Augmentation of Anti-tumor Immunity by Adoptive T-cell Transfer After Allogeneic Hematopoietic Stem Cell Transplantation

Marie Bleakley; Cameron J Turtle; Stanley R Riddell


Expert Rev Hematol. 2012;5(4):409-425. 

In This Article

Strategies to Improve the Persistence of Donor T Cells Transferred After Allogeneic HCT

One of the challenges in autologous adoptive T-cell immunotherapy is elucidating the factors that govern the persistence of transferred T cells. Although formidable in the autologous setting, the challenges to achieving durable engraftment may be even greater when in vitro expanded allogeneic T cells are transferred after HCT to patients who may be receiving immunosuppressive drugs to prevent or treat GVHD. Strategies to generate and support T cells that are not exhausted from prolonged in vitro culture and restimulation are being developed; however, the allogeneic setting provides many pitfalls that hinder the implementation of the same approaches used in autologous T-cell transfer. Here, the authors will review some of the current approaches to improve in vivo survival of adoptively transferred T cells and discuss their applicability after allogeneic HCT.

Selection of T-cell Subsets for Therapy

Heritable cell-intrinsic qualities are increasingly being recognized for their importance in determining the capacity of effector T cells to persist in vivo after in vitro stimulation, culture and adoptive transfer. In conjunction with gene transfer techniques to redirect T-cell specificity to tumor antigens, this now makes tumor immunotherapy using distinct subsets of engineered T cells both desirable and feasible. Conventional TCR αβ+ T cells for adoptive transfer can be isolated from the naive T-cell subset (TN; CD45RA+ CD45RO CD62L+) or from the antigen-experienced memory T cell (TM) subset, which, in turn, can be broadly partitioned into central memory (TCM; CD45RO+ CD62L+) or effector memory (CD45RO+, CD62L) subsets.[81] Studies in the laboratory in nonhuman primates and in mouse models have demonstrated that CMV-specific CD8+ effector T cells (TE) derived from TCM cells have a superior capacity to persist in vivo after adoptive transfer compared with those derived from T effector memory cells.[82,83] CD8+ TN cells have also been proposed as an effective precursor population from which to derive TE cells for adoptive transfer.[84] However, TN cells may be more prone to induce GVHD after allogeneic HCT than their TM counterparts, suggesting this subset is best avoided in the allogeneic setting in the absence of a proven suicide mechanism.[85,86] Recent studies have also identified a distinct subset of CD8+ memory T cells, termed memory stem cells that like TCM, express CD62L and CD95, but have retained expression of CD45RA, and demonstrated that TE cells derived from this subset may be superior to TCM.[87] This subset clearly has promise for use in gene transfer and adoptive therapy but clinical applications in allogeneic HCT will require improved methods for rapidly isolating these rare T cells, as well as careful analysis of the potential for alloreactivity. Taken together, these data suggested that CD8+ TCM-derived TE cells selected for a defined endogenous TCR specificity to avoid alloreactivity may be optimal for initial studies in the allogeneic HCT setting.

Much of the focus in cancer immunotherapy has been on CD8+ T cells; however, animal models show that CD4+ T cells facilitate the persistence and function of adoptively transferred CD8+ T cells in part through production of IL-2.[88] Encouraging clinical responses to coinfusion of CAR-modified CD4+ and CD8+ T cells have been observed in human trials, although the role of individual CD4+ subsets in anti-tumor activity has not been dissected.[78,79,89] We have evaluated the coinfusion of distinct subsets of CD4+ T cells with CD8+ CAR-modified T cells in vitro and in animal models and demonstrated differences in their capacity to promote proliferation of tumor-specific CD8+ cells [Hudecek M, Pers. Comm.]. The capacity to genetically retarget T cells with either a CAR or TCR provides the opportunity to selectively endow defined subsets of T cells with the ability to recognize a common cognate tumor antigen,[90] facilitating the analysis of cell products that contain both CD8+ and CD4+ T-cell subsets in future immunotherapy protocols.

