚB-α (Santa Cruz Biotechnology); and mouse anti-CD63 and mouse anti-CD9 (BD Bioscience). The rabbit polyclonal anti-Gal-1 antibody was obtained as previously described (90). Recombinant human Gal-1 (rhGal-1) and its monomeric variant (V5D) were produced and purified as described previously (90, 91). To obtain the oxidized form, rhGal-1 was treated with H₂O₂ as previously described (61). In all cases, biotinylation was carried out by incubating 1 mg of rhGal-1 or its variants with 112 nmol of biotin for 30 min at room temperature. Dialysis was performed overnight in 1 mM 2-mercapthoethanol.

The human CD4⁺ T cell line J-LAT (clone 10.6) was obtained from the NIH AIDS Reagent Program, Division of AIDS (NIAID, NIH). J-LAT clones 6.3, 8.4, 9.2, and 15.4 were kindly provided by Kenneth W. Witwer. HEK 293T cells were obtained from the ATCC (CRL-11268). J-LAT cells were cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS).

NL4-3-IRES-eGFP, encoding full-length HIV-1 in the pBR322 backbone under the control of a viral long terminal repeat promoter, was kindly provided by F. Kirchhoff (Institute of Molecular Virology, Ulm University Medical Center, Ulm, Germany). shRNA-expressing lentiviruses were obtained from Sigma (Mission shRNA).

Lentiviruses were produced as described elsewhere (92). In brief, 2.5 × 10³ HEK 293T cells were seeded on a flat-bottom 96-well plate. Twenty-four hours later, cells were transfected with a mix of 100 ng of pCMV-dR8.2 DVpr, 100 ng of the target’s specific shRNA in the pLKO.1 backbone, and 10 ng of pCMV–VSV-G per well, using the X-tremeGENE HP DNA transfection reagent (Roche), following the manufacturer’s recommendations. Twenty-four hours later, the medium was replaced, and supernatants containing lentiviral particles were collected at 72 h after transfection and precleared by centrifugation.

J-LAT cells (clones 6.3, 8.4, 9.2, 10.6, and 15.4) were seeded at 3 × 10⁵ cells/mL in 96-well plates and stimulated with different rhGal-1 concentrations or with PMA/PHA as positive controls. PHA and PMA were used at concentrations of 1 μg/mL and 0.5 ng/mL, respectively, unless otherwise specified. After 24 h, cells were harvested and cell viability was assessed by annexin V and 7-AAD staining. Latency reversal was quantified by flow cytometry as the percentage of cells positive for enhanced green fluorescent protein (eGFP) in the annexin V⁻/7-AAD⁻ gate. Alternatively, cells were pretreated with tunicamycin (4 μg/mL) overnight and then incubated as previously described (93).

Monocytes were isolated from buffy coats of healthy anonymous donors from the Blood Center of Julio Méndez Hospital in Buenos Aires, Argentina, by Ficoll density gradient centrifugation and positive magnetic isolation using anti-CD14-coated beads (Miltenyi). Isolated monocytes were plated for 2 h without serum and subsequently differentiated into macrophages by culturing in complete RPMI 1640 supplemented with 50 ng/mL of GM-CSF for 5 days.

Resting CD4⁺ T cells were negatively purified using immunodensity negative selection cocktail (RosetteSep human T cell enrichment cocktail immunodensity negative-selection cocktail; StemCell) from buffy coats of healthy donors. Cells were maintained in RPMI 1640 supplemented with 10% FBS (Gibco) in the presence of IL-2.

Silencing of gene expression was performed as described (94). Briefly, 1 × 10⁶ MDMs were seeded in the presence of GM-CSF and lentiviral vectors plus simian immunodeficiency virus (SIV)-like particles in the presence of 8 μg/mL of Polybrene. Forty-eight hours later, transduced cells were selected by the addition of 3 μg/mL of puromycin. Before functional studies, SCR and Gal-1-knocked down (Gal-1-KD) MDMs were harvested and reseeded in 96-well plates at 8,000 cells per well. Conditioned medium from SCR and Gal-1KD MDMs stimulated with EV_HIV or EV_HD for 24 h was used to analyze HIV latency reversal in J-LAT cells. Conditioned medium was titrated by 2-fold dilutions on 30,000 J-LAT cells as described above (Fig. S3A and B). Then a 1:2 dilution was selected for further assays.

