All animal experiments were approved by Institutional Animal Care and Use Committees (IACUCs) prior to study initiation, including the NCI/NIH Animal Care and Use Committee and the Advanced BioScience Laboratories, Inc. (ABL) Animal Care and Use Committee. Rhesus macaques were maintained at the NCI Animal Facility (Bethesda, MD) under protocol VB012 and VB001. Following the challenge phase of the study, macaques were housed at ABL (Rockville, MD) under protocol AUP526. Each of these facilities is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International. The standard practices closely follow recommendations made in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Animals were housed in accordance with the recommendations of the AAALAC Standards and with the recommendations in the Guide for the Care and Use of Laboratory Animals. Details of animal welfare and steps taken to ameliorate suffering were in accordance with the Guide for the Care and Use of Laboratory Animals and the recommendations of the Weatherall report on the use of NHPs in research, as approved by the relevant IACUCs.

A total of 20 young Male (6 to 8 years old) Indian Rhesus macaque (Macaca mulatta), 10 SIV naïve, and 10 chronic SIVmac₂₅₁ infected were included in this study. SIV naïve animal’s tonsils were obtained from the control arms of a previous, unpublished vaccine immunogenicity study. Control animals received saline intra-rectally during the challenge phase. None of these animals were exposed to SIV antigens. SIV-infected animals were obtained from a previously published study (10). chronic SIV infected macaques were challenged with repeated low doses of SIV_mac251 intra-rectally, and tonsils were obtained when animals were sacrificed 48 weeks after primary SIV⁺ results.

Tonsils were all collected at necropsies and lymphocytes isolate using assays optimized in our laboratory and according to the published protocol (11). Briefly, using forceps, place tonsils in a large plastic petri dish and cover with 4°C HBSS. Tonsil cut into 3- to 10-mm fragments with scissors and push lymphoid cells through the mesh using the flat end of a 60-ml plastic syringe plunger. Rinse the remaining tissue two to three times with HBSS until clear. Transfer equal volumes of cell suspension from the petri dish to 50-ml centrifuge tubes. Add colder HBSS and process the remaining tissue fragments by repeating steps twice to free as many lymphocytes as possible from the stroma. After transferring all cells to centrifuge tubes, break up any cell clumps and suspend the cells by repeated pipetting. Underlay cell suspension with 10 ml Ficoll-Hypaque. Centrifuge 20 min in a Sorvall H-1000B rotor at 1800 rpm (1000 × g), room temperature. Accelerate slowly and decelerate without braking. Collect mononuclear cells from the interface and discard cell pellets containing fibroblasts and other cell debris.

Tonsil single-cell suspensions were thawed, washed, and resuspended in R10 medium and incubated overnight at 37°C and 5% CO2 prior to staining for flow cytometry or activation for the intracellular staining. 2 million cells were stained with non-human primate cross-reactive fluorescent tag antibodies ( Supplemental Table 1 ). LIVE/DEAD Aqua dye (Invitrogen) and isotype-matched controls and/or fluorescence-minus-one (FMO) controls were included in all assays. Acquisitions were made on a Cytek Aurora spectral flow cytometer (Cyteck Biosciences, CA) and analyzed using FlowJo version 10.6.1 (Becton Dickinson, Ashland, OR) for high dimensional data analysis following standardized workflow. The first step was integrating the flow cytometry marker data from all the infected and uninfected samples for implemented compensation and scaling using OMIQ (Dotmatics, Boston, MA). Then, Non-linear dimensionality reduction was performed using t-Distributed Stochastic Neighbor Embedding (tSNE) (12)to visualize multi-dimensional expression landscapes in two dimensions (2D). Next, cell subsets were visualized by overlaying defined immune subset information onto the 2D tSNE plots. Next, unbiased clustering was performed on the expression values of the lineage markers using the FlowSOM algorithm in OMIQ, which uses a self-organizing map followed by hierarchical consensus meta-clustering to detect cell populations. Finally, a reduced two-dimensional space was implemented to visualize unbiased clusters using the Uniform Manifold Approximation and Projection (UMAP) algorithm (13) in the OMIQ. The next step was to annotate cell subsets by visually investigating heat maps of median marker expressions across clusters and expressions of these markers in the UMAP space.

To assess innate and adaptive immune cell function, Tonsillar (1 x10⁶ cells) were resuspended in RPMI 1640(Gibco) containing 10% FBS and either stimulated with PMA (50 ng/ml) and ionomycin (1 μg/ml) or cultured in medium alone. Anti-CD107a was added to each tube at a concentration of 20 μl/ml and, Golgiplug (brefeldin A) [Cat No: 555029, BD biosciences] and Golgistop (monensin) [Cat No: 554724, BD biosciences] were added at final concentrations of 6 μg/ml; then all samples were cultured for 12 h at 37°C in 5% CO2. After culture, samples were surface stained using markers to delineate live cells (LIVE/DEAD Aqua dye), leukocytes (CD45), and innate and adaptive immune cell populations (complete list of antibodies in the Supplemental Table 1 ). Cells were then permeabilized using Caltag Fix & Perm, and intracellular cytokine staining was performed for combinations of antibodies to measure IFN-γ, TNF-α, IL-22.

