Regulatory and Helper Follicular T Cells and Antibody Avidity to Simian Immunodeficiency Virus Glycoprotein 120

The generation of long-lived plasma cells and high-affinity Abs is largely dependent on T cell/B cell interaction in the B follicles of secondary lymphoid organs (1–3). Ag-activated B cells making contact with a specialized subset of CD4⁺ T cells, called T follicular helper cells (T_FH), can enter the germinal centers (GC) to undergo somatic hypermutation and affinity maturation (4). T_FH home to B follicles and GC (5–9) by upregulating CXCR5 and downregulating CCR7 (5, 6, 10). T_FH express high levels of programmed death 1 (PD-1), ICOS, and Bcl-6, a master transcriptional regulator that orchestrates T_FH differentiation (11–13). In the GC, T_FH provide signals for B cell survival and differentiation (5, 10) via IL-21 production and CD40L expression, and they promote the generation of Abs with high affinity (11–16).

GC reactions are tightly regulated to prevent the emergence of B cell clones that are specific or cross-reactive against self-antigens, while selecting for high-affinity Abs to microbes (17, 18). The maintenance of the appropriate number of T_FH is crucial (19); the absence of T_FH has a negative impact in the generation of the GC (20, 21), whereas their excessive accumulation leads to increased GC reactions and the onset of some autoimmune diseases (4, 22–24).

CD4⁺ T follicular regulatory cells (T_FR) contain T_FH numbers and, in doing so, they control the magnitude of GC responses (25, 26). Similarly to T_FH, T_FR migrate to the GC by expressing CXCR5 and downregulating CCR7 during T cell activation (6, 25, 27–29). T_FR differentiate from natural CXCR5⁻Foxp3⁺CD25⁺ T regulatory cells (T_REG) and express high levels of the typical T_REG markers (i.e., Foxp3, CD25, CTLA-4) and T_FH canonical markers such as ICOS, PD-1, and Bcl-6 (25, 26). Whereas Bcl-6 is essential for CXCR5 expression on B cells and T_FH and for their localization to the GC (25, 26), T_FR coexpress Blimp-1, which is known to repress CXCR5 expression (25, 30). Ablation of NFAT-2 in mice results in reduced expression of CXCR5 on T_FR, but not on T_FH, suggesting that this transcriptional factor may enable the proper localization of T_FR within B cell follicles, possibly by inhibiting Blimp-1–mediated repression of CXCR5 expression (31). T_FR restrict T_FH numbers and help to maintain a steady ratio of IgM⁺ to IgM⁻ (switched) B cells (32) via IL-10 production (29); in vivo depletion of CD4⁺ T cells with suppressive activity including T_FR, or in vivo blockade of IL-10 or TGF-β receptors results in T_FH expansion, loss of normal proportion of IgM⁻ B cells, and in increased levels of high-affinity Abs (26, 29, 33). A hallmark of HIV and SIV infection is the immune dysfunction of humoral responses characterized by loss of memory B cells and hypergammaglobulinemia (34, 35). T_FH frequency is significantly increased in the lymph nodes of HIV-infected individuals and chronically SIV_mac251-infected macaques (8, 36). Production of the IL-21 cytokine by T_FH is significantly reduced during HIV/SIV infection, possibly affecting GC homeostasis and the development of effective humoral responses to the virus (37). The HIV/SIV-associated changes in T_FH number and function may contribute to the impairment of B cell responses (9, 36, 38); however, other studies have found associations between the levels of functional T_FH and broadly neutralizing Abs in chronic HIV patients (39). Although the relative role of T_FH in HIV pathogenesis needs further investigation, it would be important to understand the molecular and cellular mechanisms that regulate T_FH expansion.

T_FR dynamics in HIV infection has not been investigated yet. We identified T_FR as CXCR5⁺ T_REG in the lymph nodes of rhesus macaques, a well-established model of HIV infection. We show that 1) T_FR are infected by SIV_mac251, 2) there is an apparent decrease in T_FR levels, particularly during chronic infection, 3) T_FR levels are associated with the levels of T_FH and the total frequency of IgG⁺ B cells, and 4) T_FR levels are inversely correlated with the avidity of Abs to SIV gp120 protein. Taken together, these findings suggest a potential role for T_FR in modulating humoral responses against HIV/SIV.

