Richard W. Childs, MD
Principal Investigator, Stem Cell Transplant Unit, Hematology Branch, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892, USA.

e-mail:  ChildsR@NHLBI.NIH.GOV

For more than 3 decades now, allogeneic stem cell transplantation (SCT) has been used as a successful therapeutic modality to cure patients with hematologic malignancies that have become resistant or refractory to conventional chemotherapy approaches. Intensive chemoradiotherapy, intended to completely eradicate neoplastic cells, is followed by an infusion of HLA-compatible donor hematopoietic stem cells (either bone marrow cells or peripheral blood stem cells mobilized from the donor by granulocyte colony-stimulating factor [G-CSF]) in order to rescue the patient from regimen-induced bone marrow aplasia.^[1]

In the first 2 decades of its use, dose-intensive chemotherapy alone was credited with curing those patients who achieved sustained remission following this procedure. More recently, investigators have come to recognize that antineoplastic effects can be induced against hematologic malignancies mediated by immunocompetent donor T-cells transplanted with the stem cell allograft. The observation of a decreased risk of leukemic relapse in patients with a history of acute or chronic graft-vs-host disease (GVHD), and the ability to induce patients with relapsed chronic myelogenous leukemia (CML) back into a durable molecular remission following the infusion of donor lymphocytes, confirmed what investigators had speculated for years; namely, that a donor immune-mediated antimalignancy effect, called graft-vs-leukemia or graft-vs-tumor (GVT), occurred following such procedures.^[2]

Indeed, the GVT effect that follows allogeneic SCT arguably is the most potent form of cancer immunotherapy currently in use. The powerful and potentially curative nature of this immunotherapy in hematologic malignancies has recently led investigators to explore the use of allogeneic SCT as immunotherapy for treatment-refractory solid tumors. The development and early results of investigational trials exploring this therapeutic approach are discussed in this review.

Immunotherapy of Cancer:
 

Since the early 1980s, there has been a growing interest in the use of immunotherapeutic strategies to treat patients with chemotherapy-refractory solid tumors. Early trials used cytokines such as interferons and interleukin-2, with the intent of enhancing innate host immune responses against tumor. Pioneering studies by Rosenberg and others in the 1980s defined, for the first time, the utility of immune manipulations in the management of some patients with metastatic melanoma and renal cell carcinoma (RCC).^[3] While few patients were actually cured with cytokine-based treatment, these early trials were important because they established proof of concept of the efficacy of immunotherapy for cancer and laid the foundation for the development of novel immunotherapeutic approaches.^[4]

Recently, a number of promising vaccine-based strategies have been developed to target host immune cells to antigens expressed exclusively or overexpressed on the surface of cancer cells. Unfortunately, at present, only a minority of subjects with cancer have had tumor regression following such vaccine therapies. The limited clinical impact that conventional immunotherapy approaches have made in the field of oncology has lead investigators to explore new and alternative immune-based therapies.

Allogeneic SCT is an effective and accepted therapeutic option for many patients with hematologic malignancies that are incurable with standard chemotherapy. The demonstration in the late 1980s that patients with CML who had relapsed following an allotransplant could be induced back into a durable remission by the single act of infusing donor lymphocytes established conclusively the existence and curative potential of the GVT effect.

Similar anticancer immune effects have also been demonstrated in other hematologic malignancies, including acute leukemia, posttransplant Epstein Barr virus-associated lymphoproliferative disorder, chronic lymphocytic leukemia, Hodgkin's and non-Hodgkin's lymphoma, and multiple myeloma.

Early evidence affirming the existence of GVT in nonhematologic cancers has largely been taken from a few case reports and small case series. Eibl and colleagues^[5] reported the case of a woman with breast cancer treated with an allogeneic SCT from an HLA-identical sibling. A GVT effect was inferred from regression of extensive liver metastasis, which occurred concomitantly with the onset of acute GVHD. Ueno and colleagues^[6] subsequently reported a small series of patients with metastatic breast cancer treated with a conventional myeloablative allogeneic transplant in which responses occurred in the context of acute skin GVHD and the withdrawal of immunosuppression, also implicating that these responses might have been the consequence of a GVT effect.

A recent report of a patient with relapsed ovarian cancer achieving disease regression following a myeloablative allogeneic SCT potentially identified this malignancy as being susceptible to GVT effects as well.^[7] Unfortunately, the 20% to 35% risk of regimen-related mortality associated with conventional "myeloablative" allotransplantation has largely precluded any systematic investigation for GVT effects in patients with chemotherapy-refractory metastatic solid tumors.

