In Human Immunodeficiency Virus primary infection, early combined antiretroviral therapy reduced γδ T‐cell activation but failed to restore their polyfunctionality

Summary

Primary and chronic human immunodeficiency virus (HIV) infection alters γδ T‐cell features. However, there is no evidence about early combined antiretroviral therapy (cART) and γδ T‐cell dynamics. In the present study, HIV‐positive individuals were divided into those with early primary infection (EPI) and those with late primary infection (LPI). The analysis of γδ T cells was performed by flow cytometry before and after therapy. Polyfunctional profile was assessed after in vitro peripheral blood mononuclear cell (PBMC) exposure to specific antigens. The results show that primary infection induced an expansion of Vδ1 T cells in LPI. Before treatment, a massive activation of γδ T‐cell subsets was observed in both groups of patients, that correlated with disease progression and was significantly reduced after cART introduction. Despite this, CD107A‐expressing Vδ1 T cells in both groups were significantly fewer than in healthy donors, but were restored by therapy introduction. Polyfunctional analysis of Vδ1 T cells from HIV‐positive individuals revealed a lower frequency of CD107A⁺ CCL‐4⁺ Vδ1 T‐cell subsets than healthy donors that persists after therapy. Functional profile of Vδ2 was similar to that in healthy donors before therapy but, at 6 months, a lower frequency of CD107A, interferon‐γ‐ or tumor necrosis factor‐α‐producing Vδ2 T cells was observed in the EPI group. Finally, individuals with LPI showed a lower frequency of quadruple‐functional Vδ2 T‐cell subset. In conclusion, during primary HIV infection, the baseline Vδ1 T‐cell activation is correlated with immune reconstitution potential. Moreover, an altered γδ polyfunctional profile occurred, persisting after cART. Further studies are needed to understand whether a longer treatment of primary infection may increase γδ T‐cell functionality.

Abbreviations

cART

combined antiretroviral therapy

EPI

early primary infection

HD

healthy donors

HIV

human immunodeficiency virus

LPI

late primary infection

PBMCs

peripheral blood mononuclear cells

Introduction

γδ T cells represent unconventional T lymphocytes that are considered as a bridge between innate and adaptive immunity. In humans, there are two major subsets of γδ T cells, of which Vδ1 T cells predominate in the thymus and peripheral tissues and Vδ2 T cells constitute the majority of γδ T cells in peripheral blood.1, 2 Vδ1 T cells recognize stress‐related molecules, whereas Vδ2 T cells recognize phosphorylated non‐peptide antigens, which are intermediates of the mevalonate pathway.1, 2 Both Vδ1 and Vδ2 T cells can exert a potent cytotoxic capability against pathogens.3, 4, 5

Human immunodeficiency virus (HIV) induces CD4 T‐cell depletion and immune suppression. Both Vδ1 and Vδ2 T cells are altered in chronic HIV infection. Vδ1 T cells increase while Vδ2 T cells are depleted, determining in peripheral blood an inverted Vδ2 : Vδ1 ratio.6, 7, 8

Vδ1 T cells might play a major role in the control of HIV infection and in defense against opportunistic infections by producing cytokines and proliferating.9, 10 Conversely, Vδ2 T cells decrease during HIV disease and show limited proliferation and cytokine responses after phosphoantigen stimulation.6, 11

The role of γδ T cells in primary infection is not well elucidated. Early γδ T‐cell activation was indirectly associated with the progression of AIDS.12 Nevertheless, Vδ2 T cells act as T regulatory cells during primary HIV infection, participating in the control of immune activation thought transforming growth factor‐β production.13 Later, during the chronic phase, γδ T cells modify their functionality from an anti‐inflammatory to a pro‐inflammatory cytokine profile, so maintaining the immune activation, suggesting different roles in different phases of the disease.13

Although combined antiretroviral therapy (cART) restores the CD4 T‐cell count, there is an incomplete recovery of T‐cell effector functions.14 Increasing evidence suggests that early cART is associated with a higher capacity for immune reconstitution15 but there is no body of evidence describing the impact of early cART on γδ T‐cell dynamics during primary infection.

