Disclosure
Over the past 40 years, steady progress has been made in the field of pediatric oncology. This has resulted largely from the stepwise integration of multimodality therapies (chemotherapy, surgery, radiation therapy) into carefully designed complex treatment regimens tested sequentially in multicenter randomized trials. The identification of chemotherapeutic agents that can eradicate micrometastases has been the backbone of this success; however, over the past decade, an understanding of the limitations of chemotherapy has grown. First, clinical and experimental evidence have shown that it is not uncommon for tumors to develop chemoresistance to multiple drugs concomitantly, resulting in cross-resistance to agents that may not have been received by the patient previously and that may show good activity when used in patients in whom cancer is newly diagnosed (Kaye, 1998). Second, although many pediatric tumors show better clinical response to intermediate-dose intensities than to low-dose intensities (Smith, 1991), response appears to plateau at high doses, thus limiting the effectiveness of continued chemotherapy dose escalation (Smith, 1996). Finally, although several new cytotoxic agents that are active against adult carcinomas (eg, taxanes) have been developed over the last decade (Choy, 2001), these new chemotherapy agents have limited efficacy in a wide variety of tumor types. The need is growing for the development of effective alternative anticancer therapies for use in children with tumors. The 1990s were marked by dramatic progress in the understanding of the basic biology of both cancer and tumor immunologies. Evidence currently suggests that the mechanisms responsible for resistance to cytotoxic agents generally do not confer resistance to immune-mediated mechanisms of tumor cell killing (Cullen, 2001; Shtil, 2000; Kleinerman, 1995; Giavazzi, 1984; Kontny, 1998). Immunotherapy can be divided broadly into the categories of innate immunity stimulation, T-cell–based therapy, and monoclonal antibody (MAB)–based therapy. This article begins with a brief review of the historical background of immunotherapy for cancer, then addresses the central scientific principles that serve as the basis for the development of immunotherapies and reviews clinical experience in pediatric oncology using each approach.
William Coley, MD, generally is regarded as the father of cancer immunotherapy. Coley was a pioneering surgeon who practiced medicine at Memorial Hospital in New York City from 1890-1936 (Hall, 1997). During this time, he focused on the treatment of sarcomas, and he developed a firm belief that activation of endogenous immune responses could induce remissions of these tumors. This conviction arose from reports and direct observations of spontaneous tumor remissions that were temporally related to bacterial infections in patients with sarcoma. Coley subsequently produced crude bacterial extracts that eventually were termed Coley toxins and were administered to patients with cancer. In some, albeit relatively infrequent, cases, dramatic response was observed in tumors. Ewing sarcoma was one of the tumors in which Coley observed antitumor responses using this approach (See, 1946; Coley, 1906); however, this finding is not widely appreciated. During the same period that Coley was observing dramatic tumor responses using Coley toxins, the spectacular radiosensitivity of Ewing sarcoma was observed by James Ewing, MD, who was the physician-in-chief at Memorial Hospital at that time. This led to a rivalry between advocates of immunotherapy in the form of Coley toxins and radiotherapy as advocated by Ewing (Hall, 1997). For a variety of reasons, not the least of which was difficulty in standardizing the toxins, immunotherapy eventually was abandoned in favor of cytotoxic radiotherapy, a legacy that persists to this day. Perhaps the greatest limitation of Coley's work was the prevailing lack of understanding of the mechanisms by which the toxins might exert their effects. Without the basic knowledge that could guide experimental studies and optimize the use of toxins, further development of Coley's therapy was bound to stall. Today, bacteria are known to be potent inducers of the first line of defense, which is termed innate immunity. Innate immunity can induce antitumor effects, such as those observed using liposomal muramyl tripeptide phosphatidylethanolamine (MTP-PE) in canine osteosarcoma (Kleinerman, 1995). Furthermore, innate immune responses are known to boost adaptive immunity, especially T-cell responses. One of the most potent adjuvants known for inducing T-cell immune responses is Freund complete adjuvant, which contains pieces of heat-killed mycobacteria (Matthys, 2000). Subsequent investigations have begun to elucidate the cellular basis for these early clinical observations. However, much speculation remains regarding the mechanisms responsible for these initial observations as the immunology community continues to attempt to harness immune effectors based on an ever-increasing understanding of basic immunologic processes.
