Applicability, safety, and biological activity of regulatory T cell therapy in liver transplantation

Regulatory T cells (Tregs) are a lymphocyte subset with intrinsic immunosuppressive properties that can be expanded in large numbers ex vivo and have been shown to prevent allograft rejection and promote tolerance in animal models. To investigate the safety, applicability, and biological activity of autologous Treg adoptive transfer in humans, we conducted an open‐label, dose‐escalation, Phase I clinical trial in liver transplantation. Patients were enrolled while awaiting liver transplantation or 6‐12 months posttransplant. Circulating Tregs were isolated from blood or leukapheresis, expanded under good manufacturing practices (GMP) conditions, and administered intravenously at either 0.5‐1 million Tregs/kg or 3‐4.5 million Tregs/kg. The primary endpoint was the rate of dose‐ limiting toxicities occurring within 4 weeks of infusion. The applicability of the clinical protocol was poor unless patient recruitment was deferred until 6‐12 months posttransplant. Thus, only 3 of the 17 patients who consented while awaiting liver transplantation were dosed. In contrast, all six patients who consented 6‐12 months posttransplant received the cell infusion. Treg transfer was safe, transiently increased the pool of circulating Tregs and reduced anti‐donor T cell responses. Our study opens the door to employing Treg immunotherapy to facilitate the reduction or complete discontinuation of immunosuppression following liver transplantation.


Regulatory T cells (Tregs) are a subset of cluster of differentiation
(CD)4-positive T cells that constitutively express the Forkhead Box P3 (Foxp3) transcription factor and have the capacity to migrate to sites of inflammation and exert a wide range of immunosuppressive effects. Animal studies indicate that Tregs play a key role in maintaining immune homeostasis and preventing autoimmunity. 1 Furthermore, they can recognize allogeneic major histocompatibility complex (MHC) molecules and suppress allograft rejection, and are essential for the induction and maintenance of transplantation tolerance through the mechanisms of "linked suppression" and "infectious tolerance." 2 Although human Tregs constitute a small proportion (5%-7%) of circulating CD4 + T cells, they are attractive candidates for immunotherapeutic purposes given that they can be isolated and expanded in large numbers in vitro without losing their immunoregulatory properties. 3 Clinical studies have demonstrated the safety of Treg adoptive transfer in graft-versus-host disease and type 1 diabetes mellitus. [4][5][6][7] Furthermore, a number of trials have been initiated both in kidney and in liver transplantation. 8,9 Liver transplantation constitutes an appealing clinical setting to evaluate the effects of Treg transfer given the lower immunogenicity of liver allografts and the substantial clinical experience that has been derived from trials of complete immunosuppression discontinuation. 10 In this setting, infusion of a single dose of a Treg-enriched autologous leukocyte cell product (generated by culturing peripheral blood mononuclear cells (PBMCs) with irradiated donor leukocytes in the presence of co-stimulation blockade), was recently shown to successfully induce operational tolerance in 7 of 10 splenectomized living donor liver transplant recipients treated with cyclophosphamide and conventional immunosuppression. 11 Despite these encouraging early results, key questions regarding the overall clinical applicability of Treg immunotherapy, the optimal clinical design, and the immunological effects of Treg infusion in human liver transplant recipients remain to be answered.
We recently described the first good manufacturing practices (GMP)-compliant protocol for the ex vivo expansion of polyclonal Tregs from prospective liver transplant recipients. 12 This protocol, which included up to three rounds of stimulation in the presence of rapamycin, was successful in expanding circulating Tregs >100fold, maintained their Foxp3 expression levels, and increasing their suppressive function. It is important to note that expanded Tregs exhibited a stable noninflammatory phenotype even after being challenged with a cocktail of inflammatory cytokines. We report here the results of a First-in-Human Phase I clinical trial evaluating the safety and immunological effects of purified, ex vivo expanded and adoptively transferred autologous polyclonal Tregs in adult liver transplant recipients.

| Participants
Patients were initially enrolled while awaiting liver transplantation and their participation was confirmed on the day of transplantation. Inclusion criteria at the time of transplantation were the following: (1) age 18-70 years; (2) Model for End-Stage Liver Disease (MELD) score ≤25; (3) no previous transplantation or need recruitment was deferred until 6-12 months posttransplant. Thus, only 3 of the 17 patients who consented while awaiting liver transplantation were dosed. In contrast, all six patients who consented 6-12 months posttransplant received the cell infusion.
Treg transfer was safe, transiently increased the pool of circulating Tregs and reduced anti-donor T cell responses. Our study opens the door to employing Treg immunotherapy to facilitate the reduction or complete discontinuation of immunosuppression following liver transplantation.

