Search

Neoantigen discovery in pediatric brain tumors is hampered by their low mutational burden and scant tissue availability. Here we develop a proteogenomic approach combining tumor DNA/RNA sequencing and mass spectrometry proteomics to identify tumor-restricted (neoantigen) peptides arising from multiple genomic aberrations to generate a highly target-specific, autologous, personalized T cell immunotherapy. Our data indicate that aberrant splice junctions are the primary source of neoantigens in medulloblastoma, a common pediatric brain tumor. Proteogenomically identified tumor-specific peptides are immunogenic and generate MHC II-based T cell responses. Moreover, polyclonal and polyfunctional T cells specific for tumor-specific peptides effectively eliminate tumor cells in vitro. Targeting tumor-specific antigens obviates the issue of central immune tolerance while potentially providing a safety margin favoring combination with other immune-activating therapies. These findings demonstrate the proteogenomic discovery of immunogenic tumor-specific peptides and lay the groundwork for personalized targeted T cell therapies for children with brain tumors., Targeting tumor-associated antigens in paediatric medulloblastomas (MB) is challenging due to their low mutational burden. Here, the authors develop a sensitive proteogenomic approach to identify tumour specific neoantigens, which may enable personalised T cell immunotherapy in paediatric MB.

Background The germline genetic events underpinning medulloblastoma (MB) initiation, and therefore the ability to determine who is at risk, are still unknown for the majority of cases. Microsatellites are short repeated sequences that make up ~3% of the genome. Repeat lengths vary among individuals and are often nonrandomly associated with disease, including several cancers such as breast, glioma, lung, and ovarian. Due to their effects on gene function, they have been called the “tuning knobs of the genome.” Methods We have developed a novel approach for identifying a microsatellite-based signature to differentiate MB patients from controls using germline DNA. Results Analyzing germline whole exome sequencing data from a training set of 120 MB subjects and 425 controls, we identified 139 individual microsatellite loci whose genotypes differ significantly between the groups. Using a genetic algorithm, we identified a subset of 43 microsatellites that distinguish MB subjects from controls with a sensitivity and specificity of 92% and 88%, respectively. This microsatellite signature was validated in an independent dataset consisting of 102 subjects and 428 controls, with comparable sensitivity and specificity of 95% and 90%, respectively. Analysis of the allele genotypes of those 139 informative loci demonstrates that their association with MB is a consequence of individual microsatellites' genotypes rather than their hypermutability. Finally, an analysis of the genes harboring these microsatellite loci reveals cellular functions important for tumorigenesis. Conclusion This study demonstrates that MB-specific germline microsatellite variations mark those at risk for MB development and suggests mechanisms of predisposition.

Genomic characterization has begun to redefine diagnostic classifications of cancers. However, it remains a challenge to infer disease phenotypes from genomic alterations alone. To help realize the promise of genomics, we have performed a quantitative proteomics investigation using Stable Isotope Labeling by Amino Acids in Cell Culture (SILAC) and 41 tissue samples spanning the 4 genomically based subgroups of medulloblastoma and control cerebellum. We have identified and quantitated thousands of proteins across these groups and find that we are able to recapitulate the genomic subgroups based upon subgroup restricted and differentially abundant proteins while also identifying subgroup specific protein isoforms. Integrating our proteomic measurements with genomic data, we calculate a poor correlation between mRNA and protein abundance. Using EPIC 850 k methylation array data on the same tissues, we also investigate the influence of copy number alterations and DNA methylation on the proteome in an attempt to characterize the impact of these genetic features on the proteome. Reciprocally, we are able to use the proteome to identify which genomic alterations result in altered protein abundance and thus are most likely to impact biology. Finally, we are able to assemble protein-based pathways yielding potential avenues for clinical intervention. From these, we validate the EIF4F cap-dependent translation pathway as a novel druggable pathway in medulloblastoma. Thus, quantitative proteomics complements genomic platforms to yield a more complete understanding of functional tumor biology and identify novel therapeutic targets for medulloblastoma. Electronic supplementary material The online version of this article (10.1186/s40478-018-0548-7) contains supplementary material, which is available to authorized users.

T cell immunotherapies are promising new tools to combat high-risk subgroups of medulloblastoma without increasing the late effects burden. The ideal target antigen of an effective antitumor T-cell response is abundantly expressed by tumor cells but not by normal tissues, in order to limit off-target effects. Tumors translate a host of highly novel transcripts that are the result of aberrations in tumor DNA and the unmasking of alternative or novel exons. We developed a novel proteogenomic approach to identify tumor-restricted peptides and used them to select and expand T cells capable of mounting a tumor-specific cytotoxic immune response. Using RNA-seq and WGS data, we created personalized custom searchable databases containing predicted novel proteins from somatic mutations, novel junctions and fusion transcripts from 56 medulloblastoma tumors. By searching these databases with raw mass spectrometry data from paired medulloblastoma tumors, we identified tens of neoantigen peptides arising from the translation of tumor-specific transcripts; novel isoforms being the predominant source. We tested these peptides for their ability to select and expand autologous polyclonal populations of T cells and tested the immunogenicity of each individual peptide. Flow cytometry revealed populations of CD4+ and CD8+ cells with an activation profile marked by IFN-γ and TNF-α. Immunosuppressive marker profiles were also characterized. Using cytotoxicity assays, we demonstrated that tumor specific T cells can eliminate neoantigen bearing tumor cells. Thus, proteogenomics can identify immunogenic tumor specific peptides that can be used to create a personalized, targeted T cell therapy for children with high risk medulloblastoma.