The aim of this study was to analyze the phenotype and polyfunctionality of γδ T cells in longitudinally enrolled patients starting cART during primary infection.

Materials and methods

Study population

HIV‐positive patients (n = 27) and healthy donors (HD, n = 7) were enrolled at the National Institute for Infectious Diseases “Lazzaro Spallanzani” in Rome, Italy. Patients were divided into two groups according to Fiebig classification16 as follows: the first groups included 11 patients classified as Fiebig II to IV (referred to hereafter as early primary infection, EPI) and the second group included 16 patients classified as Fiebig V to VI (referred to hereafter as late primary infection, LPI). Patient characteristics are described in Table 1.

Table 1.

Patient characteristics

	HD	EPI (Fiebig II–IV)	LPI (Fiebig V–VI)
No. of subjects	7	11	16
Sex F/M	7/0	1/10	3/13
Age (years), mean ± SD	43·43 ± 3·55	38·27 ± 10·71	33·50 ± 10·82
Therapy
Emtricitabine + Tenofovir + Dolutegravir		7	8
Emtricitabine + Tenofovir + Darunavir 800 + Ritonavir 100 + Raltegravir 400 bid		1	2
Emtricitabina + Tenofovir + Darunavir 800 + Ritonavir 100		0	1
Opportunistic infections		0	0

EPI, early primary infection; HD, healthy donors; LPI, late primary infection.

The study was approved by the Institutional Review Board of the National Institute for Infectious Diseases “Lazzaro Spallanzani” (SIREA study), and signed written informed consent was obtained from all patients.

Blood samples processing

Blood samples were obtained before initiation of cART (T0) and after 6 months of therapy introduction (T6). Peripheral blood mononuclear cells (PBMCs) were purified by density gradient centrifugation (Lympholyte; Cedarlane Laboratories, Burlington, ON, Canada) immediately after collection and frozen for phenotypic and function experiments.

Antibodies

The following anti‐human monoclonal antibodies were used for staining surface antigens and detection of intracellular cytokines: fluorescein isothiocyanate (FITC)‐conjugated anti‐Delta TCS1, purchased from Life Technologies (Thermo Fisher Scientific, Waltham, MA); FITC‐conjugated anti‐Vδ2, Horizon™ V500‐conjugated anti‐CD3, allophycocyanin (APC)‐conjugated anti‐CD38, phycoerythrin (PE)‐conjugated CD107A, peridinin‐chlorophyll‐protein complex cyanine 5·5 (PerCP‐Cy5·5)‐conjugated anti‐CCL‐4, PE‐Cy7)‐conjugated anti‐interferon‐γ (IFN‐γ), purchased from BD Pharmingen (San Diego, CA); and APC‐conjugated anti‐tumor necrosis factor‐α (TNF‐α), purchased from Miltenyi Biotec S.r.l (Bologna, Italy). The isotype‐matched antibodies were purchased from the corresponding company.

Phenotype and functional analysis

For phenotypic analysis, PBMCs were stained soon after thawing with two different cocktails of surface antibodies prepared in 1× phosphate‐buffered saline, 1% bovine serum albumin, 0.1% NaN₃ buffer. One containing FITC‐TCS1, the other containing FITC‐Vδ2, and both containing Horizon V500‐CD3 and APC‐CD38. After 15 min of incubation at 4°C, cells were washed twice and fixed with 2% formalin.