The immune system can be broadly divided into 2 components, the innate immune system and the adaptive immune system. Innate immunity refers to cellular components that serve as the first line of immunologic defense but without resulting immunologic memory. Critical components include monocytes and macrophages, neutrophils, eosinophils, natural killer (NK) cells, and the complement cascade. Several animal models have demonstrated that specific activation of innate effectors (eg, monocytes, macrophages, eosinophils) can induce potent killing of tumor cells (Giavazzi, 1984; Hung, 1998; Kleinerman, 1983). NK cells also are known to play important roles in tumor killing and tumor cell immune surveillance (Rees, 1999; Berke, 1995). An intricate interaction exists between innate immunity and adaptive immunity, both at inception and at execution of the immune response. For instance, the release of inflammatory mediators by cells such as macrophages and dendritic cells plays a central role in amplifying and directing specific T-cell responses. Based on such observations, many attempts to induce T-cell responses toward tumors have incorporated approaches to up-regulate innate immunity. Perhaps the best examples of this are studies showing that granulocyte-macrophage colony-stimulating factor (GM-CSF)–transduced tumor cells can induce potent T-cell immunity. This was shown by Dranoff et al to result in local recruitment of dendritic cells, which then present tumor antigens to T cells (Dranoff, 1993). Cooperation between adaptive and innate immune responses also occurs at later stages of the immune response. Activated cytolytic T cells exert some of their antitumor effects through recruitment of innate effectors (eg, monocytes, eosinophils) to the site of the tumor (Hung, 1998). Thus, innate immunity functions as a critical component in the induction of immune-mediated antitumor activity. Current studies are being directed toward more effective ways to manipulate this arm of the immune system to stimulate antitumor immune responses. |
|
|||||||||||||||||||||||||||||||||||||
Activation of monocytes and tissue macrophages occurs following administration of the lipophilic agent MTP-PE. MTP-PE is an analog of muramyl dipeptide, a substance that is contained within the cell wall of mycobacteria and has immune-activating activity. When MTP-PE is encapsulated in multilamellar liposomes (L-MTP-PE), it is delivered efficiently to pulmonary macrophages and circulating monocytes following intravenous infusion (Kleinerman, 1983). Uptake of L-MTP-PE by monocytes and macrophages leads to activation of the cells, thus allowing tumoricidal effects after subsequent encounters with tumor cells. Presumably, the methods by which monocytes/macrophages activated by MTP-PE are able to kill tumor cells relates to the induction of multiple cytokines, including interleukin (IL)–1a, IL-1b, IL-7, IL-8, IL-12, and tumor necrosis factor (TNF). Systemic administration of L-MTP-PE leads to its relative selective delivery to the liver, spleen, lung, nasopharynx, and thyroid (Kleinerman, 1995). An important issue that may limit the ability of monocyte/macrophage-activating agents, such as MTP-PE, to eradicate tumors relates to the quantitative relationship between the number of monocyte/macrophage effectors and the number of tumor cells. Administration of monocyte/macrophage-activating agents does not necessarily increase their numbers, in part because monocytes and macrophages are postmitotic cells. Even with efficient activation of all resident monocyte/macrophages contained within the lung, the number of tumor cells that could be eradicated by MTP-PE is estimated to be fewer than 109 (Kleinerman, 1995). For this reason, MTP-PE has been studied for its effects on small amounts of residual tumor localized to lungs. In particular, MTP-PE administration has been studied in patients with osteosarcoma, a disease in which micrometastases cause treatment failure in almost 40% of patients despite administration of multiagent chemotherapy. Administration of adjuvant MTP-PE combined with cisplatin chemotherapy following resection of the primary tumor was tested in 2 randomized double-blind clinical trials in dogs with spontaneous osteosarcoma. In the first trial, MTP-PE was administered following amputation and 4 cycles of cisplatin. The dogs that received MTP-PE showed a modestly prolonged median survival time of 14.4 months compared to 9.8 months in the control group (P <0.01). In the second trial, L-MTP-PE was administered concomitantly with cisplatin, and no significant effects of the agent were observed (Kurzman, 1995). Subsequently, phase I and II human clinical trials have been performed in patients with osteosarcoma. Phase I trials identified an optimal biologic dose, rather than a maximum tolerated dose, which was chosen based on elevations of monocyte tumoricidal activity and induction of serum IL-1b plus other acute phase reactants. In a phase II trial, MTP-PE was used in the optimal biologic dose chosen from the phase I trial, in combination with surgical resection for recurrent pulmonary osteosarcoma. In the first patient cohort, duration of therapy was 12 weeks; in the second cohort, the MTP-PE dosing schedule was extended to 24 weeks. Since MTP-PE was administered for minimal residual disease, which could not be measured, activity was assessed based on disease-free survival comparisons to historical control subjects. Although no significant improvement was observed in disease-free survival compared to historical control subjects following 12 weeks of adjuvant MTP-PE therapy, prolongation in disease-free survival was observed in the group that received 24 weeks of therapy (9 mo vs 4.5 mo, P <0.03) (Kleinerman, 1995). In addition, peripheral fibrosis, inflammatory cell infiltration, and neovascularization were observed in metastases from MTP-PE recipients but not in control subjects. Thus, this trial provided evidence that L-MTP-PE is an active biologic agent that can produce a survival benefit in this patient population. Further investigation of MTP-PE in osteosarcoma has been undertaken in a clinical trial in which improved survival rates were noted in patients who received standard cytotoxic drug therapy (ie, cisplatin, methotrexate, adriamycin) plus ifosfamide and MTP-PE (Meyers, 2001). No benefit was observed when MTP-PE was used in the absence of ifosfamide. The mechanisms responsible for the improved outcome in patients receiving ifosfamide plus MTP-PE are not understood fully and currently are under investigation. Future studies of MTP-PE and similar agents that can activate innate immunity currently are underway. The fact that such therapies can be administered simultaneously with chemotherapy and/or immediately following completion of chemotherapy has provoked much interest in the use of these relatively nontoxic approaches to harness innate immune effector mechanisms to eradicate residual chemoresistant cells. Furthermore, although the ability of such therapies to eradicate bulky tumors may be limited, the ability to incorporate immunomodulatory agents into existing treatment regimens raises the hope that eradication of small numbers of chemoresistant cells can be accomplished early in the disease process before the development of recurrent bulky chemoresistant tumors. The challenges of developing agents to accomplish this goal include difficulties in identifying optimal biologic doses and in accurately assessing tumor response or other clinical and biologic outcomes. Creative variations on the typical phase II trial schema may be useful to randomize small numbers of patients to look for activity in early phase trials (Simon, 2001). However, it is important to note that definitive evidence for the efficacy of any agent must await results showing improved clinical outcome in an appropriately designed randomized phase III trial.
Graft versus leukemia in bone marrow transplantation Development of bone marrow transplantation (BMT) as a clinical approach to malignancy initially rested on the assumption that high doses of chemotherapy were necessary to eradicate relatively chemoresistant leukemia. In this paradigm, the marrow graft served only to rescue marrow function that was ablated irreversibly by high-dose therapy or irradiation. However, clinical evidence accumulated over the past 20 years has shown that an important component of the antileukemic effect in BMT is related to an immunologic reaction that occurs between donor T cells contained in the marrow graft and residual tumor cells that remain following high-dose chemotherapy (Horowitz, 1990; Rassam, 1993; Sanders, 1985; Passweg, 1998; Bacigalupo, 1991; Locatelli, 2000). Several observations support this conclusion. First, patients who receive bone marrow from an identical twin (ie, the histocompatibility leukocyte antigen [HLA] and all other minor antigen types are matched) experience higher leukemic relapse rates than recipients of HLA-matched allografts that are disparate at minor histocompatibility loci. Second, patients who develop some evidence of graft versus host disease (GVHD) experience a lower incidence of leukemic relapse (Sullivan, 1989). Increased intensity of immunosuppression used to prevent GVHD also is associated with increased leukemic relapses post-BMT (Bacigalupo, 1991; Locatelli, 2000). T-cell depletion of the marrow graft increases the rate of leukemic recurrence. Finally, leukemic relapse sometimes can be treated successfully by the infusion of donor leukocyte infusions (DLIs), which comprise peripheral blood T cells that can re-induce remission, directly showing the capacity for mature T cells to induce antileukemic responses (Porter, 1999; Porter, 2000; Collins, 1997; Slavin, 1995; Singhal, 1999; Kolb, 1995; Atra, 1997). Such allogeneic responses are likely to be directed toward differences of major or minor histocompatibility antigens between the host and donor; currently, only a few examples of specific antileukemic immune responses have been documented. Perhaps the most important principle gleaned from the clinical experience stated above is the potency with which T cells can permanently eradicate aggressive, recurrent, and chemoresistant leukemic cells. Several other lessons also are pertinent. Immune responses sometimes can occur at a relatively slower tempo than that observed with cytotoxic therapies. For example, in chronic myelogenous leukemia (CML), detection of molecular evidence of leukemia for at least 6-9 months post-BMT is not uncommon, with gradual resolution of molecular evidence of residual leukemia occurring over 1-2 years (Cross, 1993). The ability of transferred T cells to eradicate leukemia is highly dependent on the tumor burden and the rate of disease progression. For instance, although DLIs induce complete responses in 60-80% of patients with stable chronic-phase CML, the response rate is approximately 30% in accelerated-phase CML and less than 20% in patients in blast crisis of CML (Collins, 1997). Although these findings are pertinent to persons who are involved in the treatment of CML post-BMT, they also provide an important lesson because immune-based therapies are developed for other tumors, and they emphasize that such therapies are likely to be effective only in patients with low tumor burdens. This conclusion also has been demonstrated experimentally in many preclinical animal models. Third, all leukemias are not equally susceptible