Depletion of polyfunctional CD26highCD8+ T cells repertoire in chronic lymphocytic leukemia

Abstract

Background: CD8⁺ T cells play an essential role against tumors but the role of human CD8⁺CD26⁺ T cell subset against tumors, in particular, haematological cancers such as chronic lymphocytic leukemia (CLL) remains unknown. Although CD4⁺CD26^high T cells are considered for adoptive cancer immunotherapy, the role of CD8⁺CD26⁺ T cells is ill-defined. Therefore, further studies are required to better determine the role of CD8⁺CD26⁺ T cells in solid tumors and haematological cancers.

Methods: We studied 55 CLL and 44 age-sex-matched healthy controls (HCs). The expression of CD26 on different T cell subsets (e.g. naïve, memory, effector, and etc.) was analyzed. Also, functional properties of CD8⁺CD26⁺ and CD8⁺CD26^- T cells were evaluated. Finally, the plasma cytokine/chemokine and Galectin-9 (Gal-9) levels were examined.

Results: CD26 expression identifies three CD8⁺ T cell subsets with distinct immunological properties. While CD26^negCD8⁺ T cells are mainly transitional, effector memory and effectors, CD26^lowCD8⁺ T cells are mainly naïve, stem cell, and central memory but CD26^high T cells are differentiated to transitional and effector memory. CD26⁺CD8⁺ T cells are significantly reduced in CLL patients versus HCs. CD26^high cells are enriched with Mucosal Associated Invariant T (MAIT) cells co-expressing CD161TVα7.2 and IL-18Rα. Also, CD26^high cells have a rich chemokine receptor profile (e.g. CCR5 and CCR6), profound cytokine (TNF-α, IFN-γ, and IL-2), and cytolytic molecules (Granzyme B, K, and perforin) expression upon stimulation. CD26^high and CD26^low T cells exhibit significantly lower frequencies of CD160, 2B4, TIGIT, ICOS, CD39, and PD-1 but higher levels of CD27, CD28, and CD73 versus CD26^neg cells. To understand the mechanism linked to CD26^high depletion, we found that malignant B cells by shedding Galectin-9 (Gal-9) contribute to the elevation of plasma Gal-9 in CLL patients. In turn, Gal-9 and the inflammatory milieu (IL-18, IL-12, and IL-15) in CLL patients contribute to increased apoptosis of CD26^high T cells.

Conclusions: Our results demonstrate that CD26⁺ T cells possess a natural polyfunctionality to traffic and exhibit effector functions and resist exhaustion. Therefore, they can be proposed for adoptive cancer immunotherapy. Finally, neutralizing and/or inhibiting Gal-9 may preserve CD26^highCD8⁺ T cells in CLL.

Keywords: Adoptive T cell therapy; Galectin-9; MAIT cells; T cells exhaustion.

Fig. 1

CD26^low and CD26^high CD8⁺ T cells are reduced in CLL. A Representative flow cytometry plots, and (B) Cumulative data comparing the frequency of CD26⁺CD8⁺ T cells in PBMCs from HCs (n = 44) and CLL (n = 55) patients. C Cumulative data comparing the frequency of CD26^low and CD26^high CD8⁺ T cells in HC and CLL. D Bar plots illustrating the proportion of CD26^neg, CD26^low and CD26^highCD8⁺ T cells in HC and CLL (E) Representative histogram plots, and (F) Cumulative data comparing the Mean Fluorescence Intensity (MFI) of CD26 in CD8⁺ T cells in HCs and CLLs. G The pie charts represent the median frequency of CD26^neg/low/high in different subsets of CD8⁺ T cells (e.g. Naïve, Stem cell memory, Central memory, Transitional memory, Effector memory, and Effector) in HCs versus CLLs. H Representative flow cytometry plots of CD26^low/high in different CD8⁺ T cell subsets of HCs (black) and CLLs (red). I Cumulative data of CD26^low/high in naïve, (J) stem cell memory, (K) central memory, (L) transitional memory, (M) effector memory, and (N) effector CD8⁺ T cells in HC and CLL. Statistics are assessed by Mann–Whitney or the Kruskal–Wallis multiple comparison tests. P-value < 0.05 was considered as significant. Error bars represent the median with an interquartile range. Each dot represents an individual human sample

