Induced Human Regulatory T Cells Express the Glucagon-like Peptide-1 Receptor

Abstract

The glucagon-like peptide-1 receptor (GLP-1R) plays a key role in metabolism and is an important therapeutic target in diabetes and obesity. Recent studies in experimental animals have shown that certain subsets of T cells express functional GLP-1R, indicating an immune regulatory role of GLP-1. In contrast, less is known about the expression and function of the GLP-1R in human T cells. Here, we provide evidence that activated human T cells express GLP-1R. The expressed GLP-1R was functional, as stimulation with a GLP-1R agonist triggered an increase in intracellular cAMP, which was abrogated by a GLP-1R antagonist. Analysis of CD4⁺ T cells activated under T helper (Th) 1, Th2, Th17 and regulatory T (Treg) cell differentiation conditions indicated that GLP-1R expression was most pronounced in induced Treg (iTreg) cells. Through multimodal single-cell CITE- and TCR-sequencing, we detected GLP-1R expression in 29-34% of the FoxP3⁺CD25⁺CD127^- iTreg cells. GLP-1R⁺ cells showed no difference in their TCR-gene usage nor CDR3 lengths. Finally, we demonstrated the presence of GLP-1R⁺CD4⁺ T cells in skin from patients with allergic contact dermatitis. Taken together, the present data demonstrate that T cell activation triggers the expression of functional GLP-1R in human CD4⁺ T cells. Given the high induction of GLP-1R in human iTreg cells, we hypothesize that GLP-1R⁺ iTreg cells play a key role in the anti-inflammatory effects ascribed to GLP-1R agonists in humans.

Keywords: CD4+ T cells; GLP-1R expression; human.

Figure 1

Activated CD4⁺ T cells express functional GLP-1R. (A) Relative GLP-1R expression in naïve CD4⁺ T cells (0 h) or CD4⁺ T cells activated for 24, 48, or 72 h. Data are normalized to GLP-1R expression in naïve CD4⁺ T cells (0 h) (mean + SEM, n = 6, one-way ANOVA with post-hoc test (Tukey’s), *** p ≤ 0.0005). (B) Relative GLP-1R expression in CD4⁺ T cells activated for 24, 48, or 72 h in Th1 or Th2 polarizing medium in the absence or presence of 100 nM 25(OH)D₃ as indicated. Data are normalized to GLP-1R expression in naïve CD4⁺ T cells (mean + SEM, n ≥ 6, one-way ANOVA with post-hoc test (Tukey’s) for each time group, * p < 0.05, ** p < 0.005, *** p < 0.0005). (C) cAMP in CD4⁺ T cells activated for 60 h in Th2 polarizing medium plus 25(OH)D₃ (100 nM) and subsequently not treated or treated with exendin-4 (10 nM) for 30 min as indicated (estimation plot, n = 6, Student’s t-test, (paired, two-tailed), * p < 0.05). (D,E) cAMP accumulation in CD4⁺ T cells activated for 60 h in (D) Th1 or (E) Th2 medium in the absence or presence of 100 nM 25(OH)D₃ and treated with increasing amounts of exendin-4 as indicated. Naïve, unstimulated CD4⁺ T cells were included as controls. Representative graphs from three independent experiments with six donors run in duplicates (mean ± SEM). (F) EC₅₀ values from the nonlinear fits (mean ± SEM, n ≥ 4, one-way ANOVA with post-hoc test (Tukey’s), **** p < 0.00005).