Isolation of EVs from human plasma was performed as described previously (32), following an adaptation of the original protocol (95). Briefly, whole blood drawn by venipuncture was collected in citrate-containing Vacutainer tubes (BD). Platelet-rich plasma (PRP) was obtained by centrifugation (300 × g, 10 min). PRP was supplemented with 200 nM Prostaglandin I2 (PGI2) to avoid platelet activation and centrifuged (2,000 × g, 15 min) to obtain cell-free plasma. Plasma (2 mL) was loaded onto a homemade size exclusion chromatography (SEC) column (resin, CL-2B; GE Healthcare; support, 12-mL empty cartridges with 20-μm hydrophobic frits; Applied Separations). Twelve fractions (1 mL each) were eluted using 0.9% NaCl–0.38% sodium citrate. Each fraction was analyzed for the presence of the EV markers CD63 and CD9 and soluble IgG by immunoblot analysis. Additionally, equal volumes of extracts were separated by 12% SDS-PAGE under nonreducing conditions and analyzed for CD63, CD81, and IgG, or under reducing conditions for ALIX, HSP70 and APOA1, and blotted onto polyvinylidene fluoride (PVDF) transfer membranes (Thermo Fisher Scientific). Primary antibodies used were as follows: CD63 (BD Pharmingen; catalog number 556019, clone H5C6), CD81 (BD Pharmingen; catalog number 555675, clone JS-81), ALIX CST (catalog number 92880, clone E6P9B), HSP70 (Enzo; catalog ADI-SPA-810, clone C92F3A-5), and APOA1 (CST; catalog number 3350S, clone 5F4). Primary antibody dilutions used for blotting were 1:10,000 for IgG and 1:1,000 for the rest of the antibodies. Blots were revealed using SuperSignal West Pico PLUS chemiluminescent substrate (Thermo Fisher Scientific), and images were acquired with the BioSpectrum-815 imaging system (UVP). Finally, EV-containing fractions were pooled, concentrated by centrifugation, resuspended in 50 μL of phosphate-buffered saline (PBS), and used within the same day for functional studies.

MDMs (250,000 in 300 μL) were cultured in 48-well plates and stimulated for 24 h with 25 μL of EV_HIV/HD concentrate. SCR or Gal-1-silenced MDMs were stimulated with 10 μL of EV_HIV/HD concentrate per well. To detect albumin as a possible contaminant in EV samples, a sensitive immunoturbidimetric method was performed using a commercial kit for microalbuminuria diagnosis (Wiener; Microalbumin Turbitest AA). Diluted plasma samples (1:500) were used as positive controls.

Plasma samples were obtained from EDTA-anticoagulated whole-blood venipuncture from 9 healthy and 9 cART-treated HIV⁺ fasted donors. Plasma EVs were isolated by SEC, followed by centrifugation at 30,000 × g for 90 min as described above and resuspension in PBS. The size distribution and concentration of EVs were analyzed using NanoSight NS300 equipment (NanoSight, Amesbury, UK). The analysis was performed using the 532-nm laser and 565-nm long-pass filter, with a camera level of 12 and detection threshold of 5. EVs were diluted in PBS before the analysis.

EV size distribution by dynamic light scattering (DLS) was analyzed as described previously (96). Briefly, EVs from healthy and HIV⁺ donors were purified from 2 mL of pooled plasma (n = 4 for each condition) by SEC and concentrated by centrifugation. Samples were diluted in PBS to a final volume of 1,000 μL (1:100). The hydrodynamic diameter (D_h) and the size distribution (polydispersity index [PDI]) of the EV samples were assayed by DLS (Zetasizer Nano-ZS; Malvern Instruments, Worcestershire, UK) at a scattering angle of 173°. The Nano-ZS contains a 4-mW He-Ne laser operating at a wavelength of 633 nm, a digital correlator (ZEN3600), and noninvasive back scatter (NIBS) technology. All samples were analyzed at 4°C. The refractive index (RI) was 1.48, and viscosities were between 1.544 and 1.569 cP (4°C). Results are expressed as means ± standard deviations (SD) from at least six runs for each specimen.