Fig. 2

Phenotypic profile of CD26^neg/low/high CD8⁺ T cells in CLL. A Representative flow plots, and (B) cumulative data showing the frequency of CD161⁺ TV $\alpha$ 7.2⁺ in CD26^neg/low/high CD8⁺ T cell subsets in CLL. C Cumulative data of the frequency CD161/ TV $\alpha$ 7.2 co-expressing cells in CD26^highCD8⁺ T cells in CLL. D Representative flow plots, and (E) cumulative data of the frequency of IL-18R $\alpha$ expressing cells among CD26^neg/low/high CD8⁺ T cells in CLL. F Cumulative data of the frequency of IL-18R $\alpha$ ^high expressing cells among CD26^neg/low/high CD8⁺ T cells in CLL. G Representative plots, and (H) the frequency of CD160⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. I Representative plots, and (J) cumulative data of the frequency of 2B4⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. K Representative plots, and (L) cumulative data of the frequency of TIGIT⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. M Representative plots, and (N) cumulative data of the frequency of ICOS⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. O Representative plots, and (P) cumulative data of the frequency of CD28⁺ among CD26^neg/low/high CD8⁺ T cells in CLL. Q Representative plots, and (R) cumulative data of the frequency of CD27⁺ cells among CD26^neg/low/high among CD8⁺ T cells in CLL. S Representative plots, and (T) cumulative data of the frequency of PD-1⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. U Representative plots, and (V) cumulative data of the frequency of CD39⁺ cells among CD26^neg/low/high among CD8⁺ T cells in CLL. W Representative plots, and X cumulative data of the frequency of CD73⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. Y Cumulative data showing the frequency of co-inhibitory/co-stimulatory expressing cells in CD26^lowCD8⁺ T cells in HC and CLL. Error bars represent the median with an interquartile range. Each dot represents an individual human sample. Florescence minus one (FMO)

Fig. 3

Cytotoxic properties of CD26^neg/low/high CD8⁺ T cells in CLL. A Representative flow plot of the gating strategy for CD26^neg, CD26^low, and CD26^high CD8⁺ T cells. B Representative flow plots of the co-expression of Granzyme-B (GzmB)/Perforin in CD26^neg/low/high CD8⁺ T cells in CLL. C Cumulative data showing the frequency of GzmB⁺, (D) Perforin⁺, and (E) GzmB⁺Perforin⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. F Representative flow plots, and (G) cumulative data of the frequency of GzmB⁺GzmK⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL. H Representative flow plots, and (I) cumulative data of the frequency of CD107a⁺ cells among CD26^neg/low/high CD8⁺ T cells in CLL either unstimulated (unstim) or stimulated (stim) with anti-CD3/CD28 (3 g$\mathrm{\mu }$/ml, 1 g$\mathrm{\mu }$/ml) in the presence of protein transport inhibitor (1/1000). J Representative flow plots, and cumulative data of the frequency of (K) GzmB⁺ (L) GzmK⁺, and (M) GzmB⁺GzmK⁺ cells among CD26^neg/low/high CCR7⁻CD8⁺ T cells in CLL either unstimulated or stimulated with anti-CD3/CD28(3 g$\mathrm{\mu }$/ml, 1 g$\mathrm{\mu }$/ml), and a cocktail of IL-18 + IL-12 + IL-15 (100 ng/ml of each). Error bars represent median with interquartile range. Each dot represents an individual human sample

Fig. 4

Cytokine production and proliferation ability of CD26^neg/low/high CD8⁺ T cells in CLL. A Representative flow plots, (B) the frequency of TNF-$\alpha$, (C) IFN-$\gamma ,$ and (D) TNF-$\alpha$⁺IFN-$\gamma$⁺ cells among CD26^neg/low/high CCR7⁻CD8⁺ T cells. E Representative plots, (F) cumulative data showing the frequency of TNF-$\alpha$⁺, (G) IFN-$\gamma$⁺, and (H) TNF-$\alpha$⁺IFN-$\gamma$⁺ among CD26^neg/low/high CCR7⁻CD8⁺ T cells in unstimulated or stimulated with anti-CD3/CD28 (3 g$\mathrm{\mu }$/ml, 1 g$\mathrm{\mu }$/ml) and a cocktail of IL-18 + IL-12 + IL-15 (100 ng/ml of each). (I) Representative flow plots and, (J) cumulative data of the frequency of IL-2 expressing cells among CD26^neg/low/high CD8⁺ T cells in unstimulated (black color) versus 5 h after in-vitro stimulation with PMA/ionomycin cocktail (Biolegend, 2 ng/ml) in the presence of Brefeldin A (1 g$\mu$/ml). (K) Representative flow plots and, (L) cumulative data of the frequency of CFSE^low (proliferated) cells among CD26^neg/low/high CCR7⁻CD8⁺ T cells, unstimulated (black color) versus 72 h stimulation. (M) Representative flow plots, and (N) cumulative data showing the MFI for ROR $\gamma \delta$ in CD26^neg, CD26^low, and CD26^high CD8⁺ T cells in CLL. Error bars represent the median with an interquartile range. Each dot represents an individual human sample

