Selective modulation of the recipient pro-inflammatory microenvironment
The strategy of minimizing the pro-inflammatory environment in the peri-transplant period with the goal of reducing GvHD whilst potentially retaining GvT effects has already been established. The introduction of RIC transplants and DLI already utilizes this concept with some considerable success. This is the only strategy to separate GvHD and GvT effects that has entered routine clinical practice. However, the use of RIC and DLI lacks specificity and efficacy, as both may be associated with unpredictable and severe GvHD, and GvT effects are still strongly associated with chronic GvHD (Valcarcel et al, 2008). RIC has a non-specific effect in reducing pro-inflammatory cytokines and a more targeted approach to modulate specific cytokines might be more fruitful. Early studies from Cooke et al (2001) targeted lipopolysaccharide (LPS), a component of cell wall of normal gut bacteria and key trigger for acute GvHD. The use of LPS antagonists reduced gut GvHD without reducing GvT effects in mice (Cooke et al, 2001). However, the success of this strategy has not translated to clinical practice. Other approaches have targeted broadly pro-inflammatory cytokines such as tumour necrosis factor-α (TNF-α). Blocking murine donor T cell-derived TNF-α reduced GvHD without affecting GvT effects (Borsotti et al, 2007). However, several human studies of TNF-α blockade to treat acute GvHD were limited by severe off-target toxicity in the form of immunosuppression and infection (Herve et al, 1992; Couriel et al, 2004). Similarly, strategies utilizing Interleukin (IL)1 blockade employed in the peri-transplant period have lacked specificity, failing to reduce overall transplantation-related toxicity (Antin et al, 2002).
Given the failure of these early approaches due to lack of specificity and broadly immunosuppressive effects, it seems unlikely they will provide avenues by which GvHD and GvT effects can be successfully separated. Research has therefore more recently been directed towards the identification and modulation of cytokine pathways acting locally in GvHD target tissues that may not be so important in priming GvT responses. A major advance in our understanding of the role of cytokines in the immunopathogenesis of GvHD was the identification of secretion of both IL23 (a pro-inflammatory member of the IL12 cytokine family) and IL21 (a pro-inflammatory cytokine produced by Th17 cells) by donor APCs as a key event in intestinal GvHD (Das et al, 2009; Hanash et al, 2011). It was subsequently hypothesized that interruption of these mechanisms may provide a means to prevent or reduce gut GvHD without impairing GvT effects of donor T cells at other anatomical sites. Indeed, when IL23 or IL21 signalling was blocked in a mouse transplant model, this resulted in protection of the intestine from GvHD while sparing GvT responses (Das et al, 2010; Hanash et al, 2011). The strategy of IL21 blockade has more recently been shown to ameliorate GvHD mediated by human lymphocytes in a murine model of xenogeneic GvHD, although this study did not address the retention of GvT responses (Hippen et al, 2012). Interestingly, in the human setting, a study examining the role of IL23 receptor single nucleotide polymorphisms (SNPs) has identified a gene variant, which, when present in donors, was associated with decreased risk of developing severe acute GvHD after AHSCT. Unfortunately, this study was not powered to examine the effect of the presence of the genotype on GvT effects and disease relapse (Elmaagacli et al, 2008). Intriguingly, the same SNP was found to have a protective phenotype in patients with Crohn's disease, underlining the role of IL23 in selective intestinal inflammation (Duerr et al, 2006).
