The manufacturing process for personalized T-cell therapies hardly begins before it stalls. Why? Right at the start, there is a severe bottleneck: the need to identify patient-derived, tumor-reactive T-cell receptors (TCRs).

To overcome this bottleneck, scientists at the German Cancer Research Center (DKFZ) and the University Medical Center Mannheim have developed predicTCR, a machine learning classifier. According to the scientists, it can identify individual tumor-reactive tumor-infiltrating lymphocyte (TILs) in an antigen-agnostic manner based on single-TIL RNA sequencing.

The scientists also assert that prediTCR can halve the time it takes to get past the bottleneck, helping to reduce the overall time needed to make a personalized T-cell therapy for cancer patients. Since the overall time is at least six months, any reduction in the time needed to complete any manufacturing step is welcome.

Details about predicTCT and its application recently appeared in Nature Biotechnology, in an article titled, “Prediction of tumor-reactive T cell receptors from scRNA-seq data for personalized T cell therapy.”

The paper makes it clear that the scientists’ approach applies to personalized transgenic T-cell therapies, which seek to identify and reinfuse defined tumor-reactive TCRs, either in patient-autologous T cells or in induced pluripotent stem cell–derived, hypoimmunogenic (allogeneic) T cells. These therapies are not produced via the enrichment of tumor-reactive T cells, so they can avoid the problem of T-cell exhaustion. However, they pose a “needle in a haystack” problem, simply because identifying tumor-reactive TCRs is so difficult.

The development of personalized transgenic T-cell therapies is a complicated process. First, doctors isolate TILs from a sample of the patient’s tumor tissue. This cell population is then searched for T-cell receptors that recognize tumor-specific mutations and can thus kill tumor cells. This search is laborious and has so far required knowledge of the tumor-specific mutations that lead to protein changes that are recognized by the patients’ immune system. During this time, the tumor is constantly mutating and spreading, making this step a race against time.

“Finding the right T-cell receptors is costly and time-consuming,” said Michael Platten, MD, head of brain tumor immunology at the DKFZ and director of the department of neurology at the University Medical Center Mannheim. “With a method that allows us to identify tumor-reactive TCRs independently of knowledge of the respective tumor epitopes, the process could be considerably simplified and accelerated.”

To develop such a method, a team led by Platten and co-study head Edward W. Green, PhD, group leader, immunogenomics, began by isolating TILs from a melanoma patient’s brain metastasis and performed single-cell sequencing to characterize each cell. The T-cell receptors expressed by these TILs were then individually tested in the laboratory to identify those that were recognized and killed patient tumor cells. The researchers then combined these data to train a machine learning model to predict tumor-reactive T-cell receptors. The resulting classifier could identify tumor reactive T cells from TILs with 90% accuracy, works in many different types of tumor, and accommodates data from different cell sequencing technologies.

“PredicTCR identifies tumor-reactive TCRs in TILs from diverse cancers better than previous gene set enrichment-based approaches, increasing specificity and sensitivity (geometric mean) from 0.38 to 0.74,” the authors of the Nature Biotechnology article wrote. “By predicting tumor-reactive TCRs in a matter of days, TCR clonotypes can be prioritized to accelerate the manufacture of personalized T-cell therapies.”

“We are now focusing on bringing this technology into clinical practice here in Germany,” Platten added. “To finance further development, we have founded the biotechnology start-up Tcelltech. predicTCR is one of the key technologies of this new DKFZ spin-off.”