Introduction

The therapeutic value of IL-2 was recognized early for renal cancer and melanoma, two malignancies known to be amenable to immunomodulation as well as highly resistant to cytotoxic agents. Other tumor types, including hematologic malignancies, were also included in Phase I trials and occasionally appeared to benefit from high-dose IL-2 or alternative IL-2 containing regimens that were tested subsequently. Since most other malignancies are more amenable to cytotoxic agents or combinations, may be less sensitive to immunotherapeutic interventions, and are rarely treated successfully with investigational regimens until after failing a sequence of prior cytotoxic therapies, enthusiasm for the use of IL-2-based approaches has been low. Nevertheless, recognition of a potential role for immunomodulation, together with new concepts regarding the mechanisms of cytotoxic agents and how they may interact with the immune system, has promoted a re-examination of the potential role for this molecule and related immunomodulators, some of which are also new since the early studies of high-dose IL-2 for multiple malignancies. Moreover, rapid developments in the field of tumor antigen recognition, vaccine development, and the mechanisms and interventions for tumor immunoediting and immune system “escape” from immune-mediated control will provide the foundation for further exploration of combination strategies that take advantage of the powerful effects of IL-2 and its well-known clinical safety and efficacy profile.

High-dose IL-2 in melanoma

Like renal cancer, melanoma is a tumor with a long history of immune-based therapy trials resulting from extensive preclinical models, encouraging results from interferon and vaccine therapies, and such poor outcomes from cytotoxic agents that new paradigms are desperately needed. Initial studies in melanoma, begun at the Surgery Branch of the National Cancer Institute (NCI), were designed exactly like those used for renal cancer. The starting doses and planned therapy were based on the results of Phase I studies with and without lymphokine-activated killer (LAK) cells produced from leukapheresis of patients following the first 5-day cycle of IL-2 and re-infused over 3 days during the second 5-day cycle [1–3] and the number of doses administered to each patient was adjusted to the patients’s own dose-limiting toxicities. While patients with melanoma tend to be younger and in better medical condition than those with renal cancer, the overall tolerance of IL-2 appears similar, although the specific pattern of toxicities in renal cancer patients may more likely reflect the uninephric state with the resulting predisposition to oligoanuria and metabolic acidosis. Patients enrolled in early Phase II studies at the NCI received regimens that included LAK cells, but the latter was abandoned among patients with renal cancer after the demonstration in small randomized trials that LAK cells generated ex vivo did not produce superior outcomes to IL-2 alone [4, 5]. The results in melanoma patients suggested a trend for better outcomes using LAK cells than with IL-2 alone [4], which supported the further development of more antigen-specific efforts in this disease, with more complex immunomodulatory regimens that are detailed later in this review. The early experience at the NCI Surgery Branch using high-dose IL-2 for melanoma was reviewed by Rosenberg et al. Among 270 patients with advanced melanoma, 66 received LAK cells, 60 were treated with IL-2 alone, and the rest participated in trials to study other drug combinations, including α-interferon, tumor necrosis factor, cyclophosphamide, or tumor-infiltrating lymphocytes (TIL). While the response rates were reported for only a fraction of subjects (the other patients had participated in Phase I trials and were not evaluable for this endpoint), the 24% objective response rate, which included several durable complete remissions, was similar to the level of activity in renal cancer [6]. A number of subsequent reports from the NCI Surgery Branch Group as well as other centers confirmed the original Surgery Branch reports, and an estimate of the objective response rate in the range of 15–20% for advanced melanoma treated with high-dose intravenous bolus IL-2, with the achievement of durable complete responses in about one-third of responders providing the basis for the 1998 US Food and Drug Administration approval of high-dose IL-2 for melanoma [7]. This low but fairly reproducible level of antitumor efficacy that provided major benefit, possibly even cure, to a very small fraction of patients has turned out to be a seemingly insurmountable “ceiling” for IL-2-based therapies in melanoma, possibly due to a unique immunobiological relationship between tumor, host, and therapy that might be amenable to further manipulation with more innovative strategies and better understanding of the mechanisms of IL-2 and its targets.

