New Hope for Immunologic Approaches to Metastatic Melanoma


Synopsis: In two recent clinical reports, novel vaccines were used to enhance host immunological response to metastatic melanoma. In one, the immunodominant epitope of a specific tumor antigen (gp-100) was modified to allow it to bind more avidly with HLA class I molecules, and this was injected, with IL-2, to patients with metastatic melanoma. In the second, dendritic cells derived from patients with melanoma were injected with tumor antigen directly into lymph nodes. Both approaches resulted in demonstrable immunity to the injected tumor antigens and a substantial number of clinical responses.

Sources: Rosenberg SA, Nature Med 1998;4:321-327; Nestle FO, et al. Nature Med 1998;4:328-332; Timmerman JM, Levy R. Nature Med 1998;4:269-270.

The treatment of malignant melanoma remains a major challenge for oncologists, as it has proven time and again to be refractory to chemotherapeutic agents and radiation therapy. However, immunologic approaches have achieved some responses. For example, tumor regression had previously been demonstrated when tumor infiltrating lymphocytes (TILs) were isolated, stimulated in vitro with interleukin-2, and reinfused intravenously.1 This technology allowed the identification of several melanoma-associated antigens including gp-100, MART-1, tyrosinase, MAGE-1, and MAGE-3, which bind HLA class I molecules on antigen presenting cells and activate cytotoxic T lymphocytes (CTLs). The immunodominant epitopes of these antigens have been prepared and modified, and clinical trials have been undertaken with the goal of inducing anti-tumor immunity in melanoma-bearing patients.

Two papers recently published in Nature Medicine provide early results from these clinical trials and afford new hope that such vaccine strategies may ultimately be of great clinical value for patients with this disease. Rosenberg and colleagues at the National Cancer Institute modified the immunodominant epitope of the gp-100 antigen in such a way as to make it bind to the HLA-A1 molecule more avidly. They demonstrated that such modification was associated with enhanced CTL generation in vitro and, in contrast to the unmodified peptide that produced no demonstrable in vivo response, injection in patients with melanoma was associated with demonstrable T-cell immunity in 91% of cases. More importantly, 13 of 31 patients (42%) with measurable metastatic disease had objective evidence of tumor regression when injected with the modified peptide with incomplete Freund's adjuvant and IL-2. Furthermore, these tumor regressions were seen not only in skin disease but also in lymph nodes, liver, lung, and brain. This compares quite favorably to the response rate of 17% in patients treated with IL-2 alone at the same institution.2

Rather than use modified peptide antigens, Nestle and colleagues from Switzerland and Germany tested a different vaccine strategy. Dendritic cells, which are resident cells in the skin that are the body's most efficient antigen presenting cells, were isolated, cultured in vitro with processed melanoma antigens, and injected with keyhole limpet heomcyanin (KLH) as a vaccine. The rationale here is that the dendritic cells with tumor antigen bound to both HLA class I and II molecules will more effectively deliver antigen to T cells and thereby produce a more potent immunization. Dendritic cells were prepared from the peripheral blood of 16 patients and cultured in vitro with a variety of melanoma antigens (or, in some cases, tumor cell lysates) and then injected directly into lymph nodes. Tumor regression was seen in five of the 16 patients, and, in two of these, a complete response has lasted for longer than one year. Of note, two of the five responders were the recipients of the tumor cell lysate-prepared vaccine.


These are two remarkable papers, not only because they introduce new techniques for treatment of a very difficult malignant disease, but because they expand our understanding of tumor immunology and biology. By altering peptide structure, Rosenberg and colleagues enhanced the immunogenicity of a melanoma tumor antigen and witnessed both in vitro and in vivo immunologic responses. Furthermore, they demonstrated that this novel peptide vaccine, when combined with IL-2 administration, produced anti-tumor responses in a substantial number of cases. Metastatic disease regression was observed in liver and brain, areas that have been considered refractory to other immunological approaches. Thus, the modification of tumor antigen to enhance binding to HLA molecules offers a new and potentially exciting approach to enhancing the immunotherapy of melanoma. Rosenberg et al speculate that other antigens may also prove amenable to such modification-enhancement, and future trials using these, either alone or in combination, are underway. As Timmerman and Levy point out in their accompanying editorial, application of these approaches needs to be undertaken with great caution because under some circumstances, peptide antigens have resulted in the induction of immunologic tolerance and enhanced tumor growth.3 Nonetheless, by careful and controlled clinical investigation, this approach may result in a major step forward in melanoma biotherapy.

Perhaps some cautionary note should be sounded. As I recall, the response rate reported here is remarkably similar to prior response rates obtained by these authors-first with lymphokine-activated killer (LAK) cells plus IL-2 and then with tumor-infiltrating lymphocytes (TIL) plus IL-2. Those 40% response rates have not been replicated. Yet, the vaccine approach reported here certainly seems better grounded scientifically than the prior approaches. LAK cells were found not to traffic to tumor in vivo and TIL were not easily generated and were often found by others to be functionally exhausted from the constant exposure to tumor and IL-2 over many weeks in vitro. By contrast, the modification of a tumor antigen to bind to MHC molecules more avidly could make it a better antigen and could be more successful at eliciting antitumor immune responses.

Similarly, although the concept of dendritic cell vaccines had been previously promoted, their use in melanoma and the demonstration of clinical response is a major step forward.4 The use of KLH as adjuvant and the injection directly into lymph nodes may both prove to be useful technical advances as well. KLH was used in a prior study in which anti-idiotype immunity was successfully generated. It is of particular interest to note that two of the five responses were seen in patients who had received dendritic cells prepared with tumor lysates rather than purified tumor antigens. This raises the possibility that similar vaccine strategies could be used against tumors for which purified antigen preparations are not available. Furthermore, it permits the possible generation of several clones of T cells that recognize several different tumor antigens.


    1. Rosenberg SA, N Engl J Med 1988;319: 1676-1680.

    2. Rosenberg SA, JAMA 1994;271:907-913.

    3. Toes REM, J Immunol 1996;156:3911-3918.

    4. Schuler G, Steinman RM. J Exp Med 1997;186: 1183-1187.