For functional studies, thawed PBMCs were rested for 4 hr, then cultured in 96‐well plates (200 000 cells/well) with RPMI‐1640 (EuroClone, Pero (MI), Italy) supplemented with 10% heat‐inactivated fetal bovine serum (EuroClone), 2 mm l‐glutamine, 10 mm HEPES buffer (N‐2‐hydroxyethylpiperazine‐N‐2‐ethane sulfonic acid), 2 mm penicillin and 50 mg/ml streptomycin (EuroClone). PBMCs were stimulated with anti‐human T‐cell receptor Vδ1 purified monoclonal antibody (Thermo Fisher Scientific) plus MICB Fc Chimera (R&D Systems, Inc. Minneapolis, MN) or with Picostim (Innate Pharma, Marseille, France) antigens for Vδ1 or Vδ2 T‐cell activation, respectively. In some conditions, PBMCs were not stimulated as controls. To analyze degranulation activity, CD107A antibody was added to all conditions (anti‐human T‐cell receptor Vδ1/MICB‐stimulated PBMCs, Picostim‐stimulated PBMCs and unstimulated PBMCs). After 1 hr, Brefeldin A (10 μg/ml; Sigma‐Aldrich, Gallarate, Milan, Italy) was added to block cytokine/chemokine release. Cells were maintained overnight in humidified air at 37°C with 5% CO₂ atmosphere and then stained first for surface antigens (FITC‐TCS1 in one mixture and FITC‐Vδ2 in the other) as described previously, and then for intracytoplasmatic staining a cocktail of antibodies (PerCP‐Cy5·5‐CCL‐4, PE‐Cy7‐IFN‐γ, APC‐TNF‐α), prepared with 1× phosphate‐buffered saline, 1% bovine serum albumin, 0.1% NaN₃, 0.5% saponin buffer, were added to fixed cells. After 20 min at room temperatures, cells were washed twice. Both phenotypic and function labeled cells were acquired with a FACSCanto II flow cytometer (BD Biosciences, San Jose, CA).

Data and statistical analysis

Data analysis was performed using flowjo software version 9.3.2 (Tree Star, Ashland, OR). The gating strategy for phenotype and functional analysis is depicted in Fig. 1. For functional analysis the Boolean gate platform, allowed for creating the full array of possible combinations, equating to 15 (2⁴–1) response patterns for four functions, excluding the fully negative subset. The background frequencies of the 15 combinations were individually subtracted from each of the 15 response patterns. Subsequently, using pestle version 1.7 (available at http://www.drmr.com/pestle.zip) and SPICE version 5.3,17 data were graphed. graphpad prism version 5.0a for Macintosh (GraphPad Software, San Diego, CA) was used to perform statistical analysis and produce graphs. Kruskal–Wallis with Dunn's correction test was used to compare groups and a non‐parametric Wilcoxon test was used to compare variables throughout follow up. The correlation were performed using the non‐parametric Spearman test. A P‐value <0·05 was considered statistically significant.

Figure 1.

Gating strategy. Representative dot plots of gating strategy to analyze (a) activation parameter (CD38). A forward scatter area (FSA) versus forward scatter weight (FSW) plot was used to discriminate singlet from doublet cells. Then lymphocytes were identified using side scatter area (SSA) versus FSA plot. Vδ1 or Vδ2 T cells were analyzed inside the CD3‐positive cells and then mean fluorescence intensity of CD38 expressing Vδ1 or Vδ2 T cells was analyzed. (b) Functional parameters (CD107A, IFN‐γ CCL‐4, and TNF‐α). Specifically, after the exclusion of doublets, singlet was gated and Vγ9Vδ2 T cells were identified inside the CD3⁺ lymphocytes. Then gates for each respective function were made using combinations that provided optimal separation. Single function gates were set based on the negative control (unstimulated) samples and were placed consistently across samples. We used the Boolean gate platform to create the full array of possible combinations, equating to 15 (2⁴–1) response patterns when testing four functions. Data are reported after background correction. Numbers in the plots indicate the percentage of functional Vγ9Vδ2 T cells among the total Vγ9Vδ2 T cells. Isotype control of antibodies were used for all the experiments.

Results

Viral and immunological features

We analyzed viral load in the EPI and LPI groups and, as expected, the EPI group showed a significantly higher viral load than LPI (P < 0·05). Nevertheless, viral load decreased to undetectable levels in both groups after 6 months of therapy introduction (Fig. 2a).

Figure 2.

Viral and immunological features. Viral load (a), CD4 T cells absolute number (b) and CD4/CD8 ratio (c) in healthy control (HD) and in HIV ⁺ patients [early (EPI) and late (LPI) primary infection groups] before therapy (T0) and 6 months after (T6) therapy introduction were anonymously abstracted from clinical data. graphpad prism 5 was used for graphs and statistical analysis. The results are shown as median and interquartile range (box plot), and vertical lines show the minimum and maximum values. Comparison between groups were performed using Kruskal–Wallis with Dunn's correction and a non‐parametric Wilcoxon test was used to compare variables throughout follow up. P < 0·05 was considered statistically significant.

