Distinct localization, transcriptional profiles, and functionality in early life tonsil regulatory T cells

Human samples

Sample Processing

Flow Cytometry and Cell Sorting

Whole transcriptome profiling by RNA sequencing

Gene and Pathway Analysis

Gene Set Enrichment analysis (GSEA) was performed using GSEA v4.3.2 on differentially expressed gene (DEG) lists preranked by log fold change in normalized expression (28). Previously published gene ontology gene sets were accessed from the Molecular Signatures Database (29, 30). Additional gene ontology (GO) analysis was performed on DEG lists using PANTHER v17.0 overrepresentation test with Fisher’s exact test and GO database version 2022-07-01 (29, 31). Pathway analysis was performed on DEG lists via Enrichr, quantifying enrichment level of every pathway in the Kyoto Encyclopedia of Genes and Genomes (KEGG) 2021 human gene set (32–34). Pathway analysis figures were generated using the Enrichr Analysis Visualizer Appyter (35, 36).

T Cell Stimulation

Immunofluorescence Imaging

5μm sections were made into immunoblank slides and provided to the Human Immune Monitoring Core at CUIMC for immunofluorescence imaging. Sections were stained using the Opal 7-Color automated IHC detection kit (Akoya Biosciences) as previously described (37). The multiplex panel included DAPI, CD4 (clone EPR6855, Abcam), CD8 (clone 4B11, Leica), CD19 (clone BT51E, Leica), CD163 (clone 10D6, Leica), and FoxP3 (clone 236A/E7, Abcam). Images were acquired on the Vectra 3 platform and imported into QuPath – 0.3.0 (38) for analysis. Cell detection feature was used to define individual cell perimeters within the image. Single measurement classifier feature was used to classify markers and define cell subtypes; B cell, CD19⁺, Macrophage CD163⁺, CD8⁺ T cell, CD8⁺/CD4⁻/CD19⁻/CD163⁻, CD4⁺ conventional T cells, CD4⁺FoxP3⁻/CD8⁻/CD19⁻/CD163⁻, CD4⁺ Treg, CD4⁺FoxP3⁺/CD8⁻/CD19⁻/CD163⁻. Phenotypic classifications for each image were then exported for additional analysis.

Identification of follicles in human tonsils

We used the imcRtools package in R, a package commonly used to analyze imaging mass cytometry data (39, 40). X and Y coordinates of cells identified in each image were concatenated into a singleCellExperiment data container. All cells within a distance radius of 12.5μm were considered interacting using buildSpatialGraph with “expansion”. Using patchDetection function, spatial clusters of CD19⁺ cells were detected and defined as follicles if a minimum of 250 cells were in the patch. Non-CD19⁺ cells were included into follicles if they fell within 1μm of a CD19⁺ patch (expansion = 1). Cells outside these patches were classified as extrafollicular.

Measuring Treg interactions in the human tonsil

We measured the average number of interactions (by expansion) between Tregs and other cell types in the extrafollicular space using countInteractions in classic mode. To account for the differences in immune cell number and composition across images, we estimated randomly expected cell type interactions by randomizing the extrafollicular cellular labels within each image and recounting interactions for 100 iterations. We then compared the observed and randomly expected interactions to identify cell types with preferential colocalization or segregation. To visualize differences in interaction likelihood between children and adults, we divided observed interactions by randomly expected interactions.

Statistical Analysis

Flow cytometry: data were analyzed using FCS express (ver. 7.12.0005) and FlowJo (ver. 10.7.1). Descriptive analyses and statistical testing of compiled flow cytometry and imaging data were performed using Prism (GraphPad ver. 9.3.1).

Materials availability

This study did not generate new unique reagents.

