Research ArticleIMMUNOLOGY

Aging promotes reorganization of the CD4 T cell landscape toward extreme regulatory and effector phenotypes

See allHide authors and affiliations

Science Advances  21 Aug 2019:
Vol. 5, no. 8, eaaw8330
DOI: 10.1126/sciadv.aaw8330
  • Fig. 1 Gene expression signatures of CD4 T cells in young and old mice.

    (A) Experimental flowchart: (i) Splenocytes were harvested from young (2 to 3 months, n = 4) and old (22 to 24 months, n = 4) mice; (ii) CD4 T cells were purified using magnetic separation and sorting; (iii) cells’ mRNAs were barcoded using 10x Genomics Chromium platform and sequenced; and (iv) data were computationally analyzed. (B) Representative flow cytometry plots showing highly pure CD4+TCRβ+ T cells after magnetic enrichment and sorting, discarding cells that were positive for CD8, CD19, CD11b, and/or NK1.1. These cells were used for the scRNA-seq experiments. (C) Analysis of the sorted young and old CD4 T cells stained for CD44 and CD62L surface markers. Top: Representative flow cytometry plots of cells from young and old mice. Bottom: Cells from old mice show a shift toward effector-memory identity. Data from two different experiments (n = 2 in each age group, per experiment). Each dot represents a mouse, bars represent mean ± SEM (unpaired t test, ****P < 10−4). (D) t-SNE projections of CD4 T cells including 13,186 and 10,821 cells from young (turquoise) and old (brown) mice, respectively. Each dot represents a single cell. (E) MA plot showing differentially expressed genes between age groups. Each dot represents a gene, with significantly up-regulated genes [ln(fold change) > 0.4, adjusted P < 10−3] in young and old mice colored turquoise and brown, respectively. (F and G) t-SNE projections with cells colored by the expression levels of age marker genes. Markers were selected as differentially expressed genes within an age group [ln(fold change) > 0.4] that best distinguish between age groups according to a receiver operating characteristic analysis [(F) AUC > 0.61, power > 0.23 and (G) AUC > 0.66, power > 0.33].

  • Fig. 2 CD4 T cell subset composition in aging.

    (A) t-SNE projection of all 24,007 CD4 T cells presenting seven different subsets, identified via shared nearest neighbor modularity optimization-based clustering algorithm, followed by merging of similar clusters. (B) Heatmap of gene expression z scores across cells. All CD4 T cells were grouped by subset and age (horizontal bars). Genes shown were up-regulated significantly [ln(fold change) > 0.4, combined P < 10−3] in at least one subset compared to all other cells. Genes were ordered by significance and associated with the subset with higher detection rates. (C) Violin plots showing the expression (z score) of selected canonical marker genes across all seven subsets. (D) Representative pie charts showing the percentage of cells belonging to each of the seven subsets in a young mouse and an old mouse. (E) Enrichment was computed as the log odds ratio between the frequency of each subset in old versus young mice, across all pairs. Naïve subsets were enriched in young mice [Naïve: log(median ratio) = −0.27, P = 0.03 and Naïve_Isg15: log(median ratio) = −0.23, P = 0.03]. rTregs subset was equally distributed [log(median ratio) = 0.02, P = 0.89]. Four subsets were enriched in every old mouse: TEM [log(median ratio) = 0.51, P = 0.03], aTregs [log(median ratio) = 1, P = 0.03], exhausted [log(median ratio) = 1.32, P = 0.03], and cytotoxic [log(median ratio) = 1.46, P = 0.03] subsets. Mann-Whitney U test. (F) A heatmap showing Spearman correlation coefficient (rho) between the frequency of each subset and cytokine concentration (in micrograms per milliliter) measured in serum of old mice. Because of the limited number of mice, we considered as meaningful only strong correlations (absolute Spearman correlation coefficient > 0.6).

  • Fig. 3 The genes and proteins characterizing cytotoxic, exhausted, and aTregs subsets.

    (A to C) MA plots showing differentially expressed genes between (A) cytotoxic and TEM cells, (B) exhausted and TEM cells, and (C) aTregs and rTregs cells. Only cells from old mice were considered in the analysis. Each dot represents a gene; genes that were up-regulated consistently across mice [ln(fold change) > 0.4] were colored by the corresponding subset. (D) Left: Representative t-SNE plots of CD4 T cells coming from young (top) and old (bottom) mice colored by subset identity. Right: Analysis was based on the expression of 10 marker proteins chosen according to the gene expression profiles. The mean fluorescence intensities (MFIs) of each marker are presented. (E) Representative flow cytometry plots gated on FOXP3+ cells showing the MFI of selected marker proteins that relate to Tregs activation, projected on CD81 (rTregs) and CD81+ (aTregs) populations. (F) Violin plots quantitatively showing the MFI of each marker in rTregs and aTregs. Each dot represents a mouse (n = 6, from two different experiments). Paired t test, *P < 0.05, **P < 0.01, and ****P < 10−4.

