Transcriptome Analysis of CCR9+ T Helper Cells From Primary Sjögren's Syndrome Patients Identifies CCL5 as a Novel Effector Molecule

Abstract

Introduction: CCR9+ Tfh-like pathogenic T helper (Th) cells are elevated in patients with primary Sjögren's syndrome (pSS) and indicated to play a role in pSS immunopathology. Here we delineate the CCR9+ Th cell-specific transcriptome to study the molecular dysregulation of these cells in pSS patients.

Methods: CCR9+, CXCR5+ and CCR9-CXCR5- Th cells from blood of 7 healthy controls (HC) and 7 pSS patients were FACS sorted and RNA sequencing was performed. Computational analysis was used to identify differentially expressed genes (DEGs), coherent gene expression networks and differentially regulated pathways. Target genes were replicated in additional cohorts.

Results: 5131 genes were differentially expressed between CCR9+ and CXCR5+ Th cells; 6493 and 4783 between CCR9+ and CCR9-CXCR5- and between CXCR5+ and CCR9-CXCR5-, respectively. In the CCR9+ Th cell subset 2777 DEGs were identified between HC and pSS patients, 1416 and 1077 in the CXCR5+ and CCR9-CXCR5- subsets, respectively. One gene network was selected based on eigengene expression differences between the Th cell subsets and pathways enriched for genes involved in migration and adhesion, cytokine and chemokine production. Selected DEGs of interest (HOPX, SOX4, ITGAE, ITGA1, NCR3, ABCB1, C3AR1, NT5E, CCR5 and CCL5) from this module were validated and found upregulated in blood CCR9+ Th cells, but were similarly expressed in HC and pSS patients. Increased frequencies of CCR9+ Th cells were shown to express higher levels of CCL5 than CXCR5+ and CCR9-CXCR5- Th cells, with the highest expression confined to effector CCR9+ Th cells. Antigenic triggering and stimulation with IL-7 of the Th cell subsets co-cultured with monocytes strongly induced CCL5 secretion in CCR9+ Th cell cocultures. Additionally, effector CCR9+ Th cells rapidly released CCL5 and secreted the highest CCL5 levels upon stimulation.

Conclusion: Transcriptomic analysis of circulating CCR9+ Th cells reveals CCR9-specific pathways involved in effector T cell function equally expressed in pSS patients and HC. Given the increased numbers of CCR9+ Th cells in the blood and inflamed glands of pSS patients and presence of inflammatory stimuli to activate these cells this suggests that CCR9-specific functions, such as cell recruitment upon CCL5 secretion, could significantly contribute to immunopathology in pSS.

Keywords: CCL5; CCR9+ Th cells; Primary Sjögren’s syndrome; autoimmunity; transcriptomics.

Figure 1

Transcriptomic profiling robustly separates CCR9+, CXCR5+ and CCR9-CXCR5- Th cell subsets and identifies differentially expressed genes in pSS patients. RNA sequencing was performed on FACS sorted CCR9+, CXCR5+ and CCR9-CXCR5- (double negative, DN) T helper cells from PBMCs from healthy donors (HC, n=7, CCR9+ subset n=5) and primary Sjögren’s syndrome patients (pSS, n=7). Representative flow cytometry plot of CCR9 and CXCR5 expression on CD4+ Th cells used for FACS sorting (A). Confirmation of elevated mRNA expression of CCR9 and CXCR5 in the sorted CCR9+ and CXCR5+ Th cells, respectively (B). NRC: normalized read counts, log2 normalized. Multidimensional scaling plot (MDS) shows all the differentially expressed genes (DEGs) between the Th cell subsets (C). Venn diagrams show number and percentages of DEGs between the CCR9+ vs CXCR5+ vs DN Th cell subsets (D) and between HC and pSS patients within the CCR9+, CXCR5+ and DN Th cell subsets (E). Heatmap representation of hierarchical clustering analysis using the top 100 DEGs of all comparisons, shows distinct gene expression profiles between the subsets and between HC and pSS patients (F).

