Biology of the graft-versus-tumor effect following hematopoietic cell transplantation

INTRODUCTION — The majority of patients with malignancy who undergo hematopoietic cell transplantation (HCT) are effectively treated, thereby resulting in minimal residual disease. However, this response is frequently not maintained since relapse ultimately occurs in 40 to 75 percent of patients who undergo an autologous transplant and 10 to 40 percent of those who undergo an allogeneic transplant. Further, with the development of non-myeloablative or reduced intensity allogeneic transplantation there is increased reliance on immune-mediated effects to control the underlying disease.

The rationale for using immunotherapy to prevent and/or treat the reemergence of malignancy is based upon the following observations:

●Evidence indicates that the graft-versus-tumor (GVT) effect plays a major role in reducing the risk of relapse following an allogeneic transplant.

●Significant advances have been made in our basic understanding of both the cellular populations responsible for potential antitumor activity and the cellular interactions and cytokines required for their activation and expansion.

The cell populations capable of recognizing and lysing malignant targets can be divided into two broad categories based upon the mechanism of cellular recognition: cytotoxic T cells (CTLs) and natural killer (NK) cells. Significant insights have been made into the functional mechanisms of these two populations.

This topic review will discuss the basic principles underlying the pathogenesis of the GVT effect after HCT. An overview of some of the clinical trials examining immunotherapy to prevent and treat relapse following HCT is presented separately. (See "Immunotherapy for the prevention and treatment of relapse following allogeneic hematopoietic cell transplantation".)

CYTOTOXIC T CELLS — Before discussing how cytotoxic T cells (CTLs) recognize and become activated in response to malignant cells, it is helpful to first briefly review the major histocompatibility complex (MHC), and how T cells recognize and are activated by any antigen. Genes of the MHC (called HLA in humans) encode two distinct classes of cell surface molecules, I and II. Class I MHC molecules are expressed on the surfaces of virtually all nucleated cells at varying densities, while class II MHC molecules are more restricted to cells of the immune system, primarily B lymphocytes, monocytes, and dendritic cells.

Since a major function of the immune system is to distinguish self from non-self, HLA molecules provide the crucial surface upon which the antigen receptors on T lymphocytes (T cell receptors or TCRs) recognize foreign (non-self) antigens. On antigen presenting cells (APCs), such as macrophages and dendritic cells, class II MHC molecules present antigenic fragments (in the form of linear peptides) to the CD4+ inducer (or helper) T cells, while class I MHC molecules function at the effector phase of immunity by presenting antigens to CD8+ T cells. This process of antigen presentation consists of the binding of a single T cell receptor to a complex on the surface of an antigen-presenting cell consisting of the MHC molecule and a peptide fragment derived from the host or foreign antigen. Due to the central role of MHC molecules in initiating an immune response, it is not surprising that matching of these genes is the major predictor of graft-versus-host disease in the setting of allogeneic transplantation. (See "Major histocompatibility complex (MHC) structure and function" and "Pathogenesis of graft-versus-host disease (GVHD)".)

In addition to engagement of the T cell receptor with the MHC molecule-antigenic peptide complex, several other signals and/or functions are required for full T cell activation and expansion:

●APCs must express costimulatory molecules that provide a mandatory second signal to the T cell. The major pathways are an interaction of the B7 family of molecules (B7-1 [CD80] or B7-2 [CD86]) or of CD58 (LFA-3) on antigen presenting cells with CD28 or CD2 on T cells (figure 1) [1,2]. Other co-stimulatory molecules (eg, CD30:CD30L, OX40) have also been identified. Anergy, a state in which T cells are unable to become activated in response to their normal target antigen, develops in the absence of costimulation [2-4].

●Helper T cells are also necessary for a normal immune response. They are primarily CD3+CD4+CD8-, compared with the majority of CTLs which are CD3+CD8+CD16-. Helper T cells release cytokines and provide cellular interactions (via costimulatory molecules) necessary for the appropriate stimulation of APCs (figure 1) [5].

Traditional understanding dictated that CD4+ T cells were required to produce cytokines (principally IL-2) in the local environment of the CD8+ T cells, thereby allowing for clonal expansion. This hypothesis required two populations of effector cells to be present and bind to an APC. However, studies indicate that CD4+ T cells are capable of activating APCs via engagement of CD40, which primes the APCs for more effective stimulation of CD8+ cells. These observations simplify the role of helper T cells and do not require a physical presence at the time of CD8+ T cell activation.