to the graft versus leukemia (GVL) effect. Although in the chronic phase, CML is highly susceptible to the GVL effect, acute myeloid leukemia (AML) shows intermediate susceptibility rates (Slavin, 1995) and acute lymphoblastic leukemia (ALL) shows the lowest response rates (Porter, 1999; Collins, 1997; Slavin, 1995). The reasons for the differences in susceptibility are not well understood, but no doubt exists that the differences hold important clues to the understanding of immune-based mechanisms of antitumor activity. According to one hypothesis, ALL cells, being lymphocytes (unlike myeloid cells), may not express the minor histocompatability antigens that are likely to be the antigen targets of the immune response. However, recent studies have reported potent GVL effects against non-Hodgkin lymphoma, which suggests that minor histocompatability antigens may not be the targets (Ratanatharathorn, 1994; Champlin, 1999). Alternative hypotheses suggest that higher proliferative rates in ALL and AML (compared to CML) may limit the capacity of T cells to expand to effective levels before development of large tumor burdens. In addition, lymphoblasts have been shown to lack costimulatory molecules and to be tolerogenic to T-cell populations, thus potentially limiting the potency of T-cell–mediated effects (Cardoso, 1996). Finally, the well-known capacity of ALL to relapse in sanctuary sites, such as the testes and the CNS, that are inaccessible to immune-mediated effects also may play a role. New approaches in BMT currently are focusing on ways to diminish the toxicity of BMT by reducing myelosuppression of the preparative regimen while preserving the antileukemic effect that is mediated by T cells contained within the graft. BMT appears to be evolving in this way, toward adoptive immunotherapy and away from high-dose myelosuppression, thus serving as a model for development of new approaches that can enhance the immune-mediated effects active in BMT. Recent reports suggest that, in BMT, T cells also can exert antitumor effects in some solid tumors (Champlin, 1999; Childs, 2000). Whether similar activity will be observed against pediatric solid tumors is unknown. Tumor vaccines As discussed above, GVL effects illustrate the potential of T cells as effectors of antitumor activity. However, this model clearly is not directly applicable to host responses to endogenous tumors because major and minor histocompatibility antigens, which serve as the driving force for GVL effects in allogeneic transplantation, both are absent. Therefore, to exploit T-cell reactivity to autologous malignancies, the immune system must be re-educated to respond to tumor antigens, present on the tumor cells, that can be distinguished from normal host tissue cells. The production of autologous autoimmune responses has been made possible by a series of advances in basic immunology that lead to the prediction that antitumor immunity can be harnessed through therapeutic tumor vaccination. First, T cells can recognize aberrant or overexpressed intracellular molecules and do not require cell surface proteins as targets of recognition. This is because T cells do not recognize whole proteins; instead, they recognize peptides derived from the ongoing breakdown of intracellular and cell-associated molecules. For such peptides to be recognized by T cells, the peptides must bind to major histocompatibility molecules expressed on the surface of the tumor cell (Zinkernagel, 1979). This means that intracellular molecules, even nuclear proteins, can be recognized. Proof of this principle derives from the identification of targets of responding T cells in human tumors, which have shown that endogenous tumor antigens are composed of a variety of intracellular and extracellular molecules that are either unique to the tumor or are preferentially expressed by the tumor compared to normal tissues. These studies have shown that even point mutations in normal proteins can be identified by the immune system as abnormal (Yanuck, 1993; Robbins, 1996). Second, the factors required for inducing optimal T-cell responses have been uncovered gradually over the last 2 decades. Basic studies have clearly shown that resting naive T cells require 2 signals to become activated against particular antigens. The first signal represents antigen-recognition stimulation. The second is a costimulatory signal that is provided by professional antigen-presenting cells and instructs T cells to become activated against antigens expressed in the milieu. The costimulatory signal is transduced through associated surface receptors that are expressed by activated dendritic cells and that can be up-regulated by stimulants of innate immunity (Matzinger, 1994). The ability to generate large numbers of dendritic cells that are efficient in antigen presentation (signal 1) and costimulation (signal 2) has allowed the therapeutic application of dendritic cell immunization, which is predicted to closely mirror the ways in which dendritic cells normally process and present antigen to T cells. Third, basic immune studies have shown that the immunologic tolerance induced by tumor antigens during the course of primary tumor growth is not absolute and can be overcome through specific immunization with professional antigen-presenting cells or by transducing tumor cells with cytokine or costimulatory receptor-encoding genes (Rosenberg, 1996). Together, these observations imply that creating T-cell–based therapies for human tumors is an achievable goal. Challenges that may be particularly pertinent to the development of vaccines for childhood tumors are delineated below. One challenge to the development of vaccines for pediatric tumors is the rarity of