In addition to revealing an additional layer of complexity in the cytokines involved in the first phase of priming alloresponses, recent observations have also highlighted the importance of the interplay between the adaptive and innate immunity in the induction of GvHD (Penack et al, 2010). Given the significance of dendritic cells (DC) in T cell priming, several animal studies have established a key contribution both for the induction and amplification of GvHD. DC subsets, such as host Langerhans cells, that persist in non-lymphoid tissues after conditioning are targets of alloreactive donor T cells and are key for the initiation of GvHD in the skin (Merad et al, 2004). In a study of 40 patients receiving allogeneic transplants, high levels of DC activation measured by the CMRF-44 marker had a predictive role for the development of severe acute GvHD (Lau et al, 2007). Therefore, targeting of DC may provide a therapeutically feasible approach to reduce GvHD. In a xenogeneic transplant model, administration of a depleting antibody specific for CD83, a marker for DC activation, reduced GvHD without compromising anti-viral or anti-leukaemic responses (Wilson et al, 2009). DC respond to environmental stimulants and become activated through sets of evolutionary conserved receptors, which sense danger signals produced in responses to infection by microbes or damaged tissues under inflammatory conditions. Microbial components, known as pathogen-associated and danger-associated molecular patterns, are detected by pattern recognition receptors (PRRs) present on APCs and include Toll-like receptors (TLRs) and NOD-like receptors (NLRs). Genetic studies have identified SNPs in genes encoding for NOD2 and TLR4 (sensing bacteria-derived muramyl dipeptide and LPS respectively), which have been shown in some (but not all) studies to influence the outcome of AHSCT (Lorenz et al, 2001; Holler et al, 2006; Mayor et al, 2007). In the study reported by Mayor et al (2007) examining the effect of SNPs in the NOD2 (CARD15) gene after unrelated donor AHSCT found an increased risk of relapse in patients with NOD2 SNPs without any difference in GvHD rates, suggesting that these receptors might be preferentially involved in GvT effects after AHSCT. Thus, PRRs and their ligands might represent new targets to manipulate to selectively reduce GvHD whilst maintaining GvT effects. Others have targeted danger signals to modulate GvHD and GvT effects. In a recent study, the contribution of adenosine triphosphate (ATP), an endogenous danger signal released by dying cells, to GvHD induction in both human and murine AHSCT was addressed (Wilhelm et al, 2010). Increased levels of ATP were found in peritoneal fluid from patients with GvHD and in the gut of mice after AHSCT. Mechanistically, the authors suggested that ATP released by necrotic cells after conditioning radiotherapy induced upregulation of costimulatory receptors on APCs through the function of the P2X purinoceptor seven receptor (P2XR7, previously P2X7R), promoting the pro-inflammatory cascade of events which lead to GvHD development. Neutralization of ATP with apyrase or P2XR7 antagonists reduced GvHD severity in mice. More importantly, when animals were challenged with A20 lymphoma cells, GvT responses remained intact despite P2XR7 blockade (Wilhelm et al, 2010). P2XR7 antagonism remains to be tested in human AHSCT. Interestingly, polymophisms in the human P2XR7 gene are associated with reduced survival after AHSCT but this resulted from a differential incidence of infection rather than GvHD (Lee et al, 2007).
Efforts have also been made to selectively protect GvHD target tissues from injury either by treatment with cytokines or growth factors that can promote local wound repair. Studies initially identified IL11 as a potential candidate for such an approach. Administration of IL11 in murine transplant models reduced levels of TNF-α and protected the small intestine from injury without affecting GvT responses (Hill et al, 1998; Teshima et al, 1999). However, administration of IL11 to patients in the setting of AHSCT proved highly toxic and was accompanied by severe fluid retention and early mortality (Antin et al, 2002). Nevertheless, given the importance of gastrointestinal tract injury in the initiation and amplification of GvHD, investigators have continued to try to identify tractable pathways selectively involved in gut damage and/or healing. A recent study showed that administration of R-spondin, an activator of the Wnt signalling pathway which regulates the intestinal cell proliferation and regeneration, promoted restoration of injured intestinal epithelium and ameliorated systemic GvHD in an MHC-mismatched GvHD murine model (Takashima et al, 2011). However, although the authors did not directly address the effect of R-spondin on GvT responses, they reported reduced donor T cell activation and expansion in R-spondin-treated mice that could potentially unfavourably affect GvT (Takashima et al, 2011). Thus, the retention of GvT responses after R-spondin treatment needs to be specifically addressed before this strategy moves into human trials.