All of the animals used in this study were colony-bred rhesus macaques (Macaca mulatta) obtained from Covance Research Products (Alice, TX). The animals were housed and maintained in accordance with the standards of the Association for the Assessment and Accreditation of Laboratory Animal Care. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All surgery was performed under general anesthesia, and all efforts were made to minimize suffering. All macaques were negative for simian retrovirus, simian T cell leukemia virus type 1, and herpesvirus B. Macaques were infected with a single high dose (6300 50% tissue culture-infective dose) (40) or with 10 repeated low doses (120 50% tissue culture-infective dose) of SIV_mac251 given rectally (41) (see Table I).

Cells were isolated from blood, lymph nodes, and spleen by density gradient centrifugation. Tissues from the rectum, jejunum, and colon were treated with 1 mM ultrapure DTT (Invitrogen Life Technologies) for 30 min followed by incubation in calcium/magnesium-free HBSS (Invitrogen Life Technologies) for 60 min with stirring at room temperature to remove the epithelial layer. Lamina propria lymphocytes were separated by cutting the tissue into small pieces and incubating in 10% FBS IMDM (Invitrogen Life Technologies) with collagenase D (400 U/ml; Boehringer Mannheim) and DNase (1 μg/ml; Invitrogen Life Technologies) for 2.5 h at 37°C. Mononuclear cells were placed over 42% Percoll (GE Healthcare) and centrifuged at 800 × g for 25 min at 4°C. Lamina propria lymphocytes were collected from the cell pellet (42).

We used the following Abs: CD3–Alexa Fluor 700 (SP34-2), CD4-PerCP-Cy5.5 (L200), CD95-PE-Cy5 (DX2), CD197-PE-Cy7 (CCR7, clone 3D12), CD25-allophycocyanin-Cy7 (M-A251), CD195-PE (CCR5, clone 3A9), CD14–Alexa Fluor 700 (M5E2), CD16–Alexa Fluor 700 (3G8), CD56–Alexa Fluor 700 (B159), IgM-PerCP-Cy5.5 (G20-127), IgG-Qdot605 (G18-145), Ki67-PE (B56), and CD21-PE-Cy7 (B-ly4), all from BD Biosciences; Bcl-6-PE (IG191E/A8), CD278–Pacific Blue (ICOS, clone C398.4A), CD25–Pacific Blue (BC96), PD-1–allophycocyanin (EH12.2H7), CD39-BV421 (MOCP-21), and CD39 (A1) from BioLegend; Foxp3-FITC (PCH101), CXCR5-PE–eFluor 610 (MU5UBEE), and CD20-Qdot650 (2H7), all from eBioscience; and CD103-FITC (α_Ε integrin, clone 2G5), CD19-PE-Cy5 (J3-119), and CD127-PE (eBioRDR5) from Beckman Coulter. The α₄β₇ Ab (Act-1) was obtained through the National Institutes of Health Nonhuman Primate Reagent Resource Program (AIDS Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health). Vivid amine-reactive dye was used to discriminate live/dead cells (Invitrogen). IgA-Texas Red (polyclonal) was obtained from SouthernBiotech, and CD38-FITC (clone AT-1) was from StemCell Technologies.

For phenotypic characterization of CD4⁺ T cells subsets, cells were stained with surface markers CD3, CD4, CD95, CD25, CCR7, CXCR5, ICOS and Vivid. Cells were then fixed and permeabilized according to eBioscience’s instructions and stained with anti-Foxp3 and anti–Bcl-6 for 30 min. The appropriate isotype-matched control Ab was used to define positivity. T_FR cells were gated as live CD3⁺CD4⁺CD95⁺ T cells, and their percentage was calculated as the frequency of CXCR5⁺ and CCR7⁻ within Foxp3⁺CD25⁺ cells (percentage of T_REG) or within CD95⁺CD4⁺ T cells (percentage of memory CD4⁺ T cells) (26). Similarly, double-positive (DP) cells were CXCR5⁺ and CCR7⁺ cells and CCR7⁺ T_REG were CXCR5⁻ and CCR7⁺. Finally, T_FH were gated as CXCR5⁺PD-1^hi cells within the Foxp3⁻ region or the memory CD4⁺ T cells population (see Fig. 1A).