Plasma samples were obtained from EDTA-anticoagulated whole-blood venipuncture from 3 healthy and 3 cART-treated HIV⁺ fasted donors. Plasma EVs were isolated by SEC followed by centrifugation at 30,000 × g for 90 min as described above. Purified EVs were resuspended in 50 μL PBS. Microfluidics resistive pulse sensing (MRPS) measurements were conducted using an nCS1 instrument (Spectradyne, Torrance, CA) (97). EVs were diluted 10⁵ to 10⁷ times in 0.1% Tween 20-PBS, and 5 μL was loaded onto polydimethylsiloxane cartridges (diameter range, 65 nm to 400 nm). About 5,000 events were recorded for each sample. All acquired results were analyzed using the nCS1 data analyzer (Spectradyne). For all samples, user-defined filtering was applied by defining two-dimensional (2D) polygonal boundaries based on transition time and diameter to exclude false-positive signals. An additional 80-nm filter was applied in order to compare samples with different background levels.

Plasma samples were obtained from EDTA-anticoagulated whole blood from healthy and cART-treated HIV⁺ fasted donors. Plasma EVs were isolated by SEC followed by centrifugation at 30,000 × g for 90 min as described above. Purified EVs were resuspended in 50 μL of PBS.

A droplet of each EV suspension fixed in 2% paraformaldehyde was mounted on a collodion-coated copper grid (400 mesh) for 20 min. Then grids were washed three times with PBS (pH 7.4) and once with distilled water. Samples were incubated with 4% uranyl acetate for 40 s. Finally, grids were visualized under a JEM 1200 EX II transmission electron microscope (JEOL Ltd., Tokyo, Japan) and photographed by an Erlangshen ES1000W camera (model 785; Gatan Inc., Pleasanton, CA) at the Electron Microscopy Central Service of the School of Veterinary Sciences, University of La Plata. Micrographs were analyzed with the ImageJ software for EV size determination.

Plasma EVs were obtained as described above and resuspended in 1 mL of PBS. HIV-1 RNA was quantified by Cobas HIV-1 test (Roche Diagnostics) in a Cobas 4800 system, which is a fully automatized system for extraction, amplification, detection, and quantification of nucleic acids by automated real-time HIV-1 assay (Abbott).

To analyze cell surface binding of Gal-1, a total of 200,000 cells were incubated with 1.6 mg of biotinylated rhGal-1 for 60 min at 4°C. Cells were subsequently incubated with conjugated streptavidin for 45 min at 4°C. When indicated, binding of Gal-1 to the cell surface was competed by adding 30 mM lactose to the culture medium. To analyze the functionality of this lectin, cells were incubated with increasing concentrations of rhGal-1 (2.5, 3, 5, and 7 μM) in the absence or presence of lactose (30 mM). Gal-1 concentrations were chosen based on those reaching different dimerization ratios, abundance in different pathophysiologic settings, and previously published studies (33, 35, 36, 38, 60, 61, 91).

The HIV latency model was established using primary CD4⁺ T cells from healthy donors as previously described (65, 66). Briefly, resting CD4⁺ T cells were cultured for 24 h in RF10 with 100 nM CCL19. Cells were incubated with NL4-3-IRES-eGFP for 2 h at 37°C, washed, and cultured for 5 days in RF10 plus 2 U/mL of IL-2. At day 5 postinfection, T cells were analyzed for eGFP expression (productive infection) and eGFP⁻ T cells were sorted (FACSAria; BD Biosciences). Aliquots of 1 × 10⁵ sorted cells were cultured for an additional period of 3 days (day 8) in 96-well plates in 200 μL of RF10 plus 2 U/mL of IL-2 either alone (spontaneous eGFP expression) or stimulated with anti-CD3/CD2/CD28 MAb-coated beads (Miltenyi). IL-7 (50 ng/mL) was also added to culture media as a latency reactivation positive control. Additionally, cells were cultured in the presence of PHA-L (1 μg/mL) and rGal-1 (7 μM), either alone or in combination. Cells were analyzed for eGFP expression by flow cytometry (FACSCanto; BD Biosciences) at day 8.

