Figure S16: Expansion of FRM-specific CD8T cells after priming with relevant peptide in PBMC from healthy individuals.

Figure S9: Results of NOP identification using simulated sequencing data.

Figure S15: Tetramer gating strategy.

Figure S13: In silico determined immunogenic properties of NOPs.

Figure S5: Long-read transcript sequencing statistics.

Figure S2: Long RNA splice correction, isoform identification, and translation prediction.

Figure S1: RNA-guided tumor genome reconstruction.

Figure S3: Identifying FRM0469-dextramer-positive CD8+ T cells.

Figure S4: Overview of somatic mutation statistics for tumor samples analyzed by WGS in this study.

Figure S12: Example of a genomic rearrangement resulting in the expression of multiple hidden NOPs in tumor sample BRE007.

Figure S6: Comparison of long and short read RNA gene expression quantification and transcript coverage bias.

Figure S11: Example of a hidden NOP resulting from a complex chromosomal rearrangement in tumor sample LUN022.

Figure S7: Protein mass spectrometry results for A375 NOPs.

Combined spreadsheet with Supplementary Tables S1 - S12

Figure S14: A375 immunopeptidomics.

Figure S10: A) NOP expression compared to normal genes.

Abstract Identification of immunogenic cancer neoantigens as targets for therapy is challenging. Here, we integrate the whole-genome and long-read transcript sequencing of cancers to identify the collection of neo-open reading frame peptides (NOP) expressed in tumors. We termed this collection of NOPs the tumor framome. NOPs represent tumor-specific peptides that are different from wild-type proteins and may be strongly immunogenic. We describe a class of hidden NOPs that derive from structural genomic variants involving an upstream protein coding gene driving expression and translation of noncoding regions of the genome downstream of a rearrangement breakpoint, i.e., where no gene annotation or evidence for transcription exists. The entire collection of NOPs represents a vast number of possible neoantigens particularly in tumors with many structural genomic variants and a low number of missense mutations. We show that NOPs are immunogenic and epitopes derived from NOPs can bind to MHC class I molecules. Finally, we provide evidence for the presence of memory T cells specific for hidden NOPs in peripheral blood from a patient with lung cancer. This work highlights NOPs as a major source of possible neoantigens for personalized cancer immunotherapy and provides a rationale for analyzing the complete cancer genome and transcriptome as a basis for the detection of NOPs.

Background Neoantigens expressed and presented by tumor cells can be recognized by the immune system and facilitate destruction of tumors by cytotoxic CD8+ T cells. The main source of neoantigens has been single nucleotide variations with only one amino acid difference from a wildtype protein. We have instead focused on a class of neoantigens that arise from insertions and deletions (indels) or other gene re-arrangements, leading to formation of neo-open reading frame peptides (NOPs). NOPs are stretches of novel amino acid sequences containing numerous potentially highly immunogenic epitopes and therefore provide a valuable source of neoantigens with the capability to improve efficacy of cancer vaccines. Methods The analysis of The Cancer Genome Atlas (TCGA) database and in-house sequenced human tumor samples was used to predict expression of shared neoantigens resulting from indels and other gene re-arrangements. Several selected candidates underwent thorough in vitro experimental validation to confirm tumor cell presentation and immunogenicity of NOPs. Additionally, we tested immunogenicity of mouse surrogate NOPs identified via the same pipeline in the melanoma cell line B16F10 by encoding ten different NOP sequences in an mRNA vaccine. Results We identified several different NOP sequences with predicted expression in human tumor samples that were selected for further validation. From each tested NOP, one or multiple epitopes were binding in vitro to either HLA-A*02:01, HLA-A*01:01, or HLA-A*24:02 allele. Presentation of epitopes on the tumor cell surface was confirmed by eluting the epitopes from HLA alleles followed by mass spectrometry analysis. We also detected presence of antigen specific CD8+ T cells in blood of healthy donors by tetramer staining. In vitro expanded NOP specific CD8+ T cell clones were able to recognize target positive tumor cell lines as measured by increased activation and degranulation markers. In vivo immunogenicity assessment after vaccination of mice with a mRNA vaccine encoding B16F10-derived NOPs showed large number (up to 60%) of polyfunctional CD8+ T cells, and to lower extent also CD4+ T cells after in vitro re-stimulation of splenocytes with NOP peptide pools. Deconvolution confirmed that 3/10 of NOPs induced CD8+ T cell responses and 2/10 NOPs mounted CD4+ T cells responses. Conclusions The neo-open reading frame peptides represent a class of experimentally validated, highly immunogenic neoantigens, suitable to induce robust in vivo immune responses when encoded in an mRNA vaccine format.

Stage III–IV non-small cell lung cancer (NSCLC) is a devastating disease characterized by a poor prognosis. NSCLC tumors carry genetic mutations, which can lead to the expression of altered protein sequences. Peptides originating from mutated proteins and bound to MHC molecules on the tumor cell surface are referred to as neoantigens, as they are tumor-specific and not expressed in normal cells. Due to their tumor specificity, neoantigens have a strong potential to induce an anti-tumor immune response and have been investigated for development of personalized therapeutic cancer vaccines. The current study describes the development of a clinical grade neoantigen vaccine formulation (FRAME-001) intended as immunotherapy in advanced NSCLC in combination with the immune checkpoint inhibitor pembrolizumab. The detection of aberrant tumor-specific transcripts as well as an algorithm to select immunogenic neoantigen peptides are described. Subsequently, selected neoantigen peptides were synthesized with a high throughput synthesis platform and aseptically formulated under good manufacturing practice (GMP) conditions into four aqueous peptides mixtures that each contained six neoantigen peptides. A validated stability-indicating analytical method was developed in which we considered the personalized nature of the formulation. An extensive stability study performed either at −25 °C or −80 °C showed that the formulation was stable for up to 32 weeks. The formulation was mixed with the vaccine adjuvant Montanide ISA 51 VG, which yielded the final vaccine emulsion. The stability of the vaccine emulsion was demonstrated using microscopic examination, differential light scattering, and the water-drop test. The presented data show that FRAME-001 is a feasible personalized vaccine formulation for the treatment of stage III–IV NSCLC. The presented data may give guidance in the development of novel personalized therapeutic vaccines since this formulation strategy could be used for any cancer indication.