B cells were gated as live/lineage (lin)⁻ (CD3⁻CD14⁻CD16⁻CD56⁻) positive for CD20 and or CD19 markers. For plasmablasts, cells were stained with lineage markers CD20, CD38, CD39, IgM, IgG, and IgA. Cells were treated with Cytofix/Cytoperm (BD Biosciences) and stained with Ki67. Plasmablasts were gated as lin⁻ (lin⁻CD20⁺ and/or CD19⁺CD21⁻Ki67⁺CD38⁺CD39⁺) (43).

Marker expression was analyzed with an LSR II flow cytometer using FACSDiva software (BD Biosciences). FACS analysis was performed using FlowJo software (Tree Star, Ashland, OR). A minimum of 10,000 cells per tube were analyzed.

The absolute number of CD4⁺ T cells was calculated as previously described (44).

To determine the RNA levels for IL-10, TGF-β, and SIV DNA, cells from lymph nodes were stained with Vivid, CD3, CD4, CD25, CCR7, and CXCR5. T_FR were defined as live CD3⁺CD4⁺CD25⁺CXCR5⁺CCR7⁻; T_REG were defined as CXCR5⁻CCR7⁺.

For proliferation, CD25⁺CD4⁺ T cells (live CD3⁺CD4⁺CD25⁺) were sorted. Sorting was performed on a FACSAria (BD Biosciences).

Sorted live CD3⁺CD4⁺CD25⁺ cells were migrated for 1 h to 1 μg/ml CXCL13 (R&D Systems, catalog no. 801-CX/CF) using 5-μm-pore polycarbonate membrane inserts (Millipore).

Cell proliferation was determined by dilution of CFSE (Life Technologies). Briefly, CD25-depleted (CD3⁺CD4⁺) cells were stained with CFSE for 10 min and were then placed in a 24-well plate in the presence or absence of CD3 (10 μg/ml; clone FN18) with soluble anti-CD28 (1 μg/ml; clone CD28.2), in the presence or absence of autologous CD3⁺CD4⁺CD25⁺ cells migrated to CXCL13 (10:1 ratio), for 4 d. Cells were then stained and analyzed by FACS as described above.

Cells from lymph nodes were incubated 30 h with 50 μg cyclosporin A (Sigma-Aldrich) and incubated for 6 h with or without PMA and in the presence of brefeldin A (GolgiPlug; BD Biosciences).

Total RNA was extracted from whole tissue with RNeasy Plus (Qiagen) and reverse transcribed with a high-capacity cDNA reverse transcription kit (Applied Byosystems). After reverse transcription, the relative amounts of transcripts were determined by real-time PCR with the SYBR Green quantitative PCR master mix (Promega) using 0.2 μM PCR primers for IL-10 (forward, 5′-AGAACCACGACCCAGACATC-3′, reverse, 5′-GGCCTTGCTCTTGTTTTCAC-3′) and TGF-β. The TGF-β was described elsewhere (45). Quantification of cDNA was normalized in each reaction according to the internal β-actin control (forward, 5′-GGCACCCAGCACAATGAAG-3′, reverse, 5′-GCTGATCCACATCTGCTGG-3′). A real-time nucleic acid sequence–based amplification assay was used to quantitative SIV RNA in plasma (46). SIV DNA was quantified as previously described (40).

The primary Abs included anti-Foxp3 (Abcam, rabbit), anti-CD20 (Dako, Carpinteria, CA; mouse Ig2a), and anti-CD3 (UCD, rat). TBS with 0.05% Tween 20 was used for all washes. Ab diluent (Dako) was used for all Ab dilutions. For all primary Abs, slides were subjected to an Ag retrieval step consisting of incubation in AR10 (BioGenex, San Ramon, CA) for 2 min at 125°C in a digital decloaking chamber (Biocare Medical, Concord, CA) followed by cooling to 90°C before rinsing in water. Primary Abs were replaced by normal rabbit IgG, mouse IgG (Invitrogen, Grand Island, NY), and rat IgG (Vector Laboratories, Burlingame, CA) and were included with each staining series as the negative control. Nonspecific binding sites were blocked with 10% goat serum and 5% BSA (Jackson ImmunoResearch Laboratories, West Grove, PA). Binding of the primary Abs was detected simultaneously using Alexa Fluor 568 goat anti-rat, Alex Fluor 647 goat anti-mouse IgG2a, and Alex Fluor 488 goat anti-rabbit. All slides were coverslipped using ProLong Gold with DAPI (Molecular Probes, Grand Island, NY) to stain nuclei. All the control experiments gave appropriate results with minimal nonspecific staining.