Quantitative real-time PCR for cell-associated (CA) HIV RNA and DNA was performed as follows. CA HIV DNA and unspliced RNA (US-RNA) were determined by real-time PCR on samples obtained once study subjects were on cART. CD4⁺ T cells were isolated from frozen PBMCs using an immunomagnetic selection kit (StemCell Technologies, Canada), and only samples with 90% purity (determined by flow cytometry) were assayed. DNA and RNA were extracted using a Qiagen minikit (AllPrep DNA/RNA minikit; Qiagen, Germany), quantified, and stored at −80°C until use. All samples from each participant were extracted at the same time, run on the same PCR plate, and subsequently analyzed together. Total HIV DNA was determined as described previously (94). Briefly, a single-step real-time PCR was performed using 5 μL of extracted DNA as input per 50 μL of reaction mixture, in triplicate. HIV DNA copy numbers were standardized to cellular equivalents using a CCR5 SYBR green real-time PCR. The lower limit of detection (LLOD) was 1 copy per well. For determination of US-RNA, Pasternak’s protocol was used (58). Briefly, a heminested PCR was performed with 16 cycles of amplification followed by a second amplification round of quantitative real-time PCR. US-RNA copy numbers were standardized to cellular equivalents using an 18S RNA real-time PCR (Invitrogen). The LLOD was 1 copy per well. Each sample was assayed in quadruplicate, and a non-reverse transcriptase control was used to detect DNA contamination. Quantitative real-time PCR assays were run for 40 cycles.

The HIV latency reversal assay was performed as described previously (47), with minor changes. Briefly, CD4⁺ T cells were isolated from peripheral blood samples (20 mL) from HIV-infected individuals under cART using immunodensity negative-selection cocktail (RosetteSep; StemCell). Cells were plated at 10⁶/mL and cultured for 16 h as follows: no stimulus (basal), PMA (50 ng/mL) plus ionomycin (1 μg/mL) as a positive control, suboptimal PMA (2 ng/mL), and suboptimal PMA (2 ng/mL) plus rGal-1 (7 μM). RNA was extracted using a PureLink RNA minikit (Invitrogen), and US-RNA was quantified as described above.

Cellular RNAs were isolated using the RNeasy Plus minikit (Qiagen), and 1 μg of RNA was reverse transcribed with Moloney murine leukemia virus (MMLV) reverse transcription (Invitrogen). Only 1:10 cDNA was used for each PCR, performed with SYBR green (Applied Biosystems) on a real-time thermal cycler (Step One Plus; Applied Biosystems). Cycle thresholds (C_Ts) were normalized to the C_T of β-actin.

Primer sequences were as follows: β-actin Fw, 5′-AGGCATCCTCACCCTGAAGT-3′; β-actin Rev, 5′-GCGTACAGGGATAGCACAGC-3′; LGALS-1 Fw, 5′-TCGCCAGCAACCTGAATCTC-3′; and LGALS-1 Rw, 5′-GCACGAAGCTCTTAGCGTCA-3′.

Cytokines (IL-6 and IP-10) were detected by enzyme-linked immunosorbent assay (ELISA) in cell culture supernatants and serum samples, according to the manufacturer’s instructions (BD Biosciences). sCD14 and sCD163 in serum samples were assessed by ELISA according to the manufacturer’s instructions (Thermo Scientific).

A cytokine bead array (CBA; BD Biosciences) was used to determine TNF-α, IL-1β, IL-6, IL-10, IL-12p70, and IL-8 concentrations in SCR and Gal-1-KD MDM conditioned media, according to the manufacturer’s instructions.