Bispecific T Cells

Adoptive transfer of allogeneic donor T cells with specificity for viral antigens can reconstitute persistent immunity to pathogens after allogeneic HCT and does not increase the incidence of GVHD.[91,92] This suggests that virus-specific T cells that are isolated from the donor and genetically retargeted to a tumor antigen may provide durable tumor immunity with minimal risk of GVHD compared with infusion of unselected TN or TM subsets. An additional advantage is that such T cells could receive proliferative and survival signals due to in vivo viral reactivation that are independent of tumor-derived antigen and the tumor microenvironment. Methods for rapidly selecting and transducing virus-specific T cells have been developed by our group, and this approach will soon be tested after allogeneic HCT using donor EBV- and CMV-specific TCM-derived TE cells that are engineered to express a CD19-specific CAR (Figure 1). This approach could also be used in conjunction with TCR gene transfer to confer MHC-dependent tumor recognition.[80]

Optimizing HCT Platforms to Reduce Post-HCT Immunosuppression

An obstacle that could impede the persistence, function and efficacy of adoptively transferred tumor-reactive T cells after allogeneic HCT with T-cell replete bone marrow or peripheral blood stem cell (PBSC) grafts is the pharmacological immunosuppression that is required to prevent and treat GVHD. This limitation could be overcome by utilizing stem cell grafts that are depleted of alloreactive T cells, thus eliminating the risk of GVHD and the need for immunosuppressive drugs. One approach is to employ adoptive transfer of tumor-reactive T cells in settings where complete T-cell depletion of the stem cell graft is used to prevent GVHD. However, complete T-cell depletion results in delayed immune reconstitution and a high risk of opportunistic infection. The authors are conducting a clinical trial to evaluate the selective depletion of CD45RA+ TN cells from PBSC grafts for the prevention of GVHD. This trial was prompted by results in several different murine HCT models that showed that the allogeneic stem cell grafts that contained TN cells caused severe GVHD compared to those that contained only TM cells.[85,93–97] Limiting dilution analysis of human T-cell subsets showed that the frequency of minor H antigen-specific T cells was significantly higher in the TN compared with the TM subset, suggesting that selective TN-depletion may similarly reduce GVHD in humans while preserving the transfer of pathogen-specific immunity contained in the TM pool.[86] The authors have developed a clinical scale procedure for rigorously depleting PBSC of TN and demonstrated that CD8+ and CD4+ T-cell responses to a range of pathogens are retained in the TN depleted PBSC. The strategy of CD45RA+ T-cell depletion for reducing GVHD in humans is now being tested in a clinical trial, and if successful, will provide an improved platform for adoptive immunotherapy with tumor-specific T cells.

Several other approaches to selective depletion of alloreactive T cells from stem cell grafts have been evaluated in clinical trials.[98–106] In vitro allodepletion strategies typically involve allostimulation of the T-cell component of the HCT graft in vitro followed by the depletion of cells expressing activation markers using immunomagnetic beads, immunotoxins or florescent-activated cell sorting or by photodynamic purging.[107] In two clinical trials, donor lymphocyte infusions were allodepleted using an anti-CD25 immunotoxin and administered following a T-cell-depleted haploidentical HCT. The incidence of significant acute and chronic GVHD was low and functional T-cell responses were observed in patients receiving ≥105 allodepleted T cells/kg.[98,101] In HLA-identical donor HCT, the administration of approximately 108 anti-CD25 allodepleted T cells/kg with CD34+ stem cells resulted in acute GVHD in 46% of patients suggesting that more efficient allodepletion strategies may be required in this setting.[100] Photodynamic purging of alloactivated T cells was also found to be inefficient and incompletely effective in preventing GVHD following HLA-identical HCT.[103–105] Further improvement in the efficacy of selective in vitro allodepletion strategies may be achieved by using immunomagnetic beads or immunotoxins conjugated to multiple activation markers,[108] or by combining in vitro allodepletion with a second safety mechanism such as transfection of the T cells in the PBSC product with a suicide gene (see the section 'Strategies to eliminate transferred T cells').[106]