Regulatory T cells — Other populations of T cells regulate immune responses. The best characterized population is that of regulatory T (Treg) cells, which express CD4 and CD25 as well as the transcription factor FoxP3 [6]. Treg cells reduce T cell proliferation and may play a major role in controlling graft-versus-host disease (GVHD) in murine model systems [7] as well as in man [8]. Importantly, Treg cells did not inhibit graft-versus-tumor (GVT) reactions in mouse models [9,10]. Further, following the adoptive transfer of regulatory T cells in the setting of haploidentical transplantation, relapse rates remain low [11].  

Invariant natural killer T (iNKT) cells — A unique subset of natural killer like T cells that express an invariant T cell receptor have also been recognized that have both cytotoxic and immunoregulatory properties. Several groups have indicated that this population of cells is capable of exerting an anti-tumor effect and also regulate immune responses such as the suppression of GVHD [12,13]. Subsets of iNKT cells exist that have been shown in murine models to suppress GVHD on the one hand (iNKT2, and iNKT17 subsets) or have cytotoxic function (iNTK1) [14]. These cells may also stimulate endogenous immunity which could be enhanced by the introduction of a CAR construct [15]. Clinical translation of these concepts are underway.

Natural killer (NK) T cells also play an important immunoregulatory role under certain biological conditions, as discussed below. (See 'Natural killer cells' below.)

Recognition of malignant cells — For a donor CTL to recognize and lyse a malignant host target cell, antigen(s) expressed by the tumor must be recognized as foreign. Based upon the type of transplant, this identification principally occurs via the following two mechanisms:

●Donor CTLs recognize alloantigens, possibly minor histocompatibility antigens, on host cells. This is of principal importance with allogeneic transplants.

●CTLs also react against unique antigens expressed by the malignant host cells [16]. This can occur with any donor-recipient pair, as well as outside the transplant setting.

Self versus non-self — With nonidentical transplants, donor CTLs may recognize neoplastic (and normal) cells as foreign due to the expression of epitopes unique to the host. CTLs may subsequently become activated, lysing such cells.

This mechanism for a GVT effect is very similar to that which underlies GVHD. When the recipient and nonidentical donor are not matched at the MHC (eg, with mismatched related or unrelated, haploidentical or cord blood donors) GVHD (or GVT) may result because of differences within the major MHC antigens. However, when the donor-recipient pair are matched for MHC antigens, GVHD (or GVT) is initiated and propagated via recognition by the T cell and its receptor of additional antigens called minor histocompatibility antigens [17]. A major question in the field is whether the repertoire of T cells responsible for GVHD is the same as those responsible for GVT, and if not how much overlap between these two populations exists. Another question is whether there are truly disease-specific determinants that can be recognized by donor derived T cells. The use of high throughput sequencing of the TCR may help clarify how these different T cell populations interact in these complex biological reactions [18]. (See "Pathogenesis of graft-versus-host disease (GVHD)".)

The importance of this effect for the development of GVT is suggested by the different outcomes with grafts from identical twin and nonidentical donors. Patients who have an identical twin donor do not develop GVHD. However, they are at a higher risk of relapse of the underlying malignant disease than similar patients transplanted with HLA-matched but nonidentical sibling donors [19]. This is presumably due to the lack of a GVT reaction [20].

These relationships were illustrated in a review from 163 transplant centers that compared the results of 103 identical twin (syngeneic) and 1030 HLA-identical sibling (allogeneic) transplants for leukemia [21]. The three-year probability of relapse of leukemia was substantially higher in the syngeneic than allogeneic transplants in acute myeloid leukemia (52 versus 16 percent) and chronic myeloid leukemia (40 versus 7 percent).

The importance of the pathogenic link between GVT and GVHD is also suggested by the finding that GVHD can be induced by donor lymphocyte infusions. With this technique, lymphocytes from the original donor are administered to patients with relapsing disease to attempt to elicit a GVT reaction; this benefit is tempered by the enhanced risk of developing GVHD. (See "Immunotherapy for the prevention and treatment of relapse following allogeneic hematopoietic cell transplantation".)

However, the observation that some patients clearly have GVT without GVHD suggests that these two processes may result via different mechanisms (see 'Possible separation of GVT from GVHD' below).

Expression of tumor specific antigens — CTLs, independent of origin, can also recognize malignant cells as foreign when unique antigens are expressed by the tumor [22,23]. Potential antigenic targets include the following (table 1):

●Unique idiotypes or TCRs, which result from rearranged immunoglobulin or TCR genes, respectively; this may occur in some B and T cell malignancies [24].

