CD26 Expression on T Helper Populations and sCD26 Serum Levels in Patients with Rheumatoid Arthritis

Abstract

We studied dipeptidyl peptidase IV (DPP-IV, CD26) expression in different T helper cells and serum soluble DPP-IV/sCD26 levels in rheumatoid arthritis (RA) patients, correlated these with disease activity score (DAS), and examined how they were affected by different therapies, conventional or biological (anti-TNF, anti-CD20 and anti-IL6R or Ig-CTLA4). The percentage of CD4+CD45R0+CD26- cells was greatly reduced in patients (up to 50%) when compared with healthy subjects. Three other subsets of CD4 cells, including a CD26high Th1-associated population, changed variably with therapies. Data from these subsets (frequency and staining density) significantly correlated with DAS28 or DAS28 components but different in each group of patients undergoing the different therapies. Th17 and Th22 subsets were implicated in RA as independent CCR4+ and CCR4- populations each, with distinct CD26 expression, and were targeted with varying efficiency by each therapy. Serum DPP-IV activity rather than sCD26 levels was lower in RA patients compared to healthy donors. DPP-IV and sCD26 serum levels were found related to specific T cell subsets but not to disease activity. We conclude that, according to their CD26 expression, different cell subsets could serve to monitor RA course, and an uncharacterized T helper CD26- subset, not targeted by therapies, should be monitored for early diagnosis.

Fig 1. Major CD4⁺ T cell subsets defined by surface CD45R0 and CD26 expression.

A) Representative flow cytometry dot-plot showing lymphocytes gated physically. B) Representative dot-plot of gated CD4⁺ T lymphocytes. C) and D) Dot-plots showing the differential expression of CD45R0 and CD26 in the whole lymphocyte region gated in A (C) or in the CD3⁺CD4⁺ region gated in B (D). Naïve CD4 cells are mostly CD26⁺ in contrast to the CD8⁺ cells (compare C and D). From D) four major subsets were selected by their differential expression of CD45R0 and CD26: CD4⁺CD45R0^low/− CD26⁺ (naïve T cells); and effector/memory CD4⁺CD45R0⁺CD26^- (CD26 negative), CD4⁺CD45R0⁺CD26⁺ (intermediate) and CD4⁺CD45R0⁺CD26⁺⁺ (CD26high).

Fig 2. Effect of different therapies on CD4⁺ T cell subsets and cell surface CD26 levels.

Frequencies of different CD4⁺ T cell populations defined by CD45R0 and CD26 expression, and CD26 levels (MFI) in those cell populations, were determined by flow cytometry in RA patients undergoing different therapies. Confidence interval can be found in Table 1. Asterisks show statistically significant differences (p<0.05; Student’s t-test).

Fig 3. Correlation between disease activity and CD26⁺ naïve T cells.

DAS28 (Disease Activity Score 28) correlations with CD26 cell surface density of circulating CD4 CD45R0^- CD26⁺ naïve T lymphocytes in the whole cohort of patients (upper left panel), in the No BT group (patients without biological activity, upper right panel), and in two groups of patients with biological therapies (lower panels). Pearson’s coefficient (r) and significance (p) data are insert in the panels.

Fig 4. Correlations of CD4⁺ T cell subsets defined by surface CD45R0 and CD26 expression with non-DAS28 RA activity variables hemoglobin (Hb), hematocrit and the platelet count.

A) In the whole cohort of patients a low positive correlations seen between the percentages of CD45R0⁺CD26⁺ and CD45R0⁺CD26^- subsets with Hb (white bars) (numbers represent Pearson’s coefficient, r, asterisks for statistical significance, *p<0.05, ** p<0.005) or % erythrocytes (dark bars), whereas the CD45R0^-CD26⁺ subset showed a negative correlation and the CD26high subset did not correlate. B) Correlations of the four cell subsets in two patients’ groups. In the anti-TNF group some statistically significant correlations with Hb are shown for some T cell subsets (white bars). In the anti-CD20 group those cell populations show correlations with the platelet count (dark bars), generally in the opposite direction compared to the anti-TNF group. Non-significant correlations in the panels are shown only when the total cell number of the subset (instead of the percentage) was significant.

Fig 5. Serum dipeptidyl peptidase IV activity correlates with particular T cell populations in RA patients.

As an example, only correlations for the anti-CD20 therapy group are shown. A) Negative correlation between percentages of effector/memory T cells (CD4⁺CD45R0⁺CD26⁺) and DPP-IV enzymatic activity. B) Positive correlation between percentages of naïve T cells (CD4⁺CD45R0^-CD26⁺) and DPP-IV enzymatic activity. Pearson’s coefficient (r) and significance (p) data are insert in the panels.

Fig 6. T helper 17 and 22 cell subsets defined by surface CD4, CCR6 (CD196), CD161 and CCR4 (CD194).

A) Representative flow cytometry dot-plot showing lymphocyte gating strategy. B) Representative dot-plot showing that CCR6 is expressed mostly by CD4⁺ T lymphocytes. C) and D) Dot-plots showing the differential expression of CD161 and CCR4 in the whole lymphocyte region gated in A (C) or in the CD4⁺CCR6⁺ region gated in B (D). The upper left and right subsets of CD4⁺CCR6⁺CD161⁺, CCR4^- and CCR4⁺ respectively, were recorded for each patient as different Th17 subsets. E) and F) Dot-plots showing the differential expression of CCR10 and CCR4 in the whole lymphocyte region gated in A (C) or in the CD4⁺CCR6⁺ region gated in B (D). The upper left and right subsets of CD4⁺CCR6⁺CCR10⁺, CCR4^- and CCR4⁺ respectively, were recorded for each patient as different Th22 subsets.

Fig 7. Correlation between Th17 and Th22 and CD26 cell subsets.

A) In the whole cohort, the frequency of CD45R0⁺CD26^- subset correlates negatively with the Th17 CCR4^- cells but positively with the other Th17 and Th22 subsets (Th22CCR4^- did not correlate in the whole cohort but it did among different patients’ groups, data not shown). B) The number of Th17CCR4^- cells correlates with the percentage of CD26high cells (upper graph) and strongly with absolute cell number (lower graph). C) The cell surface CD26 (MFI) in the CD26high population positively correlates with the number of Th17CCR4^- cells and negatively with both the Th17CCR4⁺ and Th22CCR4⁺ cells (Th22CCR4⁺ did not correlate in the whole cohort but it did among different patients’ groups, data not shown). Numbers represent Pearson’s coefficient, r, (*p<0.05, ** p<0.005).