Analysis of T cell repertoires of CD45RO CD4 T cells in cohorts of patients with bullous pemphigoid: A pilot study

Abstract

Autoimmune diseases develop over years - starting from a subclinical phenotype to clinically manifest autoimmune disease. The factors that drive this transition are ill-defined. To predict the turning point towards clinical disease and to intervene in the progress of autoimmune-mediated dysfunction, the establishment of new biomarkers is needed. Especially CD4 T cells are crucially involved in autoimmunity: first, during the initiation phase, because they lose their tolerance towards self-peptides, and second, by the subsequent ongoing presentation of self-peptides during the active autoimmune disease. Accordingly, changes in the degree of diversity of T cell receptor (TCR) repertoires in autoimmunity have been reported. These findings led to the hypothesis that transition from pre-disease to autoimmune disease is associated with an increase of abnormally expanded T cell clones that occupy large portions of the TCR repertoire. In this pilot study, we asked whether the ratio and the diversity of the TCR repertoires of circulating memory (CD45RO) and naïve (CD45RA) CD4 T cells could serve as a predictive factor for the development of autoimmunity. To find out, we analyzed the TCRβ repertoires of memory and naïve CD4 T cells in a small cohort of four gender- and age-matched elderly patients having the autoimmune blistering disease bullous pemphigoid or non-melanoma skin cancers. We found that the extent of clonal expansions in the TCRβ repertoires from the circulating memory and naïve CD4 populations did not differ between the patient groups. This result shows that the diversity of TCR repertoires from peripheral CD4 T cells does not reflect the manifestation of the skin-associated autoimmune disease BP and does not qualify as a prognostic factor. We propose that longitudinal TCR repertoire analysis of younger patients might be more informative.

Keywords: T cell receptor; autoimmunity; naive and memory CD4 T cells; next generation sequencing; repertoire diversity; skin diseases.

Figure 1

The number of memory T cells prevails in NMSC patients. PBMC from BP and NMSC patients were sorted for CD4 CD45RA and CD45RO T cells. (A) The gating strategy of one typical example is shown. Because CD45RO T cells were sorted by negative selection (middle panel), they appear as a specific population in contrast to the CD45RA population (right panel). (B) The size of the circulating CD4 T cell population is similar in BP and NMSC patients as shown as percentages of PBMC (Data are mean ± SD, p=0.43, not significant (n. s.) Mann-Whitney test). (C) NMSC patients harbor significantly more CD45RO T cells (Data are mean ± SD, ***p < 0.001, 2-way ANOVA with turkeys’ multiple comparison test). (D) The ratio of the CD45RO/RA population is significantly lower in BP patients. (Data are mean ± SD, ***p < 0.001, Mann-Whitney U test), (in B–D, n=14 (BP) and n=6 (NMSC)).

Figure 2

BP patients harbor fewer CD45RO clonotypes but the frequency of the high-abundant TCRβ clonotypes do not differ compared to the NMSC patients. (A) The numbers of the unique TCRβ clonotypes from the CD45RA (black symbols) and CD45RO (red symbols) CD4 T cell population are shown. The values obtained from the same patient are connected. (n = 4, *p<0.05, One-way ANOVA with turkeys’ multiple comparison test). (B) The ratio of the unique TCRβ clonotypes from the CD45RO/RA population was calculated. Data are mean ± SEM. No significant difference between BP and NMSC patients was found. (n = 4, Mann-Whitney test). (C) The clonal distribution of the TCRβ clonotypes is shown. TCRβ clonotypes were divided regarding their clonal abundance. Each bar represents the clonotypes of the CD45RA and CD45RO population from one patient. Each bar is divided into 4 sections. Each section represents the fraction of clonotypes that exist at a certain frequency (yellow: > 1 x 10^-5, bright gray: 1 x 10^-5 - 1 x 10^-4, gray: 1 x 10^-4 - 1 x 10^-3, dark gray: < 1 x 10^-3). (D) Inverse Simpson Index shows no difference between the BP or NMSC patients. (Data are mean ± SEM, gray bars: CD45RA, white bars: CD45RO no significance, 2-way ANOVA).

Figure 3

No difference in the degree of overlapping TCRβ clonotypes between BP and NMSC patients. (A) To assess the portion of overlapping TCRβ clonotypes between the CD45RO and CD45RA populations within each patient quantitatively, the Jaccard Index was calculated by either including all TCRβ clonotypes or including only the top 500 most abundant TCRβ clonotypes. (Data are mean ± SEM, ***p < 0.001, **p < 0.01, 2-way ANOVA with turkeys’ multiple comparison test). (B) The 20 most frequent TCRβ clonotypes present in the CD45RO population (black dots, left panel) are compared to their presence in the CD45RA population (red dots, left panel) and vice versa, the 20 most frequent TCRβ clonotypes present in the CD45RA population (black dots, right panel) are compared to their presence in the CD45RO population (red dots, right panel) within BP patient 1 (upper panel) or NMSC patient 1 (lower panel). One representative comparison out of all patients is shown.

Figure 4

A trend toward a reduced length of the CDR3 region in BP patients was observed but no difference in the V gene usage between the patient groups. (A) Relative frequencies of Vβ genes were compared between the CD45RO (red bars) and CD45RA (black bars) population in BP patients (1-4, left) and NMSC patients (1-4, right). White arrows indicate the high expression of the Vβ genes TRB20-1 in all samples. The high usage of TRBV20-1 has been described previously and was associated with aging (35, 36). (B) Relative length distribution of the CDR3 regions of CD45RO and CD45RA TCRβ clonotypes were compared between BP and NMSC patients (Data are means ± SD). The black arrow indicates the increased frequency of TCRβ clonotypes with a CDR3 length of 13 amino acids in BP patients (black, gray bars) compared to NMSC patients (bright and dark red bars).

Figure 5

The overlap between the skin-residing and the circulating TCRβ clonotypes is negligible. (A, B) To assess the percentage of overlapping TCRβ clonotypes between the blood-derived CD45ROand CD45RA populations and the skin-resident T cell clones of BP patient 4, the Jaccard Index was calculated and displayed as a heat map with a range from 1e-006 to 0.1 (A) or from 1e-006 to 0.0007 (B). (C) The top 100 most frequent TCRβ clonotypes present in the skin of BP4 (white dots, left panel) are compared to their presence in the CD45RA (black dots, left panel) or CD45RO population (red dots, left panel) of BP patient 4. Vice versa, the top 100 most frequent TCRβ clonotypes of in the CD45RA population (black dots, right upper panel) or the CD45RO population (red dots, right lower panel) are compared to their presence in the skin and could not be found. (D, E) Relative frequencies of Vβ genes (D) and Jβ genes (E) were compared between the skin (blue bar), the CD45RO (red bars), and CD45RA (black bars) populations of BP patient 4. The TRBV genes and TRVJ genes that are preferentially expressed in the skin are underlined. Indicated are means ± SEM. (F) Relative length distribution of the CDR3 regions of skin-residing TCRβ clonotypes were compared to the CD45RO and CD45RA populations of BP patients 4. (Data are means ± SEM). Differences can be observed at a length of 11, 13, 14, and 17 amino acids.