Regulatory T cell expression of CLA or α₄β₇ and skin or gut acute GVHD outcomes

Abstract

Regulatory T cells (Tregs) are a suppressive subset of CD4⁺ T lymphocytes implicated in the prevention of acute GVHD (aGVHD) after allo-SCT (ASCT). To determine whether increased frequency of Tregs with a skin-homing (cutaneous lymphocyte Ag, CLA⁺) or a gut-homing (α₄β₇⁺) phenotype is associated with reduced risk of skin or gut aGVHD, respectively, we quantified circulating CLA⁺ or α₄β₇⁺ on Tregs at the time of neutrophil engraftment in 43 patients undergoing ASCT. Increased CLA⁺ Tregs at engraftment was associated with the prevention of skin aGVHD (2.6 vs 1.7%; P=0.038 (no skin aGVHD vs skin aGVHD)), and increased frequencies of CLA⁺ and α₄β₇⁺ Tregs were negatively correlated with severity of skin aGVHD (odds ratio (OR), 0.67; 95% confidence interval (CI), 0.46–0.98; P=0.041) or gut aGVHD (OR, 0.93; 95% CI, 0.88–0.99; P=0.031), respectively. This initial report suggests that Treg tissue-homing subsets help to regulate organ-specific risk and severity of aGVHD after human ASCT. These results need to be validated in a larger, multicenter cohort.

Introduction

Acute GVHD (aGVHD) occurs unpredictably after allo-SCT (ASCT), affects over 50% of patients during the first 100 days of transplant and is directly or indirectly responsible for about 50% of ASCT mortality.^1,2 Previous studies have suggested that variability in T-cell subsets,^3–7 lymphocyte trafficking^8–10 and early immune interactions between donor lymphocytes with recipient APCs^11,12 could help explain the clinical heterogeneity of aGVHD.

Regulatory T cells (Tregs) are a naturally occurring subset of suppressive CD4⁺ T lymphocytes involved in the prevention of aGVHD in murine models of transplantation and in human ASCT.^3–7 Tregs are characterized by the expression of IL-2 receptor α chain, CD25 and a unique forkhead/winged helix transcription factor, Foxp3.^13–15 The exact location where Tregs exert their suppressive effect after ASCT is still unclear. In human ASCT, skin and gut biopsy data suggest that decreased localization of Foxp3⁺ Tregs in target tissues is associated with increased severity of aGVHD at those sites.^16,17 Furthermore, the stereotypical involvement of specific organs by aGVHD suggests that lymphocyte trafficking is crucial to the pathophysiology of aGVHD. In addition, Tregs are thought to maintain immune homeostasis through various contact-mediated mechanisms and through the expression of distinct adhesion molecules including the skin-homing marker cutaneous lymphocyte Ag (CLA) and the gut-homing integrin, α₄β₇.^18–20 Therefore, we hypothesized that after ASCT, the ability of Tregs to migrate to cutaneous or gut tissues through the expression of CLA or α₄β₇ could be important in the prevention of skin or gut aGVHD, respectively.

Patients and methods

Study design

Patients with hematological malignancies undergoing myeloablative or reduced intensity conditioning followed by a T-cell-replete transplant were enrolled after obtaining informed consent. Individuals receiving in vivo T-cell depletion with thymoglobulin were excluded. All preparative regimens were a part of a standard-of-care or institutional review board-approved protocol. GVHD prophylaxis consisted of a calcineurin inhibitor and either MTX or mycophenolate mofetil. Sirolimus was not used in any of the aGVHD prophylaxis regimens. aGVHD was diagnosed clinically and confirmed by biopsy when feasible. Biopsy data and clinical features of aGVHD including patterns of skin/gut involvement, severity and recurrence rates were assessed weekly for the first 100 days after ASCT. Recurrent aGVHD was defined as any increase in symptoms or therapy for aGVHD after initial response or during steroid taper. Recurrent aGVHD was analyzed only for the first 100 days of transplant. The severity of aGVHD was determined by the overall grade (0–IV) and the individual organ stage (0–4), using standard guidelines, as outlined by Glucksberg and colleagues.^21,22 All staging/grading of aGVHD was performed prospectively by a single individual (MJ) who was blinded to the results of the Treg data.

Cell isolation

A 60mL sample of heparinized blood was collected from subjects at neutrophil engraftment. PBMCs were isolated by density gradient centrifugation using Ficoll-Hypaque (Sigma-Aldrich, St Louis, MO, USA) and were cryopreserved in aliquots of 5–10×10⁶ cells per mL of CryoStor CS-10 freezing medium (VWR International, Westchester, PA, USA). Before analysis, cryopreserved cells were thawed in a 37 °C water bath, incubated with 20 μg/mL DNase (Roche, Mannheim, Germany), washed with cold PBS and suspended in 1% fetal bovine serum before staining. All samples were handled uniformly.