The role of adoptive cell therapy with IL-2 in melanoma

Although the contribution of autologous LAK cells produced ex vivo from patients exposed to IL-2 and re-infused with additional IL-2 was not sufficiently active to justify the added procedures, expense, and toxicities associated with their inclusion in IL-2-based regimens, the use of expanded populations of T cells with melanoma antigen specificity emerged soon after the initial experience with IL-2 and LAK cells. Using the tumor itself as a source of tumor antigen-specific T cells and a culture system in which IL-2 promotes the expansion of these effector cells while allowing the simultaneous attrition of tumor cells, investigators at the NCI Surgery Branch observed remarkable activity in a small number of patients with metastatic melanoma who received infusions of IL-2-expanded TIL cells followed by systemic IL-2 [8]. These results, together with the demonstration of markedly superior activity for TIL cells compared with LAK cells in the preclinical tumor models [9, 10], provided strong justification for further trials of antigen-directed approaches to adoptive cell therapy. A long series of additional modifications of this elegant strategy has taken advantage of rapid developments in antigen identification, and understanding of the essential elements of successful adoptive cell therapy (particularly the impact of IL-2 on regulatory T cells [11] and methods to enhance homeostatic proliferation of adoptively transferred, antigen-specific effector cells [see below]). The participation of centers within and outside of the NCI to better define the role of therapeutic components has added further insight as well as a “reality check” on the NCI-reported results, since patients referred for therapy at major centers like the NCI are often highly-selective for overall fitness, motivation, and other factors that are known to contribute to favorable outcomes independent of—or interacting with—the therapeutic intervention. As an example, the US Multicenter Cytokine Working Group reported recently the results of a Phase II randomized trial to assess three different schedules of high-dose IL-2 in combination with a vaccine containing three peptides from melanoma antigens that have been used extensively by the NCI investigators for patients with melanoma whose histocompatibility ty** (HLA type) carries at least one copy of HLA-A2 (required for the ability of T cells to recognize the peptides used in the vaccine) [12]. The results of this study, which randomized approximately 40 patients to each of the three schedules, did not show any of the treatment arms to be dramatically superior to any other or to the level of efficacy expected from high-dose IL-2 alone [13]. The group from the NCI Surgery Branch recently reported 684 consecutive patients with melanoma who received IL-2 alone or with vaccine in one of the multiple trials over 11 years and found a very slight advantage associated with vaccine therapy [14], supporting the continued investigation of strategies to optimize the antigen-specific component of IL-2-based and other immunostimulatory strategies for advanced melanoma. Success in this arena is also likely to lead to promising options for adjuvant therapy of high-risk disease following surgical resection.

In addition to assessing the contribution of HLA-restricted, common melanoma antigen-derived peptides to the control of advanced melanoma, other major modifications of IL-2-based therapies have shown great promise and may be combinable with antigen-specific strategies. In particular, the use of moderately high-doses of chemotherapy to induce a transient state of profound lymphopenia that promotes the homeostatic proliferation of effector T cells once they have been expanded ex vivo and reinfused [15] is undergoing testing in many centers both within and outside of the NCI. A common theme in the design of immunomodulatory regimens for malignancy has been the recognition of a need to “suppress the suppressors”, which has long been achieved with the use of various doses of cyclophosphamide, a drug with potent immunosuppressive properties that can be used in a wide range of doses and schedules. Since IL-2 itself can promote existing antigen-specific immunity that is found in many melanoma patients [16], the Cytokine Working Group recently evaluated the use of high-dose cyclophosphamide plus fludarabine, a lymphodepleting but minimally myelosuppressive agent, prior to high-dose IL-2 for patients with advanced melanoma. Patients also received granulocyte–monocyte colony stimulating factor for both hematopoietic support and as an immunomodulator [17]. The preliminary analysis of data from this study did not suggest superiority over the activity of high-dose IL-2 alone. Of interest is that IL-2 toxicities during the first cycle (IL-2 days 1–5), which corresponded to the chemotherapy-induced hematologic nadir, were markedly less severe than expected from IL-2 alone, while second-cycle toxicities (IL-2 days 15–19), occurring after hematopoietic recovery, were similar to those reported with IL-2 alone (Ernstoff et al., manuscript in preparation).

