Cancer Immunity 3 Suppl 2:11 (2003)

Engineered T cells as reagents to understand and overcome the limitations of naturally- or vaccine-induced T-cell responses

Philip D. Greenberg^1*, Joseph Blattman¹, Laurence Cheng¹, Michelle Dossett¹, Carla Fowler¹, Yamina Hamel¹, William Ho¹, David Kranz², Adrian Ochsenbein¹, Claes Ohlen¹, Max Topp¹, and Cassian Yee¹

¹University of Washington and Fred Hutchinson Cancer Research Center, Seattle, WA
²University of Illinois, Urbana, IL
^*Presenting author

Abstract

CD8+ T cells have the capacity to recognize and kill virally infected or transformed cells, but generating responses capable of eliminating HIV infection or eradicating established tumors has proven very difficult. Despite improvements in vaccine design, the magnitude of responses achievable with most vaccines remains limited, often making it difficult to determine if therapeutic failures represent quantitative and/or qualitative obstacles. To address this problem, our lab has studied in both clinical trials and in murine models the adoptive transfer of large numbers of virus- or tumor-specific CD8+ T cells to identify the obstacles to efficacy, and has utilized molecular strategies to modify CD8+ T-cell function and fate as a means to determine the requirements for successful immunotherapy.

Patients with malignant melanoma have been treated by administration of autologous CD8+ T-cell clones specific for melanosomal proteins such as MART1. Very limited anti-tumor activity was observed in the initial trial, and immunologic analysis suggested this might reflect a very short in vivo survival of the transferred CD8+ T cells due to apoptosis following encounter with tumor targets. Therefore, a second trial was initiated in which low doses of IL-2 to provide growth and survival signals were administered in association with infusion of the CD8+ clones - under these conditions the CD8+ T cells both persisted and mediated detectable anti-tumor effects. However, another obstacle to efficacy became apparent - the selection of antigen loss variants. This likely reflected consequences of targeting proteins not essential to the tumor for viability or maintenance of the malignant phenotype, and suggests that antigen selection will be a critical element for success in vaccine trials.

Expression profiling has revealed many pro-oncogenic proteins that are over-expressed in malignant cells and contribute to oncogenicity, and might be targeted by CD8+ T-cell responses. However, such proteins are also expressed at lower levels in normal tissues, which can result in deletion of the high affinity repertoire. Thus, vaccination may elicit only low affinity CD8+ T cells, which might not be sufficient to eliminate the tumor cells. Therefore, using WT1 as a model tumor antigen, we have explored, in collaboration with the Kranz lab, methods to generate high affinity TCRs from low affinity responses by in vitro mutagenesis and selection of cloned TCRs. Such modified TCRs can be re-introduced via retroviral vectors into CD8+ T cells to establish a high affinity response, and will be studied in tumor therapy.

In HIV infection, by contrast, the affinity of CD8+ responses for foreign viral antigens is not an obstacle. However, it has been suggested that the virus might decrease the activity of responding CD8+ T cells by inducing a qualitatively abnormal response - in particular, generating CD8+ cells that do not fully differentiate to effector cells, as reflected by retention of CD27 expression. Therefore, the consequences of retained CD27 expression by HIV-specific CD8+ T cells was examined in detail. This TNFR family member was found to deliver a potent pro-survival and proliferative costimulatory signal to CD8+ T cells following antigen encounter, and CD27+ cells demonstrated normal effector functions in vitro, suggesting no inherent disadvantage to CD27 expression. Moreover, following adoptive transfer into an HIV-infected individual, CD27+ CD8+ HIV-specific T cells were found to persist long-term whereas CD27- CD8+ cells derived from the same original clone exhibited only brief in vivo survival. Thus, the observed accumulation of CD27+ CD8+ T cells in HIV-infected individuals, rather than evidence of abnormal differentiation, likely reflects a selection process with preferential in vivo survival, particularly in a setting in which CD4+ T cell help is limiting.

However, the CD8+ response to HIV, as with the response to tumors, is not effective, and one major problem in both settings is the absence of an adequate CD4+ T helper cell response. Administration of exogenous IL-2 can be used to augment or transiently sustain the CD8+ response, but even this approach can be problematic. Doses of IL-2, that are safe and non-toxic and can enhance CD8+ responses at times when antigen burden is low, can cause fatal toxicity at times when the antigen burden is high. This toxicity results directly from stimulation of the responding CD8+ T cells, and suggests that providing better regulated responses may be important. Therefore, we have examined the introduction of genes that might provide regulated autocrine growth signals following target recognition. Studies with two molecular strategies will be discussed - enforced expression of CD28 in effector CD8+ cells to provide a costimulatory signal capable of inducing regulated IL-2 production, and introduction of chimeric IL-2R chains that permit CD8+ T cells to utilize the regulated autocrine production of GM-CSF as an alternative growth signal.

One obstacle common to both tumors and HIV infection is that the targets are not optimal antigen presenting cells, and may provide only weak triggering signals due to many factors, including down-regulation of peptide/MHC complexes and of accessory molecules. The magnitude of this problem for vaccine responses is difficult to directly assess. Therefore, several strategies are being pursued to determine if enhancing signal strength in responding CD8+ T cells can improve therapeutic activity. In particular, expression of dominant negative molecules and/or siRNA molecules targeting regulatory molecules in CD8+ T cells that normally function to dampen responses is being examined. Preliminary studies targeting Cbl-b and Shp-1 will be presented.