  • Fig. 4 Dynamics of RECs across age and immune sites.

    (A to C) Prevalence of RECs and naïve cells measured via flow cytometry (representative panels, left) in mice at the age of 2, 6, 12, 16, and 24 months (right). Color-shaded area of each graph represents ± SEM. Spearman correlation coefficients (r) are shown. (A) Exhausted cells (pink), defined as PD1+CD62L cells, and naïve cells (blue), defined as CD62L+PD1 out of CD4+FOXP3EOMESCCL5cells. (B) aTregs cells, defined as CD81+ cells out of CD4+FOXP3+ cells. (C) Cytotoxic cells, defined as EOMES+CCL5+ out of CD4+ cells. Gating was based on unstained samples and fluorescence minus one. Data include two (A) or three (B and C) different experiments, n = 5 to 15 per time point. (D and E) Absolute numbers of total CD4 (D) and RECs (E) cells out of total live cells in each spleen, calculated as specific cell count per gram of spleen in young (2 months, n = 7) and old (24 months, n = 9) mice. (F to H) Percentages of RECs in different immune sites: (F) exhausted cells, (G) aTregs cells, and (H) cytotoxic cells. LNs represent the mean cell frequencies of four LN sites. Data include three different experiments, n = 2 to 3 per experiment. P values computed via one-way analysis of variance (ANOVA) test with Tukey correction for multiple comparisons (*P < 0.05, **P < 0.01, and ****P < 10−4).

  • Fig. 5 The regulatory and functional properties of RECs.

    (A) A heatmap of old CD4 T cells showing 27 high-confidence regulons that were active consistently across all old mice. Active regulons per cell appear in black; the horizontal color bar indicates the subset associated with each cell. Numbers in parentheses represent the number of genes comprising the regulon. (B) Radared-balloon plot showing regulons’ activity per subset per mouse. Each circle corresponds to a single subset and is divided into four slices, one per mouse. Slice size reflects the fraction of subset cells with active regulon, normalized to the maximal fraction of that regulon across mice and subsets (fig. S6A shows other regulons). (C) Ex vivo suppression activity of sorted CD25highCD81 (yellow, cells derived from young mice), referred to as rTregs, or CD25highCD81+ (brown, cells derived from old mice), referred to as aTregs, after 72 hours of coculture with activated naïve CD4 T cells from young CD45.1 mice. The reduction in the proliferation of the activated CD4 T cells was measured via flow cytometry and calculated as percentage of suppression (Materials and Methods). Left: Representative flow cytometry plots showing reduced proliferation in the presence of aTregs. Right: Violin plot showing the suppression ability (%) of rTregs versus aTregs. Each dot represents cells pulled from three mice (n = 8, from two independent experiments). Unpaired t test (*P < 0.05). (D) The percentage of cells positive to pro-inflammatory and cytotoxic cytokines in TEM, exhausted, and cytotoxic subsets after 48 hours of activation, measured by flow cytometry. Lines connect measurements within the same mouse. Data from two independent experiments, n = 7 mice. Paired t test (*P < 0.05, **P < 0.01, and ****P < 10−4). (E) Schematic illustration of the accumulation of RECs with age, showing their key transcription factors and markers (within and to the right of each cell, respectively), which point to a dysregulated immune response.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/8/eaaw8330/DC1

    Fig. S1. Purification of CD4 T cells from young and old mice.

    Fig. S2. Dimensionality reduction of scRNA-seq data.

    Fig. S3. Subset identity and abundance in mice used for scRNA-seq.

    Fig. S4. RECs activation state and correlation with serum cytokines.

    Fig. S5. RECs abundance with time and across different immunological sites.

    Fig. S6. Gene regulatory networks and cytokine secretion in RECs.

    Fig. S7. Characterization of T helper polarizations within RECs.

    Table S1. The number of cells per mouse before and after quality control.

    Table S2. Subset and age-group expression of genes.

    Table S3. RECs comparison to closely related subsets.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. Purification of CD4 T cells from young and old mice.
    • Fig. S2. Dimensionality reduction of scRNA-seq data.
    • Fig. S3. Subset identity and abundance in mice used for scRNA-seq.
    • Fig. S4. RECs activation state and correlation with serum cytokines.
    • Fig. S5. RECs abundance with time and across different immunological sites.
    • Fig. S6. Gene regulatory networks and cytokine secretion in RECs.
    • Fig. S7. Characterization of T helper polarizations within RECs.

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Table S1 (Microsoft Excel format). The number of cells per mouse before and after quality control.
    • Table S2 (Microsoft Excel format). Subset and age-group expression of genes.
    • Table S3 (Microsoft Excel format). RECs comparison to closely related subsets.

    Files in this Data Supplement:

Navigate This Article