We are writing in response to a letter by Manjili et al. (1) regarding our recent report demonstrating that tumors can recur in patients despite the induction, by peptide vaccines, of large numbers of T cells capable of recognizing antigens expressed by the tumor (2). Manjili et al. (1) suggest new mechanisms for tumor escape, which would add to those described previously (3).
However, new evidence using adoptively transferred antitumor lymphocytes indicates that the major problem in cancer immunotherapy may not be tumor escape but rather inadequate or inconsistent function of the peptide-induced T cells.
Preclinical and clinical studies indicate that vaccines can increase the frequency of antitumor CD8+ T cells. Indeed, investigators often report the induction of CD8+ T cell responses as evidence for vaccine efficacy. This conclusion seemed logical because of incontrovertible evidence for the existence of antitumor T cells recognizing tumor-associated Ags. Many workers in academia and industry expected that the elicitation of strong cellular immune responses would result in tumor cell destruction, but current vaccine approaches have proven to be disappointing (4). Data indicate that the mere presence of antitumor CD8+ T cells may not be a “surrogate marker” for tumor regression in mouse or in man (2, 5). Although increases in frequencies are often reported as small fractions of a percent, our own recent study (2) indicated that inducing very large numbers of antitumor T cells might be insufficient to prevent tumor growth, even in patients with minimal residual disease. Studies in TCR transgenic mice where tumor-specific CD8+ T cells exceeds 90% have similarly shown that antitumor T cells do not necessarily affect the growth or lethality of even small tumor implants (5).
Although tumor-escape mechanisms seem like a plausible explanation of the coexistence of tumor growth and large numbers of avid antitumor T cells, recent advances in the use of adoptive cell transfer (ACT) casts doubt on the notion that tumor escape is the primary mechanism underlying the failure of current therapeutic vaccines. About 50% of patients experience objective clinical responses when they receive lymphodepletion prior to ACT and IL-2 (6, 7), and many of these patients have previously failed therapeutic vaccine-based immunotherapies. The success of ACT may indicate that successfully treated patients did not bear tumor escape variants that were invulnerable to immune destruction.
It seems likely that ex vivo culture of T cells allows their more complete activation. Transfer into the lymphoablated setting results in further CD8+ T cell activation and tumor destruction. Mouse models reveal that lymphodepletion can remove T regulatory cells and “sinks” for homeostatic cytokines capable of activating antitumor T cells in ACT-based regimens (8). Thus, the regression of large established tumors after ACT indicates that tumors are vulnerable to immune attack by robustly activated T lymphocytes.