Before treatment, both EPI and LPI groups showed a lower CD4 T‐cell count than HD (EPI: P = 0·001; LPI: P < 0·05, Fig. 2b) and the CD4 cell count was lower in the EPI group than the LPI group (P < 0·05). After 6 months of therapy, the LPI group, but not EPI (P < 0·01), was able to restore CD4 count to HD levels, probably because the EPI group started therapy with a lower CD4 cell count than LPI (Fig. 2b). Interestingly, the CD4/CD8 T‐cell ratio was significantly lower than that in HD in both groups before (EPI: P < 0·01; LPI: P < 0·001) and after (EPI: P < 0·05; LPI: P < 0·01) therapy, although a significant increase in the CD4/CD8 T‐cell ratio was observed in the LPI group at T6 (P < 0·05) (Fig. 2, panel c).

Influence of cART on γδ T‐cells frequency in EPI and LPI

We then analyzed the frequency of Vδ1 and Vδ2 T cells in the EPI and LPI groups. Before treatment, Vδ1 T‐cell frequency was comparable in both groups of HIV patients and with HD. Interestingly, in the LPI group, we observed an increased frequency of Vδ1 T cells after 6 months of therapy when compared with baseline (P < 0·05) (Fig. 3a), reaching values significantly higher than in HD (P < 0·01) (Fig. 3a).

Figure 3.

Human immunodeficiency virus (HIV) infection affects Vδ1 T‐cell distribution and activation that is associated with worst clinical outcome. Frequency of Vδ1 (a) and Vδ2 (b) T cells and mean fluorescence intensity (MFI) of CD38 on Vδ1 (c) and Vδ2 (d) T cells was assessed in healthy control (HD) and in HIV ⁺ patients [early (EPI) and late (LPI) primary infection groups] before (T0) and 6 months after (T6) of therapy introduction by flow cytometry. Correlation of Vδ1 T cells expressing CD38 with viral load at T0 (e) or with CD4 T‐cell count at T0 (f) and at T6 (g) was also determined. graphpad prism 5 was used for graphs and statistical analysis. The results are shown as median and interquartile range (box plot), and vertical lines show the minimum and maximum values. Comparisons between groups were performed using Kruskal–Wallis with Dunn's correction and a non‐parametric Wilcoxon test was used to compare variables throughout follow up. The Spearman's non‐parametric correlation was used to estimate the association of two continuous variables of interest. P < 0·05 was considered statistically significant.

In contrast, the frequency of Vδ2 T cells was comparable in both groups of HIV patients and with HD before and after treatment, although, in the LPI group, a significant decrease of the Vδ2 T‐cell frequency was observed after treatment when compared with T0 (P < 0·05) (Fig. 3b).

Activated Vδ1 T‐cell frequency correlates with HIV viral load and CD4 T‐cell count

During HIV infection there is a general activation of the immune system.18 We investigated the activation profile of γδ T cells in EPI and LPI groups by flow cytometry. CD38 expression on Vδ1 and Vδ2 T cells was comparable in both groups of patients before and after treatment, although at T0, the EPI group, showed a significantly higher expression of CD38 in both Vδ1 and Vδ2 T cells when compared with HD (P < 0·001) (Fig. 3c,d). After treatment, CD38 expression significantly decreased in both groups of patients on Vδ1 and on Vδ2 T cells (P < 0·05) (Fig. 3c,d, respectively). Before treatment the frequency of activated Vδ1 T cells significantly correlated with HIV viral load (P < 0·05, r = 0·56) (Fig. 3e) and CD4 T‐cell count (P < 0·05, r = −0·57) (Fig. 3f), suggesting a role for HIV replication in activating Vδ1 T cells. Finally, before treatment, the frequency of activated Vδ1 T cells correlated, although at the limit of significance, with CD4 T‐cell count analyzed after 6 months of therapy (P = 0·05) (Fig. 3g), suggesting a role for activated Vδ1 T cells as biomarker of worst immune reconstitution. We did not observe any correlation when Vδ2 T cells were considered (data not shown).