Serum Gal-1 was determined using an in-house ELISA as described previously (35). In brief, high-binding 96-well microplates (Costar; Corning) were coated with capture antibody (2 μg/mL of purified rabbit anti-Gal-1 polyclonal IgG) in 0.1 M sodium carbonate (pH 9.5). After incubation for 18 h at 4°C, wells were rinsed three times with washing buffer (0.05% Tween 20 in PBS) and incubated for 1 h at room temperature with blocking solution (2% bovine serum albumin (BSA) in PBS). One-hundred-microliter volumes of samples and standards were diluted in 1% BSA and incubated for 18 h at 4°C. Plates were then washed and incubated with 100 ng/mL of biotinylated detection antibody (purified rabbit anti-Gal-1 polyclonal IgG) for 1 h. Plates were rinsed three times before addition of 0.3 μg/mL horseradish peroxidase (HRP)-labeled streptavidin (Sigma-Aldrich) for 30 min. After a washing, 100 μL of TMB solution (0.1 mg/mL of tetramethylbenzidine and 0.06% H₂O₂ in citrate phosphate buffer [pH 5.0]) was added to the plates. The reaction was stopped by adding 4 N H₂SO₄. Optical densities were determined at 450 nm in a Multiskan MS microplate reader (Thermo Fisher Scientific). A standard curve ranging from 2.5 to 160 ng/mL of rGal-1 was run in parallel.

For surface staining, cells were washed once in PBS, followed by incubation for 30 min at 4°C with the corresponding antibody. Prior to FACS analysis, cells were washed twice with PBS. For intracytoplasmic staining, cells were fixed with 4% paraformaldehyde (PFA), washed, and permeabilized according to the manufacturer’s instructions (BD CytoFix and CytoPerm kit). Cells were subsequently incubated with the conjugated antibody for 45 min at room temperature. Cells were acquired on a FACSCanto instrument (BD) and analyzed using FACSDiva software (BD).

A total of 10⁵ cells were seeded on poly-l-lysine-coated glass coverslips for 30 min, fixed in 4% paraformaldehyde, quenched with 0.1 M glycine, permeabilized in ice-cooled methanol for 7 min, and incubated with a primary mouse anti-human p65 NF-κB antibody (BD Biosciences) overnight. After extensive washing, cells were incubated with Alexa Fluor 594-labeled donkey anti-mouse secondary antibodies (Jackson ImmunoResearch Laboratories, Inc.). The coverslips mounted with 4′,6-diamidino-2-phenylindole (DAPI) Fluoromount-G (SouthernBiotech) were examined under a Zeiss LSM800 confocal microscope using a Plan Apochromat 60 × 1.42 numerical-aperture (NA) oil immersion objective. Images were analyzed using NIS-Element software.

Cells were lysed in precooled radioimmunoprecipitation assay buffer (1% Triton X-100, 0.1% SDS, 50 mM Tris [pH 7.5], 150 mM NaCl, and 0.5% sodium deoxycholate) supplemented with a cocktail of protease inhibitors (Roche) and were cleared from nuclei by centrifugation at 15,000 × g for 5 min. Equal amounts of protein extracts were separated by 4 to 12% SDS-PAGE and blotted onto a PVDF transfer membrane (Thermo Fisher Scientific) under reducing conditions. Blots were revealed using SuperSignal West Pico chemiluminescent substrate (Thermo Fisher Scientific).

Studies were selected to analyze and compare LGALS1 transcript levels in uninfected healthy donors and HIV-infected patients. Studies involving coinfection and/or comorbidity cohorts, analysis of clinical outcomes based on drug interventions, and SIV infection were excluded. All studies included an uninfected control sample. A total of 14 studies including 15 data sets that qualified for our meta-analysis were analyzed. Detailed information can be found in Table S3. Data were analyzed with the GXB software available at http://hiv.gxbsidra.org/dm3/geneBrowser/list.

Data were analyzed using Prism (GraphPad Software). Normality of the data was tested using the Kolmogorov-Smirnov test. Based on the normality test, either one-way analysis of variance (ANOVA) followed by the Tukey’s honestly significant difference (HSD) posttest or the Kruskal-Wallis test followed by Dunn’s posttest was used for multiple-comparison analyses. Due to the disparity in the FACS values obtained in different experiments, data were normalized to control conditions to show pooled results from several experiments when indicated.

All samples from HIV-infected subjects were collected after written informed consent was obtained. For in vitro experiments, CD4⁺ primary T cells were isolated from buffy coats of healthy anonymous donors from the Blood Center of the Julio Méndez Hospital in Buenos Aires, Argentina. All donors were >18 years. In this case, consent was not obtained [exemption 45 CFR 46.101(b), HHS] because samples had not been collected specifically for this research study and were supplied without personal identifiable information. None of the investigators on this research project have any ready means to link the materials back to living individuals.