Immunotherapy of Myelodysplastic Syndrome: You Can Run, but You Can't Hide

In this issue of Clinical Cancer Research, Griffiths and colleagues conduct a phase I clinical trial in patients with myelodysplastic syndrome (MDS) of a combinatorial immunotherapy (Fig. 1) comprising the hypomethylating agent decitabine plus a vaccine against the “cancer/testis” antigen NY-ESO-1 (1). This strategy addresses a critical, unmet need in cancer immunotherapy: the treatment of cancers with few available immunologic targets.

The success of the immunologic checkpoint inhibitors (CI) and of chimeric antigen receptor–modified T cells, or CAR T cells, has raised the level of enthusiasm for cancer immunotherapy to a fever pitch. Studies on the mechanism of the antitumor effect of the anti–CTLA-4 antibody ipilimumab or the anti–PD-1 antibodies nivolumab or pembrolizumab have shown that these CIs stimulate endogenous (i.e., patient-derived) T cells specific for tumor “neoantigens” (2), tumor-specific antigens that result from somatic mutation. The reactivity of CAR T cells against tumor cells is achieved by inserting into T cells a receptor whose extracellular domain consists of the antigen-binding portion of an antibody reactive to a molecule on the tumor cell surface while the intracellular portion triggers the cytolytic machinery of the T cell. Thus, CAR T cells combine the exquisite specificity of antibodies for protein antigens with the lethality of cytotoxic T cells.

Still, there are major obstacles to the broad application of immunotherapy to the treatment of several common tumor types. For example, the efficacy of CIs is limited primarily to tumors with large numbers of clonal mutations (3). Even in tumors with a high mutational burden, such as melanoma or lung cancer, response rates to CIs rarely exceed 50%, and clonal escape from immunosurveillance or active resistance mechanisms are common (4). Tumors with low numbers of somatic mutations, such as pediatric cancers or cancers of the breast, prostate, ovary, or pancreas, generally respond poorly to CIs. The major limitation to CAR T cells is that the target of the chimeric receptor must be expressed on the cell surface, yet there are no cell surface molecules that distinguish a cancer cell from its normal tissue counterpart. Thus, CAR T cells are tissue specific rather than tumor specific and can only be used to treat cancers deriving from tissues, such as B lymphocytes, that are not essential to the survival of the patient.

So, what can be done to stimulate an effective antitumor immune response against cancers with no unique targets on the cell surface and a low mutational burden? This is the challenge that Griffiths and colleagues confronted in developing an approach to the immunotherapy of MDS (1). MDS is a clonal hematopoietic disorder characterized by ineffective hematopoiesis and an inexorable progression to acute myeloid leukemia (AML). Because the molecules on the surface of MDS cells are all expected to be expressed on normal white blood cells, it would be difficult if not impossible to target MDS with CAR T cells without unacceptable toxicities. Furthermore, MDS is characterized by a low mutational burden, and CIs have only modest activity against this disease (5).

The only approved therapies for MDS are the hypomethylating agents azacitidine and decitabine. Although their mechanism of action has not been fully defined, several lines of evidence suggest that demethylation of promoter regions enables the reexpression of potentially immunogenic proteins, leading to a therapeutic antitumor immune response. One such protein with inducible expression on myeloid cells is the cancer/testis antigen NY-ESO-1. Cancer/testis antigens are a heterogeneous group of proteins that are normally expressed in germ cells and in diverse types of cancers, but not in differentiated tissues of the adult. However, NY-ESO-1 is constitutively expressed in up to one third of patients with melanoma and lung, esophageal, liver, gastric, prostate, ovarian, or bladder cancers (6), and antibodies against NY-ESO-1 can be found in the serum of such patients. In patients with MDS or AML, the promoter region of the NY-ESO-1 gene is heavily methylated, and so the protein is not usually made in the malignant cells.

In a previous study, the authors showed that the expression of cancer/testis antigens NY-ESO-1 and MAGEA3/A6 is induced on AML blasts in patients treated with decitabine (7). This finding raised the possibility that decitabine could be used to increase NY-ESO-1 expression in blasts while a vaccine against NY-ESO-1 could be administered to generate a cytotoxic T-cell response from autologous T cells specific for this antigen. In the current study, Griffiths and colleagues conducted a phase I trial in which nine patients with MDS received an HLA-unrestricted NY-ESO-1 vaccine (CDX-1401 + poly-ICLC) every 4 weeks with standard-dose decitabine. NY-ESO-1 expression was induced in all seven patients who reached the end of the study, and NY-ESO-1–specific CD4⁺ and CD8⁺ T-cell responses were seen in six of seven and four of seven vaccinated patients, respectively (1). Myeloid cells, isolated from a patient at different time points during decitabine therapy, expressed NY-ESO-1 and activated cytotoxicity from autologous, NY-ESO-1–specific T cells. The intensity of the T- and B-cell response to NY-ESO-1 vaccination was somewhat heterogeneous but appeared to correlate with the prevaccination frequency of dendritic cells (DC) that express the cell-surface marker CD141. Perhaps this correlation is not surprising, as these DCs express high levels of DEC-205, which is targeted by the vaccine, and TLR3, which is targeted by the vaccine adjuvant. Finally, a clinical response to decitabine plus NY-ESO-1 vaccination was associated with an increase in the frequency of CD141^HI DCs in the bone marrow.

Although the results of this combinatorial strategy are encouraging, some observations need further characterization and explanation. First, there was not a consistent correlation between immunologic and clinical responses. The patient with the best immunologic response to treatment achieved a durable complete remission, but the next best immunologic responder (patient 2) only had stable disease following treatment. Second, one patient (patient 6) expressed NY-ESO-1 in myeloid cells prior to decitabine treatment, but this was not associated with an endogenous immune response. Finally, another patient (patient 7) experienced a durable complete remission despite a relatively weak induction of CD4⁺ T cells and no induction of CD8⁺ T cells. One explanation for dissociation between NY-ESO-1 expression and response to vaccination is that some patients may be immunologically tolerant of NY-ESO-1. Although NY-ESO-1 is generally not expressed in adult tissues, it has been found to be expressed in the medullary thymic epithelium (8), where it can induce tolerance in developing T cells. Indeed, cancer/testis antigen–specific T cells isolated from cancer patients often have suboptimal affinity and do not mediate significant killing (9), suggesting that high-affinity clones have been physically or functionally deleted from the repertoire. Although this finding may bode poorly for strategies of cancer vaccination, decitabine-induced expression of cancer/testis antigens could be combined effectively with adoptive immunotherapy with high-affinity T cells specific for these antigens.

In summary, Griffiths and colleagues (1) have provided an elegant tool for sensitizing cancer cells to tumor-specific T cells by using a demethylating agent to express otherwise cryptic antigens. Further studies will be required to determine the generalizability of this approach.

No potential conflicts of interest were disclosed.