Modulation of priming and migration of alloreactive T cells to GvHD target organs
In the last decade, the process of migration of alloreactive donor T cells after AHSCT has been shown to be essential for their priming and effector function. Modulation of migration of infused donor lymphocytes, proposed by some investigators to be valid addition to Billingham's pre-requisites for functional alloreactivity represents an exciting potential opportunity to separate GvHD and GvT (Sackstein, 2006). Shortly after infusion, alloreactive donor T cells home to secondary lymphoid organs (SLOs, e.g. spleen, lymph nodes and Peyer's patches), where, following encounter and activation by donor APCs, they migrate to different anatomic sites and cause tissue damage (Panoskaltsis-Mortari et al, 2004). These processes of migration to sites of antigen priming and end-organs where tissue damage is mediated, represent opportunities to separate GvHD and GvT effects, particularly as migration to different organs are mediated by different molecular pathways. Further evidence for the rationale for development of such strategies is provided by several published studies that demonstrate the benefit of non-absorbable steroids in the treatment of gut GvHD by provision of localized immunosuppression (Ibrahim et al, 2009).
Homing of alloreactive donor T cells to SLOs is mediated by chemokine-chemokine receptor interactions and adhesion molecules that facilitate the migration to SLOs and egress of T cells from peripheral circulation via high endothelial venules. Several of these pathways have been modulated to reduce the priming of alloreactive donor T cells in murine transplant models. CC-chemokine receptor (CCR)7 is a protein receptor for Chemokine C-C motif ligands 19 L21, which plays a key role in the entry of naive T cells into SLO prior to antigen-specific priming. In a murine model of AHSCT, CCR7−/− T cells were incapable of inducing GvHD whilst displaying intact GvT responses (Coghill et al, 2010), suggesting that chemokine receptor blockade might be useful new strategy. However, concerns remain that such strategies might impair the priming of pathogen-specific immune responses reduce functional immunity after AHSCT. Similarly, it was shown some years ago that blockade of CD62 ligand (CD62L), which facilitates entry of T cells to SLOs, can reduce GvHD in murine models (Li et al, 2001). In addition, human alloreactive T cells are predominantly CD62L+ naive and/or central memory cells (Foster et al, 2004). Consistent with this hypothesis, Chen et al (2004) observed reduced GvHD after adoptive transfer of CD62L− T cells in MHC-mismatched recipients. The strategy of CD62L+ cell depletion is currently being tested in an ongoing clinical trial using naive T cell-depleted DLI following allogeneic HSCT (ClinicalTrials.gov Identifier NCT01627275). Whether this study retains clinically relevant GvT effects will be key, as a recent in vitro study using human leucocyte antigen (HLA)-matched donor-patient pairs demonstrated that leukaemia-reactive CD8+ cytotoxic T-lymphocytes were mainly derived from subsets enriched for naïve T cells (Distler et al, 2011). Alternatively, blockade of CD62L might prevent GvHD via inhibition of priming in SLOs. However this strategy might be also expected to impede antigen-specific priming that is important for pathogen-specific immunity post-transplant and might also diminish GvT responses. Furthermore, recent murine transplant studies utilizing donor T cells deficient in CD62L have shown that the expression of this molecule is not essential for migration of donor T cells into SLOs or for GvHD, casting doubt on the clinical utility of blockade of this pathway (Anderson et al, 2008).
