Investigational CRISPR-Cas9-edited T cells for metastatic colorectal cancer treatment: Clinical insights for future development

Main text

Significant efforts in immunobiology and tumor cell biology have led to important advances in incorporating immunotherapy into treatment for patients with metastatic colorectal cancer (CRC) (mCRC). However, these successes have for the most part been achieved in the relatively small number of patients with tumors harboring a mismatch repair deficiency¹ resulting in microsatellite instability-high (MSI-H) disease; notably, in CRC this number is lowest (∼5%) in the metastatic setting than in earlier stages (∼10%–15%).^1,2,3 The remaining 90%–95% of cases of mCRC are classified as microsatellite stable, and the field of gastrointestinal oncology has been intensively investigating how to induce immunogenicity in this large subset of patients, while also exploring other biomarkers or conditions that could effectively enable successful immune checkpoint blockade.

The strategic use of unmanipulated (endogenous, non-edited) cell therapeutic products for patients with gastrointestinal cancers, including those with mCRC, has gained ground in recent years^4,5,6 but has yet to result in complete responses. Partial responses, when reported, are relatively few and far between.⁷ At the 2025 Annual Meeting of the American Association for Cancer Research, our institutional team reported results from a first-in-human clinical trial using CRISPR-Cas9-based editing to knock out the gene encoding for an intracellular immune checkpoint called CISH (cytokine-inducible SH2-containing protein),⁸ utilizing patient tumor-derived autologous tumor-infiltrating lymphocytes (TILs).⁹ This treatment resulted, most remarkably, in a complete clinical response in a young adult female patient with mCRC refractory to multiple prior lines of therapy, including immune checkpoint blockade. An additional patient gained highly stable disease for several months after experiencing an initial period of significant disease progression.

This trial provided an extremely promising signal, providing an unprecedented and durable cure (nearly three years to date) for a young adult patient with long-standing mCRC resistant to chemotherapies, biologic therapies, and combination as well as single-agent forms of immunotherapy. The key challenge now is to build upon this unique success to fulfill the potential and promise of gene-edited immune cell therapies. In this era of tissue-based molecular biomarkers, a biomarker-independent treatment such as CRISPR-engineered cell therapy offers the potential for broader clinical applicability.

How do we convert promising clinical signals to the predominant form of clinical response? An insight to consider for future development and design of such trials is chronobiology, an aspect increasingly being investigated for immune checkpoint inhibitors¹⁰ and other forms of cancer-directed therapy,^11,12,13 and one that would benefit considerably from mathematical modeling for dosing regimen optimization.¹⁴ A recent preprint by Lyons et al. found that morning administration of chimeric antigen receptor (CAR)-T cell infusions is associated with higher overall survival and lower toxicity, with overall survival depending on the time of day of infusion in an hour-by-hour manner.¹⁵ A recent opinion piece by Tellinga et al. tackled the potential utility in considering the time of day when using adoptive cell therapies specifically.¹⁶ The authors provided an interesting premise, supported by preclinical data showing reduced toxicity and improved efficacy of this treatment strategy. The authors surmise that circadian rhythms affect aspects as far-reaching as the extent and ability of adoptive cell therapy tissue infiltration, cell therapy product metabolism, and overall patient toxicity; the last point is especially profound when considering the potential for cytokine release syndrome or similarly harmful treatment-related adverse events that require cautious evaluation and management.

Biomarkers that predict successful editing of cell products and, in the post-treatment setting, identify biological response or minimal residual disease (MRD) in cases of exceptional responses are an emerging clinical need. The growing role of cell-free/circulating tumor (ct) DNA in solid tumor oncology, particularly in CRC oncology, presents major opportunities to further refine this technology and enhance its efficacy. The technology for measuring and evaluating ctDNA has taken hold over the past decade and improved substantially in terms of sensitivity and specificity, and ctDNA testing is now accepted as a strong indicator of tumor burden dynamics.^17,18,19,20 Some consensus guidelines and trials are now helping to define how ctDNA can be optimally used. However, knowledge gaps in the field remain, regarding the sensitivity, specificity, timing, and frequency of ctDNA collection to most effectively guide clinical decision making and impact patient survival. Detecting MRD remains one of the most pressing and actively investigated aspects of ctDNA analysis. Additional challenges include integrating correlative studies and financial costs of ctDNA testing and attaining uniformity between testing platforms. The short half-life of ctDNA suggests that high-frequency sampling may be more predictive than low frequency, due to its ability to detect change rather than absolute values. Earlier this year, our team published a mathematical model that suggested that testing for ctDNA within days to weeks—rather than months, as is the current standard—of any given treatment intervention may be more predictive of objective response.²¹ We propose that ctDNA can be leveraged effectively to further improve upon multiple aspects of CRISPR-engineered cell therapies, including adding a novel biomarker of potential response beyond dependence on radiographic imaging (Figure 1). This could potentially allow clinicians to more rapidly identify responders and non-responders. Furthermore, genomic evaluation by ctDNA may provide an added benefit of capturing tumor heterogeneity, which is commonly missed by biopsy specimens. This heterogeneity may introduce selection biases and confound the interpretation of neoantigen reactivity assays currently used to select TILs in clinical trials.

Figure 1.

Mathematical modeling for refining CRISPR-based gene editing for cell therapy for colorectal cancers: incorporating quantitative measures of tumor burden, ctDNA, and TIL dynamics

With mathematical modeling, we can connect data across different modalities and frequencies (e.g. more frequent ctDNA, less frequent tumor burden, etc.) using underlying models of the biological processes. This approach would allow us to map the treatment design and i implementation to individual patient characteristics, TIL design features, and other aspects of this therapeutic strategy. Modeling will in turn uncover events in real-time related to latent previously unobserved dynamics, and allows us to infer the underlying parameters governing TIL efficacy with the ultimate goal of improving efficacy and safety.

In summary, there is a strong future and a solid foundation for ongoing pursuit of CRISPR-Cas9-edited T cells for metastatic colorectal cancer treatment. Despite improved survival rates for patients with mCRC as compared to decades past, the needle has moved surprisingly little over the past few decades in terms of adding novel therapeutic strategies that can induce high rates of objective clinical response while prolonging survival and quality of life by years, rather than by mere weeks to months. The time is now to integrate mathematical modeling, ctDNA-based observational biomarkers, and other novel approaches toward future development of this emergent technology.

Acknowledgments

The first-in-human clinical trial using CRISPR-Cas9-based editing to knock out the gene encoding for the intracellular immune checkpoint CISH and related correlative scientific work was funded by Intima Bioscience, Inc.

Declaration of interests

The authors declare no competing interests.