Data and code availability

Population RNA sequencing data has been uploaded to GEO (GSE262622). Available at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE262622

Tonsil tissue from pediatric subjects provides insight into early life regulatory T cells

We have previously acquired blood and/or tonsil tissue during pediatric ambulatory visits and elective surgical procedures to study immune system development (3, 41, 42). We obtained tonsil tissue and/or blood samples from pediatric (n= 88) subjects aged 0–10 years (see methods). Using flow cytometry, regulatory T cells (Tregs) were identified by cell surface marker expression (CD25⁺/CD127⁻) on CD4⁺ T cells along with intracellular detection of the canonical Treg transcription factor FoxP3 (Fig. 1A). In blood, the highest proportions of Tregs were noted in the first years of life whereas stable proportions were noted in tonsillar tissue (Fig. 1B). In paired tonsil and blood samples, proportions of Tregs and FoxP3 expression, as measured by mean fluorescence intensity, was highest in tonsil (Fig. 1C). The higher proportions and increased expression of FoxP3 seen in tonsil Tregs during childhood suggest tonsillar tissue is a unique tissue environment for investigating the role of Tregs in adaptive immunity.

Tregs preferentially localize with extrafollicular CD8⁺ T cells during early life

As prior studies have shown differences in Treg proportions and phenotype in tissues when comparing infants to older children and adults (3, 18), we procured tonsils from adult organ donors (n=16) to define age related differences. We chose to compare our youngest subjects (< 3 years of age) to adults, eliminating the impact of ongoing early life immune development and maximizing age related differences in our analysis. We sought to define Treg localization to determine if spatial orientation differences existed. Multiplex immunofluorescent imaging was performed on tonsil tissue sections to identify immune cell subsets including conventional T cells (CD4⁺/FoxP3⁻ or CD8⁺), Tregs (CD4⁺/FoxP3⁺), B cells (CD19⁺), and macrophages (CD163⁺) (Fig. 2A, Fig. S1A and methods). Using computational tissue segmentation analysis we identified follicular structures, enabling further categorization of immune cell subsets into discrete tissue areas; follicular vs extrafollicular (Fig. 2A, Fig. S1B and see methods). There were notable differences in cellular and tissue composition when comparing adult and pediatric tonsil. Pediatric tissues had increased overall cellular density, B cells, follicular area, and Tregs whereas adult tonsils had increased macrophages and CD8⁺ T cells (Fig. S1C). To determine if differences in tissue composition impacted Treg localization we analyzed Treg density by tissue structure. Despite pediatric samples having significantly increased follicular area, Treg density within follicles was similar to that in adults (Fig. 2B). Strikingly, Treg density was significantly higher in the extrafollicular space in pediatric tissue compared to adults (Fig. 2B).

The observed bias in tissue localization led us to investigate whether additional differences existed in interactions between Tregs and immune cell subsets. We quantified total Treg to cell interactions across the adult and pediatric groups and found no significant differences (Fig. 2C). We performed additional analysis on interactions occurring in the extrafollicular space given the observed bias of Treg density in early life (Fig. S1D and see methods). We found that adult Tregs mainly had observed interactions which were similar to what would be randomly expected, except for a decreased interaction with macrophages (Fig. S1D). In contrast, early life Tregs demonstrate decreased interactions with B cells and macrophages but increased interactions with conventional CD4⁺ and CD8⁺ T cells (Fig. S1D). Finally, we compared the Treg to cell interactions (observed divided by randomly expected) across age. While both adult and pediatric Tregs had been shown to have decreased interactions with macrophages, this difference was most apparent in adult tissue (Fig. 2D). Early life Treg interactions with B cells (decreased) and CD8⁺ T cells (increased) were both significantly different compared to adults (Fig. 2D). These observations were quite striking given the significantly reduced number of CD8⁺ T cells in pediatric tissue (Fig. S1C). Together these localization patterns reveal a bias of pediatric Tregs towards extrafollicular CD8⁺ T cells, suggesting a critical role for Tregs in shaping CD8⁺ T cell responses in early life.