●Foreign proteins encoded by the viral genome in malignancies that occur in association with viral infections, such as EBV-associated B cell lymphomas.

●Specific peptides derived from proteins expressed on tumor cells such as the her-2 neu protein, a product expressed by a subset of breast cancers and neuroblastomas [25].

●Products of specific oncogenes (the importance of which is unclear). Peptides derived from breakpoint regions of bcr/abl, an oncogene found in the majority of patients with chronic myeloid leukemia, have been used to stimulate a productive T cell immune response in vitro [26,27].

●A variety of lineage-specific antigens, including minor histocompatibility antigens (which are found on hematopoietic tumors), the MAGE proteins (on melanomas), proteinase III, WT-1 and Wnt (on certain leukemias), and carcinoembryonic antigen (on some solid tumors).

Ongoing investigations are focusing on those characteristics of the antigen that would best enable CTLs to recognize and lyse malignant targets. No consensus has emerged concerning the optimal nature and source of these antigens, particularly whether they should be in the form of peptide(s), a whole protein, DNA, or cDNA, or whether they should be introduced into dendritic cells by gene transfer [28].

Other therapeutic maneuvers include the following:

●Donor lymphocyte infusion (DLI). (See "Immunotherapy for the prevention and treatment of relapse following allogeneic hematopoietic cell transplantation", section on 'Donor lymphocyte infusion (DLI)'.)

●Expansion of dendritic cells from leukemic precursors, based upon the assumption that these cells will present leukemic specific antigens [29].

●Adoptive T cell transfer by infusion of activated/expanded T cells, which may promote a cytolytic response against tumor cells and enhance vaccination strategies, especially in the lymphopenic environment early after transplantation [30,31].

Introduction of a chimeric antigen receptor (CAR) construct into the donor T cells have been shown to have clinical activity in those patients who relapse following allogeneic HCT [15].

Malignant cells and costimulation — Tumor cells may lack the costimulatory molecules necessary for such an interaction [32], which, as noted above, may lead to anergy [2-4]. A variety of strategies have been employed in an attempt to overcome this problem:

●In one study, follicular lymphoma cells were treated with CD40L, a protein known to upregulate the expression of costimulatory molecules of the B7 family [33]. Using this approach, autologous tumor-infiltrating CTLs could be generated in the presence of interleukin (IL)-2; they could be further expanded in vitro with IL-4, IL-7 and interferon gamma (IFNg).

●Another approach stimulates malignant B cells through the toll-like receptor complex using CpG or another immunostimulant to induce more effective antigen presentation and then to re-inject these cells following irradiation as a therapeutic vaccine.

●Another strategy has been to transfect tumor cells with costimulatory genes. In animal models, injection of transfected tumor cells resulted in an effective immune response against both the transfected and wild type tumor cells [34,35].

The identification of dendritic cells and their role in stimulating an immune response is a major scientific advance. Clearly, much work is required to identify not only the appropriate clinical situation but also to develop the necessary approach that is effective and technically feasible.

Checkpoint inhibitors — T cell activation, proliferation, and function are influenced by a number of different receptor ligand interactions. Some of these interactions downregulate T cell responses and can be broadly considered physiologic "brakes" or "checkpoints" on the T cell activation that is normally triggered by antigen-presenting cells. Release of this brake with checkpoint inhibitors may augment the native or graft T cell anti-tumor response. Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and programmed cell death 1 (PD-1) are examples of checkpoints that have been targeted clinically. (See "Principles of cancer immunotherapy", section on 'Checkpoint inhibitor immunotherapy'.)

Checkpoint inhibitors have demonstrated efficacy in solid tumors, and initial studies in hematologic malignancies (eg, Hodgkin lymphoma) have demonstrated responses in highly pretreated patients. (See "Treatment of relapsed or refractory classic Hodgkin lymphoma", section on 'PD-1 blockade'.)

There are limited data regarding the use of checkpoint inhibitors in the transplant setting. A phase 1 study evaluated the CTLA-4 inhibitor ipilimumab in 28 patients with relapsed hematologic malignancy after allogeneic HCT [36]. Administration was feasible, and durable responses were observed in some patients. Six patients (21 percent) had immune-related adverse events, including one death. These included GVHD, immune thrombocytopenia, colitis, and pneumonitis.