All statistical and graphical analyses were performed using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA). The Mann-Whitney-Wilcoxon test was used to compare continuous factors between two groups. Correlation analyses were performed using the Spearman rank correlation method with exact permutation P value.

To elucidate the immune cell composition in adult rhesus macaques’ tonsils, we performed high dimensional flow cytometry on cryopreserved lymphocytes from tonsils. We first assessed the percentage of different innate and adaptive immune cells in tonsils using manual gating ( Supplemental Figure 1 ) and unsupervised raw data clustering of single live CD45⁺ cells. We defined innate (monocytes and NK), adaptive (T and B cells), and unconventional (NKB) immune cells based on receptor expression and compared the abundance of each subset. We found that CD3^-CD20^-NKG2A^-NKp44^- cells are predominant in macaque tonsils, followed by B cells (CD20⁺) and CD3⁺CD4^-CD8^- (Double Negative) T cells respectively ( Figure 1 ). Similar to peripheral blood, we identified three major monocyte populations, classical monocytes (CD14⁺⁺CD16^-), non-classical (CD14⁺CD16⁺⁺) and intermediate (CD14⁺⁺CD16⁺) monocytes. The classic subset was the dominant monocyte population in the tonsils ( Figure 1C ). We found that CD3^-CD20^-CD14^-NKG2A^-NKp44^- cells are more common than classical cytotoxic NK cells (CD3^-CD20^- CD14^-NKG2A⁺) and regulatory NK (CD3^-CD20^- CD14^-NKp44⁺) cells in tonsils from naïve animals ( Figure 1D ). B cells were more common than CD3 cells in the tonsils. However, the difference was not statistically significant ( Figure 1E ). CD4 and CD8 T cells were more common than CD3⁺CD4⁺ CD8⁺ T (DP) cells. However, DN cells (CD3⁺CD4^-CD8^-) were the dominant population of CD3⁺ cells in the tonsils ( Figure 1F ). Additionally, we found that NK-T-like cells, as well as the recently described B-cell population, termed NKB cells, present in the naïve macaque’s tonsils ( Figures 1G–I ). NKB cells express CD20 (B cell receptor) and NKG2A, NKp44, CD16, and CD56 receptors on their surface.

Next, we utilized an unbiased clustering approach based on a multivariant t-mixture model to analyze our high-dimensional spectral flow cytometric data comparing cell populations from naïve to chronically SIV infected animal’s tonsils. We identified eight distinct immune cell clusters out of 15 that we visualized in two dimensions using the UMAP algorithm ( Figure 2 ). Using clustered heatmap analysis. We correlated expression intensity with clusters to annotate T cells (clusters 1,2,3,4,5), B cells (clusters 11, 12), unconventional cells (clusters 6: T cells expressing B cells receptor CD20 and NK receptor NKp44, clusters 7: NKT-like cells (CD3+CD56+CD161+CRTH2+), NK cells (clusters 8, 13,15), other immune cells (cluster 9: CD20+CD3+CD161+), cluster10: CD20+CD3+CD4+CD56+CD161+, cluster14: CD20+CD8+CD27+). Comparing immune population dynamics, we found significant higher CD161⁺ CD4 T cells (cluster 2), CD27⁺ CD4 T cells (cluster 3), NKG2A⁺ NK cells (cluster 8), and CD161⁺ CD3⁺CD20⁺ cells (cluster 9) in chronic SIV infected macaques. Conversely, we observed significantly lower CRTH2⁺ NKp44⁺ cells (cluster 4), CD56⁺NKG2A⁺ NK cells (cluster 13), and CD117⁺ CD27⁺ innate cells (cluster 15) in chronic SIV compared to uninfected animals.

It is well-established that HIV infection impacts monocytes subpopulations (14). Thus, we compare the frequency and functional difference of the monocyte populations between tonsils from naïve and chronic SIV infected animals ( Figure 3 ). We observed a slight reduction of total monocytes ( Figure 3A ), with classical monocytes (CM) being the dominant population in infected tonsils ( Figure 3B ). In contrast, inflammatory monocytes (intermediate/ITM) were significantly depleted when compared to healthy animals ( Figure 3C ). Concurrently, we found lower IFN-γ production in intermediate monocytes ( Figure 3E ) but not in the other two monocytes subsets. In contrast to IFN-γ, the production of TNF-α ( Figure 3F ) was higher in non-classical and intermediate monocytes in tonsils from SIV-infected animals. We further found a significantly higher CD107a ( Figure 3G ) in all three subsets indicating a more activation, potentially skewed cytokine profile, in SIV infection. While we found considerable IL-22 production by monocyte subsets in tonsillar monocytes from uninfected animals, IL-22 expression in non-classical and intermediate monocytes were significantly reduced in chronic SIV infected animals ( Figure 3H ).