Tissue location drives expression of regulatory T cell transcriptional signature

As circulating T cells enter tonsil tissue through high endothelial venules located in extrafollicular zones (43), we hypothesized that blood Tregs could be contributing to the increased frequency of extrafollicular Tregs in children (Fig. 2B). To assess similarity between circulating and tonsil Tregs we recruited healthy adult blood donors (n=4) and performed population RNA sequencing on sorted blood and tonsil conventional CD4⁺ CD25⁻/CD127^+/− T cells (TConv) and Tregs from pediatric (n= 8, aged < 3 years) and adult (n=7) donors/subjects (see methods, Fig. 1A). Principal component analysis (PCA) demonstrated that tissue location and cell subset were the major drivers of variance (Fig. S2A) with canonical Treg (FOXP3 and IL2RA) and conventional (IL7R and THEMIS) genes differentially expressed between the subsets (Fig. S2B/C). Analysis of the top 50 DEGs revealed enrichment of genes associated with Treg identity, including, BATF (44, 45), IKZF4 (46, 47), TNFRSF9 (48), TNFRSF18 (49), LRRC32 (50), LAYN and RTNK2 (51) (Fig. S2C) demonstrating that our sorting strategy resulted in isolation of targeted subsets. Relative expression of individual genes within the DEG analysis suggested additional differences within the Treg population were likely to exist based on age and tissue site.

To identify whether Tregs derived from blood and tonsillar tissue of adult and pediatric donors exhibited differentiated identities, we directly compared their gene expression profiles. PCA performed on all Tregs revealed tissue location as the main driver of transcriptional variance, with donor age related differences visible to an extent in blood, but markedly apparent within tonsillar Tregs (Fig. 3A). DEG analysis similarly demonstrated that gene expression patterns segregate primarily by tissue source (Fig. 3B). Notably, compared to circulating Tregs, tonsillar Tregs exhibit upregulation of key genes involved in Treg immunosuppressive activity such as IL10 and LAG3 (52, 53). To further characterize this tonsil-specific signature, we performed independent analyses comparing tonsillar to circulating Tregs within the two separate age cohorts. We found a considerable degree of overlap between their DEG lists (697 genes) with 923 and 1562 genes differentially expressed in only the pediatric or adult cohorts respectively (Fig. 3C). The set of 697 shared genes representing a conserved tonsil Treg signature was analyzed via GSEA to infer age-independent tissue Treg specialization and was found to be enriched for multiple functional processes including regulation of T cell activation (GO:0050863), regulation of interleukin-10 production (GO:0032733), and regulation of T cell cytokine production (GO:0002724) (Fig. 3D and Table S1). These results support tissue as a distinct location of Treg activity throughout human life and suggest that within tissue, Tregs adapt to the unique challenges posed by discrete phases of human life.

Increased functional signature defines early life tonsillar regulatory T cells

As the majority of immune responses occur within tissue and notable differences were observed across age in tonsil Treg localization and transcriptional signatures, we performed additional analysis on tonsil Tregs. DEG between pediatric and adults tonsillar Tregs were identified and used to perform pathway analysis (see methods, Fig. 4A, and Table S2). In the pediatric group, a dramatic increase in pathways indicating cellular turnover was noted, as demonstrated by the nearly tenfold overrepresentation of Cell Cycle (MAP04110) and DNA Replication (MAP03030) pathways. In addition, pathways for expression of Cell Adhesion genes (MAP04514), phosphatidylinositol 3-kinase (PI3K)-Akt signaling (MAP04151) and TNF signaling (MAP04668) were also preferentially upregulated, indicating subtler tuning of tonsillar Treg function and phenotype in early life. Fewer pathways were significantly upregulated in adult tonsil Tregs compared to children, and included Calcium Signaling Pathway (MAP04020), MAPK signaling pathway (MAP04010), and Cytokine-Cytokine receptor interaction (MAP04060) (Table S2). To further explore differences related to Treg biology we removed genes appearing in the Gene Ontology Biologic Process “Cell_Cycle_Process” gene set (GO 0022402) and repeated GSEA on the remaining 1,061 DEG (549 pediatric, 512 adult). Pediatric Tregs were enriched for Negative Regulation of Immune System Process (GO 002683) and Regulation of Lymphocyte Activation (GO 0050870) (Fig. 4B) suggesting a potentially enhanced effector identity in early life.