Slides were visualized with epifluorescent illumination using a Zeiss Axioplan 2 microscope (Carl Zeiss, Thornwood, NY) and appropriate filters. Digital images were captured and analyzed by using Openlab software (Inprovision, Waltham, MA). Alexa Fluor 647 was captured in a black-and-white channel whereas other fluorescence dyes were pictured in color channels. Five high-power (×40) microscope fields were randomly chosen and captured digitally with the system described above. Each captured field includes an area of ∼0.04 mm². Only clearly positive cells with distinctly labeled nuclei (DAPI) and bright staining were considered positive. Individual positive cells in the five captured high-power microscope fields of the immunohistochemical stained tissue sections were counted manually by a single observer. The numbers of positive cells are presented as cells per square millimeter.

Avidity was analyzed as previously described (41). Briefly, recombinant SIV gp120 protein made from codon-optimized SIV_mac239 gp120 fused to the C-terminal tag of HIV-1 gp120 was used as an Ag for the capture ELISA to detect SIV Abs against conformational epitope. Ab avidity was determined by parallel ELISA. Heat-inactivated plasma samples were serially diluted and applied to a 96-well plate capturing SIV_mac239 gp120. After 1 h of incubation, the plate was washed and half the samples were treated with TBS, whereas the paired samples were treated with 1.5 M sodium thiocyanate (Sigma-Aldrich) for 10 min at room temperature. The plate was washed and a goat anti-monkey IgG-detecting Ab (Fitzgerald) was used. The avidity index (%) was calculated by taking the ratio of the sodium thiocyanate–treated plasma dilution giving an OD of 0.5 to the TBS-treated plasma dilution giving an OD of 0.5 and multiplying by 100. Plasma of uninfected normal macaques served as negative controls.

Tests of two groups of animals for differences between cell types, tissues, or stages of infection were performed using the exact Wilcoxon rank sum test. Differences before and after infection within the same group of animals were assessed using the Wilcoxon signed rank test. Differences across three stages of infection were modeled using repeated measures ANOVA when distributional assumptions were met. Correlation analyses were performed using the exact Spearman rank correlation method. Trends across three groups were assessed by the Jonckheere–Terpstra test. Owing to the exploratory nature of this study, the p values reported were not corrected for multiple comparisons. The p values ≤ 0.05 were considered statistically significant, and we note that for outcomes where all pairwise comparisons of three groups are possible, the p values ≤ 0.02 remain significant after correction for the multiple tests.

T_FR localize in the GC of mice and humans (25, 26, 29, 47). We confirmed the presence of Foxp3⁺CD3⁺ T cells in healthy macaque’s B cell follicles of lymph nodes by immunohistochemistry (Fig. 1A). We then characterized T_FR in cell suspension by flow cytometry and compared their frequency, phenotype, and localization to those of CCR7⁺ T_REG and T_FH. The gating strategy used to define these three cell subsets is shown in Fig. 1B. T_FR and T_REG were gated within the live Foxp3⁺CD25⁺CD95⁺CD4⁺ T cells and defined as CXCR5⁺ and CCR7⁻, consistent with GC location, and as CXCR5⁻ and CCR7⁺, consistent with T zone location (Fig. 1B). Of note, the T_REG population identified by this strategy only includes a specific subset based on the expression or the lack thereof of the two considered chemokines. We refer to this subset as CCR7⁺ T_REG or T_REG for simplicity.

In agreement with Sage et al. (48) and Linterman et al. (25), we used the Foxp3 marker to distinguish between T_FH and T_FR subsets because both populations express CXCR5, PD-1, ICOS, and Bcl-6. T_FH were defined as Foxp3⁻CXCR5⁺PD-1^hiCD4⁺ T cells (Fig. 1B).