Transplant strategies for the depletion of alloreactive T cells in vivo are also being investigated. One such approach is the administration of cyclophosphamide early after allogeneic HCT to promote tolerance in alloreactive host and donor T cells leading to suppression of both graft rejection and GVHD after allogeneic HCT. High-dose cyclophosphamide on days 3 and 4 after infusion of unmanipulated bone marrow results in low rates of severe acute and chronic GVHD both in nonmyeloablative haploidentical HCT, in which case additional pharmacological immunosuppression is required following cyclophosphamide, and in myeloablative HLA-matched related or unrelated donor bone marrow transplant, in which case additional anti-GVHD pharmacotherapy is not essential.[109–111] Building on the posttransplant cyclophosphamide backbone, additional novel methods of inducing immunological tolerance in the allogeneic HCT setting are under investigation. A recent publication reported minimal GVHD and concurrent acceptance of renal allografts, among five patients who received haploidentical hematopoietic stem cells and 'facilitating cells' with a single dose of post-HCT cyclophosphamide and additional short-term pharmacological immunosuppression, although unfortunately the derivation or composition of the stem cell or facilitating cell products were not described in detail.[112]

Generation of T Cells That are Resistant to Post-HCT Immune Suppression

Several groups, including the authors', are developing strategies to endow T cells with resistance to immune-suppressive drugs including calcineurin inhibitors, mycophenolate mofetil and corticosteroids to enable their use in post-HCT settings where immunosuppression is necessary. T cells have been rendered resistant to calcineurin inhibitors using siRNA to knock down the 12 kDa FKBP12 or by engineering T cells to express calcineurin mutants.[113,114] Similarly, introduction of a variant inosine monophosphate dehydrogenase (IMPDH2) gene has been used to render T cells resistant to mycophenolate mofetil.[115] The authors' research group has employed zinc-finger nuclease technology to knock out the glucocorticoid receptor in T cells to render them resistant to corticosteroids used in the treatment of GVHD [Gardner R, Pers. Comm.]. These approaches, which entail some risk, continue to be evaluated in animal models and have not yet been subjected to clinical evaluation.

Improving Persistence Through Cytokine Support & Lymphodepletion

Once activated and expanded ex vivo, T cells are susceptible to apoptosis, and many clinical trials have administered IL-2 in vivo after cell transfer to promote T-cell survival and persistence. However, IL-2 administration is nonselective and promotes expansion of regulatory T cells that could interfere with anti-tumor activity of transferred T cells. Strategies to increase autocrine cytokine production, such as inclusion of 'in-line' costimulatory domains in CAR-modified T cells may be preferable to exogenous cytokine support.[116] Alternative approaches for preferential signaling in transferred T cells include transduction of T cells with cytokine encoding genes,[117] chimeric cytokine receptors designed to amplify cytokine signaling and provision of cytokines in cell-bound nanoparticles.[118]

The γ-chain cytokines other than IL-2 may also have a role, as suggested by the improvement in T-cell persistence observed after the deliberate induction of lymphopenia to augment endogenous levels of IL-7 and IL-15 prior to autologous T-cell therapy.[119,120] In the HCT setting, conditioning therapy and infusion of allogeneic grafts should increase the availability of cytokines; however, in contrast to the autologous setting, additional exogenous systemic cytokine support after allogeneic HCT has greater potential to be detrimental due to the risk of cytokine-induced proliferation of donor alloreactive T cells and induction of GVHD.


The capacity of T cells to proliferate and persist in response to antigen in vivo is in part dependent upon their expression of costimulatory molecules, such as CD28, 4-1BB or OX40. To optimize the persistence of adoptively transferred T cells, one can select T-cell subsets that constitutively express costimulatory molecules or alternatively, costimulatory molecules alone or with their corresponding ligands can be introduced by genetic modification. Human CAR-modified T cells that incorporate a CD28 costimulatory domain exhibit better persistence after autologous T-cell transfer than CAR-modified T cells expressing only a CD3z endodomain.[121] The incorporation of a 4-1BB costimulatory domain in a CD19-specific CAR may be one of the factors responsible for the impressive expansion of transferred T cells and clinical efficacy in advanced CLL.[78,79] Much remains to be done to define the optimal design of CARs to increase survival, proliferation and effector function, but the ability to evaluate signals from molecules such as CD28, 4-1BB, OX40 and NKG2D to overcome the absence of costimulatory molecule expression on tumor cells may be a key advantage for CAR-modified T cells.