Figure 2

GLP-1R expression directly correlates with the percentage of CD25⁺FoxP3⁺ T cells. (A) Relative GLP-1R expression, (B) percentage of CD25⁺FoxP3⁺ T cells and (C) representative flow cytometric analysis of CD4⁺ T cells activated for 120 h in the indicated polarization medium in the absence or presence of 25(OH)D₃ (100 nM). GLP-1R expression was normalized to GLP-1R expression in CD4⁺ T cells activated for 120 h in Th0 medium in the absence of 25(OH)D₃ (mean + SEM, n ≥ 4, one-way ANOVA with post-hoc test (Dunnett’s), ** p < 0.005, **** p < 0.00005. The mean of each column was compared with the mean of the column for Treg minus vitamin D). Data presented in (B,C) were obtained by gating on live, single lymphocytes as shown in Figure S2. The percentage of CD25^hiFoxP3⁺CD4⁺ T cells correlated with the expression level of the GLP-1R in CD4⁺ T cells activated for 120 h in the indicated differentiation media in both the absence (D) and presence (E) of 25(OH)D₃ (100 nM) as determined by linear regression analysis.

Figure 3

The combination of TGFβ, rapamycin and retinoic acid induces expression of FoxP3 and the GLP-1R in parallel in T cells. Relative FoxP3 (A) and GLP-1R (B) expression in CD4⁺ T cells activated for 120 h in the presence of the indicated components of the Treg polarizing medium in the absence or presence of 25(OH)D₃ (100 nM) (RA = retinoic acid, Rapa = Rapamycin). Data were normalized to FoxP3 (A) and GLP-1R (B) expression in CD4⁺ T cells activated in the absence of any of the components (mean + SEM, n = 4, one-way ANOVA with post-hoc test (Dunnett’s), * p < 0.05, ** p < 0.005, **** p < 0.00005. The mean of each column was compared with the mean of the first column).

Figure 4

Approximately 30% of iTreg cells express the GLP-1R. (A) Visualization of scRNA-seq from CD4⁺ T cells activated in Th1, Th17 or Treg cell-polarizing medium using uniform manifold approximation and projection (UMAP). Cells from three donors—Treg_D1, Treg_D2 and Treg_D3—were included for Treg cells. (B,C) The iTreg phenotype was confirmed by expression of (B) FOXP3 and IRF4 as well as high expression of (C) IL2RA (CD25) and low expression of IL7R (CD127) and DPP4 (CD26) at the mRNA (top) and protein (bottom) level using CITE-seq. mRNA expression is shown at the log-normalized UMI count; protein expression was de-noised and normalized using the ‘dsb’ R package. (D) Alignment coverage of cDNA reads from the single-cell RNA-seq transcriptome library across the 5′ end of the GLP1R locus. Expression of GLP-1R after targeted amplification shown as (E) quantification of unique GLP-1R transcripts across donors, (F) distribution of GLP-1R⁺ cells across UMAP-space, and (G) percentage of cells co-expressing targeted FOXP3 transcripts including percentage of cells within each quadrant shown in the corners.

Figure 5

GLP-1RA inhibit proliferation of CD4⁺ and CD8⁺ T cells and increase PD-1 and PD-L1 expression in iTreg cells. (A–D) Proliferation indexes of CD4⁺ (A) and CD8⁺ (B) T cells, live cell frequency (C) and IFNγ production (D) from PBMC cultures stimulated with anti-CD3 mAb for 120 h in the presence of the indicated concentration of the GLP-1RA liraglutide (Lira) or forskolin (FSK) (mean + SEM, n = 6, one-way ANOVA with post-hoc test (Dunnett’s), * p < 0.05, **** p < 0.00005. The mean of each column was compared with the mean of the first column). (E,F) PD-1 (E) and PD-L1 (F) expression in CD4⁺ T cells activated for 120 h in Treg polarizing medium and subsequently not treated or treated with exendin-4 (10 nM) for 24 h, as indicated (estimation plot, n = 6, Student’s t-test, (paired, two-tailed), * p < 0.05, ** p < 0.005).

Figure 6

GLP-1R⁺CD4⁺ T cells are found in the skin. First quadrant GLP-1R (red), second quadrant CD4 (green), third quadrant DAPI (grey) and fourth quadrant merged stained fluorescent microscopy images of vehicle-exposed (Vehicle) and nickel-exposed (Nickel) skin from patients with allergic contact dermatitis to nickel. Representative images from three patients (P1, P2, P3).