Epigenetically quantified immune cells in salivary glands of Sjögren's syndrome patients: a novel tool that detects robust correlations of T follicular helper cells with immunopathology

Abstract

Objective: To investigate whether epigenetic cell counting represents a novel method to quantify immune cells in salivary glands of patients with different forms of Sjögren's and sicca syndrome and to capture immunopathology and potentially aid in diagnosis.

Methods: DNA from frozen salivary gland tissue sections of sicca patients was used for bisulphite conversion of demethylated DNA cytosine residues, followed by cell-specific quantitative PCR to calculate cell percentages in relation to total tissue cell numbers as quantified by housekeeping gene demethylation. The percentages of epigenetically quantified cells were correlated to RNA expression of matched salivary gland tissue and histological and clinical parameters.

Results: The percentages of epigenetically quantified CD3, CD4, CD8, T follicular helper (Tfh) cells, FoxP3+ regulatory T cells and B cells were significantly increased in the salivary glands of patients with SS. Unsupervised clustering using these percentages identified patient subsets with an increased lymphocytic focus score and local B cell hyperactivity and classifies patients different from conventional classification criteria. In particular, Tfh cells were shown to strongly correlate with the expression of CXCL13, lymphocytic focus scores, local B cell hyperactivity and anti-SSA positivity.

Conclusion: Epigenetic cell counting is a promising novel tool to objectively and easily quantify immune cells in the labial salivary gland of sicca patients, with a relatively small amount of tissue needed. In view of the potential of this technique to include a huge number of (cell-specific) biomarkers, this opens up new standardized ways of salivary gland analysis with high relevance for patient classification, understanding of immunopathology and monitoring of drug responses in clinical trials.

Keywords: Sjögren’s syndrome; Tfh cells; epigenetic cell counting; salivary gland.

Fig. 1

ECC detects increased CD3 lymphocytes in salivary gland biopsies correlating with gene expression and IHC (A) ECC is based on cell type–specific epigenetic biomarkers marked by the absence of CpG methylation (demethylated regions). See Supplementary Methods. (B) Only in T cells, the CD3 gene region is epigenetically active (yellow) and will be amplified by PCR. In other cell types, this region is epigenetically inactive (blue). The number of epigenetically active CD3 gene copies translates into the number of T cells vs the total number of cells with epigenetically active reference gene GAPDH. (C) ECC detects increased CD3 percentages in salivary glands of iSS, pSS and sSS patients vs nSS patients. (D) Epigenetically quantified CD3 T cells strongly correlate with CD3 gene expression and (E) CD3 expression as digitally quantified following IHC. Medians are shown. Kruskal–Wallis test with Dunn’s multiple comparison test was used. **P < 0.01 and ***P < 0.001 vs nSS patients. Spearman correlation coefficients (r) are given for all sicca patients and for pSS patients.

Fig. 2

ECC clusters patients with severe lymphocytic infiltration and B cell hyperactivity (A) Using unsupervised hierarchical clustering, three clusters of sicca patients are identified based on ECC counts (blue to yellow represents Z scores from −2 to +2 standard deviations). (B) Lymphocytic focus scores as diagnostically assessed (clinical) and of ECC-matched tissue are shown for all three clusters, with cluster 3 identifying patients with the highest LFS scores. (C) ECC-based clustering identifies patients with high B cell hyperactivity as indicated by percentages of IgG⁺ and IgM⁺ plasma cells as assessed by routine clinical practice by the pathologist and (D) by mRNA expression of the heavy chain constant region from IgG and IgA isotypes and IgM. Boxplot whiskers indicate the 5–95% interval. Medians are shown. Kruskal–Wallis test with Dunn’s multiple comparison test was used. *P < 0.05, **P < 0.01 and ***P < 0.001.

Fig. 3

Salivary gland Tfh cells quantified by ECC are increased in SS patients and correlate with LFS (A) Tfh frequencies as assessed by ECC are elevated in pSS and sSS as compared with nSS patients. (B) Tfh cell frequencies correlate with clinical LFS scores and (C) strongly correlate with LFS scores of tissue sections that were paired to tissue sections that were used for ECC. (D) Tfh cells also robustly correlated with CXCL13 RNA expression (DNA/RNA isolated from the same tissue sections). Medians are shown. Kruskal–Wallis test with Dunn’s multiple comparison test was used. **P < 0.01 and ***P < 0.001. Spearman correlation coefficients (r) are given for all sicca patients and for pSS patients.

Fig. 4

Salivary gland Tfh cells quantified by ECC correlate with B cell hyperactivity in SS patients Tfh cell frequencies correlate with (A) the percentage of IgG- and IgM-expressing plasma cells, (B) serum IgG levels and (C) the presence of autoantibodies anti-Ro/SSA (median is shown; Mann–Whitney U was used). (D) In addition, Tfh cell frequencies correlate with B cell frequencies as assessed by ECC and (E) are modestly associated with the presence of CD21⁺ FDCs containing lymphoid structures in sicca patients. *P < 0.05, **P < 0.01 and ***P < 0.001. Spearman correlation coefficients (r) are given for all sicca patients and for pSS patients.