Further studies are underway regarding the use of checkpoint inhibitors to stimulate a more effective GVT response. A significant concern in the setting of allogeneic transplantation is whether this approach could result in worsening of GVHD.

NATURAL KILLER CELLS — Natural killer (NK) cells are commonly referred to as nonspecific or major histocompatibility complex (MHC)-nonrestricted. However, these terms are misnomers, since NK cells express receptors, which recognize MHC molecules.

The cell surface phenotype of NK cells is CD3-CD56+CD16+, although other cell populations have been described. These cells can be activated by a variety of different cytokines, particularly IL-2 [37,38].

Cell receptors — The mechanism of target cell recognition by NK cells is not completely understood [39]. Several groups have identified cell surface receptors expressed by NK cells, which recognize MHC molecules and, upon productive interaction, initiate an inhibitory signal that prevents lysis.

In humans, two distinct families of inhibitory receptors have been characterized:

●One group of proteins is referred to as the killer inhibitory receptors (KIRs); they are composed of either two or three extracellular immunoglobulin-like domains, and recognize specific class I HLA molecules [40-43]. The role of peptides in class I recognition by KIRs has been controversial, although it appears that some peptide specificity exists.

●A second structurally distinct class of receptors (CD94/NKG2A) has a C-lectin domain and recognizes peptides derived from certain class I alleles in the context of HLA-E [44,45].

NK cells also express activation receptors, which, upon triggering, engage the cytolytic machinery. The best characterized activation receptor is a molecule termed NKG2D. Ligands for NKG2D include molecules such as MICA, MICB and the ULB binding proteins [46]. NKG2D ligands are upregulated by cellular stress, including viral infection and transformation. Other natural cytotoxicity receptors (NCRs) have been identified that, upon stimulation, result in NK cell activation and more effective cytolysis.

In the setting of bone marrow transplantation, donor NK cells are inhibited upon interaction with an appropriate MHC class I molecule expressed by a normal host cell. As a result, all donor NK cells tolerate normal (eg, not malignant or infected) host cells expressing normal levels of MHC class I molecules. By comparison, tumor or virally infected cells, which either downregulate class I expression or have altered peptide expression, are recognized and killed [47].

Several aspects of NK function are unclear, including the exact requirements for an interaction to occur and what role activator receptors play [48]. Adhesion molecules appear to be required since monoclonal antibodies directed against LFA-1 and intercellular adhesion molecule (ICAM)-1 inhibit NK cytolytic activity.

NK cells also express the Fc-gamma receptor (CD16); this allows cells bound by immunoglobulin to be lysed by antibody dependent cellular cytotoxicity mechanisms. This mechanism of tumor cell removal is thought to play an important role in the mechanism of therapeutic antibodies. Individuals who express different polymorphisms of the Fc-gamma receptor may differ in their response to rituximab and presumably other therapeutic monoclonal antibodies [49,50].

Autologous and MHC-matched transplants — The mechanisms by which tumor cells are recognized by NK effector cell populations with autologous and MHC-matched transplants are not completely understood. As previously mentioned, recognition and killing of such cells may occur if the associated KIRs or CD94/NKG2A molecules are not bound by the appropriate HLA-class I molecules. By comparison, nonmalignant autologous or MHC-matched cells presumably express HLA class I molecules with the appropriate peptide; this permits binding to the inhibitory receptor expressed on the surface of the NK cell, thereby inhibiting NK cell function.

This hypothesis assumes that HLA class I molecules on the surface of tumor cells are either downregulated or alternative peptides are displayed, characteristics that block the ability of HLA class I molecules to signal through the appropriate inhibitory receptor. Experimental evidence for the lack of HLA class I expression has been found in a number of different tumor types. Alteration in peptide expression is more difficult to evaluate, but it has been observed following certain viral infections. Alternatively, expression of NKG2D ligands has been observed on a number of solid tumors and hematopoietic cell malignancies which may allow for appropriate cell recognition and lysis.

MHC-mismatched transplants — A low incidence of relapse has been observed with MHC-mismatched transplants, and NK cell-mediated cytotoxicity may contribute to GVT in MHC-mismatched transplants without associated GVHD.

Depending upon the donor-recipient pair, host cells may lack the appropriate HLA class I molecule for binding and inhibiting inhibitory molecules expressed on donor NK and NK-like T cells. Such alloreactive NK cells, which recognize recipient tumor cells based upon the lack of inhibition through the appropriate KIR molecules, have been observed both before and after mismatched transplantation [51]. Alloreactive NK cells that do not induce (and may even suppress) GVHD are observed in animal models of adoptive transfer and in human transplantation with HLA-mismatched NK cells, including haploidentical transplants [52-56].