Other chemokine pathways mediate migration of donor T cells from SLOs to end-organs, and some, such as the CCR5 pathway mediate migration to both SLOs and target organs. CCR5 modulation has been extensively investigated in murine AHSCT models, but data suggest that the effect of this strategy on GvHD is variable and exacerbation of GvHD has been seen in some circumstances. In a murine AHSCT model without conditioning, prevention of CCR5-signalling resulted in significant reduction of GvHD (Murai et al, 2003). In contrast, in a model that used lethal conditioning, donor T cells from CCR5−/− mice resulted in more severe GVHD than wild type donor T cells (Wysocki et al, 2004). In addition, data supporting preservation of GvT effects after CCR5 blockade are somewhat lacking, and CCR5 is also expressed on NK cells; thus, CCR5 blockade might be expected to reduce NK-cell mediated GvT effects (Weiss et al, 2011). There is, however, recent evidence that CCR5 is important in human AHSCT. CCR5+ T cells have been shown to infiltrate the skin of patients with acute GvHD (Palmer et al, 2010) and CCR5 polymorphisms influence risk of GvHD after transplant (Bogunia-Kubik et al, 2006). A phase I/II study examining the effect of Maraviroc, an oral CCR5 antagonist, on visceral acute GvHD after AHSCT has recently been published with promising early results, although longer follow-up is needed to assess whether this strategy impacts adversely on GvT effects (Reshef et al, 2012).
After priming, alloreactive T cells must migrate from SLO into GvHD target tissues. Alloreactive T cell trafficking to target tissues is imprinted by the expression of specific sets of cell surface molecules including selectins, integrins and chemokine receptors driven by the spatio-temporal expression patterns of their cognate ligands on inflamed tissues. A short time after AHSCT several pro-inflammatory chemokines are up-regulated in lymphoid tissues. These prompt T cells to reprogramme homing receptors so that they can migrate to non-lymphoid tissue targets, such as the skin, lungs and gut (Serody et al, 2000; New et al, 2002; Wysocki et al, 2004). Thus, interfering with the trafficking of donor T cells to target organs may be another attractive option to separate GvT from GvHD. In line with this hypothesis, the sphingosine-1-phosphate receptor inhibitor FTY720, which decreases egress of lymphocytes to peripheral sites of inflammation, diminished GvHD while retaining anti-lymphoma responses in a haploidentical murine AHSCT model (Kim et al, 2003).
The approach of blockade of pathways governing migration of alloreactive T cells into either organ-specific SLOs or from SLOs into target organs of GvHD has the advantage of providing potential for effective GvT responses directed towards other end-organs and the retention of priming of pathogen-specific responses post-transplant.
Many murine AHSCT studies using knockouts for integrins, chemokines and other molecules involved in distal T cell migration have identified other pathways by which GvHD and GvT might be separated. The chemokine receptor CCR2 is expressed on most haematopoietic cell types and controls leucocyte migration during inflammation. Its cognate ligand CCL2 is expressed in the gut, liver, skin and lung early after transplantation. In a murine AHSCT model, mice receiving CCR2−/− CD8+ T cells had less GvHD than wild-type cell recipients. This was attributed to a functional defect of CCR2−/− CD8+ T cells in their capacity to migrate to end-organ targets of GvHD, because such cells displayed intact proliferation, activation and cytotoxicity. Furthermore, CCR2−/− CD8+ T cells protected mice challenged with P815 mastocytoma cells inferring retention of GvT effects (Terwey et al, 2005). Migration of donor T cells to gut-associated lymphoid tissues is dependent on the integrin α4β7 and its ligand Mucosal Addressin Cell Adhesion Molecule (MadCAM). In vivo, α4β7−/− donor T cells administered to allogeneic recipient mice, demonstrated defective infiltration to intestinal mucosa and reduced gut GvHD while retaining GvT responses by conferring protection against circulating EL4 murine leukaemia/lymphoma cells (Petrovic et al, 2004; Waldman et al, 2006). Similarly, β2 integrins have been recently shown to play an important role in T-cell trafficking after AHSCT. Murine integrin β2−/− T cells failed to infiltrate the liver and intestine in a murine AHSCT model and higher numbers of such cells were found in SLOs, suggesting impaired migration from SLOs to target end-organs. However, β2−/− T cells retained their GvT potential in vivo in mice infused with the A20 lymphoma cell line (Liang et al, 2008).