The genetic events leading to medulloblastoma initiation, and therefore the ability to determine who is at risk and why, are still unknown. Microsatellites are short tandem repeated sequences that comprise ~3% of the genome. The number of repeats varies between individuals and are often non-randomly associated with disease, including several cancers such as breast, glioma, lung and ovarian. We have developed a novel approach for identifying and validating a microsatellite signature set for medulloblastoma tumors using germline DNA. Analyzing germline DNA sequence from a training set of 120 medulloblastoma subjects and 425 healthy controls, we have identified a set of 139 microsatellite loci whose genotype are significantly different between medulloblastoma subjects and controls. Using an iterative classification algorithm, we found a subset of 43 microsatellites able to distinguish medulloblastoma subjects from controls with a sensitivity and specificity of 0.88 and 0.92, respectively. This signature set was validated in an independent dataset with greater sensitivity and specificity. Analysis of the minor alleles of those 139 informative loci demonstrates that these loci are more mutable in medulloblastoma patients than controls. Finally, a downstream analysis of the genes harboring these microsatellite loci revealed significantly enriched pathways (Autophagy and mTOR signaling) and genes associated with cell cycle and DNA replication, recombination, and repair. Interestedly, we identified two exonic microsatellites in the transcription factors RAI1 and BCL6B associated with medulloblastoma. This study demonstrates that medulloblastoma-specific microsatellite variations in the germline DNA can distinguish individuals with medulloblastoma as well as provide insight into mechanisms of predisposition.

Despite sharing the same sequence specificity in vitro and in vivo, CCCTC-binding factor (CTCF) and its paralog brother of the regulator of imprinted sites (BORIS) are simultaneously expressed in germ cells. Recently, ChIP-seq analysis revealed two classes of CTCF/BORIS-bound regions: single CTCF target sites (1xCTSes) that are bound by CTCF alone (CTCF-only) or double CTCF target sites (2xCTSes) simultaneously bound by CTCF and BORIS (CTCFBORIS) or BORIS alone (BORIS-only) in germ cells and in BORIS-positive somatic cancer cells. BORIS-bound regions (CTCFBORIS and BORIS-only sites) are, on average, enriched for RNA polymerase II (RNAPII) binding and histone retention in mature spermatozoa relative to CTCF-only sites, but little else is known about them. We show that subsets of CTCFBORIS and BORIS-only sites are occupied by several testis-specific transcriptional regulators (TSTRs) and associated with highly expressed germ cell-specific genes and histone retention in mature spermatozoa. We also demonstrate a physical interaction between BORIS and one of the analyzed TSTRs, TATA-binding protein (TBP)-associated factor 7-like (TAF7L). Our data suggest that CTCF and BORIS cooperate with additional TSTRs to regulate gene expression in developing male gametes and histone retention in mature spermatozoa, potentially priming certain regions of the genome for rapid activation following fertilization.

The genomic characterization of medulloblastoma at the level of the chromosome, epigenome and transcriptome has radically redefined diagnostic classifications and reoriented disease definitions toward a biological underpinning. However, it still remains a challenge to leverage this wealth of data toward a true understanding of disease biology and subsequently to capitalize on the promise of biologically based therapies. One reason for this difficulty is the loose correlation between the genetic and the functional aspects of the cancer cell resulting from multilayered regulatory mechanisms governing protein production, the predominant functional moiety of the cell. To better actualize the promise of genomics, it is necessary to add an understanding of the proteomic dimension to cancer cell biology. We have undertaken a super-SILAC based quantitative proteomics survey of 42 tissue samples spanning the 4 genomic subgroups of medulloblastoma and control cerebellum. We have identified and quantitated thousands of proteins across these groups and find that we are able to recapitulate the genomic subgroups based upon subgroup restricted and differentially abundant proteins while also identifying one additional subgroup. Adding array gene expression data to this analysis, we calculate a relatively poor correlation between mRNA and protein abundance with the best concordance found among stable proteins. EPIC 850k methylation array data performed on the same samples extend this analysis to the epigenome yielding chromosomal copy number variations and correlations to gene promoter methylation. From the quantitative proteomic data, we are able to discern networks of functional pathways with differential activation between the medulloblastoma subgroups and normal control cerebellum yielding potential avenues for clinical intervention. This work demonstrates that quantitative proteomics can further elucidate the biology inherent in a genomically based classification regime.

Background CTCF and BORIS (CTCFL), two paralogous mammalian proteins sharing nearly identical DNA binding domains, are thought to function in a mutually exclusive manner in DNA binding and transcriptional regulation. Results Here we show that these two proteins co-occupy a specific subset of regulatory elements consisting of clustered CTCF binding motifs (termed 2xCTSes). BORIS occupancy at 2xCTSes is largely invariant in BORIS-positive cancer cells, with the genomic pattern recapitulating the germline-specific BORIS binding to chromatin. In contrast to the single-motif CTCF target sites (1xCTSes), the 2xCTS elements are preferentially found at active promoters and enhancers, both in cancer and germ cells. 2xCTSes are also enriched in genomic regions that escape histone to protamine replacement in human and mouse sperm. Depletion of the BORIS gene leads to altered transcription of a large number of genes and the differentiation of K562 cells, while the ectopic expression of this CTCF paralog leads to specific changes in transcription in MCF7 cells. Conclusions We discover two functionally and structurally different classes of CTCF binding regions, 2xCTSes and 1xCTSes, revealed by their predisposition to bind BORIS. We propose that 2xCTSes play key roles in the transcriptional program of cancer and germ cells. Electronic supplementary material The online version of this article (doi:10.1186/s13059-015-0736-8) contains supplementary material, which is available to authorized users.