Altered functional γδ T‐cell profile in primary infection

To define the impact of HIV primary infection on the functional profile of γδ T cells, we analyzed IFN‐γ, TNF‐α and CCL‐4 production and CD107A expression in stimulated Vδ1 and Vδ2 T cells by flow cytometry. Before treatment, the EPI and LPI groups showed a significantly lower frequency of CD107A‐expressing Vδ1 T cells when compared with HD (P < 0·01) (Fig. 4a,b). Moreover, at T0, the LPI group showed a significantly lower frequency of CCL‐4‐producing Vδ1 T cells when compared with HD (P < 0·01) (Fig. 4b). Therapy introduction restored CD107A‐expressing Vδ1 T cells to the HD levels although only in the EPI group was there a significantly increased frequency when compared with T0 (P < 0·01) (Fig. 4a). After 6 months of therapy, the frequency of CCL‐4‐producing Vδ1 T cells significantly decreased in the EPI group compared with T0 (P < 0·05) and it was significantly lower than HD (P < 0·05) (Fig. 4a). In contrast, after 6 months of therapy, in the LPI group, CCL‐4‐producing Vδ1 T cells were comparable to HD (Fig. 4b).

Figure 4.

Magnitudes of Vδ1 and Vδ2 T‐cell responses. Peripheral blood mononuclear cells (PBMCs) were stimulated with anti‐Vδ1/MICB or Picostim to analyzed Vδ1 or Vδ2 T‐cell responses, respectively. Magnitudes of CD107A, IFN‐γ, CCL‐4 and TNF‐α responses of Vδ1(a, b) or Vδ2 (c, d) T cells were analyzed in early primary infection (EPI) (a, c) and later primary infection (LPI) (b, d) groups before (T0) and 6 months after (T6) therapy introduction and compared with healthy donors (HD). spice software was used to make graphs. The results are shown as median and interquartile range (box plot), and vertical lines show the minimum and maximum values. Comparison between groups were performed using Kruskal–Wallis with Dunn's correction and a non‐parametric Wilcoxon test was used to compare variables throughout follow up. P < 0·05 was considered statistically significant.

Referring to Vδ2 T cells, before treatment no differences were observed when comparing the EPI and LPI groups with HD (Fig. 4c,d). After 6 months of therapy, a decreased frequency of IFN‐γ‐ or TNF‐α‐producing Vδ2 T cells was observed in the EPI group when compared with baseline (P < 0·05) (Fig. 4c); a similar decrease, although at the limit of significance (P = 0·05), was found for CCL‐4 (Fig. 4c). After 6 months of therapy, TNF‐α‐producing Vδ2 T cells were significantly lower in the EPI group when compared with HD (P < 0·05) as well as CD107A and IFN‐γ‐producing Vδ2 T cells (P = 0·05) (Fig. 4c). No differences were found in the LPI group for all the Vδ2 T‐cell subsets (Fig. 4d). Interestingly, these data showed that HIV infection affected both Vδ1 and Vδ2 T‐cell functionality, and therapy introduction has only a marginal role in restoring T‐cell function.

Specific individual γδ T‐cell subsets were affected soon after HIV infection

To specifically study the distinct subsets affected by HIV infection, we analyzed the 15 individual γδ T‐cell populations, as described in the Materials and methods section. As shown in Fig. 5, in the EPI group the frequency of the double (CD107A⁺ CCL‐4⁺) Vδ1 T‐cell subset was significantly lower compared with HD (P < 0·001) (Fig. 5a) and was not restored by therapy introduction (P < 0·01) (Fig. 5a). Moreover, the frequency of the single (CCL‐4⁺) Vδ1 T‐cell subset significantly decreased after therapy when compared with baseline (P < 0·05) reaching a level lower than HD (P < 0·05) (Fig. 5a). Finally, the frequency of the single (CD107A⁺) Vδ1 T‐cell subset was significantly lower when compared with HD (P < 0·05) but significantly increased after therapy introduction when compared with T0 (P < 0·05) reaching the HD level (Fig. 5a). Similarly, in the LPI group, the frequency of double (CD107A⁺ CCL‐4⁺) Vδ1 T‐cell subset was significantly lower when compared with HD (P < 0·001) (Fig. 5b) and was not restored by therapy introduction (P < 0·001) (Fig. 5b). Moreover, we observed a significantly increased frequency of the single (CD107A⁺) Vδ1 T‐cell subset after therapy introduction when compared with T0 (P < 0·05) (Fig. 5b).

