Skip to main content

Single cell analysis of tumor cells for personalized cancer vaccines

Final Report Summary - SCA4PCV (Single cell analysis of tumor cells for personalized cancer vaccines)

Personalized cancer vaccines are therapeutic vaccines custom tailored to target tumor-specific mutations unique to a given patient. Such vaccines may be the ultimate treatment against cancer with the potential to specifically target and destroy lesions all over the body, provide sustained, potentially lifelong immunity to relapse and offer an effective treatment for every patient. A key step in generation of personalized cancer vaccines involves identification and characterization of single nucleotide variations (SNVs) in the patient’s tumor and subsequently selection of a subset of neoepitopes most likely to be immunogenic and confer tumor control. Understanding intratumor heterogeneity can potentially improve the efficacy of such vaccines. The focus of the SCA4PCV program was to develop experimental and computational methods to improve our ability to measure intratumor heterogeneity and use this knowledge to enhance the efficacy of vaccines. A central goal of the SCA4PCV program was to translate findings into the clinic through the host’s clinical personalized cancer vaccine program.

Analysis of the first tumor samples provided to the host as part of its early clinical personalized cancer vaccine trials during the first year of the SCA4PCV project showed that state-of-the-art mutation callers suffer from high false positive rates and exhibit sample-to-sample variability in performance. Furthermore, it was discovered that clinical tumor samples were typically highly impure, further complicating both mutation calling and intratumor heterogeneity analysis. Consequently, the fellow developed an alternative tool to the state-of-the-art called MyMUT® that solves the strongly coupled problem of SNV detection, estimation of sample purity, estimation of absolute copy numbers and estimation of SNV subclonality while achieving an exceptionally low, clinically suitable, false discovery rate. MyMUT® utilizes a novel class of statistical algorithms that provide an exceptionally accurate analysis of the tumor’s exome while achieving reproducible performance from patient to patient regardless of the purity of the patient’s tumor sample or the complexity of the patient’s tumor genome. Since MyMUT® is intended for massive high-throughput automated clinical applications it also includes a sensitive quality control module to detect and flag faulty input. MyMUT® has been extensively tested both in silico on hundreds of publically available samples and experimentally on clinical samples and has proven to be a significant improvement over the state-of-the-art. Consequently, MyMUT® has been integrated into the host’s clinical vaccine platform currently being commercialized by BioNtech in collaboration with Genentech.

Within the scope of the SCA4PCV program the fellow also developed a single cell imaging technique to estimate single cell sorting accuracy by FACS required for accurate identification of cancer subpopulations. Furthermore, the fellow demonstrated the feasibility of extracting single cells and patches of single cells from tumor sections using laser microdissection, post amplifying extracted material and obtaining even genomic coverage. Further experiments to test subclonality predictions in single cells using FACS, the Fluidigm C1 single cell system and laser microdissection are ongoing.

In terms of career development, the fellow was provided with the opportunity to establish a group, mentor PhD students, and subsequently apply for professorship at Mainz University. In addition, the fellow was offered a faculty position in HI-TRON, a soon-to-be established Helmholtz Institute for translational oncology at the University Medical Center Mainz, formed as a partnership between the host, TRON, and the German Cancer Research Center (DKFZ) in Heidelberg. The primary goal of the fellow’s group will be to develop novel computational and experimental methods based on next generation sequencing, single cells and circulating tumor DNA to improve cancer diagnosis and the efficacy of personalized cancer vaccines.