these tumors; hence, the ability to effectively test vaccine strategies directed at common immunodominant antigens may be difficult. Although identification of a particular peptide that is capable of inducing antitumor effects may be possible in patients with positive results for HLA-B7 and with alveolar rhabdomyosarcoma, this is likely to occur in only a few patients per year. The ability to test such complicated and specific therapies in rare tumors is limited. Another alternative would be to develop immunization strategies that can be used for a variety of tumors with variable histologies. To this end, targeting whole proteins or the use of manipulated whole tumor cells, rather than individual peptides, as immunogens may induce immune responses across a population of variable HLA types more effectively. Second, since dose-intensive chemotherapy is a standard component of almost all therapies for pediatric tumors, integrating T-cell–based therapies into existing therapies is particularly challenging. Indeed, while pediatric patients show more rapid recovery of immunity than adults following T-cell–depleting chemotherapy, immune recovery in pediatric patients following dose-intensive chemotherapy still usually takes at least 6-12 months (Mackall, 1995; Mackall and Gress, 1997; Mackall, Immunol Today, 1997; Mackall, Semin Immunol, 1997; Mackall, 2000). In many cases, disease may recur during the interval between dose-intensive chemotherapy and immune recovery. Therefore, alternative strategies may be needed, such as the use of vaccines as part of autologous hematopoietic transplantation (Nagler, 1995). These studies have shown that antitumor responses can be generated in association with autologous transplantation and vaccination with tumor cells transduced with cytokines or costimulatory receptors (Borrello, 2000; Borrello, 1999). Clinical trials are underway using these approaches to generate antitumor immune responses in solid tumors and leukemia. The usefulness of immune-based therapies is likely to be optimal in patients with minimal residual disease. Early trials in patients with endstage disease, nutritional depletion, bulky tumors, or chemotherapy-based T-cell depletion are unlikely settings to prove efficacy for immunostimulatory approaches. Furthermore, since toxicity is likely to be based on immune reactivity to normal tissues, even phase I trials conducted in immunosuppressed populations are not likely to be informative. Thus, creative trial strategies are necessary, and the development of surrogate markers to monitor activity, such as measurement of minimal residual disease and antitumor activity, remain challenging. However, definitive evidence of activity of any agent requires findings of clinical benefit in randomized phase III trials. Regarding individual trials, Brenner et al have used gene therapy to transduce neuroblastoma cells using complementary DNA (cDNA) to encode cytokines and then to immunize patients using subcutaneous injection of the transduced cells (Brenner, 2000). The rationale for this approach is based on data demonstrating that immunostimulatory cytokines can generate immune responses that show cytolytic activity for both cytokine-secreting and non–cytokine-secreting cells bearing the same antigens. Clinical evidence for antitumor activity was observed using autologous neuroblastoma cells transduced to secrete IL-2 (20% objective tumor response, 30% stable disease in 10 patients treated) (Bowman, Blood, 1998). As a result of the technical difficulties of reliably generating gene-transfected tumors in patients, these investigators also have explored the use of allogeneic neuroblastoma cells transduced with IL-2 (Bowman, Hum Gene Ther, 1998). Neither treatment produced any toxicity aside from local induration at the site of tumor injection. Ongoing studies are attempting to improve the outcome of this approach by cotransfecting the cDNA encoding cytokines and chemokines in autologous neuroblastoma cells (Brenner, 2000). Another approach for tumor vaccines involves the use of peptide antigens. Using current technologies for producing peptides, clinical-grade material can be manufactured relatively easily. Such an approach requires identification of a tumor-specific or tumor-associated peptide, which is immunogenic and is present in sufficient amounts, by major histocompatibility complexes (MHCs) on tumor cells to allow T-cell targeting. Identifying optimal tumor-specific peptides for vaccine therapy remains a significant challenge. For neuroblastoma, peptides derived from the amplification MYCN oncogene have been shown to be capable of inducing immune responses ex vivo and lysing neuroblastoma cells; thus, this target could potentially be exploited clinically in future studies (Sarkar, 2000). Animal studies have shown immune-mediated lysis of tumor cells bearing the fusion protein encoded by the t(2;13) translocation in alveolar rhabdomyosarcoma following immunization with a peptide derived from the breakpoint region (Goletz, 1995). Peptide-based dendritic cell vaccines are being developed to target the translocation breakpoints in Ewing sarcoma and alveolar rhabdomyosarcoma. In a phase I/II trial that involved patients with recurrent and measurable disease, evidence for an immune response was detected in one patient and a mixed clinical response was observed in one patient. Ongoing studies are underway to optimize the antigen-presenting cell used in these studies. In addition, immunorestorative approaches currently are being combined with tumor peptide vaccination to study whether such tumor vaccines can be administered effectively immediately following chemotherapy in the setting of minimal residual disease (Mackall, 1999).