Consistent with their mouse counterpart, macaque T_FR expressed comparable levels of Foxp3 (Fig. 1C), equal intensity and frequency of CD25, and frequency of CD39 to T_REG (Fig. 1D, Supplemental Fig. 1A, 1B) (25, 26). T_FR were also negative for CD127, the marker for the IL-7 receptor, and expressed common T_FH markers, such as PD-1, Bcl-6, and ICOS (Fig. 1E, Supplemental Fig. 1C) (28). Only a subset of T_REG but not of T_FR or T_FH was positive for the α_E (CD103) and α₄β₇ integrin (25, 26) (Fig. 1E and data not shown).

Within the Foxp3⁺CD25⁺ population we identified an additional CXCR5⁺CCR7⁺ DP cell subset (Fig. 1B). DP cells expressed intermediate levels of PD-1 as compared with T_FR and T_REG and had equal levels of Bcl-6 to T_FR (Supplemental Fig. 1D).

We looked at the distribution of T_FR, T_REG, and T_FH in blood, peripheral lymph nodes, and in the GALT (colon, jejunum, and rectal mucosa) obtained from 6, 21, and 8 naive macaques, respectively (Fig. 1F–H). Representative flow plots obtained from blood, lymph node, and rectal mucosa tissue from one healthy animal are shown in Supplemental Fig. 1E. T_FR were mainly in the GALT and lymph nodes and only a few were detected in blood (Fig. 1F, Supplemental Fig. 1E). Within the Foxp3⁺CD25⁺ population, cells that expressed only CXCR5 were less frequent in lymph nodes than in the GALT (p < 0.0001 by the Wilcoxon rank sum test; Fig. 1F, upper panel), and the opposite was observed for cells that expressed only CCR7 (p < 0.0001 by the Wilcoxon rank sum test; Fig. 1G, upper panel). The apparent difference in tissue distribution in lymph nodes and GALT was lost when we looked at the frequency of T_FR and T_REG within the memory CD4⁺ T cell population (lower panels in Fig. 1F, 1G). DP cells were equally distributed among all the tissues analyzed, including the blood (Supplemental Fig. 1F). Finally, T_FH frequency was significantly higher in lymph nodes and in the GALT than in the blood, as previously described (PBMC versus lymph node, p < 0.0001; PBMC versus GALT, p = 0.0031) (49) (Fig. 1H).

Because we could not determine whether DP cells home exclusively to the B zone, we excluded them from the rest of the analysis.

In mice, T_FR control T_FH numbers and decrease their proliferation in vivo and in vitro (26). We studied whether macaque lymph nodes also contained a suppressive CD4⁺ T cell population that homes to the B follicles. We sorted CD4⁺CD25⁺ live cells from the lymph nodes of two naive animals and isolated those capable of migrating in response to CXCL13, the ligand for CXCR5 (Fig. 2). Although we could not use Foxp3, an intracellular marker, to discriminate suppressor CD4⁺ T cells, sorted CD25⁺CD4⁺ T cells from lymph nodes consisted primarily of Foxp3⁺CD4⁺ T cells (Fig. 2A). Regulatory CD4⁺ T cells that migrated to CXCL13 had higher levels of CXCR5, lower levels of CCR7, and expressed higher levels of Bcl-6 than did those that did not migrate (Fig. 2B). This strategy allowed us to obtain a population of CD4⁺ T cells, highly enriched for T_FR, which could be used in downstream functional assays. Unsorted and sorted cells were stimulated with or without CD3 and CD28, in the presence or absence of migrated CD25⁺ CD4⁺ T cells (follicular T_REG–enriched population). We assessed proliferation (percentage of CFSE^dim cells) within CXCR5⁺PD-1^hiCD4⁺ T cells (T_FH-gated cells) by FACS analysis. Representative plots of the gated CXCR5⁺CD4⁺ T cells (gate 1) are shown in Supplemental Fig. 2A. Interestingly, T_FH proliferated less in the presence of T_FR-enriched cells following stimulation than in the absence of CD25⁺ cells, as shown in the representative plots in Fig. 2C and graphically in Fig. 2D. As expected, T_FR also reduced non-T_FH CD4⁺ T cell subset proliferation (gate 2) (Supplemental Fig. 2B).

Consistent with their regulatory function and similar to T_REG, T_FR produce IL-10 and TGF-β, which together suppress the proliferative potential and function of CD4⁺ T cells in mice (50, 51). Thus, we measured the levels of IL-10 and TGF-β in enriched populations of T_FR (CD3⁺CD4⁺CD25⁺CXCR5⁺CCR7⁻) and CCR7⁺ T_REG (CD3⁺CD4⁺CD25⁺CXCR5⁻CCR7⁺) obtained from peripheral lymph nodes of three naive macaques. T_FR had equivalent IL-10 and TGF-β mRNA levels, by RT-PCR, than CCR7⁺ T_REG (Fig. 2E, 2F).

