Abstract:
Despite 50 years of dedicated efforts in the "War on Cancer", the results achieved in
treating cancer have been regrettably unsatisfactory. Particularly challenging are
tumors characterized by high inter- and intratumoral heterogeneity and an advanced
stage. These complex cancers, including ovarian, breast and glioblastoma, present
significant barriers to effective therapy. Nowadays, the healthcare system faces two
major issues in cancer treatment: low success rates of newly approved anti-cancer
drugs and ineffective treatments leading to adverse side effects in patients. The ability
to pre-select and pre-determine individualized treatment options prior to clinical care,
as envisioned in the context of personalized medicine, could thus facilitate therapeutic
decision making and ultimately improve patient outcomes. This will require advances
in the implementation of diagnostic tools for detailed and accurate patient stratification,
and advances in the prediction of patient-specific response to treatment. To enable
preclinical validation of anticancer drug efficacy in personalized cancer therapy, it is
crucial to develop patient-derived tumor models that mirror the unique complexity of
individual tumors and account for the significant impact of the tumor microenvironment
and cellular diversity on drug response. In this context, this study presents an ex vivo
tumor model composed of patient-derived 3D microtumors (PDM) and autologous
tumor-infiltrating immune cells (TIL), established and validated for ovarian cancer,
breast cancer, and glioblastoma patients to identify individual tumor vulnerabilities. By
limited digestion and subsequent culture in defined media, PDM and TIL cultures with
high viability were successfully generated from freshly resected primary tumors. In-
depth histopathological, immunohistological and proteomic analyses of PDM and
corresponding primary tumors were performed and confirmed conserved subtype-
specific histology, tumor marker expression, and the presence of tumor
microenvironment components including extracellular matrix, tumor-associated
macrophages, and cancer-associated fibroblasts. Comprehensive protein profiling of
up to 200 analytes was performed in both primary tumors and PDM with limited sample
material using advanced technologies such as DigiWest® and RPPA immunoassay
screening. The preservation of molecular protein signatures and molecular
heterogeneity of the original primary tumor in PDM was confirmed by the extensive
protein data obtained. Functional drug testing on PDM and PDM-TIL co-cultures with
small molecules, chemotherapeutic as well as immunotherapeutic agents identified
tumors sensitive to specific treatments, enabling the prediction of individual therapeutic susceptibility. In combination with the collected proteomic data, molecular protein
signatures have been revealed that correlate with treatment response and resistance.
The clinical utility of PDM is based on their efficient isolation process, time-saving
generation, ethical non-animal culture conditions, patient-specific representation,
preservation of tissue architecture and TME components, and compatibility with
various downstream readout technologies. These combined advantages position PDM
as a powerful and versatile tool that holds great promise for drug mode of action
analyses, biomarker identification and personalized therapeutic sensitivity prediction.