Somatic mutations in clonally expanded cytotoxic T lymphocytes in patients with newly diagnosed rheumatoid arthritis

Abstract

Somatic mutations contribute to tumorigenesis. Although these mutations occur in all proliferating cells, their accumulation under non-malignant conditions, such as in autoimmune disorders, has not been investigated. Here, we show that patients with newly diagnosed rheumatoid arthritis have expanded CD8+ T-cell clones; in 20% (5/25) of patients CD8+ T cells, but not CD4+ T cells, harbour somatic mutations. In healthy controls (n=20), only one mutation is identified in the CD8+ T-cell pool. Mutations exist exclusively in the expanded CD8+ effector-memory subset, persist during follow-up, and are predicted to change protein functions. Some of the mutated genes (SLAMF6, IRF1) have previously been associated with autoimmunity. RNA sequencing of mutation-harbouring cells shows signatures corresponding to cell proliferation. Our data provide evidence of accumulation of somatic mutations in expanded CD8+ T cells, which may have pathogenic significance for RA and other autoimmune diseases.

Figure 1. Clonal expansions of CD8+ T cells occur in RA patients and clonality increases with age.

(a) The study workflow, presented as a flow chart. (b) The V-gene usage of the largest flow-antibody-stained population correlated well with the corresponding V-gene usage in TCRB sequencing. The comparison was performed using the 61 patients for whom both flow cytometry and TCRB-sequencing data were available. Correlation was tested with Spearman correlation. (c) CD8+ clonality index calculated on the basis of TCRB sequencing was compared between patients (n=65) and healthy controls (HC, n=20) with the Mann–Whitney test, but there was no statistically significant difference. The horizontal line represents the group median, and error bars interquartile range (IQR). (d) The CD8+ TCRB-sequencing-based clonality index increases with age in RA patients (tested with Spearman correlation, n=65). (e) The percentage of patients and controls harbouring at least one CD8+ clone larger than 1%, 5%, 10%, and 20% of CD8+ is shown. (f) The V-gene usage in CD8+ clones comprising over 1% of all CD8+ in patients and controls is shown. The data were obtained via TCRB sequencing, and are presented as the number of clones using the denoted TCRBV gene adjusted for the number of individuals (RA=65, HC=20). (g) Odds ratios for CD8+ clones comprising over 1% of CD8+ for each TCRBV gene. The error bars represent 95% confidence intervals (CI), the arrow showing that the upper limit does not fit the scale. There were no differences between patients and controls (tested with Fisher’s test), except for TRBV07, which may be slightly underrepresented in patients compared with controls (Fisher’s test, P=0.00664, not adjusted for multiple comparisons, error bars 95% CI).

Figure 2. Somatic mutations are detected in 5 patients and in 1 healthy control.

(a) Somatic mutations were discovered in newly diagnosed RA patients and in one healthy (HC) by using the immunogene panel (presented on the left side), and exome sequencing (presented on the right side of the figure). Proliferation-associated and immune-related genes are highlighted with different colours. Percentages below the gene names denote the VAFs in CD8+ cells (immunogene panel) or expanded Vβ cells (exome). (b) The frequency of non-synonymous coding, nonsense-, frameshift-, and splice-site mutations shown as percentage of all identified mutations. Silent mutations were excluded from the bioinformatics analysis.

Figure 3. Somatic mutations occur in expanded CD8+ T-cell clones that persist during follow-up.

