Tumor Antigen Heterogeneity: The “Elephant in the Room” of Adoptive T-cell Therapy for Solid Tumors

Adoptive transfer of T cells expressing chimeric antigen receptors (CAR) has revolutionized the treatment of many hematologic malignancies. Unfortunately, T-cell transfer for the treatment of solid tumors has not yet achieved similar success (1). One of the most important yet understudied challenges is the issue of tumor antigen heterogeneity (2), because not all tumor cells will express the targeted antigen, and even when present, the levels may be quite variable. In addition, therapy-induced immune editing could lead to antigen escape. Successful therapy will, thus, need to overcome this critical hurdle.

Tumor antigen heterogeneity could be overcome by the presence of “bystander effects,” the ability of the transferred T cells to also kill tumor cells that are not expressing the targeted antigen(s). This could occur by release of cytokines from the activated T cells or by engagement of death receptors on the nontargeted tumor cells. However, it is hoped that the most powerful bystander effects would be due to epitope (or antigen) spreading. In this process, transferred T cells kill targeted tumors cells, resulting in the release of tumor antigens in an immunostimulatory environment that are then taken up by antigen-presenting cells that have the ability to cross-present tumor antigens, ultimately generating endogenous B-cell and T-cell immunity against tumor antigens not originally targeted by the transferred T cells (3).

The extent of bystander killing that occurs after CAR or T-cell receptor (TCR) T-cell therapy has not been well studied. In this context, the study by Xin and colleagues in this issue (4), whose major goal was defining and augmenting epitope spreading, is timely and important. Using a syngeneic mouse model with defined tumor antigens, this group provide data that epitope spreading after treatment with T cells expressing a TCR targeted to the melanoma tumor antigen gp100 was minimal, resulting in the inability to eradicate tumor cells lacking gp100. However, by cotransducing the adoptively transferred T cells with a second TCR that recognizes a bacterially expressed antigen (ovalbumin) and treating tumors with modified Listeria bacteria that expresses ovalbumin, epitope spreading [mediated by BATF3⁺ dendritic cells (DC)] and bystander effects were markedly enhanced, along with the generation of central memory and tissue-resident memory CD8⁺ T cells. Although this study examined T cells transduced with TCRs and may have some issues relating to practicality and safety of intratumoral injections, it serves as an important proof of principle showing the importance of augmenting epitope spreading. Our group has similar data showing that CAR T cells are also very limited in the degree of epitope spreading that they induce without augmentation.

Finding a way to enhance bystander effects will be critical for success in treating most solid tumors where less than 100% of tumors express targeted antigens and in preventing immune editing–induced escape after therapy. Increasing epitope spreading is one strategy. This could include the approach described by Xin and colleagues in this issue (4) or DC activators delivered by the transferred T cells (i.e., IL12, flagellin, or CD40 ligand) or administered systemically (i.e., agonistic anti-CD40 antibody, FLT3 ligand, or STING agonists). Other potentially useful cargo could be cytokines or secreted bispecific T-cell engagers that would kill nonantigen-expressing cells. It may also be possible to address tumor heterogeneity by injecting mixtures of T cells that target multiple tumor antigens, transducing T cells with multiple CARs or TCRs targeting different antigens, or designing CARs that target multiple antigens.

Disclosure of Potential Conflicts of Interest

S.M. Albelda reports receiving commercial research grants from Novartis and Tmunity. No other potential conflicts of interest were disclosed.