Sorted T_FR and T_REG from lymph nodes of four chronically infected macaques expressed comparable levels of CCR5 (Fig. 3A) and harbored equivalent levels of SIV DNA (Fig. 3B). Similar SIV DNA levels were also found in T_FH cells.

We analyzed the effect of SIV_mac251 infection on the frequency of T_FR and CCR7⁺ T_REG in the lymph nodes of 10 acutely (2 or 3 wk postinfection) and 23 chronically infected (12–15 wk postinfection) macaques (Fig. 3C, 3D, Table I). Representative dot plots for two macaques before and after infection are shown in Supplemental Fig. 3A. Cells expressing only CXCR5⁺ were significantly reduced within the Foxp3⁺CD25⁺ population during chronic infection (chronic versus negative, p < 0.0001; chronic versus acute, p < 0.0001) (Fig. 3C), whereas CCR7 single-positive cells simultaneously increased (chronic versus negative, p < 0.0001; chronic versus acute, p = 0.0003) when compared with noninfected and chronically infected animals (Fig. 3D).

We looked at the levels of T_FR and CCR7⁺ T_REG with respect to the memory CD95⁺CD4⁺ T population. T_FR showed a downward trend from negative to acute to chronic (p = 0.0005 by the Jonckheere–Terpstra test for trend), with marginal differences in acute (negative versus acute, p = 0.049) and a significant decrease in chronic infection (negative versus chronic, p = 0.0003) (Fig. 3E). CCR7⁺ T_REG levels showed an increasing trend after infection and significance was reached between the values detected in acute and in chronic phase (p = 0.0018) (Fig. 3F).

We performed immunohistochemistry in lymph nodes at 3 and 12 wk postinfection from seven and eight animals, respectively (Fig. 3G, 3H). The numbers of CD3⁺ cells expressing Foxp3 in the B cell follicles was significantly reduced in acute infection (p = 0.0022, n = 7) and contracted even further in chronic infection (p < 0.0001, n = 8), when compared with naive (n = 8) animals (Fig. 3G, Supplemental Fig. 3B). The number of Foxp3⁺CD3⁺ cells in the T zone did not change, as described by others (Fig. 3H) (52).

We did not see any significant correlation between the frequency of T_FR or T_REG and the SIV RNA plasma levels (data not shown).

To further explore possible mechanisms for the SIV-associated decrease of T_FR, we looked at markers of immune activation. We could not find any association with the frequency of Ki67⁺CD4⁺ T cells in lymph nodes of 10 chronic animals (data not shown).

In mice NFAT-2 is critical for the upregulation of CXCR5 on T_FR (31), hence for their migration to the GC. Thus, NFAT-2 may be involved in the reduction of T_FR during chronic infection. To determine whether CXCR5 expression on macaque T_FR was also dependent on NFAT activity, we treated lymph nodes cells from two naive animals with cyclosporin A and measured changes in the CXCR5 and CCR7 levels within Foxp3⁺CD25⁺CD4⁺ T cells (Supplemental Fig. 3C, 3D). Although cyclosporin A treatment had no effect on CCR7, it decreased CXCR5 expression levels on Foxp3⁺CD25⁺CD4⁺ T cells.

The decrease of CXCR5⁺ regulatory T cells may be associated with T_FH expansion during SIV_mac251 infection. We measured the frequency of T_FH in lymph nodes of infected macaques as shown in Fig. 1B. The percentage of T_FH cells within the memory CD4⁺ T cell population did not change during acute infection, but this population significantly expanded during chronic infection (weeks 12–15 after infection), as described by others (8) (negative versus chronic and acute versus chronic, p < 0.0001) (Fig. 4A, 4B). Of note, we found a significant inverse correlation between the levels of T_FR and T_FH on memory CD4⁺ T cells during acute and chronic infection (R = −0.82, p = 0.0058 and R = −0.69, p = 0.0010 by the Spearman rank test) (Fig. 4B, 4C).