When Kohler and Milstein reported the technology for generating MABs in 1975 (Kohler, 1975; Kohler, 1976), many tumor biologists expected the rapid evolution of a variety of antibodies that could act as “magic bullets” to target and kill tumors. Although progress in the field of MAB therapy for neoplasia has proceeded much slower than was anticipated initially, recent clinical trials have demonstrated antitumor activity in a variety of malignancies. The first criterion that must be met for effective MAB therapy in patients with tumors is the identification of appropriate tumor-specific targets. The targets must be (1) tumor specific (eg, show no binding to normal tissues), (2) highly overexpressed in tumors (compared to normal tissues) so that toxicity and targeting of the tumor can be accomplished, or (3) coexpressed on normal tissues that can be regenerated. In addition, the antigen target of the MAB should not be shed from the tumor following MAB binding, but, preferably, the antigen should be internalized by the tumor cell. Effectively targeted proteins have been categorized into 3 groups, namely, oncofetal antigens (eg, carcinoembryonic antigen [CEA]), differentiation antigens (CD19 and CD20 on malignant B cells, CD25 on malignant T cells), and growth factor receptors/oncogene products (eg, HER2 or NEU). Pediatric ALL can be targeted with anti-CD19 MABs, neuroblastoma can be targeted with anti-GD2 MABs, and AML can be targeted with anti-CD33 MABs. The second criterion is that the MAB must be rendered sufficiently nonimmunogenic to prevent development of neutralizing antibodies when administered to immunocompetent hosts. MABs initially were produced by fusing murine myeloma cells with B cells from mice immunized with specific antigens. MABs generated in this manner are murine proteins; they are recognized as foreign in immunocompetent humans, and, thus, they generate neutralizing antimurine antibodies, which are termed human antimouse antibodies (HAMAs). Following development of these antibodies in patients, the half-life of MABs is reduced greatly, which significantly reduces their biologic activity. Recent advances in genetic engineering have allowed the creation of antibodies in which the murine component of the variable region of the antibody is retained but the rest of the antibody is converted to a human protein. Chimeric antibodies use human constant regions and retain the entire murine variable region, while humanized antibodies retain only the murine hypervariable region responsible for epitope binding; the rest of the variable region and the entire constant region are human derived. Chimeric and humanized antibodies are much less immunogenic than HAMAs, essentially eliminating the problem of HAMA generation. Alternatively, the use of murine MABs in highly immunosuppressed populations sometimes can be accomplished without HAMA induction (Cheung, 2000). The success of MAB therapy in tumors also depends on the ability of the MAB, following targeting of the tumor, to induce death in tumor cells. In some cases, MABs can transmit a death signal directly after binding so that targeting the tumor is all that is required. Examples of this include anti-CD20 in lymphoma cells (Wilson, 2000) and anti-CD99 in Ewing sarcoma (Scotlandi, 2000). (Unfortunately, CD99 also is expressed on hematopoietic progenitors and T cells, thus limiting its clinical potential for tumor targeting.) Recent evidence has suggested that the capacity to induce cell death as a result of MAB binding can be synergistic with cytotoxic chemotherapy. One example of this is trastuzumab, which targets the HER2 or NEU protooncogene in breast cancer and has increased the chemotherapy response rate substantially in a variety of clinical trials (Murray, 1997). For other molecules, simple binding of a MAB is not sufficient to induce cell death. In this case, indirect immunologic effector mechanisms must be induced to kill tumor cells. Such mechanisms include antibody-dependent cell-mediated cytotoxicity (ADCC) in which the Fc portion of the antibody binds to NK cells or monocyte/macrophages, which then are responsible for clearing the targeted tumor cells. Alternatively, antibodies can induce complement fixation with subsequent tumor cell lysis. Another benefit of chimeric human/murine antibodies over murine antibodies is the higher capacity of the human Fc receptor region than the murine Fc region to generate ADCC. Thus, the production of chimeric and humanized MABs has not only eliminated the problem of HAMA generation, but it also has improved efficacy of these agents