Fig. 5

CD26⁺CD8⁺ T cells exhibit a greater migratory capacity in CLL. A Representative plots of the frequency of CCR5⁺CD8⁺ T cells among CD26^neg/low/high subsets. B The graphical illustrates the tendency of CCR5⁺CD8⁺ T cells homing to secondary lymphoid organs or inflamed tissues in response to CCL3, MIP-1 $\alpha$/$\beta$, and RANTES. C Cumulative data of the frequency of CCR5⁺CD8⁺ T cells among CD26^neg/low/high CD8 + T cells in CLL. D Representative flow plots of the frequency CCR6⁺CD8⁺ T cells among CD26^neg/low/high subsets. E The graphical illustrates the tendency of CCR6⁺CD8⁺ T cells homing to colon, and mucosal tissues in response to CCL20. F Cumulative data of the frequency of CCR6⁺CD8⁺ T cells among CD26^neg/low/high. G Representative plots of the frequency of CCR7⁺CD8⁺ T cells among CD26^neg/low/high subsets of CD8⁺ T cells. H The graphical illustrates the tendency of CCR7⁺CD8⁺ T cells homing to lymph nodes in response to CCL19. I Cumulative data of the frequency of CCR7⁺CD8⁺ T cells among CD26^neg/low/high subsets. J Representative plots of the frequency of Cutaneous Lymphocyte Antigen (CLA)⁺CD8⁺ T cells among CD26^neg/low/high cells. K The graphical illustrates the tendency of CLA⁺CD8⁺ T cells homing to skin following E/P Selectin binding on endothelial cells. L Cumulative data of the frequency of CLA⁺CD8⁺ T cells among CD26^neg/low/high cells. M Representative plots, and the frequency of migrated CD26^neg, CD26^low, and CD26^high CD8⁺ T cells at the baseline and after 18 h in response to (N) Fetal Bovine Serum (FBS-10%), (O) RANTES (10 nM), and (P) IL-18 (100 ng/ml). Error bars represent the median with an interquartile range. Each dot represents an individual human sample.

Fig. 6

Gal-9 preferentially promotes CD26^high CD8⁺ T cells apoptosis in CLL. A Representative plots, and (B) cumulative data of the frequency of CD127⁺/KLRG1⁻ cells among CD26^neg/low/high subsets of CD8⁺ T cells. C Representative plots, and (D) cumulative data of the intensity of Annexin-V expression (MFI) among CD26^neg/low/high CD8⁺ T cells. E Representative plots, and (F) cumulative data of the MFI of CD57⁺ among CD26^neg/low/high CD8⁺ T cells. G Cumulative data of the plasma Gal-9 concentrations in HCs and CLL patients. H Representative plots, and (I) cumulative data of the frequency of CD26^highCD8⁺ T cells following stimulation with anti-CD3/CD28 in the presence or absence of recombinant human Gal-9 (0.02 g$\mu$/ml). J Representative plots, and (K) cumulative data of the intensity of Annexin-V expression in CD26^high CD8⁺ T cells following stimulation with anti-CD3/CD28 in the presence or absence of Gal-9. L Cumulative data of Gal-9 concentration in the supernatants of isolated non-B cells versus malignant B cells (B-CLL) after 12 h culture. M Representative plots, and (N) cumulative data ofthe intracytoplasmic MFI of Gal-9 in B cells from HC versus B-CLLs. O Representative plots, and (P) cumulative data of the frequency of Annexin-V expressing CD8⁺ T cells in CD26^high cells at the baseline versus stimulation with a cytokine cocktail (IL-18 + IL-12 + IL-15) (100 ng/ml of each) for 18 h. Error bars represent the median with an interquartile range. Each dot represents an individual human sample

Fig. 7

Visual summary. A CD26 expression defines three distinct populations of CD8⁺ T cells in CLL with discrete properties. The table summarizes different properties of CD26^neg, CD26^low, and CD26^high CD8⁺ T cells ranked as high (∎∎∎) moderate (∎∎), and low (∎) B The illustration depicts the proposed mechanism of decreased frequency of CD26^low and CD26^high subsets and the expansion of CD26^neg CD8⁺ T cells in CLL. Migration towards inflamed tissues, elimination by apoptosis (Gal-9, Inflammatory cytokines), change of character/exhaustion due to chronic antigenic stimulation in CLL are proposed as potential mechanisms leading to the depletion of CD26⁺CD8⁺ in CLL