(a) The figure shows the CD8+ clonal architecture of patient 1, on the basis of TCRB sequencing and V-J genes. The Vβ13.1+ population detected by flow cytometry corresponded to clones using TCRBV06-05, 06-06, or 06-09 genes. (b) TCRB sequencing showed that patient 1 harboured two very large CD8+ T-cell clones. The two expanded clones of patient 1 persisted at a similar level during follow-up. The amino-acid TCRB sequences and the V genes of these unique clones are shown in the figure. (c) Amplicon sequencing of FACS-sorted cell fractions confirmed the identified mutations. The table presents the VAFs in each cell fraction. (d) The clonal architecture of patient 2's CD8+ pool as shown via V and J genes. Clones using TCRBV09-01 correspond to Vβ1+ antibody in flow cytometry. (e) Similarly presented TCRB sequencing results of flow-sorted Vβ1+ cells. When examining unique clones (‘unique’ defined by a unique nucleotide sequence) the largest clone composed 73% of all CD8+ Vβ1+ cells. (f) Patient 2 harboured several unique CD8+ T-cell clones at diagnosis. The TCRBV09-01 clone, in which mutations were observed, increased slightly during the follow-up. Amino-acid sequences derived from these unique clones are shown. The clone using TCRBV09-01 in this panel had the exact same nucleotide TCRB sequence as the largest clone in sorted Vβ1+ cells. (g) Amplicon sequencing results on FACS-sorted cells show the VAFs in each cell fraction. The low (<1%) VAFs found in CD4+ and Vβ13.6+ cells are considered sorting impurities, and thus the mutations in patient 2 occur exclusively in CD8+Vβ1+ cells. The mutation VAFs in CD8+Vβ1+ cells correspond well with the TCRBV09-01 clone size in sorted cells (h) The clonal landscape of CD8+ cells from patient 3. Clones using TCRBV04-03 gene correspond to Vβ7.2 usage in flow cytometry. (i) Patient 3 harboured a very large CD8+ T-cell clone at diagnosis, which persisted at a similar level during follow-up. (j) Amplicon sequencing of flow-sorted cell populations showed that the mutations detected in exome sequencing exist only in Vβ7.2 CD8+ T cells and not in other T cells.

Figure 4. Expression levels of mutated genes and a survival signature in mutation-harbouring cells.

CD8+ cells from 3 healthy controls as well as mutation-harbouring CD8+Vβ+ and polyclonal CD8+ T cells from patients 1, 2, and 3 were sorted with flow cytometry. (a) RNA sequencing was performed on the sorted fractions. The normalized mean expression levels calculated from the healthy CD8+ T cells and from the patients’ polyclonal CD8+ fractions are presented in the control expression column. In the following columns, each sorted population containing the mutated clone is presented individually. (b) The upper panel presents proliferation/survival -associated genes identified by paired analysis of the mutation-harbouring cell populations and their polyclonal background populations. Immunologically important transcripts are presented in the lower panel. The transcripts were filtered with threshold expression level 0, FC>1,5 or <−1,5 and P<0.01. Abbreviations: Aa, amino acid; fs, frameshift; *, stop-codon gained; -, splice-site acceptor mutation. The function of the gene was retrieved from the RefSeq database and noll-hypothesis was tested using ANOVA in the Qlucore software package.

Figure 5. Expanded CD8+ T-cell clones are effector-memory T cells.

Multicolour-flow cytometry was used to determine the phenotype of patients’ T cells in samples taken during the follow-up. CCR7 and CD45RA were used to determine whether the cells were central memory (CM), naive, effector-memory (EM) or terminal effector-memory RA-positive (TEMRA) cells. (a) The CD8+ Vβ13.1+ cells of patient 1 comprise 50% of all CD8+ T cells. The phenotype of CD8+Vβ13.1+, other CD8+, and CD4+ T cells are displayed in separate plots. (b) The phenotype of CD8+Vβ7.2+, other CD8+, and CD4+ T cells of patient 3 are displayed in separate plots. (c) The phenotype of expanded CD8+Vβ1+, other CD8+, and CD4+ T cells of patient 2 are presented in separate plots.

Figure 6. Hypothesis for the accumulation of somatic mutations.

Mature T cells undergo thymic selection, and during the expansion phase after antigen stimulation, mutation(s) occur (1st mutation). When the cell population continues to grow, additional mutations may arise in the same clone, thereby resulting in different VAFs in the same clonal population. In our patient cohort, CD4+ cells presented a more equally distributed T-cell repertoire than did CD8+ T cells (analysed with flow cytometry), and we did not find any somatic variants in CD4+ cells. APC, antigen-presenting cell; DN, double negative; DP, double positive; SP, single positive; MHC, major histocompatibility complex.