because of ADCC. Ongoing work in this field is focused on identifying approaches to increase ADCC using agents such as GM-CSF (Yu, 1997), IL-2 (Kendra, 1999), and macrophage colony-stimulating factor (M-CSF) (Munn, 1990). A third mechanism by which antibodies can kill tumor cells involves targeted delivery of toxins or radionuclides. Conjugating antibodies to radionuclides has shown success in lymphomas, although the bone marrow toxicity is substantial and, therefore, this approach generally is undertaken in the setting of marrow rescue (Press, 2000). Recent success with conjugation of antibodies to toxins, including ricin (eg, zolimomab), pseudomonal exotoxin, and calicheamicin (eg, gemtuzumab), also has been observed, although end-organ toxicity remains a concern (Kreitman, 1999; Onda, 2001). MAB therapy in acute leukemia Hematopoietic stem cells differentiate into a wide variety of cell types that, during the course of differentiation, express a wide variety of stage-specific proteins on their cell surfaces. Most leukemic blasts express many of the same differentiation antigens, depending on the stage in which differentiation the leukemic transformation occurred. When the leukemic blast expresses antigens that are not expressed on the normal hematopoietic precursor, a therapeutic opportunity arises that provides the possibility to eradicate the leukemic blasts by targeting the surface antigen while preserving the normal stem cell and the subsequent ability to recover trilineage hematopoiesis. MABs have been generated against such differentiation antigens, and, over the past decade, they have been conjugated with toxins or radiochemicals to target both ALL and AML. Another interesting approach to targeting differentiation antigens has been to conjugate specific ligands of the receptors to toxins or radiochemicals. Although the latter approach has been shown to be an alternative to targeting leukemia and lymphoma, it is not a strictly immunologically based method and, therefore, is not discussed further in this article. In ALL, several promising approaches have included MAB-toxin conjugates directed at CD19 and CD20 on B-lineage leukemic lymphoblasts (Uckun, 1992; Uckun, 1999). Anti-CD20 MAB has been used extensively in adult patients with lymphoma, and a form of anti-CD20 MAB was one of the first to be approved by the US Food and Drug Administration (FDA) (Treon, 2000; Witzig, 2000; Jandula, 2001). Anti-CD20 MAB currently is being tested in combination with chemotherapy in patients with lymphoma and B-lineage leukemia (Maloney, 1998). T-lineage lymphoblastic leukemia has been targeted with similar approaches, with MABs directed at thymic differentiation antigens such as CD7 (Flavell, 2001). In AML, the differentiation antigen, CD33, is expressed in almost all patients. The expression of CD33 increases during normal myeloid differentiation, but self-renewing hematopoietic stem cells do not express CD33. Several murine-derived anti-CD33–directed MABs have been generated, humanized, and conjugated to toxins (eg, calicheamicin) or radiochemicals (radioactive iodine) (Larson, 2001). Clinical trials with these agents have shown that they have significant antileukemic ability. The FDA has approved anti-CD33 conjugated to calicheamicin (Mylotarg) for the treatment of AML in adults, especially elderly persons. In patients with relapse, a response rate of approximately 37% has been observed (Sievers, 2000). Apart from some infusional allergic reactions, the primary toxicity of this approach has been bone marrow suppression caused by binding the MAB-toxin conjugate to normal hematopoietic precursors that express CD33. However, in the absence of mucositis, neutropenia and thrombocytopenia are tolerated significantly better. Another still unexplained toxicity of anti-CD33–calicheamicin conjugates causes hepatic damage, which is characterized by transient increases in liver enzymes in approximately 25% of patients and, occasionally, a more severe complication consistent with venoocclusive disease (Giles, 2001). Currently, the agent is undergoing testing as a single agent and in combination with chemotherapy in children with AML. MAB therapy in neuroblastoma The development of MAB therapy for pediatric solid tumors is illustrated by research on neuroblastoma. Several studies have shown that neuroblastoma is susceptible to complement-mediated lysis and ADCC via lymphocytes, neutrophils, and activated macrophages. Several antibodies directed against the GD2 disialoganglioside, on the surface of neuroblastoma cells, have been developed for use in