Fraunhofer Institute for Cell Therapy and Immunology IZIOptimizing immunotherapies against malignant melanoma with artificial intelligence
Together with Leipzig University Hospital and the Helmholtz-Zentrum Dresden-Rossendorf (HZDR), the Fraunhofer Institute for Cell Therapy and Immunology IZI is researching the AI-based development of a tumor-specific CAR T-cell immunotherapy for the treatment of malignant melanoma.
Currently, every year, around 24,000 people in Germany are diagnosed with malignant melanoma, i.e. black skin cancer, an extremely aggressive tumour with high metastatic potential. In spite of modern treatment methods, the 10-year survival rate of patients in the metastasized stage is below 20 percent.
So far, considerable success has been achieved in treating malignant melanoma with antibodies (so-called immune checkpoint inhibitors) which inhibit the activity of tumour-masking proteins. This, in turn, reactivates the body’s immune response to the tumour. However, a majority of patients (60 percent) does not respond to these treatments and the recurrence rate is high.
The CAR-T cell treatment, which has already been established in haemato-oncological applications, now also provides new options for the treatment of solid tumours, such as malignant melanoma: For this treatment, the patients’ immune cells are equipped with a synthetic receptor which enables the immune cells to detect potential cancer cells and specifically eliminate them in future.
As part of the “AI-CARs” project, the partners first want to define new target structures for treating malignant melanoma using human tumour samples and to subsequently develop an effective CAR-T cell treatment during the pre-clinical phase.
To this end, Fraunhofer IZI will use state-of-the-art sequencing technologies (single cell sequencing and spatial sequencing) to define potential target structures first. This involves the collection of highly complex datasets reflecting the cells’ dynamic behaviour. In cooperation with Leipzig University, effective interacting molecule pairs and specific target structures will be identified from these data, e.g., using artificial intelligence methods and computer-assisted modelling.
The Helmholtz-Zentrum Dresden-Rossendorf will develop CAR-T cells specifically for these target structures on the basis of a RevCAR-T platform and then test these pre-clinically. Unlike conventional CAR-T cells, RevCAR-T cells cannot directly detect tumour cells. They need an additional tumour-specific target model. As a result, the activity of the RevCAR-T cells and, hence, side effects can be controlled via the target modules’ availability.