Building on the promising data favoring antigen-specific adoptive T cell therapy, investigators at the NCI and elsewhere have designed elegant strategies that incorporate multiple essential elements of an antitumor immune response. The cellular element has evolved from unmanipulated TIL cells through TIL or circulating cells selected and expanded to contain high-frequency antigen-specific cytotoxic cells [18] to the current approach involving the use of gene therapy methods that transduce T cells to express an antigen receptor specific for melanoma proteins that are targets of cytotoxicity reviewed in [19]. Thus, IL-2 continues to play an important role in both the ex vivo expansion component as well as post-infusion to prolong survival and promote cytotoxicity mediated by the infused effector cells.

Biochemotherapy with IL-2 combinations for melanoma

Investigators working in the field of IL-2-based immunotherapy of malignancy have taken advantage of important differences in the biology of renal cancer and melanoma. The two most important features of melanoma that lend themselves to the development of innovations in IL-2-based therapy includes: (1) the availability of chemotherapeutic agents with activity against melanoma that possess only partially overlap** toxicities with those of IL-2, and (2) the availability of well-characterized tumor antigens in melanoma that have been studied in combination with IL-2 and other immunostimulatory agents in both the advanced disease and the adjuvant setting. Although many chemotherapy combinations with IL-2 with or without other cytokines (often called “biochemotherapy”) appeared promising when first reported, recent data from randomized studies have nearly all shown disappointing results, suggesting the lack of benefit for using complex multi-agent regimens containing one or more chemotherapies and IL-2 with or without α-interferon [20]. At present, the role of chemotherapy in combination with IL-2 appears limited to the lymphodepleting regimens mentioned above, which provide a specific niche for the homeostatic repopulation of a T cell compartment that is markedly enriched for antigen-specific effector cells with the potential for prolonged survival and function in vivo [21] (Table 1).

Table 1 IL-2-based therapy for advanced melanoma
Full size table

IL-2-based therapy of hematologic malignancies

During the time that IL-2-based therapies for solid tumors were under intense investigation, the potential of IL-2 for the treatment of hematologic malignancies was also explored. In the case of leukemias and lymphomas, it was first necessary to demonstrate that IL-2 did not act in any way to promote the growth of neoplastic cells [22–24]. Initial trials, modeled after the regimens used for solid tumors, showed promise in patients with relapsed acute myelogenous leukemia (AML) [25, 26] and lymphomas, including Hodgkin lymphoma [27–30]. Additional investigations were designed to take advantage of the unique microenvironment provided by these diseases, including the proximity of neoplastic cells to the antigen-presenting cells (APC) and immune effector cells (T and NK) in the blood and marrow [31, 32]. Because the early trials in leukemia and lymphoma were encouraging, innovative strategies followed, including the combination of antibody therapy with IL-2 to enhance antibody-dependent cellular cytotoxicity and other potential synergistic interactions. The toxicities of this combination, its low activity in the limited number of patients enrolled in trials [33], and the contemporaneous development of more promising therapies for lymphoma precluded its further use in this patient population, although the principles of combining cytokine with antibody continue to be explored in other settings. The other clinical setting in which IL-2 has been extensively evaluated for hematologic malignancies is in patients who have undergone high-dose chemotherapy with hematopoietic cell support (generally autologous peripheral cells; in some reports, autologous marrow or allogeneic cells). The additional rationale for bringing IL-2 therapy into the post-transplant setting is its potential contribution to immunologic recovery and even the stimulation of a new, less-tolerized immune response to tumor antigens, analogous to the melanoma studies detailed earlier. Despite promising early pilot data [34–37], the randomized studies to evaluate the contribution of IL-2 following autologous transplant in leukemia and lymphoma failed to demonstrate a benefit [38, 39]. Its potential role as an adjunct to post-transplant manipulations such as donor lymphocyte infusions following allogeneic transplant remains under evaluation (Table 2).