Figure 5.

Analysis of the polyfunctional responses of each individual Vδ1 and Vδ2 T‐cell subset. Peripheral blood mononuclear cells (PBMCs) were stimulated with anti‐Vδ1/MICB or Picostim to analyze the frequency of each individual functional Vδ1 (a, b) or Vδ2 (c, d) T‐cell subset, respectively. (a) + or – sign below the graphs indicates the presence or absence of the specific function. The frequencies were analyzed in patients with early primary infection (EPI) (a and c) and in those with late primary infection (LPI) (b and d) groups before (T0) and 6 months after (T6) of therapy introduction and compared with healthy donors (HD). flowjo software, through the Boolean gate platform, allowed for creation of the full array of possible combinations, equating to 15 (2⁴–1) response patterns for four functions, excluding the fully negative subset. spice software was used to make graphs. The results are shown as median and interquartile range (box plot), and vertical lines show the minimum and maximum values. Comparisons between groups were performed using Kruskal–Wallis with Dunn's correction and a non‐parametric Wilcoxon test was used to compare variables throughout follow up. P < 0·05 was considered statistically significant.

Referring to Vδ2 T cells, after therapy introduction, both the EPI and LPI groups showed a lower frequency of the quadruple (CD107A⁺ IFN‐γ ⁺ CCL‐4⁺ TNF‐α ⁺) Vδ2 T‐cell subset when compared with HD, although it reached the significant difference only in the LPI group (P < 0·05) (Fig. 5d). After therapy, the EPI group showed a decreased frequency of the triple (IFN‐γ ⁺ CCL‐4⁺ TNF‐α ⁺) Vδ2 T‐cell subset when compared with baseline (P < 0·05), although it remained comparable with HD (Fig. 5c). Moreover, after therapy, the EPI group showed a higher frequency of the single (IF‐Nγ ⁺) Vδ2 T‐cell subset when compared with HD (P < 0·001) Fig. 5c). Finally, in the LPI group, before therapy, the frequency of the single (CD107A⁺) Vδ2 T cells was significantly higher when compared with HD (P < 0·05) and was not restored by therapy introduction (P < 0·01) (Fig. 5d). Overall, these data suggest that HIV affects the frequencies of individual γδ T‐cell functional subsets soon after infection and the anti‐viral therapy only partially restores them.

Discussion

Vδ1 and Vδ2 T cells are able to exert antiviral capability against HIV‐infected cells but their phenotype, function and distribution are altered during HIV infection,6 suggesting a central role during pathogenesis.19 The study of γδ T‐cell dynamics during early PHI represents a challenging issue to be addressed, that could clarify the precocity of innate dysfunction and, on the other hand, verify the effectiveness of early cART in avoiding innate immune dysfunction.

The aim of this study was to analyze the effect of early treatment on γδ T‐cell phenotype and function during primary infection. The main findings are the following: (i) during PHI no significant changes of Vδ1 and Vδ2 T‐cell frequency has been observed; (ii) the baseline Vδ1 T‐cell activation was correlated with disease progression (viral load and CD4 T‐cell count before treatment) and with the immune reconstitution (CD4 T cells after treatment); (iii) PHI deeply impacts the polyfunctional profile of γδ T cells and early cART failed to restore it to healthy control levels.

It is well known that chronic HIV infection is characterized by low frequency of Vδ2 T cells and high frequency of Vδ1 T cells if compared with healthy donors (with a resulting inversion of Vδ1/Vδ2 T‐cell ratio), persisting also after therapy.20 In contrast, PHI (both EPI and LPI patients) did not significantly modify Vδ1 and Vδ2 T‐cell frequency, suggesting that the dysregulation of γδ T‐cell homeostasis represents a later event in the pathogenesis of HIV infection.