K E Y W O R D S
cellular transplantation (nonislet), immunosuppression/immune modulation, liver transplantation/hepatology, T cell biology, tolerance, translational research/science for simultaneous liver-kidney transplantation; (4) absence of autoimmune disease, active viral disease, Epstein-Barr virus seronegativity or hepatocellular carcinoma outside of Milan criteria; (5) leukocyte count >1500/µL and platelet count >50 000 µL; (6) recipient of a brain-dead liver donor; (7) recipient of a cardiac death liver donor if donor age <50-years-old, warm ischemia time <20 minutes, and cold ischemia time <8 hours. For Treg isolation, 250 mL of whole blood was collected during the induction of anesthesia. Participants received thymoglobulin induction (three doses of 1.5 mg/kg, i.v., between posttransplant days 1 and 7), tacrolimus (1 mg twice daily on posttransplant day 1 with doses subsequently adjusted to reach 5-8 ng/mL trough levels), and methylprednisolone (500 mg intraoperatively followed by tapering and discontinuation on posttransplant week 10). Between posttransplant weeks 6 and 8, rapamycin (5-8 ng/mL trough levels) was initiated and levels of tacrolimus (2-5 ng/mL) were decreased. Three months after transplant a liver biopsy was performed to exclude subclinical allograft damage, and patients were admitted for Treg infusion.
Due to the difficulties of enrolling patients before transplantation when following the protocol described, 26 months after its initiation the trial design was amended and all subsequent patients were recruited 6-12 months after transplant. Otherwise, the same inclusion/exclusion criteria were maintained. Immediately after enrollment, patients had their immunosuppressive regimen switched to combined tacrolimus and rapamycin (trough levels 2-5 ng/mL and 2-8 ng/mL, respectively), and 2 months afterward they underwent leukapheresis to collect the starting material for Treg manufacture.
This was followed by a protocol liver biopsy and by the infusion of Tregs 4 months after enrollment. The amended study protocol did not require thymoglobulin induction. This Phase I trial did not include attempts at immunosuppression discontinuation.

| Study endpoints
The primary endpoint was the rate of dose-limiting toxicities within the 4 weeks following infusion. Dose-limiting toxicities were defined as: (1) occurring in the first 72 hours postinfusion, including: National Cancer Institute's Common Terminology Criteria for Adverse Events (CTCAE; Version 4.0) grade 2 or higher cytokine release syndrome, grade 2 or higher injection site reaction, grade 2 or higher fever and/or rigors, grade 3 or higher bronchospasm, grade 3 or higher hypoxia; (2) occurring in the first 30 days of the infusion including: grade 3 or higher infection, grade 3 or higher hematological complication, any CTCAE grade 3 or higher toxicity not clearly related to underlying disease, moderate or severe acute rejection.
Secondary endpoints were: acute and chronic toxicity associated with Treg infusion; incidence of major/opportunistic infections; malignancy; rejection; graft loss; patient mortality; sequential liver and renal function tests; immunosuppressive drug doses and levels; and changes in immunological parameters following Treg infusion.

| Isolation and manufacture of polyclonal Tregs
Two hundred fifty milliliters of whole blood or 180 mL of leukapheresis product was collected and transferred to the GMP Cell Therapy.
The manufacture protocol as well as the phenotypic characteristics, functional properties, and stability of the expanded Tregs have been reported previously 12 and are described in detail as Supplementary Information.

| Treg infusion
Patients were admitted on the day of infusion. The cryopreserved Treg product was thawed in a 37°C water bath, diluted in an infusion bag containing 50 mL of 5% human albumin solution (Albunorm, Octapharma), and infused via a peripheral cannula over 15 minutes with an additional 50 mL of 5% human albumin added to the bag to ensure delivery of the full dose. Premedication consisted of oral paracetamol (1 g) and chlorphenamine (4 mg) 30 minutes prior to infusion. All patients were monitored for 12 hours postinfusion prior to discharge.