Severe toxicities associated with myeloablative conditioning are one of the major contributors to the morbidity and mortality associated with myeloablative allogeneic SCT. This led investigators to design conditioning regimens of lesser intensity to test the hypothesis that GVT effects alone would be sufficient to induce remissions in some hematologic malignancies. To facilitate donor immune engraftment, these reduced-intensity regimens use potent immunosuppressive agents that have considerably less toxicity than their myeloablative predecessors.

The results of early clinical trials have confirmed that the GVT effects generated following NST may be curative for a variety of different hematologic malignancies. The improved safety profile and preliminary success in patients with hematologic malignancies has laid the foundation for investigating this potent form of immunotherapy in patients with solid tumors.

Initial studies of NST in nonhematologic cancers have focused on tumors that were believed to be amenable to immunologic intervention. Although a number of different transplant approaches are currently being investigated in a variety of metastatic tumors, all share the similar goal of harnessing allogeneic lymphocytes in such a manner that will lead to the generation of antitumor effect. Since the majority of patients with metastatic tumors enrolled in such trials will have disease that is refractory to cytotoxic chemotherapy, even at high doses, the ideal approach is to perform an allogeneic SCT using a modified, dose-reduced conditioning regimen that provides sufficient immunosuppressive capacity to ensure donor immune engraftment for the generation of a GVT effect. Such "low-intensity" or nonmyeloablative regimens would potentially avoid unnecessary toxicities associated with high-dose conditioning.

There are several NST approaches currently being investigated in patients with metastatic RCC.^[8] While all seem to be well tolerated, the incidence of transplant-related complications, including graft rejection, GVHD, and transplant-related mortality, varies among different approaches. Furthermore, the rate and degree of donor engraftment, which follows NST, may be highly variable and is affected by such factors as the patient's pretransplant immune status as well as the choice and dose of agents used for conditioning and GVHD prophylaxis. However, all such regimens, by definition, are nonmyeloablative, thus permitting some degree of patient (autologous) hematopoietic and lymphoid recovery.

The coexistence of donor and host cells that typically occurs following such regimens, called mixed chimerism, may be both beneficial and detrimental to the posttransplant outcome. Although mixed chimerism is associated with a higher likelihood of immunologic tolerance between host and donor tissues, potentially decreasing the incidence of GVHD, donor tolerance to the tumor may be promoted in this setting. Furthermore, since some (eg, RCC) patients may not have received prior chemotherapy, full donor immune engraftment following NST may be delayed compared with patients with hematologic malignancies undergoing similar procedures.

Additionally, since the proliferative potential of some metastatic tumors may be rapid, and since the conditioning regimen provides little to no direct antitumor effect, there is a need to quickly establish donor engraftment to a level sufficient to confer a graft-vs-tumor effect. Most regimens, therefore, incorporate methods to convert patients from mixed to full donor chimeras, such as the withdrawal of posttransplant immunosuppression and/or the infusion of donor lymphocytes.

Figure 1. Nonmyeloablative transplant approach. Patients with metastatic renal cell carcinoma (RCC) received pretransplant conditioning with cyclophosphamide (120 mg/kg) and fludarabine (125 mg/m²) followed by infusion of a granulocyte colony-stimulating factor-mobilized peripheral blood stem cell allograft on day 0. Cyclosporine (CSA) was started on day -4 to prevent graft rejection and continued thereafter as graft-vs-host disease (GVHD) prophylaxis. After transplantation, blood samples were obtained and assessed for donor engraftment in T-cell lineages using a polymerase chain reaction-based mini-satellite analysis. Patients with mixed T-cell chimerism on day 30 were tapered off CSA and, if necessary, received incremental doses of donor lymphocyte infusions monthly until GVHD or disease regression occurred. Patients with 100% donor T-cell chimerism by day 30 continued on CSA until day 60, at which point it was tapered off over the following 40 days. *Mycophenolic acid was subsequently added to CSA as GVHD prophylaxis (after RCC patient #25) and was discontinued in conjunction with CSA.

Figure 2. Computed tomography (CT) and x-ray evaluation of a renal cell carcinoma (RCC) patient receiving a nonmyeloablative transplant. Pretransplant chest CT images of a patient with metastatic RCC with bulky mediastinal and bilateral hilar disease (A1 and A2). Regression of metastatic disease, which was markedly delayed (B1 and B2 are 6 month images and C1 and C2 are 9 month images), is shown on a follow-up CT and chest x-ray images.

