Discussion
Adoptive Treg transfer with the aim to improve the control of immune responses and re-establish peripheral tolerance holds therapeutic promises in transplantation, GVHD, and autoimmune diseases. In a future perspective, antigen expanded as well as TCR engineered or CAR Treg will rapidly become available alternatives to polyclonal Treg. The clinical experience however has shown some issues that may impact the therapeutic value of this approach. Specifically, the need of a high number of Treg for adoptive therapy has been fulfilled with robust in vitro expansion protocols, but expanded Treg have shown fragilities and a limited capacity to persist in patients. Moreover, there is a need of measurable biological parameters that, beside the mere cell yield, can predict the capacity of expanded Treg to adapt to an in vivo environment in which competition with other cells for growth factors and nutrients generate a selective pressure. We report that Treg can be expanded using a combination of IL-7 and IL-2 with several advantages in term of resistance to apoptosis and stress and maintenance of a poorly differentiated phenotype. These findings are relevant to the improvements of Treg expansion protocols for adoptive Treg therapy and to overcome some issues related to the long-term persistence of Treg once infused in patients.
In vitro expansion of Treg for adoptive transfer traditionally rely on the use of high doses of IL-2 (17) (18) (19). The final Treg product is composed by a large majority of Treg with a short life-span in vivo and a long-lived subset that can persist for over a year (8). Based on studies performed mainly on conventional T cell we hypothesized the use of homeostatic cytokines for Treg expansion. Homeostatic cytokines such as IL-7 induces conventional T cell expansion (20) and provide anti-apoptotic signals (21). However, their use in Treg expansion has not been fully explored based on the low expression of the IL-7 receptor alpha chain, and assuming that Treg do not respond to IL-7.
Treg are phenotypically defined as CD4+CD25highFOXP3+T cells and sorted as CD4+CD25highCD127lowT cells for in vitro expansion (14). Compared to circulating B cells that lack the expression of CD127 (15) we found a low but significant expression of CD127 that is sufficient to trigger significant STAT5 phosphorylation upon stimulation with 10ng/ml of IL-7. We previously reported that Treg are responsive to IL-7, even though at higher concentrations than conventional T cells (12). At physiological IL-7 concentrations of 2-8 pg/ml (22) Treg are not likely to receive significant signals, but when IL-7 is used at high concentration for in vitro expansion can elicit a robust response to IL-7. Importantly, we reported that in patients experiencing lymphopenia in which IL-7 is present at supra-physiological concentrations it may contribute to their homeostasis and proliferation to reconstitute the depleted Treg compartment (13). The response to IL-7 was higher in Treg with a naïve or memory stem T cell phenotype and decline with the progression of differentiation into central-memory and effector-memory subsets. The responsiveness was not strictly related to changes in the expression of CD127, suggesting that other unidentified factors may regulate STAT-5 phosphorylation beside the mere expression of the receptor. High IL-7 responsiveness of CD45RA+CD62L+ Treg was one reason why we decided to sort CD45RA+CD62L+ for expansion. The other reason was to obtain a Treg product with an immature phenotype that apparently are the cells that survive longer in patients (8). Since IL-7 per se was not sufficient to trigger significant Treg expansion we combined IL-7 to IL-2. Surprisingly, we did not observed an additive or synergistic effect but instead the presence of IL-7 reduced the proliferation rate induced by IL-2. We experimentally proved that in the presence of IL-7 the formation of the high affinity IL-7 receptor engage a significant amount of the common-γ chain (CD132), reducing the ability of IL-2 to form the high affinity IL-2 receptor with CD25, CD122 and CD132. We have previously shown that competition of CD127 and CD25 for CD132 can occur in conventional T cells (16). Accordingly, expansion of CD45RA+CD62L+ Treg with the IL-7M resulted in a reduced final cell yield as well as a different surface phenotype in which a significant higher proportion of Treg display a CD45RA+CD62L+CD95+ phenotype. This phenotype is reminiscent of that of conventional memory stem T cells (23) and, to the best of our knowledge Treg with a stem