Table 2 IL-2 in hematologic malignancies
Full size table

IL-2 in other malignancies

The value of IL-2 or IL-2 containing combinations in other tumor types, particularly the common solid tumors, has not been established. The reasons for this relative exclusion of other histologies include the greater responsiveness of these tumors to cytotoxic (and more recently, in some cases, small molecule “targeted”) agents as well as the still-challenging dose- and schedule-dependent toxicities associated with IL-2 in a group of patients who tend to be relatively immunocompromised and in suboptimal condition due to tumor and prior therapy. Even in recent years, with greater understanding of the important interactions between chemotherapy and biological agents’ effects on the immune system, it is difficult to find a clinical setting in which the optimal results of IL-2-based therapy can be determined. At present, the role for IL-2 in other tumor types remains highly investigational and adjunctive to the other therapeutic elements. Examples of a role for IL-2 in these settings include its important activity in the expansion and activation of cells for adoptive immunotherapy, which has been developed less extensively in other solid tumors than in melanoma. At the same time, investigations directed at better defining the optimal setting for IL-2 in those diseases for which it has proven activity are likely to inform the design of new strategies in which IL-2 may play a role for other tumor histologies. The approaches to selecting patients for therapy with IL-2 has taken different approaches for renal cancer than for melanoma; the available data and principles of ongoing investigations are addressed elsewhere in this issue. For melanoma, while the current understanding of how to select patients most likely to benefit and least likely to suffer excessive toxicity remains more elusive, there are nevertheless intriguing reports from investigators studying the tumor microenvironment [40, 41], suggesting that it is possible to use sophisticated molecular techniques to better dissect the elements of tumor immunobiology (host, tumor, and therapeutic agents) and to use the results to design improved interventions.

Reducing IL-2 toxicity and enhancing its efficacy

Many investigators and biotechnology companies have endeavored to design a “better” IL-2 that can overcome both the limited activity achieved with regimens to date as well as the excessive toxicity that further limits the use of this agent in patients with cancer. In addition to structural alterations (detailed below), several agents designed to target mediators of IL-2 toxicity showed initial promise, raising hopes for both a safer IL-2 regimen and the potential for increased exposure that might lead to greater antitumor activity. These agents have included inhibitors of fatty acid signaling [42], inducible nitric oxide synthase [43], interleukin-1 [44], and tumor necrosis factor [45]—all without sufficient efficacy in Phase II or Phase III trials to warrant further use. Another method to reduce IL-2 toxicity is to alter its structure in one of several ways: (1) substitution of one or more amino acids leading to a greater binding to the IL-2 receptor of T cells over that of NK cells, expected to result in greater antigen specificity and less secondary cytokine cascade [46]; (2) chemical modification of the IL-2 molecule to covalently bind it to polyethylene glycol or human albumin which increases its half-life and potentially reduces its toxicity by avoiding peak concentrations while maintaining exposure over prolonged times [47, 48]; or (3) linking the IL-2 molecule with a tumor-targeting molecule (e.g., an antibody with specificity for a tumor antigen or a ganglioside that is differentially expressed in tumor over normal tissue), which can localize the IL-2 to bring target cells to the site of tumor and limit the incidence of systemic toxicities [49]. The addition of a low, immuno-activating dose of recombinant OKT3 (a molecule commonly used in higher doses to suppress T cells in the management of organ transplant rejection and used in low doses in vitro to activate T cells for immunotherapy strategies) was also evaluated in patients with metastatic renal cancer or melanoma undergoing high-dose IL-2 therapy. While treatment was safe with this combination, there was insufficient clinical activity or evidence of enhanced immune endpoints to justify its further clinical use, although this molecule continues to be used for in vitro T cell activation [50]. Currently, several immunocytokine molecules are undergoing evaluation that may show promise and may open the field to other malignancies, based on the target of the IL-2 fusion partner in these structures.