The onset of GVT effects were typically delayed by months (median, 5; range, 1-9) from the time of the transplant, and usually occurred following the withdrawal of CSA and conversion from mixed to predominantly full donor T-cell chimerism (either spontaneously or following the withdrawal of CSA and/or a donor lymphocyte infusion).

Figure 3 shows the typical timeline and course of common posttransplant events. Although acute GVHD was favorably associated with a disease response and occurred in association with tumor shrinkage in some patients, it was not required for the generation of a GVT effect; 1 patient had a complete response in the absence of acute GVHD and several had disease regression that occurred months after their acute GVHD had resolved. These clinical observations support the notion that distinct T-cell populations, capable of recognizing tumor-restricted antigens or antigens shared by both the tumor and normal tissues, may be involved in the rejection of RCC following this approach. The first patient transplanted with metastatic RCC remains in complete remission more than 4 and a half years since undergoing treatment.

Figure 3. Posttransplant events in renal cell carcinoma (RCC) patients receiving nonmyeloablative transplant. Time-course of events occurring in patients with metastatic RCC following nonmyeloablative transplant using cyclophosphamide/fludarabine conditioning. Boxes show range of events and arrows represent median time-point of individual event occurrence.

Several other investigators have recently reported GVT effects in RCC patients using a variety of NST regimens.^[8,10,11] While proof of concept of GVT in RCC is undeniably important, there exists a number of limitations of this approach that will need to be overcome before phase 3 trials comparing cytokine-based therapy with this technique can be pursued (Table 1). Approximately 20% of patients can be expected to develop severe, potentially life-threatening acute GVHD. Although regimen-related toxicities appears to be less than those expected with a conventional "high-dose" myeloablative transplant, the risk of transplant-related mortality remains in the range of 10% to 20%.^[12,13] Therefore, given the potential for life-threatening toxicities associated with the procedure, most pilot trials have required patients to have progressive metastatic disease and have failed prior treatment with cytokine-based therapy.

The tolerogenic effects of the early mixed T-cell chimerism that follow such low-intensity transplants as well as the initial requirement for immunosuppressive therapy (eg, CSA) to prevent GVHD likely contribute to the prolonged time interval required for the induction of an antitumor effect. As a consequence, patients with rapidly advancing metastatic RCC would be unlikely to benefit and are typically excluded from such therapy. Whether surgical tumor debulking might prolong survival for the time required for the generation of a GVT effect in such poor-prognosis patients is currently being investigated. Finally, given the increased risk of GVHD and transplant-related mortality associated with allogeneic SCT from unrelated donors (ie, HLA-matched unrelated donors), most trials have required an HLA-matched sibling to serve as a stem cell source.

Table 2 lists the types of patients who are most likely to benefit from a nonmyeloablative approach. Given these limitations, only a minority (approximately 15% to 20%) of patients with metastatic RCC can be expected to be candidates for such investigational therapy.

The observation of GVT effects in RCC has led to the initiation of similar trials of NST for patients with a variety of different treatment-refractory solid tumors. While preliminary data from several groups provide evidence for the existence of a GVT effect in RCC, only anecdotal data are available on other solid tumors. Case reports of GVT effects in patients with metastatic breast, ovary, and colon cancer have recently been described. Indeed, we recently observed evidence for a GVT effect in a patient with metastatic colon carcinoma that had become refractory to treatment with 5-fluorouracil and irinotecan.

While the advent of nonmyeloablative conditioning seems to be associated with a decreased regimen-related mortality, morbidity associated with acute GVHD remains a serious concern. A continuing challenge will be to develop better methods for preventing acute GVHD, the major toxicity associated with the procedure. Methods to enhance tumor-specific alloimmune responses through posttransplant tumor vaccination strategies could enhance the efficacy or, more importantly, expedite the onset of a GVT effect in a variety of different solid tumors. While much work needs to be done before allogeneic transplantation could ever be expected to play a routine role in the treatment of patients with nonhematologic cancers, the framework has been laid for future innovative immunotherapy approaches that will attempt to harness the powerful potential of the GVT effect in a variety of different solid tumors.

 

1. Requirement for an HLA-matched donor (exclusion of approximately 2/3 of renal cell carcinoma patients)
2. Toxicity	Incidence
Graft-vs-host disease	30% to 60%
Cytomegalovirus reactivation	20% to 40%
Graft rejection	5% to 20%
Transplant-related mortality	10% to 20%
3. Graft-vs-tumor effects delayed -- patients with rapidly progressing tumors are less likely to benefit

 

Progressive metastatic disease

Expected survival 6 months

Failed prior cytokine therapy

Absence of central nervous system involvement

HLA-compatible sibling available