cell memory phenotype have never been described before. Further studies are needed to clarify whether a subset with memory stem cell function exist also in the Treg compartment with characteristics of self-renewal and the capacity to generate the full phenotypic diversity of Treg subsets. An increased expression of CD95 was found in expanded Treg from cord-blood as compared to expanded Treg from peripheral blood (24) that also contain an increased proportion of CD45RA+CD62L+ Treg. Such phenotypically immature Treg population obtained with the IL-7M also showed relevant differences in terms of metabolic machinery and survival capacity. An improved metabolic fitness in terms of mitochondrial mass and capacity to uptake glucose can advantage Treg expanded with the IL-7M once infused in patients. Compared to conventional T cells, Treg cells are less reliant on glycolysis and use mitochondrial metabolism and oxidative phosphorylation (OXPHOS) for energy production (25). In vitro studies revealed that Foxp3 is directly responsible in reprograming T cell metabolism by suppressing glycolysis and enhancing OXPHOS (26)(27). Elevated glycolysis may be detrimental to Treg cell induction and suppressive function and deletion of HIF-1a, a transcription factor that can promote glycolysis, leads to increased Foxp3 induction (28). However, also Treg needs glycolysis to support some processes such as proliferation (29) and migration (30) when a high amount of ATP and metabolic intermediates are needed. It is reasonable to speculate that Treg cells precisely balance cellular glucose consumption, when glycolysis increases Treg cell proliferation and expansion, but this activity is balanced OXPHOS to maintain lineage stability and suppressive activity. Treg obtained with the IL-7M showed signs of both increased glycolysis (2NBDG uptake and lactate production) but also an increased mitochondrial mass. Reduction of glycolysis after a few days of resting corresponded, in our study, to a recovery of the Treg suppressive capacity. An important issue is also that Treg adoptive transfer implicates moving Treg from in vitro culture conditions with high glucose, oxygen and intense cytokine and TCR/CD28 signaling to an in vivo environment where these factors are reduced or completely missing. Therefore, a improved metabolic machinery can be helpful to Treg to adapt to changing conditions and survive in vivo overcoming substrate and grow factor signaling restrictions. We showed indeed and increased survival capacity of Treg both after growth factor deprivation but also in response to direct apoptotic stimuli possibly. Naïve Treg do not express CD95 and become susceptible to fas ligand mediated apoptosis only when start expressing CD95 upon stimulation (31). Increased resistance to fas ligand mediated apoptosis in Treg expanded with the IL-7M can be due to the increased expression of the anti-apoptotic molecule Bcl-2, which is presumably induced by the presence of IL-7 (21). We also observed a reduced telomere shortening in Treg expanded with the IL-7M. Telomere length is regulated by telomere erosion during cell division, and by the activity of telomerase which is negligible is T cells (32). Preservation of telomere length during IL-7 mediated T cell expansion has been reported (32), however in our Treg expansion model this could also be due to the reduced rate of expansion of Treg with the IL-7M. Preliminary data indicate an increased capacity of Treg expanded with the IL-7M to survive in NSG mice as well as a distribution pattern that includes migration to the bone marrow. While these observations require confirmation in relevant disease models to determine whether the increased persistence is associated to an improved therapeutic effect, our data prompted us to further characterized Treg expanded with the IL-7M. It has to be determined whether a lower number of phenotypically immature Treg but also with better performances can increased the therapeutic value as compared to a higher number of phenotypically mature Treg with fragilities in terms of resistance to stress and apoptosis.
We therefore suggest that addition of IL-7 during expansion improves the performances of Treg despite a lower final cell yield and a (reversible) reduction of suppressive function. Our expansion model needs to be further explored as a potential improvement of current Treg expansion protocols based on IL-2. The fragility of Treg used for adoptive transfer represents an issue that need to be resolved in order to increase the efficacy or this promising immune-therapy.