Intralesional therapies have long been of interest for melanoma due to its potential for immunomodulation and the frequent presence of metastatic lesions that are readily accessible (and may create serious medical morbidity) due to their location in the skin, subcutaneous, and nodal sites. Among the immunomodulatory agents that have been administered by intralesional injection is IL-2, for which the limited published experience was reviewed in 2004 [51]. While IL-2 is now rarely used for regional therapy in this way, many new protocol designs incorporate intralesional injections and melanoma as an optimal clinical setting for this approach (Table 3).

Table 3 Modulation of IL-2 toxicity
Full size table

Novel IL-2 combinations

While combinations designed to reduce the toxicities and/or enhance the efficacy of IL-2 in renal cancer and melanoma—diseases for which IL-2 continues to play an important role—have not shown great promise to date, there is reason to remain optimistic that newer investigational agents may be amenable to combination with IL-2 in regimens that have potential to overcome the obstacles to therapeutic progress. The most promising of these approaches involves the combination of IL-2 with an investigational antibody that inhibits the physiologic checkpoint imposed on antigen-activated T cells by the expression of cytotoxic T lymphocyte antigen (CTLA)-4. This form of “checkpoint blockade” has already shown single agent activity against melanoma and other solid tumors, and the safety of its combination with high-dose IL-2 has also been reported [52]. It is likely that further development of combination such as anti-CTLA4-antibody with IL-2 as well as many other 2- or 3-component immunomodulatory combinations will be possible as other agents demonstrate the possibility of safe combination and efficacy in preclinical and pilot clinical studies.

The role for combining or carefully sequencing one or more of the newer receptor kinase inhibitors with IL-2 for metastatic cancer is best addressed in the setting of renal cancer, where several of these agents have already been approved and combinations as well as sequences have started to undergo evaluation. The most challenging aspect of designing strategies to combine these agents with IL-2 may be the analysis of the relative impact of the small molecule on tumor versus effector cells and other components of the immune system. For example, sunitinib and sorafenib, two small molecule VEGFR inhibitors currently approved for therapy of renal cancer, have marked differences in their effects on T cell function [53]. Guidelines for the design of any new combinations will need to consider both the potential for IL-2 and other immunomodulatory molecules in the disease of interest as well as to carefully assess the impact of the other agent(s) on the immunologic functions and clinical effects of IL-2.

Conclusions

In the more than two decades since the discovery of IL-2, the expansion of its role in various approaches to the biological therapy of malignant disease has taken several promising directions. While the original application of IL-2 in supraphysiologic doses continues to provide remissions, sometimes durable, in a small fraction of patients with advanced renal cancer and melanoma, its mechanisms of action remains speculative, ranging from antigen-driven T cell-based effects to nonspecific activation of NK cells against tumor. IL-2 continues to be an essential element of more precisely defined strategies such as vaccines that involve dendritic cells and other methods of optimized antigen presentation to induce cytolytic T cell responses in an antigen-specific, HLA-restricted fashion. Promising combinations of IL-2 with other cytokines, chemotherapeutic agents, angiogenesis inhibitors, and small molecules with defined molecular targets are likely to find a niche in the near future. More innovative approaches such as bispecific IL-2 containing molecules that retarget effector lymphocytes and derivative molecules that provide enhanced activity and/or reduced toxicity are also in development. Experience with the design of translational studies of IL-2 over the past 20 years has provided the framework for the study of other immunotherapies, which will continue to evolve as the field expands into the 21st century.