Cytokine production — NK cells secrete a variety of cytokines, such as interferon-gamma, tumor necrosis factor (TNF), granulocyte macrophage colony-stimulating factor (GM-CSF), and transforming growth factor (TGF)-beta. These cytokines may have some direct effects on the tumor cells and may promote recruitment of other populations of effector cells.

Activation and expansion — A number of strategies have been employed to activate and expand cells with NK cell function. These cells appear to be activated with cytokines, particularly IL-2, IL-12, and IL-15 [57,58]. To enhance the antitumor function of NK cells, different methods of administering IL-2, both ex vivo and in vivo, have been used:

●Many groups have used IL-2 for the ex vivo activation of NK cells, thereby resulting in cells with enhanced anti-tumor cell activity (termed lymphokine activated killer [LAK] cells). However, since NK cells do not expand effectively ex vivo, it is difficult to obtain the numbers of cells required for treatment, thereby resulting in limited efficacy.

●The direct administration to the patient of high-dose IL-2 is associated with significant morbidity and limited enhancement in NK cell antitumor efficacy [59].

●A randomized study of IL-2 following autologous transplantation did not improve outcomes [60], perhaps because lower doses of IL-2 expand regulatory T cells, which may negatively impact anti-tumor responses [61,62]. In the absence of more definitive data such approaches should be used only in the context of a clinical trial.

Cytokine-induced killer (CIK) cells — The active expansion of cytolytic cells derived from T cells which share functional similarities with NK cells may be obtained with the use of specific cytokines [63,64]. These NK-like T cells, also called cytokine-induced killer (CIK) cells, express both T cell specific antigens (such as the T cell receptor and CD3 complex) and NK cell markers (such as CD56 or neural cell adhesion molecule) [65]. These ex vivo treated CD3+CD56+ cells rapidly expand following mitogenic stimulation through the T cell receptor (similar to T cells) and recognize tumor cell targets without prior exposure (similar to NK cells) [66]. They also produce cytokines such as GM-CSF, interferon, and tumor necrosis factor, but do not produce IL-4. NKG2D is an activating receptor on NK cells and CD8 T cells that allows for the recognition of tumor cell by these CIK cells [67,68].

Clinical trials using these cells have suggested that this approach is tolerable following autologous and allogeneic transplantation [69,70]. In one study using this methodology, improvement in recurrence-free survival was noted in patients undergoing hepatic tumor resection at high risk of relapse where those individuals who received cellular therapy had improved DFS compared with untreated controls [71]. CIK cells infected with the oncolytic vaccinia virus were particularly effective in animal models [15,72].

The different properties of these effector cell populations that may prove useful in immunotherapy posttransplantation are summarized in the table (table 2).

CYTOTOXIC CELL LINES — Several cell lines with cytotoxic activity have been developed which have phenotypic and biological characteristics very similar to NK cells and NK-like T cells. The cell line TALL-104 originated from a child with acute T cell lymphoblastic leukemia. These cells lyse a broad panel of tumor cell lines, but not normal tissues, and have in vivo activity in animal models [73,74]. The use of irradiated TALL-104 cells is in early phase clinical trials. Other NK-like cell lines have also been developed, such as NK-92, with intriguing biological activity [75-79].

Engineered T cells — A number of groups have investigated the use of engineered T cells, which express novel receptors that redirect their cytotoxicity. An example that has demonstrated efficacy is through the use of chimeric antigen receptor T cells (CAR-T) where the T cells are capable of recognizing tumor-associated antigens, such as CD19 and CD22. Early studies had limited success, primarily thought to be related to the poor survival of these engineered T cells. With novel constructions, including co-stimulatory molecules that improve survival of the adoptively transferred T cells, dramatic responses have been noted [80,81]. Use of engineered cells with defined and desired properties is an area of intense investigation.

MECHANISMS OF CYTOLYSIS — Upon appropriate stimulation, all effector cells use similar cytotoxic machinery. Two major cytolytic pathways have been described [82]:

●One involves the exocytosis of perforin and granzymes into the extracellular space at the site of target cell adhesion. Perforin induces pore formation in the membrane of the target cell, causing osmotic lysis and introducing granzymes and granulysin, which induce apoptosis [82-84]. The importance of these molecules in CTL and NK cell function has been demonstrated by the finding of greatly impaired effector cell function in mice deficient in these molecules [85,86].

