Transient increase of activated regulatory T cells early after kidney transplantation

Abstract

Regulatory T cells (Tregs) are crucial in controlling allospecific immune responses. However, studies in human kidney recipients regarding the contribution of polyspecific Tregs have provided differing results and studies on alloreactive Tregs are missing completely. In this retrospective study, we specifically analyzed activated CD4⁺CD25^highFOXP3⁺GARP⁺ Tregs in 17 patients of a living donor kidney transplantation cohort longitudinally over 24 months by flow cytometry (FOXP3: forkhead box protein 3, GARP: glycoprotein A repetitions predominant). We could demonstrate that Tregs of patients with end-stage renal disease (ESRD) are already pre-activated when compared to healthy controls. Furthermore, even though total CD4⁺CD25^highFOXP3⁺ Treg numbers decreased in the first three months after transplantation, frequency of activated Tregs increased significantly representing up to 40% of all peripheral Tregs. In a cohort of living donor kidney transplantation recipients with stable graft function, frequencies of activated Tregs did not correlate with the occurrence of acute cellular rejection or chronic graft dysfunction. Our results will be important for clinical trials using adoptive Treg therapy after kidney transplantation. Adoptively transferred Tregs could be important to compensate the Treg loss at month 3, while they have to compete within the Treg niche with a large number of activated Tregs.

Figure 1

Frequency of alloreactive Tregs after allogeneic stimulation. (a) PBMC of seven healthy volunteers (HC1–HC7) were stimulated with PBMC of five different PBMC donors. Background activation was determined in unstimulated PBMC of each healthy volunteer (unstimulated, black dots). Donor PBMCs were identified by CFSE positive staining and further excluded. Recipient’s PBMC were gated on CFSE^-CD4⁺CD25^high T cells. Allogeneically activated Tregs were further identified by their expression of FOXP3 and GARP (for detailed gating strategy see Supplementary Fig. 1). Frequency of allogeneically activated Tregs was expressed as percentage of all Tregs by calculating the ratio of CD4⁺CD25^highFOXP3⁺GARP⁺ (activated Tregs) to CD4⁺CD25^highFOXP3⁺ (total Tregs). (b) Representative dot plots of two different healthy individuals (HC5 and HC7) after allogeneic stimulation, populations are gated on CD4⁺CD25^high T cells. Activated Tregs are defined by their co-expression of FOXP3 and GARP (upper right quadrant). Left column shows unstimulated PBMC, middle and right panel show activated Tregs after allogeneic stimulation with two different allogeneic stimuli.

Figure 2

Increased Treg frequency and activation in patients with end-stage renal disease. (a) Frequency of CD4⁺CD25^high FOXP3⁺ polyspecific regulatory T cells was assessed in 13 patients with ESRD and seven healthy controls (HC) by flow cytometry. Frequency of polyspecific Tregs was significantly increased in ESRD patients (HC 3.2 ± 0,9% vs. ESRD 7.5 ± 3.4%, p = 0.005). (b) Frequency of endogenously activated Tregs in ESRD patients was compared to healthy controls. Activated Tregs were identified as CD4⁺CD25^highFOXP3⁺GARP⁺. Percentage of activated Tregs of all Tregs was calculated by the ratio of GARP⁺ Tregs of all Tregs. Patients with ESDR displayed a significantly higher activation status of Tregs than healthy controls (HC 13.1 ± 0,3.5 vs. HD 21.3 ± 7.0%, p = 0.01). Data are presented as mean ± SD.

Figure 3

Specificity of activated Tregs. (a,b) Two representative recall experiments are shown. After thawing, PBMC of kidney recipients were rested overnight in culture medium before expression of GARP in Tregs was analyzed by flow cytometry the next day (d1, left column). After five days rest samples were reanalyzed for downregulation of GARP expression (d5, middle column). At day 5 Tregs were restimulated with donor-antigen (restimulation, right column). (a) Represents an example of sufficient downregulation of GARP to baseline levels and adequate increase of activation marker GARP as donor-specific immune response. (b) Represents an example of insufficient downregulation and strong expression of activation marker GARP, which is independent from the restimulation with donor-antigen. T cells were gated on CD4⁺CD25^high, allogeneically activated Tregs were identified by their co-expression of FOXP3 and GARP. Frequency of allogeneically activated Tregs was expressed as percentage of all Tregs in analogy to Fig. 1. (c) Three representative analysis of allogeneic activation of Tregs from renal transplant recipients. The majority of the samples was susceptible to allogeneic stimulation after after five day rest in culture medium as described in section 3A (left and middle plot), whereas few samples remained unresponsive to further stimulation (right plot).

Figure 4

Activated regulatory T cell increase after kidney transplantation. PBMCs of renal transplant recipients collected pre-Tx and at month 3, 12, and 24 post-Tx were stimulated with their corresponding donor PBMC for 24 hours in vitro and subsequently stained for CD4, CD25, FOXP3 and GARP. (a) Percentage of CD4⁺CD25^highFOXP3⁺ polyspecific regulatory T cells dropped significantly from pre-Tx to 3 months after transplantation (8.9 ± 2.4% vs. 2.2 ± 1.8%, p < 0.0001) and recovered until 12 months post-Tx (6.8 ± 2.6%, p = 0.001). Healthy controls (HC) and patients on chronic hemodialysis (HD) served as controls. (b) Frequency of endogenously activated regulatory T cells was determined in unstimulated PBMC samples. After thawing, PBMC were rested for 5 days before analysis by flow cytometry was performed. Percentage of GARP⁺ Tregs of total CD4⁺CD25^highFOXP3⁺ regulatory T cells was calculated. Data are presented as mean ± SD. (c) After in vitro stimulation with donor-Ag increase of frequency of activated Tregs was noted in the early post-Tx course (pre-Tx 28.7 ± 9.3% vs 3 months 42.2 ± 17.1%, p = 0.043) and declined to pre-Tx levels until 12 months post-Tx (23 ± 9.1%, p = 0.006). From 12 to 24 months activation level of Tregs stayed stable (25.6 ± 7.3%, p = 0.66). (d) Baseline frequency of activated Tregs in unstimulated, rested samples (unstim) and after allogeneic stimulation with donor PBMC (+donor-Ag) in the longitudinal post-Tx course.

Figure 5

Frequency of alloreactive Tregs does not predict allograft tolerance Frequency of activated Tregs at time point 3 months of non-rejectors and rejectors are plotted in this panel. No significant differences have been observed when comparing frequency of activated Tregs in unstimulated or allogeneically (against donor and 3^rd party) activated Tregs in the respective groups. Data are presented as mean ± SD.

Figure 6

Frequency of alloreactive Tregs does not correlate with IFTA. Frequency of activated Tregs at time point 3 months were correlated with the occurrence of IFTA at six weeks, 3 months or 6 months. No significant differences have been observed when comparing frequency of activated Tregs in unstimulated or allogeneically (against donor and 3^rd party) activated Tregs at any time point of the occurrence of IFTA. Data are presented as mean ± SD.

Figure 7

Frequency of alloreactive Tregs does not correlate with allograft function. (a) Overall benign clinical course of renal graft function during the observation period of 24 months is depicted. Of 17 patients, 12 patients showed a stable or improving graft function. (b,c) Frequency of activated Tregs at time point 3 months of patients with stable and worsening graft function are plotted in this panel. No significant differences have been observed when comparing frequency of activated Tregs in unstimulated or allogeneically (against donor and 3^rd party) activated Tregs in the respective groups. Data are presented as mean ± SD.