In mice, T_FR cells play a role in reducing plasma cell differentiation (26). We measured the frequency of IgM⁺, IgG⁺, or IgA⁺ CD20⁺ B cells and plasmablasts, defined as lin⁻ (CD3⁻CD14⁻CD16⁻CD56⁻) CD20⁺ and/or CD19⁺ and CD21⁻Ki67⁺CD38⁺CD39⁺ in lymph nodes of SIV_mac251 chronically infected animals (n = 14) by flow cytometry. Whereas we did not find any associations with the frequency of total memory B cells or plasmablasts measured in lymph nodes, we found a weak negative correlation between the frequency of IgG⁺ B cells and the frequency of IgM-switched plasmablasts (IgG⁺ and IgA⁺) with the percentage of T_FR within Foxp3⁺CD25⁺ cells (Fig. 5A, 5B).

T_FH are associated with the avidity to influenza virus and SIV (8, 53). In SIV_mac251-infected macaques, avidity to the gp120 is low during acute infection and increases during chronic infection (Fig. 5C). We confirmed that T_FH were positively associated with gp120 avidity when all the infected macaques were considered (R = 0.88, p = 0.0031; Fig. 5D), but not when the acute and chronic phases values were looked at separately. Importantly, T_FR levels on memory CD4⁺ T cells were associated with a reduction of binding high avidity Abs to SIV gp120 in all the infected animals (R = −0.85, p = 0.0061) and in chronic phase (R = −1.0, p = 0.017), but not during the acute phase (R = −0.80, p = 0.33) (Fig. 5E, 5F and data not shown).

In this study we identified T_FR in lymphoid tissues of healthy nonvaccinated rhesus macaques. We confirmed the presence of Foxp3⁺ T cells in the B zone of intact lymph nodes of macaques. Because T_FR share markers of T_FH and T_REG, they have been isolated and functionally characterized in mice as T_FH positive for Foxp3 (25) or, alternatively, as “follicular” T_REG expressing CXCR5 (26, 47). We opted for the latter identification strategy to characterize macaque T_FR in lymph nodes. Additionally, we identified a subset of CCR7-expressing T_REG that are CXCR5⁻ and, therefore, in principle are unable to enter the GC, and another subset of CXCR5⁺ and CCR7⁺ T_REG (DP). Because it is possible that DP cells may localize at the T/B borders/mantle zone, following gradients of CXCL13 (B cell follicles, GC) and of CCL21 and CCL19 (T zone), we excluded this population from our analysis (7).

In accordance with their mouse counterpart, macaque T_FR expressed high levels of PD-1, ICOS, and Bcl-6, they were CD39⁺ and CD127⁻, and they mainly resided in the GALT and lymph nodes. In a few macaques, T_FR had similar levels of IL-10 and TGF-β mRΝΑ as did CCR7⁺ T_REG (25, 26). It is possible that IL-10 and TGF-β may play a role in the T_FR-mediated control of T_FH proliferation, as shown in mice. We did not directly assess the suppressive ability of sorted CD25⁺CXCR5⁺CCR7⁻ cells; however, lymph nodes of healthy noninfected macaques contained a CXCR5^hiCCR7^loBcl-6^hiCD25⁺CD4⁺ population displaying in vitro chemotaxis toward CXCL13, as well as suppressive activity on CD4⁺ T cells and T_FH proliferation. Further characterization is needed to confirm IL-10 and TGF-β production by macaque T_FR and their role in the apparent suppression of T_FH cells.

We took advantage of the established similarities between SIV infection of macaques and HIV-1 infection of humans to study T_FR susceptibility and dynamics during infection in comparison with CCR7⁺ T_REG. Previous studies on CD4⁺ T susceptibility have shown that CD25⁺Foxp3⁺ cells are less susceptible than other subsets to HIV/SIV infection owing to their anergic nature and to the Foxp3-mediated inhibition of HIV-1 long terminal repeat activation (54–57). Therefore, the relative frequency of CD25⁺Foxp3⁺CD4⁺ T cells increases in acute and chronic HIV/SIV infection, whereas their absolute number remains the same (56, 58, 59). Similarly, we observed a trend for increased frequency of CCR7⁺ T_REG within the memory CD4⁺ population and no differences in the number of CD3⁺Foxp3⁺ cells in the T zone of intact lymph nodes in chronic infection.