We and other previously described three major types of NK cell subsets in rhesus macaque mucosal tissues (10, 15). All three NK cell subsets were present in tonsils from naïve animals ( Supplemental Figure 1 ). There was a reduction of NKp44⁺ (NKp44⁺ ILC) cells in SIV-infected tonsils. Still, the difference was not statistically significant ( Figure 4B ). NKG2A+ NK ( Figure 4A ) and NKp44-NKG2A- (DN) cells ( Figure 4C ) did not differ between naïve and tonsils from SIV-infected animals. In addition, we measured IFN-γ, TNF-α, IL-22 and CD107a in NK cell subsets ( Figures 4D–G ). We found no statistically significant difference in IFN-γ, TNF-α, and IL-22 secretion between NK cell populations from naïve and chronic SIV-infected animals. However, CD107a (degranulation) was significantly higher at baseline (unstimulated) as well as post-stimulation in NKp44⁺ cells and NKp44^- NKG2A^- cells ( Figure 4F ) in the tonsils of chronically SIV-infected animals. SPICE analysis ( Figures 4H–J ) showed higher multifunctionality in NKp44⁺cells ( Figure 4I ) and NKG2A^-NKp44^-(DN) cells ( Figure 4J ) from SIV-infected tonsils.

NKT–like cells (CD3⁺CD56⁺) and NKB cells (CD20⁺NKG2A⁺) are poorly described but might play a role in bridging the innate and adaptive immune response. Alternatively, both cell types might higher as part of immune dysregulation/exhaustion in chronic disease. In this study, we found a significantly higher level of NKT-like cells in tonsils from chronically SIV macaques ( Figure 5A ). Although not statistically significant, NKB cells ( Figure 5B ) were also higher in infected animals. We found no difference in cytokine production by NKT cells between naïve and infected animals ( Figure 5C, E, F ). However, NKB cells from infected animals produced significantly lower levels of TNF-α ( Figure 5D ).

Although peripheral blood T cell subsets show significant changes during chronic SIV infection, we did not find substantial differences in the frequency of total T cells (CD3⁺) ( Figure 6A ), CD4 T cells ( Figure 6B ) or CD8 T cells ( Figure 6C ) in tonsils between SIV infected and healthy animals. However, CD3⁺CD4⁺CD8⁺ T cells (double-positive T cells, Figure 6D ) significantly higher in chronic SIV infection. In contrast, we found a significantly lower frequency of CD3⁺CD4^-CD8^-(DN) ( Figure 6E ) cells in infected tonsils. IFN-γ production was significantly enhanced in CD4 and CD8 T cells ( Figure 6F ) in chronic SIV infected animals. TNF-α was elevated in CD8 T cells and DN T cells ( Figure 6G ) from infected animals, while IL-22 secretion in unstimulated (baseline) was significantly lower in CD4 T cells, CD8 T cells, and CD4^-CD8^- T cells ( Figure 6H ) and respectively in CD8 and DN T-cells after stimulation. Except for CD8 T cells, degranulation marker CD107a was elevated at baseline and after stimulation in CD4 T cells, CD3⁺CD4⁺CD8⁺ (DP) cells, and CD3⁺CD4^-CD8^- (DN) T cells ( Figure 6I ) in infected animals.

Our high dimensional flow analysis suggested the presence of CD161⁺ T cells in macaque’s tonsils. These cell subsets have shown a role in chronic inflammatory conditions. However, in HIV infection, this phenotype was shown to be associated with high permissiveness for HIV-1 infection (16). We compared CD161⁺ and CD161^- CD4 and CD8 T cells in tonsils from SIV-infected and naïve animals. We found significantly elevated CD161⁺ CD4 T cells levels ( Figure 7A ) but not CD161^-CD4 ( Figure 7B ), CD161⁺CD8 T cells ( Figure 7C ), or CD161^-CD8 T ( Figure 7D ) cells. Secretion of IFN-γ and TNF-α, was not different between CD4⁺ CD161⁺ T cells from naïve, and SIV-infected animals. CD161^- CD4⁺ cells showed lower IFN-γ production at baseline with no difference for TNF-α ( Figures 7E, F , respectively). CD161^- CD8 cells secreted more IFN-γ and TNF-α, post-stimulation compared to cells from naïve animals. However, no difference was seen for CD161⁺CD8 T-cells ( Figures 7E, F ). We found a general reduction (pre- and post-activation) of IL-22 in CD161^+/- CD4 T cell subsets from infected animals. Similarly, CD161⁺ CD8 T cell showed lower levels of IL-22 at baseline ( Figure 7G ) with no difference post-stimulation between infected and naïve animals. More CD161^+/- CD4 T cells from infected animals stained positive at baseline and post-stimulation for CD107a. No difference was observed for CD8⁺ CD161⁺ T cells, and CD8⁺CD161^- cells showed higher CD017a levels only after activation ( Figure 7H ).