Flow cytometry

T-cell immunophenotyping was performed using 10-color multiparametric flow cytometry. Surface staining was performed with directly conjugated Abs for 30 min at room temperature using the following titrated Abs or dye: CD3-PerCP-Cy5.5, CD4-Alexa700, CD25-APC-Cy7, CD45ROPE-Cy7, CLA-FITC (BD Biosciences, San Jose, CA, USA), CD8-PE-Cy5, CD14-PE-TR, amine viability dye (Invitrogen, Carlsbad, CA, USA), CD127-Pacific Blue (eBioscience, San Diego, CA, USA) and α4β7-PE (kind gift from Millennium Pharmaceuticals Inc, Cambridge, MA, USA and commercially conjugated by Chromaprobe Inc, Maryland Heights, MD, USA). Cells then were fixed and permeabilized using the Human Foxp3 Buffer Set (BD Biosciences) as per the manufacturer’s instructions, before intracellular staining with Foxp3-Alexa 647 (clone 259D/C7; BD Biosciences). After washing, cells were resuspended in 500 μL of PBS containing 1% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA, USA). Data were acquired using a LSRII flow cytometer (BD Biosciences) and analyzed with FlowJo software version 8.0 (Tree Star, Ashland, OR, USA). All flow cytometric gating/analysis was performed by a single individual (MTR) who was blinded to the results of the clinical data. The frequency of CD45RO⁺CD25⁺Foxp3⁺CD127^lo Tregs was expressed as the percentage of positive cells in the total CD4⁺ gate, whereas the percentage of CLA⁺ or α₄β₇⁺ Tregs was expressed as the percentage of their respective subpopulations within the total Treg gate. Owing to significant lymphopenia at engraftment, Treg percentages were felt to be more accurate than absolute numbers of rare Treg subsets.

Statistical analysis

Data sets were summarized using descriptive statistics including median and range for continuous variables, as well as percentage and frequency for categorical variables. Comparison between independent groups was carried out using the Mann–Whitney U-test for continuous variables and χ²-test or Fisher’s exact test for categorical variables. Logistic regression and the proportional odds model was used to assess the association between Treg subset percentages at engraftment with aGVHD incidence/recurrence and severity, respectively. Specifically, logistic regression models were created using Treg subset percentages as the independent, continuous, predictor variable, whereas aGVHD (yes/no) was the dependent, categorical, outcome variable. For the proportional odds model, Treg subset percentages were used as the independent, continuous, predictor variable, whereas aGVHD severity (that is, grade 0–IV/organ stage 0–4) was the dependent, ordinal, outcome variable. For tests of severity, the maximal grade or organ stage during the first 100 days of transplant was used. Multivariable logistic regression evaluated whether the percentage of Tregs or their subsets could be an independent predictor of aGVHD incidence or recurrence after adjusting for potential confounding variables. All reported P-values were two-tailed and considered significant at P<0.05. Analyses were performed using SPSS version 16 (SPSS Inc, Chicago, IL, USA) and R version 2.7.0 (Free Software Foundation, Boston, MA, USA).

Results

Patients

From 10 February 2007 through to 14 March 2008, 43 adult patients underwent ASCT and were followed up for 100 days post transplant. The clinical characteristics are summarized in Table 1. In all, 41 (95%) patients survived until day 100. aGVHD grade II–IV occurred in 34 (79%) patients and was biopsy confirmed in 30 (88%) cases. Initial signs or symptoms of aGVHD manifested at a median of 23 days post transplant (range, 9–93 days). The most commonly involved sites were the gastrointestinal tract in 28 (65%) patients, followed by skin in 20 (47%) patients. The high incidence of grade II–IV aGVHD was likely attributable to prompt esophagogastroduodenoscopy to evaluate nausea and anorexia related to upper-gut aGVHD, which was seen in 23 patients. High-dose corticosteroids were used in 31 (72%) of these individuals, starting at a median of 26 days post transplant (range 9–87 days). A total of 19 (44%) patients previously diagnosed with aGVHD had recurrent symptoms of grade II–IV aGVHD during either initial therapy or at the time of steroid taper. Except for early mortality, there were no significant differences in clinical characteristics among patients with or without grade II–IV aGVHD (Table 1).

Table 1.