●The second pathway involves the expression of fas ligand (fasL); this receptor binds to fas on the target cell, directly inducing apoptosis in a calcium-independent manner [87,88].

MECHANISMS OF TUMOR CELL ESCAPE — Insights into tumor cell recognition have identified a number of potential mechanisms by which tumor cells can escape immunologic recognition (table 3). These include:

●The absence of donor- or tumor-specific antigens or the downregulation of HLA class I molecules, preventing recognition by CTLs.

●The lack of costimulatory molecules, resulting in anergy.

●The expression of both fas and fasL, thereby inactivating T cells due to fas-mediated apoptosis [89,90]. In animal models, for example, fasL expressing tumors progress more rapidly in wild type mice than in animals with defective fas [89].

●The heterogeneity of human tumor cells and the possibility of ongoing mutations.

●The appropriate expression of self-MHC class I molecules, thereby inactivating NK cells.

●The expression of soluble factors that inhibit NK cells, such as the NKG2D ligands MICA and MICB [91]. These factors have been described in vitro, but their clinical significance is unclear.

An additional mechanism that may undermine immune surveillance is the presence of factors that impair the ability of effector cells to survive or to successfully home to malignant cells.

These dynamic interactions complicate immunotherapy applications. However, with increased understanding of these biological processes, more effective approaches will be developed which are based upon the underlying biology of the tumor cell and the immune system. Since patients harbor a state of minimal disease and relapse rates are predictable following hematopoietic cell transplantation, such individuals are ideal candidates to test novel immunotherapeutic approaches.

POSSIBLE SEPARATION OF GVT FROM GVHD — With nonidentical transplants, donor cytotoxic T cells (CTLs) may recognize neoplastic cells as foreign due to the expression of epitopes unique to the host (and the underlying malignancy). CTLs may subsequently become activated, lysing such cells. This graft-versus-tumor (GVT) effect is very similar to that which underlies graft-versus-host disease (GVHD). This theory is supported by evidence that the timing and intensity of immunosuppressive treatment directed at preventing and treating GVHD influences the risk of recurrent malignancy [92].

However, some patients have developed a GVT response without GVHD [93]; additionally, epidemiologic studies from the European Group for Blood and Marrow Transplantation suggest that these two effects may be separable [94].

●An experimental approach involves trying to enhance GVT but not GVHD. In this model, recipients were immunized against their tumor (tumor cell vaccine) one month after HCT [95]. The immunized recipients showed improved survival and protection against tumor growth; the immune response was tumor-specific as no exacerbation of GVHD was observed.

●Administration of defined doses of regulatory T cells, natural killer cells, or cytokine-induced killer cells has also resulted in control of GVHD and maintenance of GVT responses in murine models of leukemia and lymphoma [9,10,96,97]. Early phase clinical trials support these observations [98-100].

●Memory T cells are also under investigation as they appear to not cause GVHD and may have anti-tumor properties [101-104].

This separation of GVHD from GVT, which has been accomplished in animal models, remains an attractive clinical goal [9,66,105-117]. The successful application of these principles to patients will be a major advance in the field of hematopoietic cell transplantation.

SUMMARY

●Most patients with malignancy who undergo hematopoietic cell transplantation (HCT) achieve an initial complete remission. However, relapse ultimately occurs in 40 to 75 percent of patients who undergo an autologous HCT and 10 to 40 percent of those who undergo an allogeneic HCT.

●The rationale for using immunotherapy to prevent and/or treat the reemergence of malignancy (ie, relapse after HCT) is based upon evidence that the graft-versus-tumor (GVT) effect plays a major role in reducing the risk of relapse following an allogeneic HCT. (See "Immunotherapy for the prevention and treatment of relapse following allogeneic hematopoietic cell transplantation".)

●When the recipient and nonidentical HCT donor are not matched at the major histocompatibility complex (MHC), GVT may result because of differences within the major MHC antigens. (See 'Recognition of malignant cells' above.)

●When the donor-recipient pair is matched for MHC antigens, GVT is initiated and propagated via recognition by the T cell and its receptor of additional antigens called minor histocompatibility antigens. (See "Pathogenesis of graft-versus-host disease (GVHD)" and 'Cytotoxic T cells' above and 'Natural killer cells' above.)