T_FR from naive macaques expressed comparable levels of surface CCR5 to T_REG and had equivalent levels of SIV DNA following infection. Conversely, we saw a reduction in T_FR frequency and a decrease of CD3⁺Foxp3⁺ cells in the B follicles of infected animals, in particular during the chronic phase of infection. Whereas we could not discern between CD8⁺ T cells and CD4⁺ T cells in our immunohistochemical analysis, some chronically infected animals had negligible numbers of Foxp3⁺ T cells in the GC, indicating an overall reduction of regulatory cells, likely including T_FR.

We could not determine whether T_FR are more susceptible to SIV_mac251 infection than T_REG, and we did not observe any association between the T_FR levels and viral replication levels or with immune activation.

The reduction in Foxp3⁺ cells in the GC may be driven by SIV-associated changes in the homing patterns of these cells. NFAT-2 is an essential transcriptional factor for CXCR5 expression on mice T_FR (31). HIV envelope induces NFAT-2 translocation to the nucleus, where it binds to multiple sites within the HIV long terminal repeat (60, 61). HIV/SIV may therefore alter NFAT-dependent expression of CXCR5. We showed that cyclosporin A, a calcineurin inhibitor that blocks NFAT dephosphorylation, decreases CXCR5 levels on macaque CD4⁺CD25⁺Foxp3⁺ cells, but we were unable to test this intriguing hypothesis in our model due to the lack of reagents that cross-react with macaque NFAT proteins.

T_FH numbers expand in acute and chronic viral infections, such as in influenza A virus infection in mice, and in chronic hepatitis B and HIV/SIV in humans and macaques (62–64). In particular during acute influenza A virus infection, a temporary T_FH expansion occurs 3 d after challenge (62). Differently, in HIV/SIV infection, a sustained expansion of T_FH is seen in chronic, but not during the acute phase, of infection, as we also observed in our study (8).

We found an association between the levels of T_FR and T_FH in acutely and chronic macaques infected with SIV_mac251. It is possible that a reduction or lack of expansion of T_FR may contribute to the increased T_FH number in chronic infection. Alternatively, the persistence of the Ag may lead to the increase in T_FH, resulting in higher levels of PD-1 in the GC and in T_FR reduction (47). Additionally, changes in T_FR function, other regulatory subsets in the GC (CD8⁺ and NK T-cells), imbalanced cytokine milieu (i.e., increased IL-6), and immune activation are likely to participate in the HIV-associated increase in T_FH numbers (8, 19).

To our knowledge our study is the first to describe T_FR dynamics and changes in the T_FH/T_FR ratio during SIV infection, together with the study by Chowdhury et al. (65) reporting similar findings in SIV_smE660-infected macaques.

By suppressing T_FH numbers and proliferation, T_FR modulate B cell responses in mice (26, 29, 47). No correlation was found with the frequency of plasmablasts or with the IgA⁺ B cells. We found an association between the relative frequency of CXCR5⁺ cells within the CD25⁺Foxp3⁺ population and the frequency of total IgG⁺ B cells and of overall switched IgM⁻ (IgA⁺IgG⁺) B cells and plasmablasts in lymph nodes (32).

The accumulation of T_FH in chronic SIV infection is associated with increased titers of higher avidity SIV-specific Igs (8). Interestingly, we observed an antithetic role of T_FR and T_FH in the avidity of Abs to the SIV gp120 protein throughout the infection, and only T_FR levels were strongly correlated with the increased avidity during chronic infection. It has been proposed that T_FH accumulation, together with HIV-associated changes in cell function, may lead to a reduction in affinity maturation due to a lowered competition for B cell selection. However, so far, the role of T_FH cells in HIV/SIV pathogenesis has been studied without making a clear distinction between T_FR and T_FH. Thus, the relative contribution of T_FH and T_FR in the impairment in B cell selection during HIV infection remains to be determined.

In summary, we identified a population of macaque CD4⁺ T cells with a phenotype, function, and location consistent with T_FR, and we revealed SIV-associated changes in the T_FR/T_FH ratio, adding to the complexity of humoral immunity to HIV.

We thank James Arthos and Claudia Cicala for helpful discussion, and Namal P. Liyanage, Dallas P. Brown, and Veronica Galli for critical reading of the manuscript.

The authors have no financial conflicts of interest.