patients. GD2 is expressed at high densities on nearly all neuroblastoma cells, is not shed from the cell surface, and is restricted to neuroectodermal tissues, thus representing a potentially good target for MAB therapy. Initial studies using the 3F8 MAB, a murine monoclonal antibody in the immunoglobulin G3 (IgG3) subclass, demonstrated that the primary adverse effects of therapy were limited to results of acute toxicity, such as pain, tachycardia, occasional hypotension, fever, anaphylactoid reactions, nausea, vomiting, diarrhea, and reversible neuropathy (Cheung, 2000). Despite toxicity, 3F8 MAB can be administered in the outpatient setting with symptomatic management of the toxic effects. Results of initial nonrandomized clinical trials report a long-term disease-free survival rate of approximately 50% when anti-GD2 monoclonal antibodies are used in conjunction with standard therapy for patients with stage IV neuroblastoma; these results are comparable to those of historical control subjects without MAB exposure (Cheung, Curr Oncol, 2000; Cheung and Guo, 2000). Preliminary evidence suggests that low levels of HAMA and the development of nonneutralizing antiidiotypic antibodies (antibodies that are directed to the variable region of the immunizing antibody rather than the constant region, which is the target of most HAMA) correlate with improved survival rates following adjuvant 3F8 MAB therapy. Current hypotheses suggest that improved survival rates may be the result of induction of an antiidiotypic network. An idiotypic network occurs when the original antibody (Ab1) induces an antiidiotypic antibody (sometimes termed Ab2), which then goes on to further induce an anti-antiidiotypic antibody (sometimes termed Ab3), and so on. The induction of such a network can be beneficial because of the relative similarity that exists between the original antigen (in this case, GD2) and the idiotypic portion of Ab2. The similarity is inferred because they both bind to the same variable region of Ab1 and, therefore, the likelihood is high that they are structurally similar (Cheung, Curr Oncol, 2000; Cheung and Guo, 2000). Since proteins tend to be better immunogens than carbohydrates or gangliosides, the Ab3 response induced by Ab2 may be more potent than the initial Ab1 response. Ongoing research is underway to understand idiotypic networks better and to determine whether such antiidiotypic antibodies are critical components of effective MAB-based therapies. In this regard, it is of interest that induction of antitumor responses can be observed following immunization with Ab2 alone. These results suggest that consolidation with MABs directed at the GD2 gangliosides is worth pursuing in larger randomized clinical trials. Cytokines IL-2 is active against renal cell carcinoma and malignant melanoma, and a 15% response rate is observed in both (Rosenberg, 1987). Although this is a relatively low response rate overall, the rate is particularly notable because standard cytotoxic agents are not effective in patients with renal cell carcinoma or malignant melanoma. Selecting patients who are likely to benefit from treatment with IL-2 and converting nonresponders to responders are difficult. Despite reproducible data showing activity in these diseases, evidence is lacking for IL-2 as a beneficial agent in treatment of patients with other tumors. In pediatric tumors, several trials of IL-2 were performed, and no antitumor effects were seen (Roper, 1992; Bauer, 1995). Even in neuroblastoma, in which questions regarding immune responsiveness often have been asked, systemic administration of IL-2 as a single agent has shown no benefit. Few other cytokines have been administered as single agents in patients with cancer. Activity of interferon (IFN)–alpha has been documented in the treatment of patients with CML (Kantarjian, 1997) and hairy cell leukemia (Pettitt, 1999), both of which are very rare in pediatric patients. Use of IFN-gamma has not been studied as extensively as IFN-alpha, although evidence for up-regulation of MHC in neuroblastoma was observed in one study (Evans, 1989). Preclinical evidence has suggested that the combination of IL-2 and IL-12 has antitumor effects in animal models, and early clinical trials of this combination are underway (Wigginton, 2001). Regional therapy with TNF-a has been undertaken in sarcoma, and some activity was documented, although this approach is limited by the development of systemic toxicity (Stam, 2000).
Innate immunity
T-cell immunity
Monoclonal antibodies
| ||||||||||||||||||||||||||||||||||||||
){kind=small}