Moreover, the early cART introduction was able to prevent Vδ2 T cells loss occurring later in infection. Interestingly, a Vδ1 T‐cell increase was observed after T6 of cART in LPI patients (patients starting treatment later than EPI), becoming higher than in HD. We can speculate that, despite effective early treatment, in the LPI patients a dysregulation of Vδ1 homeostasis occurred.

Primary HIV infection induced a significant Vδ1 and Vδ2 T‐cell activation, confirming previous data.13 Of note, we found that Vδ1 T‐cell activation was negatively correlated with CD4 T‐cell count and positively correlated with viral load, suggesting a strict association between viral replication and Vδ1 T‐cell activation. Therapy introduction decreased γδ T‐cell activation to normal values in both groups of patients, as still reported in patients with chronic HIV infection.13 We also found a negative correlation between Vδ1 T‐cell activation at baseline and CD4 T‐cell count after 6 months of therapy, indicating that the activation of Vδ1 T cells in primary infection could predict immune reconstitution. Overall, these data indicated that the extent of viral replication regulates Vδ1 T‐cell activation, which, in turn, may reduce the immune reconstitution.

Polyfunctionality of αβ T cells, namely the capacity to perform simultaneously more than one function at single‐cell level, is associated with better HIV control with slower disease progression and better outcome.21, 22 Polyfunctionality is indeed higher in HIV‐positive patients who naturally control HIV infection for a long time without therapy (Long Term Non Progressor and Elite controls).23 Little is known about Vδ1 and Vδ2 T‐cell polyfunctionality during HIV infection. In a cross‐sectional study, we previously demonstrated a reduction of polyfunctional Vδ2 T cells in cART‐treated patients chronically infected with HIV, which is particularly evident in patients with low CD4 T‐cell count.24 How early this dysfunction is, and the possible effectiveness of early cART in recovery of a healthy polyfunctional profile has not been clarified. Here, we first demonstrated that γδ T‐cell polyfunctionality was already deeply affected in primary HIV infection (in both the LPI and EPI groups). Moreover, treatment introduction was able to restore only a few functional subsets whereas the majority of them persisted in their altered state. An impairment of degranulation capability (CD107a expression) of Vδ1 T cells was observed in both the EPI and LPI groups and therapy introduction restored it to healthy level. Hence, although Vδ1 T cells during primary HIV infection showed a terminally differentiated phenotype,12, 13 their effective cytotoxic potential was impaired. We also observed that treatment introduction did not restore CCL‐4‐producing Vδ1 T‐cell subset in early primary infection as well as double (CD107A⁺ CCL‐4⁺) and single (CCL‐4⁺) Vδ1 T‐cell subsets, indicating that CCL‐4 production is particularly affected by HIV, soon after infection. CCL‐4 interfere with HIV infection by blocking the binding of HIV to CCR5 co‐receptor but data about CCL‐4 function are controversial. In our previous work, CCL‐4‐producing CD8 T cells are associate with worse immune reconstitution during chronic infection.25

Interestingly, our data showed a deep impact of primary HIV infection on Vδ2 T cells particularly in the EPI group. As shown in Table 1, the EPI group showed a lower CD4 T‐cell count compared with the LPI group.

A reduction of IFN‐γ‐producing Vδ2 T cells has been shown in primary infection,13 and this persists in chronic HIV infection.24 The impact of different functional Vδ2 T‐cell subsets on the pathogenesis and response to treatment during primary HIV infection remains to be elucidated; however, no correlation between cytokine functional subsets and immune reconstitution was observed.

Taken together, our data suggest that therapy introduction during primary infection can reduce γδ T‐cell activation but is not able to completely restore functional properties, at least in the short term. Further studies are needed to address whether longer‐term therapy could have a deeper impact on γδ T‐cell dynamics.

Disclosures

The authors declare that there are no competing interests.

Author contributions

RC and CA designed the study; RC and FB performed the research; CP, AM and AA managed the HIV patients; RC, CA AS, VB and GG discussed and analyzed the data; RC wrote the paper; and CA and AA revised the manuscript. All the authors approved the revised manuscript.