Characteristics of total cohort, 43 patients undergoing ASCT

Characteristic	No. of patients with characteristic for indicated grade of aGVHD (percentage)
	Grade 0–I	Grade II–IVa	P-values
Total patient no.	9	34	—
Age in years
Median	46	45	0.714
Range	34–70	21–65
Sex
Male	6 (67)	13 (38)	0.127
Female	3 (33)	21 (62)
Diagnosis
Acute leukemia+MDS	3 (33)	21 (62)	—
CML+MPD	1 (11)	2 (6)
NHL+CLL+myeloma	5 (56)	11 (32)
Conditioning regimen
Myeloablative	5 (56)	21 (62)	0.735
Reduced intensity	4 (44)	13 (38)
Donor
Related	7 (78)	19 (56)	0.232
Unrelated	2 (22)	15 (44)
Stem cell source
Peripheral blood	7 (78)	25 (74)	0.795
Other	2 (22)	9 (26)
HLA
Matched	8 (89)	31 (91)	1.00
Mismatched	1 (11)	3 (9)
Donor/recipient sex
Matched	7 (78)	22 (65)	0.457
Mismatched	2 (22)	12 (35)
CD34+ (×106 cells per kg)
Median	5.61	5.99	0.952
Range	2.44–10.1	0.04–10.1
aGVHD prophylaxis
CSA+MTX	5 (56)	20 (59)	0.860
CSA/FK506+MMF	4 (44)	14 (41)
aGVHD organ involvement
Skin	3 (33)	17 (50)	—
Gut	—	28 (82)
Liver	—	2 (5)
Day +100 malignancy status
CR or PR	7 (78)	31 (91)	0.277
Relapse or progression	2 (22)	3 (9)
Day +100 survival
Alive	7 (78)	34 (100)	0.040
Dead	2 (22)b	—

Abbreviations: aGVHD=acute GVHD; ASCT=allo-SCT; CML=chronic myeloid leukemia; CR=complete response; FK506=tacrolimus; MDS=myelodysplastic syndrome; MMF=mycophenolate mofetil; MPD=myeloproliferative disorder; NHL=non-Hodgkin’s lymphoma; PR=partial response.

^aGrade of aGVHD (0–I vs II–IV) indicates maximal grade of aGVHD during first 100 days of transplant.

^bCauses of death included relapse of malignancy (n=1) and infection (n=1).

The median time from transplant to initial Treg sample acquisition was 19 days (range, 10–31 days). Figure 1 illustrates the flow cytometric gating hierarchy used to identify Tregs and their tissue-homing subsets in these subjects.^{13–15,18–20,23,24} As per standard institutional operating guidelines, sorted chimerism data were obtained at day +30 from patients receiving reduced intensity conditioning. RFLP analysis showed that the median percentage of donor CD3⁺ cells in peripheral blood was 73% (range 47–100%) at day +30 in patients undergoing reduced intensity conditioning, suggesting that the majority of Tregs were donor derived. Generally, Treg analysis could be performed before aGVHD or the start of high-dose steroids, with median time periods from sample acquisition to aGVHD occurrence or treatment of 3 days (range, 0–73 days) and 6 days (range, 0–69 days), respectively. Six patients (14%) were on steroids for a median of 1 day (range, 1–11 days) before sample acquisition.

Figure 1.

Gating hierarchy to identify Tregs and their tissue-homing subsets using 10-color multiparametric flow cytometry. (a) CD4 gating. Viable T lymphocytes were identified by gating on cells that expressed CD3 while excluding cells expressing CD14 or marking with the amine viability dye (data not shown). CD4⁺ T cells were then selected by gating on CD4⁺CD8⁻ cells. (b) Setting gates for additional parameters. Gates for Foxp3, CD25, CD127, CD45R0, α4β7 and CLA were determined in two-parameter comparisons before use in subset analysis using the total CD4+ cell population (from panel a), to insure adequate cell numbers for accurate gating. (c) Data acquisition. Data on Treg subset frequency were acquired by sequentially applying gates established in panel b to the CD25⁺Foxp3⁺ Treg population derived from the parent CD4⁺ cell population (from panel a).