The most refined application of the strategy of modulation of donor T cell migration to separate GvHD and GvT effects would be to modulate a pathway that mediates migration to a specific end-organ after priming. Donor T cells acquire their target-organ homing properties very rapidly after the initiation of an immune response in priming SLOs (Campbell & Butcher, 2002). Organ-specific tropism is likely to be instructed by the site of priming as a result of the interaction of T cells with organ-derived dendritic cells (Kim et al, 2008); for example, mucosa-associated lymphoid tissue induces the expression of gut-specific integrins and chemokine receptors. Thus the blockade of migration of donor T cells to organ-specific SLOs might be expected to reduce the subsequent migration of donor T cells to the related end-organ. Somewhat surprisingly, a murine AHSCT model utilizing organ-specific SLO tissue radioablation showed cells primed in any SLO could result in GvHD of any target tissue. Thus, although strategies to prevent migration of donor T cells to organ-specific SLOs have been successful in reducing organ-specific GvHD, blockade of distal chemokine pathways mediating migration of donor T cell from SLOs to end-organs might be the most effective strategy to prevent GvHD in an organ-specific fashion (Beilhack et al, 2008). An example of a relatively organ-specific distal migratory pathway is that mediated by the integrin CD103 (the ligand of epithelial cell-specific E-Cadherin). CD103 has been implicated in the migration and cytolytic function of gut-infiltrating CD8+ lymphocytes during GvHD (El-Asady et al, 2005). Furthermore, a recent murine transplant model has shown that transplantation of CD103−/− CD8+ T cells mediated less intestinal GvHD with maintained tumour clearance (Liu et al, 2011). The manipulation of such organ-specific distal migratory pathways might provide the best option to selectively reduce GvHD whilst retaining GvT effects and minimizing off-target effects.
Although obviously an exciting new area, there are several key challenges to the ultimate translation of modulation of donor T cell migration to the clinic. Potential strategies include both gene therapy approaches pre-transplant to knock down expression of chemokine receptors or other molecules on donor T cells, and the use of antibody or small molecule-mediated inhibition of selective pathways in vivo post-transplant. Some of the potential target pathways are summarized in Fig 2. One significant concern is off-target inhibition of migration of pathogen-specific, and potentially tumour-associated antigen-specific T cells that share common chemokine receptor expression with alloreactive T cells. For example, the chemokine receptor CCR5 is expressed on both human cytomegalovirus-specific T cells (Palendira et al, 2008), and on alloreactive T cells that mediate skin GvHD (Palmer et al, 2010). In reality, the degree of degeneracy and functional redundancy among pathways mediating T cell migration to individual end-organs, and the potential for unanticipated off-target toxicity may prove a major challenge to these approaches.
Figure 2. Molecular targets for selective manipulation of migration of alloreactive immune effectors. GvHD, graft-versus-host disease; GvT graft-versus-tumour; GI Tract, gastrointestinal tract; SLO, secondary lymphoid organ. Red text denotes molecules involved in migratory pathways common to more than one organ whereas blue text denotes molecules involved in relatively organ-specific migration. CCR, C-C chemokine receptor; CXCR, C-X-C chemokine receptor; XCR, X chemokine receptor.
Download figure to PowerPoint
Influencing the balance of donor regulatory and effector T cells
Post-transplant immunosuppression, most commonly with calcineurin inhibitors and methotrexate, is currently used to prevent GvHD after AHSCT (Ram et al, 2009). An alternate approach is non-specific T cell depletion of grafts. However, both strategies have major off-target effects, increasing both infection and disease relapse post- transplant (Horowitz et al, 1990; Weaver et al, 1994; Chakrabarti et al, 2002).
Over the past decade there has been a resurgence of interest in T cell subsets with suppressive function, and in the role of such cells in both healthy immune responses and disease. Although thymus-derived CD4+ ‘natural’ regulatory T cells (Tregs) are best characterized, Tregs may also be induced from CD4+ effector T cells (Teffs) in the periphery by Transforming Growth Factor-β (TGF-β) or tolerogenic APCs (Sakaguchi et al, 2010). Multiple other suppressive cell subsets have also been identified, including IL10-secreting Tr1 CD4+ cells and CD8+ Ts cells (Reibke et al, 2006; Roncarolo et al, 2006). Suppressive T cells represent an exciting therapeutic opportunitiy to prevent GvHD post-transplant with potential for retention of pathogen-specific immunity and GvT effects.