Figure 2

Weighted gene co-expression network analysis (WGCNA) reveals networks of genes with coherent expression patterns involved in distinct Th cell functions. Modules with elevated eigengene expression in CCR9+ Th cells (A) or CXCR5+ Th cells (B), and modules with differential eigengene expression between HC and pSS patients (C) are shown. Each bar represents the eigengene expression of one sample in the indicated Th cell subset of a donor. Pathway enrichment analysis reveals different pathways within the modules. Pathways are categorized, and categories are indicated by the colors in the legend. The number of pathways per category is shown. The total number of pathways per module is indicated behind the module name on the X-axis (D).

Figure 3

Workflow for selection of target genes for validation per module. Genes from the modules are selected based on the criteria shown, including expression level (normalized read counts >8), module membership >0.7, differential expression between the Th cell subsets with a fold change of >1.5, or between HC and pSS of >1.5 or <0.7, and finally transcription (co) factors are selected. From the lower 4 rows, target genes that are also supported by literature evidence are selected for further analysis (A). Representative examples of genes selected based on the criteria and per module are shown (B).

Figure 4

Validation of selected target genes from the blue module in replication cohorts confirms differences between Th subsets, but not between HC and pSS patients. According to the criteria shown in Figure 3 and based on pathway enrichment analysis (Figure 2) and literature, genes were selected from the blue module and were replicated by qPCR and flow cytometry. RNA sequencing data is shown in (A, C), qPCR and flow cytometry data are shown in (B, D), respectively. HOPX, SOX4, ITGA1, C3AR1 and ABCB1 were evaluated by qPCR in an additional replication cohort (B). Genes encoding for surface proteins for which antibodies were available were replicated by flow cytometry: ITGAE (CD103), CD73 (NT5E), NCR3 (NKp30), CCR5 and ABCB1 (D). NRC, normalized read counts; FC, fold change, DN HC mean is set at 1. *, **, *** indicates statistical significance of p < 0.05, 0.01, 0.001, respectively.

Figure 5

Increased production and release of CCL5 by CCR9+ Th cells in pSS patients and controls is higher in effector cells and is enhanced by IL-7 and antigenic triggering. mRNA expression as assessed by RNA sequencing (A). In a replication cohort qPCR shows validation of significantly increased CCL5 mRNA in the CCR9+ Th cell subset as compared to CXCR5+ and CCR9-CXCR5- Th cell subsets (B). CCR9+, CXCR5+ and CCR9-CXCR5- Th cells from HC (n=4) and pSS patients (n=3) were cultured with monocytes for 72 hours followed by PMA/ionomycin restimulation (C). To determine the effect of antigen triggering (SEB=Staphylococcal Enterotoxin B) and IL-7 stimulation on CCL5 production by the Th cell subsets we pooled HC and pSS (D). CCL5 protein expression in Th cells from peripheral blood from healthy controls (n=13) and pSS patients (n=13) was analyzed ex vivo by flow cytometry in a 2^nd replication cohort (E). CCR9 expression on Th cells was also assessed in this replication cohort by flow cytometry (F). In HC CCL5 expression by effector memory and effector Th cells is compared to naive Th cells, both for CCR9- and CCR9+ Th cells (G); and in all CD45RO/CD27 compartments CCL5 expression is compared between CCR9- and CCR9+ Th cells (H). Histograms from a representative donor show CCL5 expression upon 4h of stimulation with PMA/ionomycin in the different Th cell compartments. Blue and red shows CCL5 expression in unstimulated and PMA/Ionomycin stimulated cells, respectively (I). CCL5 release expressed as the difference between unstimulated – the stimulated cultures (delta) between all four CD45RO/CD27 compartments for both CCR9- and CCR9+ Th cells (J) and within each compartment comparing CCR9- and CCR9+ Th cells (K). FC: fold change. Th cell phenotypic validation: naive (CD27+CD45RO-), central memory (CD27+CD45RO+), effector memory (CD27-CD45RO+), effector (CD27-CD45RO-) cells. MFI: mean fluorescence intensity. ΔMFI: deltaMFI (unstimulated minus PMA/ionomycin stimulated). *, **, *** indicates statistical significance of p<0.05, 0.01, 0.001, respectively.