Treg subsets and aGVHD

There was no significant association between the percentage of total Tregs in the CD4⁺ compartment at engraftment between patients with grades 0–I and II–IV aGVHD (P=0.387) (Figure 2a). Differences were detected, however, between patients with and without the disease when skin- or gut-homing phenotype was analyzed. Patients with any stage skin aGVHD had a significantly lower proportion of Tregs expressing CLA at engraftment (median, 1.7 vs 2.6%; P=0.038). A trend was noted with regard to the association between α₄β₇⁺ Tregs and the occurrence of any stage gut aGVHD (13.9% vs 20.8%; P=0.074 (gut aGVHD vs no gut aGVHD)) (Figures 2b and c). Interestingly, patients with any stage skin aGVHD had a significantly higher frequency of α₄β₇⁺ Tregs (median, 20.1 vs 9.87%; P=0.030). These data support the hypothesis of organ-specific prevention of aGVHD by CLA⁺ Tregs and also suggests that the expression of CLA and α₄β₇ by Tregs post transplant is reciprocally regulated.

Figure 2.

Association between Tregs, CLA⁺ Tregs and α₄β₇⁺ Tregs with aGVHD outcomes. During neutrophil engraftment, the frequency of Tregs, CLA⁺ Tregs and α₄β₇⁺ Tregs was measured by multiparametric flow cytometry. Box plots define the values for median, range, 25th and 75th percentiles. Two-tailed P-values were calculated using logistic regression to assess for differences in the median percentage of Tregs, CLA⁺ Tregs and α₄β₇⁺ Tregs for the following comparisons: (a) grade 0–I aGVHD vs grade II–IV aGVHD; (b) no skin aGVHD vs skin aGVHD; (c) no gut aGVHD vs gut aGVHD; (g) grade 0–I recurrent aGVHD vs grade II–IV recurrent aGVHD; (h) no recurrent skin aGVHD vs recurrent skin aGVHD; (i) no recurrent gut aGVHD vs recurrent gut aGVHD. Scatter plots indicate the percentage of Tregs, CLA⁺ Tregs and α₄β₇⁺ Tregs for patients with varying degrees of aGVHD severity. Line represents median values. The proportional odds model was used to calculate the odds of developing increasing grade of aGVHD (d), increasing stage of skin aGVHD (e) and increasing stage of gut aGVHD (f) for increasing frequencies of Tregs, CLA⁺ Tregs and α₄β₇⁺ Tregs, respectively.

Treg subsets and aGVHD severity

Next, we investigated the relationships between Treg subsets at engraftment with clinical severity of aGVHD. The maximum recorded organ stage or overall grade of aGVHD during the first 100 days of transplant was used for all statistical tests of severity. Increased frequencies of CLA⁺ or α₄β₇⁺ Tregs at engraftment were associated with lower stages of skin (odds ratio (OR), 0.67; 95% confidence interval (CI), 0.46–0.98; P=0.041) or less severe gut aGVHD (OR, 0.93; 95% CI, 0.88–0.99; P=0.031), respectively. In contrast, a significant relationship between the percentage of total Tregs at engraftment with overall grade of aGVHD was not observed (OR, 0.95; 95% CI, 0.88–1.03; P=0.227) (Figures 2d–f). The association between Treg subsets with histological severity of aGVHD could not be assessed because of the predominance of grade I/mild aGVHD on tissue sections.

Treg subsets and aGVHD recurrence

As many patients with aGVHD do not achieve a complete response to initial therapy or have a nondurable response, we compared the percentage of Tregs or their tissue-homing subsets at engraftment among patients with recurrent signs or symptoms of aGVHD.¹ Patients developing recurrent grade II–IV aGVHD (n=19), any recurrent skin aGVHD (n=17) or any recurrent gut aGVHD (n=13) had lower percentages of total Tregs, CLA⁺ Tregs and α₄β₇⁺ Tregs when compared with patients without recurrent aGVHD, (4.0 vs 8.2%; P=0.061), (1.8 vs 2.1%; P=0.095) and (9.9 vs 19.1%; P=0.028), respectively (Figures 2g–i). Patients with recurrent gut aGVHD also had a significantly lower frequency of Tregs (2.9 vs 8.0%; P=0.025). Recurrent aGVHD was analyzed only during the first 100 days and no patients received donor lymphocyte infusions during this time.

Multivariable analysis

Univariable analysis did not detect significant differences in transplant characteristics among patients with or without aGVHD (Table 1). In a multivariable logistic regression model, controlling for either the intensity of the conditioning regimen, donor type or stem cell source, a higher percentage of CLA⁺ Tregs continued to be associated with the prevention of skin aGVHD, whereas Tregs seemed to be an independent predictor for recurrent grade II–IV aGVHD (Table 2).

Table 2.