Flow cytometry and time-of-flight mass cytometry (CyTOF) immunophenotyping:
The flow cytometry reagents and staining protocols employed were designed and standardized in collaboration with the ONE Study EU Consortium and have already been described (Table   S1). 15 The antibody panel, staining protocol, and data analysis strategy for the CyTOF [16][17][18][19][20]

experiments are described as Supplementary
Information.

| Manufacture of ex vivo expanded Tregs
Tregs were isolated from 11 patients (5 from whole blood and 6 from leukapheresis). The manufacture process failed in two patients (all of them from the first cohort of patients). The first case was due to an insufficient number of Tregs (49 million Tregs), likely resulting from the very low number of Tregs isolated from blood (1.5 million Tregs as compared to 5.9 million, which was the mean from all whole blood Treg isolations). The second failure was due to a low frequency of Tregs in the final product (46% of CD4 + CD25 + Foxp3 + ). In the nine successful manufacture runs, cells were expanded 21-to 486-fold, yielding between 1250 and 22 530 million cells containing 61%-92% Tregs. As compared to whole blood, the use of leukapheresis products allowed a reduction in the duration of Treg culture (from 36 to 24 days) and the need for lower expansion rates to achieve the target dose ( Table 2).
The use of immunosuppressive drugs by the trial participants at the time of leukapheresis did not hamper the Treg manufacture process, as Tregs were successfully expanded from all six recipients recruited 6-12 months after transplant ( Table 2).

| Characteristics of manufactured Tregs and effects on the phenotype of circulating immune cells following infusion
In the six patients who received the 4.  (Tables  S3 and S4) and CyTOF ( Figures S3 and S4), but observed no significant changes in association with the infusion of Tregs.  CXCL9, and CXCL11 1 day after Treg infusion, with gradual decrease by day 3 and complete normalization by day 7 ( Figure 4B). As per the study protocol, the high-grade pyrexia was considered a dose-limiting toxicity and resulted in the expansion of the 3.0-4.5 million Tregs/kg cohort to six participants. The infusion of Tregs did not result in serum cytokine changes in the remaining five patients receiving 4.5 million Tregs/kg ( Figure 4B). Of note, the levels of IL-12p40, IL-18, IL-27, IL-33, CCL17, CCL3, CXCL10, CXCL9, and CXCL11 were already higher in P04 than in the remaining participants immediately before Treg infusion, suggesting that the adverse event may not be solely attributable to the Treg infusion.

| Impact of Treg infusion on donor-specific T cell responses
In the six recipients who received 4.5 million Tregs/kg, a gradual decrease of T cell responses (as assessed by the upregulation of CD154 on memory CD8 + T cells) directed against donor-type cells was observed (P = .066). Although these changes did not reach statistical significance, the trend was clearly different from the responses directed against third-party cells (P = .3) or the cytomegalovirus (CMV) pp65 antigen (P = .5), which remained stable throughout the study period. In contrast, in the three recipients dosed with 1 million Tregs/kg we observed no decrease in donor-specific T cell responses in association with cell infusion ( Figure 5).

| D ISCUSS I ON
Liver transplantation constitutes an optimal clinical scenario to ex-  Although it is not possible to formally establish a causal link between the development of donor hyporesponsiveness and Treg infusion, our findings could be explained by the preferential survival and/or proliferation after infusion of Treg clones with anti-donor alloreactivity, which is an observation that has been documented in experimental animal models. 26 This would be in keeping with the lack of F I G U R E 5 Sequential changes in donor and third-party alloimmune responses. donor-reactive T cell clones. 34,35 In summary, we have described here the successful expansion under GMP conditions of polyclonal Tregs isolated from both endstage liver disease patients awaiting liver transplantation and stable liver transplant recipients under maintenance immunosuppression.
Treg infusion was safe, well-tolerated, and exerted a potentially beneficial effect on donor-specific immune responses. The implementation of the clinical protocol was challenging, however, and its applicability was reliant on deferring patient recruitment and cell infusion until at least 6 months after transplant. Future studies should address the capacity of this strategy, alone or in combination with lymphodepletive therapies, to facilitate the reduction or even the complete discontinuation of antirejection medications following liver transplantation.

ACK N OWLED G M ENTS
We are grateful to Rakesh Sindhi and Chethan Ashokkumar from Plexision (Pittsburgh, PA, USA) for their critical review of the study results.

D I SCLOS U R E
The authors of this manuscript have no conflict to disclose as described by the American Journal of Transplantation.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data supporting the results in the paper will be archived in an appropriate public repository.