Natural Tregs constitute only 5–10% of the total CD4+ T cells in peripheral blood (Wang et al, 2011). These cells possess the capacity to suppress proliferative responses of Teffs and play a pivotal role in maintenance of peripheral tolerance and immune homeostasis. Studies have shown adoptive transfer of large numbers of allogeneic CD4+ Tregs and Teffs cells can prevent (but not reverse) GvHD in mice, without abrogating GvT effects (Hoffmann et al, 2002; Taylor et al, 2002; Edinger et al, 2003). More recently, human CD4+ Tregs were shown to suppress effector T cell (Teff)-mediated alloresponses in vitro (Godfrey et al, 2004) and also reduce GvHD in a xenogeneic GvHD murine model (Cao et al, 2009). Naive CD4+ Tregs appear to home to SLOs much like Teffs, where they may suppress the activation and proliferation of the latter cells (von Boehmer, 2005). Furthermore, CD4+ Tregs may also regulate the later stages of immune responses as, after priming in SLOs, these cells express chemokine receptors enabling them to migrate to inflamed tissues. Importantly, CD4+ Tregs appear to influence both GvHD and GvT effects after human AHSCT. Several studies have suggested that both acute and chronic GvHD is associated with reduced numbers of donor CD4+ Tregs in either the graft or the recipient peripheral blood post-transplant (Miura et al, 2004; Zorn et al, 2005; Rezvani et al, 2006). More recently, studies have suggested that the relative balance between donor-derived CD4+ Tregs and Teffs may be closely associated with acute GvHD (Matthews et al, 2009). In addition to quantitative differences, new data suggest there may also be qualitative defects in peripheral blood CD4+ Tregs in patients who develop GvHD after AHSCT. Transcriptome analysis revealed functional differences in CD4+ Tregs from patients with and without GvHD after AHSCT. Tregs from patients without GvHD expressed higher levels of chemokine receptors enabling migration to GvHD target organs and molecules conferring functional suppressive activity (Ukena et al, 2011). This observation is supported by a number of studies showing reduced numbers of CD4+ Tregs in tissues affected by GvHD compared to control biopsies (Rieger et al, 2006; Fondi et al, 2009).
Given these observations, attempts to utilize CD4+ Tregs to control human alloresponses after AHSCT have been made. Initial strategies have focussed on adoptive transfer of donor Tregs during or after AHSCT. These approaches rely on polyclonal expansion of Tregs ex vivo to achieve sufficient numbers (Godfrey et al, 2004; Hoffmann et al, 2004; Hippen et al, 2011a). Two clinical studies have so far reported the use of ex vivo expanded donor Tregs in the context of AHSCT, both reporting encouraging control of acute GvHD and improved immune reconstitution (Brunstein et al, 2011; Di Ianni et al, 2011). However, the potential loss of GvT effects after infusion of polyclonal donor Tregs with non-specific suppressive capacity is a cause for some concern. These studies were not designed or powered to determine if Treg infusion had an impact of GvT effects or disease relapse. However, CD4+ Treg content within DLI administered after AHSCT to patients with relapsed CML had a negative correlation with disease response rate, suggesting that Tregs can suppress the allogeneic GvT effect in this context (Hicheri et al, 2008). To further address this issue, strategies to deplete DLI of Treg have recently been used to improve the GvT effect of alloreactive T cells (Maury et al, 2010).