Logistic regression models for development of any stage skin aGVHD or recurrent grade II–IV aGVHD

Variable	Multivariable analyses
	Any stage skin aGVHD		Recurrent grade II–IV aGVHD
	Odds ratioa (95% CI)	P-values	Odds ratioa (95% CI)	P-values
Model 1b
Percentage CLA+ Tregs	0.65 (0.43–0.97)	0.034	—	—
Myeloablative conditioning	1.57 (0.42–5.90)	0.507	0.74 (0.16–3.40)	0.695
Percentage Tregs	—	—	0.90 (0.80–1.02)	0.112
Model 2b
Percentage CLA+ Tregs	0.67 (0.45–0.98)	0.041	—	—
Unrelated donor	3.27 (0.81–13.2)	0.096	1.81 (0.39–8.51)	0.452
Percentage Tregs	—	—	0.88 (0.77–0.99)	0.047
Model 3b
Percentage CLA+ Tregs	0.65 (0.43–0.98)	0.041	—	—
PBSCs	0.23 (0.04–1.20)	0.081	0.42 (0.06–2.80)	0.367
Percentage Tregs	—	—	0.87 (0.76–0.99)	0.049

Abbreviations: aGVHD=acute GVHD; CI=confidence interval; CLA=cutaneous lymphocyte Ag; OR=odds ratio; Tregs=regulatory T cells.

^aOR less than 1 indicates decreased odds of developing the event in relationship to the characteristic studied.

^bOwing to limitations in sample size and the proportion of patients with aGVHD events, regression models could be constructed using only two covariates.

Discussion

Our data suggest that aGVHD organ specificity and severity can be predicted by Treg tissue-homing subsets as early as 2–3 weeks after ASCT. The frequencies of tissue-specific homing Treg subsets seemed to predict aGVHD outcomes more accurately than the percentage of the total Treg population in the blood. The frequencies of Tregs or α₄β₇⁺ Tregs at engraftment also predicted for late episodes of recurrent grade II–IV aGVHD or gut aGVHD, respectively. These data imply that early T-cell populations create conditions favoring either long-term tolerance or alloreactivity, consistent with the clinical data showing that aGVHD is a major risk factor for chronic GVHD.^25,26

In murine models of transplantation, only the lymph node-homing (CD62L⁺) population of Tregs prevented the development of aGVHD.^4,5 Our data is similar in that it supports the importance and unique functions of Treg subsets in ASCT. The apparent contradiction (that is, the importance of lymph node-homing Tregs in mice vs tissue-homing Tregs in humans) could potentially be explained by the fact that lymph node localization is generally felt to be a prerequisite to tissue migration. Interestingly, secondary lymphoid organs have a critical role in determining lymphocyte compartmentalization with peripheral lymph nodes/DCs inducing CLA and CCR4 expression by lymphocytes, whereas mesenteric lymph nodes/DCs are associated with the upregulation of α₄β₇ and CCR9.^19,27 Thus, it seems that Treg expression of CD62L, CLA and α₄β₇ could be intimately related.

In our study, aGVHD analysis was performed in a blinded and prospective fashion and was biopsy confirmed in majority of the patients. Transplant characteristics appeared to be well distributed between patients with or without aGVHD. Notable exceptions include patient sex and donor type wherein there seemed to be a predominance of female patients and unrelated donors in those individuals who developed grade II–IV aGVHD. It was reassuring that in univariable analysis, significant differences were not detected in transplant characteristics among patients with or without grade II–IV aGVHD, although this could have been affected by sample size. Multivariable analysis showed that Tregs or their subsets continued to be independent predictors of aGVHD outcomes after controlling for possible confounding variables. Taken as a whole, these data support a role for Treg subsets in the prevention of aGVHD across a broad range of differing chemotherapy regimens, donor types and stem cell sources.

Although our study is limited by a modest sample size and a greater than previously reported incidence of aGVHD, our novel results support an important role for Treg tissue-homing subsets in the prevention of organ-specific aGVHD. An association between CLA⁺ Tregs and α₄β₇⁺ Tregs in peripheral blood with immunohistochemical evidence of Treg organ tropism would further validate this finding; however, data supporting the critical role of adhesion molecules with lymphocyte compartmentalization already exists.^{8–10,19,20,27,28} To our knowledge, this is the first report showing that increased CLA⁺ or α₄β₇⁺ expression by circulating human Tregs is associated with reduced risk of skin or gut aGVHD, respectively. If validated in a larger cohort, these laboratory parameters could enable practitioners to institute earlier and possibly more effective (organ-directed) immunosuppression, thereby improving ASCT outcomes. Furthermore, improvements in the percentages of skin- or gut-homing Tregs at engraftment could be used as an end point in future clinical trials for development of novel interventions to increase circulating tissue-specific regulatory cells.