Transcriptionally, Tregs can be defined by a set of genes associated with their identity, including FOXP3, CTLA4, and others. We selected a gene set defining an effector Treg gene signature (54, 55) and summarized core gene expression (Fig. 4C). The 39 core gene set was separated into 2 major nodes. The first node contained 22 genes of which 13 demonstrated increased expression in pediatric tonsillar Tregs. These included genes for costimulatory molecules CTLA4, ICOS, and LAG3, known to correlate with stronger immunosuppressive capabilities in Tregs (56–58) and activation of PI3K-Akt signaling (59). The pediatric Treg group displayed increased expression of CCR5, key for suppression of CD8⁺ T cell responses (60), and IL10, coding for the anti-inflammatory cytokine characteristically produced by Tregs (61). The second node of 17 genes included lineage defining genes FOXP3 and IL2RA (Fig. 4C), which were not differentially expressed. Only 2 DEGs emerged from this set including, LGALS1, which functions in cell adhesion, and UTS2, recently identified as part of a human Treg core gene set (51), with both being increased in adults. Altogether these discrete differences in early life Tregs reveal preferentially activated pathways associated with Treg suppressive functionality, distinguishing themselves from their adult counterparts.

Phenotype of early life and adult tonsillar Tregs follow transcriptomic distinctions

To validate protein expression of key DEG across age cohorts, we performed flow-cytometric phenotyping on pediatric and adult tonsil Tregs. Overall, differences in the phenotype of both groups closely followed their transcriptomic signatures. Consistent with the prominent expression of cell cycle genes (Fig. 4A) pediatric Tregs expressed high levels of Ki67⁺ (Fig. 5A). Early life Tregs were further characterized by significantly increased expression of costimulatory molecules ICOS, which results in signaling through the PI3K pathway (Fig. 4A), and CD28 (Fig. 5B). While CTLA4 was differentially expressed at the transcript level, proportions of CTLA-4⁺ Tregs did not differ (data not shown) and increased mean fluorescence intensity in the pediatric Tregs did not reach statistical significance (Fig. 5C). The transcription factor Helios, encoded by IKZF2, which is a marker of Treg stability in humans (62, 63) was also similar between groups (Fig. 5D). To examine if differential adaptations to the tissue environment were present, we assessed expression of ligand CD11a, encoded by IGTAL, and CD69, an activation and canonical tissue residency marker, which both showed similar expression across age (Fig. 5E). Similar proportions of CD39 (encoded by ENTPD1) were also observed, consistent with its variable expression across age cohorts (Fig. 4C and 5F). As homing receptor CCR5 was among the DEGs (Fig. 4C) we assessed the expression of several chemokine receptors. CCR5 and CCR2, known to promote Treg abundance in secondary lymphoid organs by enhancing CD25 expression (64) were increased in pediatric Tregs whereas CCR4 was found to be higher in adult Tregs (Fig. 5G). These data demonstrate that transcriptional signatures largely translated to protein expression with differential adaptations existing across age.

Enhanced functionality defines early life regulatory T cells

Our RNA sequencing analysis revealed DEGs associated with Treg effector function including IL10 and GZMB (Fig. 4C) and upregulated pathways related to cell cycle. We sought to determine if early life Tregs possessed differential functional ability. Tonsil Tregs were sorted (Fig. 1A and see methods) then stimulated and stained for intracellular cytokine detection using flow cytometry. Pediatric Tregs produced significantly more IL-10 compared to adults (Fig. 6A), suggesting an increased ability to produce canonical Treg cytokines. To determine if differences in proliferative ability existed, we performed a proliferation assay on sorted Treg and conventional CD4⁺ T cells (see methods and Fig. 6B). Pediatric Tregs had both a higher division index and proliferation index, which excludes cells that never divided, compared to adult Tregs (Fig. 6C). As other early life T cell subsets are known to undergo increased proliferation following stimulation (65) this ability may be a unique property of early life that is lost with aging. These results reveal an enhanced ability to perform canonical effector cytokine function and capabilities not present in later life.