It is therefore clear that to effectively modulate Treg and Teffs to reduce GvHD without abrogating GvT a more subtle strategy than infusion of polyclonal donor CD4+ Tregs will be required. Human alloantigen-specific CD4+ Tregs suppress alloresponses more effectively and specifically than polyclonal Treg (Albert et al, 2005; Peters et al, 2008). This may represent a strategy to refine the specificity of CD4+ Treg to alloantigens, potentially sparing tumour-associated antigen-specific responses that could mediate some GvT effects. A two-step protocol for the ex vivo expansion of alloantigen-specific CD4+ Tregs has recently been developed. Activation of Tregs with allogeneic dendritic cells is followed by flow cytometric-based purification of activated Tregs, which are further expanded with high doses of IL2. Using this approach Tregs could be expanded approximately 1000-fold and retained specificity for alloantigens (Sagoo et al, 2011). In a different study, Tregs were effectively expanded with allogeneic dendritic cells in the presence of IL2, IL15 and rapamycin, retaining much greater alloantigen-specific suppressive capacity than their polyclonal counterparts (Veerapathran et al, 2011). Although such approaches may refine the suppressive capacity of donor Treg to alloantigens, there still remains the potential for loss of a significant part of the GvT effect – that mediated by donor Teffs directed against alloantigens present on recipient tumour cells. One approach to mitigate this problem would be to preferentially direct the migration of allosuppressive Tregs to sites of GvHD, such as the gut, by up-regulation of gut-specific chemokine receptors or by redirection of Tregs by means of transduced organ-specific antigen receptors. One such strategy has been successfully applied in a murine model of gut inflammation, providing proof of principle that Treg migration can be directed to organ-specific sites (Menning et al, 2010).
In addition to the potential for loss of some GvT effects, there are currently several technical barriers to the use of ex vivo expanded Tregs to control alloresponses after AHSCT. These include proliferation of contaminating Teffs during ex vivo expansion, necessitating very high stringency purification of Tregs pre-expansion. Purification of human Tregs has not previously been easy due to the absence of a specific surface phenotype. Initially defined by high expression of the α chain of the IL2 receptor (CD25), expression of the nuclear transcription factor Forkhead Box P3 (FOXP3) is required for the suppressive function of both natural and induced CD4+ Tregs (Albert et al, 2005). However, both CD25 and FOXP3 are up-regulated on activated human CD4+ Teffs making these markers unsuitable for Treg identification. Relatively recently, expression of the IL7 receptor α-chain (CD127) was found to differentiate between activated Teffs and suppressive Tregs, which will greatly facilitate cell therapy approaches using Tregs (Liu et al, 2006; Seddiki et al, 2006). Addition of rapamycin to ex vivo cultures, which preferentially decreases proliferation of Teffs, provides an alternative approach to further improve Treg purity for clinical use (Hippen et al, 2011b). Loss of stable FOXP3 expression and suppressive function upon ex vivo expansion of Tregs is another significant challenge. Given these technical barriers to effective ex vivo expansion of donor Treg, several strategies are in development to manipulate donor T cells pre- or post-transplant to favour the expansion of tolerogenic donor Treg in vivo post-transplant. Such strategies include the transplantation of alloanergized donor T cells, which results in expansion of allospecific Tregs in vivo post-transplant (Davies et al, 2009), and the use of low dose IL2 to restore defective Treg numbers in patients with chronic GvHD (Koreth et al, 2011).
One remaining concern common to all strategies aiming to modulate GvHD and GvT with Treg is the recently identified phenomenon of plasticity between CD4+ Treg and Teffs. Under certain pro-inflammatory conditions, Treg can change their phenotype to Teffs (Bailey-Bucktrout & Bluestone, 2011). This obviously raises the potential for the loss of suppressive activity of Tregs post-transplant. New strategies to stabilize the suppressive phenotype of Treg by epigenetic modulation of FOXP3 are being tested in the transplant setting and an early study has shown that use of azacitidine in vivo has potential to expand CD4+ Treg numbers after AHSCT (Goodyear et al, 2012).
In summary, modulation of the balance of suppressive and effector T cells represent an exciting potential new approach to influence GvHD and GvT after AHSCT. After overcoming initial technical barriers related to ex vivo expansion, donor CD4+ Tregs have recently been used with some success to reduce GvHD. However, concerns regarding the potential for suppression of GvT effects and plasticity of T cell subsets resulting in loss of suppressive activity will represent major challenges to the further development of these approaches.