Target Selection Drives The Future Of CAR T Therapy In Solid Tumors

By Erin Harris, Editor-In-Chief, Cell & Gene
Follow Me On Twitter @ErinHarris_1

At the 2026 PMWC session on CAR T strategies for solid tumors, I had the chance to moderate a discussion with three of the visionaries who helped build this field, Dr. Carl June, Dr. Michel Sadelain, and Dr. David Barrett. They’ve taken CAR T-cells from early experiments to approved therapies and into the much tougher terrain of solid tumors.

If you want the full breakdown of the session, you can read the complete panel summary here. But it’s worth slowing down at the very beginning because before anyone talked about trafficking, persistence, or engineering tricks, the conversation zeroed in on something far more fundamental: what are we targeting?

And yet, despite all the advances in cell engineering, gene editing, and manufacturing, they kept coming back to the same starting point. Target selection isn’t just one piece of the puzzle; it’s the constraint that shapes everything else. Each leader brought a slightly different lens to the conversation (i.e. clinical, translational, and scientific), but they all agreed that if the target isn’t right, nothing downstream really works.

From CD19 to Complexity: A Shift in Target Paradigms

The modern CAR T field was built on the success of CD19, a lineage-restricted antigen that enabled transformative outcomes in leukemia. Dr. June, whose group led the development of the first FDA-approved CAR T therapy, demonstrated that targeting CD19 could produce deep and durable remissions with manageable toxicity. Similarly, Dr. Sadelain’s early work helped define CAR design principles using CD19 as a model system.

But as Dr. Sadelain emphasized during the panel, CD19 was an exception, not a template. Its near-exclusive expression on B cells created a uniquely clean therapeutic window. In contrast, solid tumor antigens are rarely tumor specific. Instead, they are often shared with essential normal tissues, making them inherently “dirty” targets. This fundamental difference explains why simply applying hematologic CAR T strategies to solid tumors has largely failed.

Dr. Sadelain’s own research at Columbia has increasingly focused on expanding the universe of viable targets. His group has explored nontraditional antigen classes, including aberrantly glycosylated proteins and intracellular-derived peptides presented on the tumor surface. These efforts reflect a growing recognition that the known tumor surface proteome is incomplete, and that better targets may still be undiscovered.

Safety as the Primary Filter

Dr. Barrett, drawing on both his clinical experience and his leadership at Kite Pharma, emphasized that target selection is ultimately constrained by safety. While many antigens are overexpressed in tumors, very few are absent from vital tissues. The key question is not whether a CAR T-cell can kill tumor cells (it almost always can), but whether it can do so without causing unacceptable collateral damage.

His example of GD2 highlights how nuanced this evaluation must be. GD2 is expressed on tumors such as neuroblastoma, but also on peripheral nerves. Antibody therapies targeting GD2 are effective but cause severe pain due to nerve activation. However, Dr. Barrett noted that CAR T-cells appear to interact differently with these tissues, resulting in a more tolerable safety profile. This insight underscores a critical principle, which is target risk cannot be assessed independently of the therapeutic modality.

Dr. Barrett also pointed to a major gap in the field, which is the lack of predictive preclinical models. His earlier work with Dr. June on CD19 CAR T-cells revealed toxicities such as cytokine release syndrome only after treating patients. In solid tumors, this uncertainty is even greater, leading to situations where promising targets are shelved because safety cannot be adequately de-risked before clinical testing.

Lessons from Clinical Trials

Dr. June’s extensive clinical work has highlighted another key barrier: tumor heterogeneity. Solid tumors are genetically and phenotypically diverse, and this diversity evolves over time. In his glioblastoma trials at the University of Pennsylvania, CAR T-cells targeting EGFRvIII successfully eliminated antigen-positive cells but failed to produce durable responses because antigen-negative cells persisted.

This experience has directly informed Dr. June’s current multi-targeting strategy. By engineering CAR T cells to recognize more than one antigen, his team has improved response rates in glioblastoma, achieving meaningful tumor regression in a subset of patients. These findings reinforce the idea that target selection must account not only for specificity and safety, but also for coverage across heterogeneous tumor populations.

Dr. Sadelain echoed this point, noting that single-antigen targeting is unlikely to succeed in most solid tumors. His work has increasingly explored combinatorial targeting approaches, including logic-gated CAR designs that require dual antigen recognition to activate. These strategies aim to both improve specificity and reduce the risk of antigen escape.

The Importance of Binding and Antigen Biology

Another key lesson from both Drs. June and Sadelain’s research is that not all targets, or even all approaches to the same target, are equal. Dr. June highlighted how differences in binding affinity and epitope selection can dramatically affect outcomes. In the case of HER2, early CAR designs using high-affinity binders led to severe toxicity, while later iterations with modified binding properties showed improved safety.

This insight has reshaped how researchers think about target selection. It is no longer sufficient to identify a tumor-associated antigen; the molecular details of how CAR T-cells engage that antigen are equally critical. Dr. Sadelain’s work has contributed to this understanding by systematically studying how CAR structure influences T-cell activation, persistence, and function.

Emerging Targets and Clinical Momentum

Despite these challenges, the panel highlighted growing momentum in solid tumor targeting. Dr. June pointed to recent successes in neuroblastoma and Claudin 18.2 (CLDN18.2)-positive gastric cancers as evidence that the field is making progress. These targets, while not perfect, demonstrate that meaningful clinical activity is achievable.

Dr. Barrett noted that some targets previously considered too risky are being revisited as CAR T technology improves. Advances in cell engineering, many pioneered by Drs. June and Sadelain, are expanding the therapeutic window, making it possible to safely target antigens with limited normal tissue expression.

Target Selection as the Foundation for the Field

The panelists agreed that target selection is not just one component of CAR T-cell therapy, it is the foundation. Dr. June’s clinical insights, Dr. Sadelain’s innovations in CAR design, and Dr. Barrett’s translational perspective all draw the same conclusion. Without the right target, improvements in trafficking, persistence, or tumor microenvironment modulation will have limited impact.

At the same time, their collective work suggests a path forward. By expanding the search for novel antigens, refining multi-target strategies, and integrating deeper biological understanding into CAR design, the field is gradually overcoming the barriers that have long limited solid tumor success.

As Dr. Sadelain noted, the next breakthroughs may come not from better engineering alone, but from discovering entirely new classes of targets. For a field that began with a single, unusually ideal antigen, the future of CAR T therapy in solid tumors will depend on finding the next generation of targets, ones that are not just effective, but truly specific.