TL;DR
Researchers have developed a new computational pipeline called SSNIP to identify "public" neoantigens—immune targets shared across different patients and cancer types—derived from errors in RNA splicing rather than DNA mutations. By targeting these widespread "neojunctions," they demonstrated a potential path toward "off-the-shelf" immunotherapies that can overcome the challenges of high tumor heterogeneity and low mutational burden.
What Problem Does This Paper Address?
Traditional cancer immunotherapies, like checkpoint inhibitors or personalized vaccines, often rely on neoantigens: "red flag" proteins on the surface of cancer cells created by DNA mutations. However, this approach faces two major hurdles:
- Low Mutational Burden: Many cancers don't have enough DNA mutations to generate recognizable neoantigens.
- Intratumoral Heterogeneity: Even when mutations exist, they may only be present in a small portion of the tumor, allowing other parts of the cancer to "escape" treatment.
The authors of this study looked beyond DNA mutations to RNA splicing. Splicing is the process where "non-coding" segments of RNA are removed to create a final blueprint for proteins. In cancer, this process often breaks, creating "neojunctions"—novel protein sequences that the immune system has never seen before. Because these splicing errors can be shared across many patients (making them "public"), they represent a goldmine for universal immune targets.
Methods
The team developed the SSNIP (Splicing-derived Shared Neoantigen Identification Pipeline) to find these targets:
- Bioinformatic Screening: They compared RNA sequencing data from TCGA (cancer) and GTEx (normal tissue) across 10 cancer types. They looked for neojunctions present in >10% of cancer samples but <1% of normal tissues.
- Conservation Analysis: To ensure the targets weren't just in small pockets of a tumor, they used an in-house glioma dataset with 535 biopsies from 56 patients to find "intratumorally conserved" junctions.
- Validation: They used algorithms to predict which of these RNA errors would actually be processed into peptides and "presented" on the cell surface via HLA molecules (specifically HLA-A*02:01).
- T-Cell Engineering: They identified T-cell receptors (TCRs) from healthy donors that could recognize these antigens and "transduced" (inserted) them into reporter cells and donor T-cells to test their ability to kill cancer cells.
Datasets
- TCGA (The Cancer Genome Atlas): Used to identify recurring splicing events across 10 different cancer types.
- GTEx (Genotype-Tissue Expression): Used as a "normal" control to ensure targets were truly cancer-specific.
- In-house Glioma Dataset: A high-resolution dataset featuring 56 patients with roughly 10 spatially distinct biopsies per patient (n=535 total) to study how conserved these antigens are within a single tumor.
Key Insights and Ideas
- The "Public" Advantage: The pipeline identified 789 public neojunctions. Unlike personalized vaccines that must be built from scratch for every patient, these shared targets could allow for "off-the-shelf" T-cell therapies.
- The GNAS Discovery: One specific neojunction in the GNAS gene was found to be highly conserved, appearing in over 90% of biopsies in about 27% of the glioma patients studied. This suggests that targeting GNAS could eliminate the vast majority of a patient's tumor cells, leaving few places for the cancer to hide.
- The Mechanistic Link: The researchers found a biological "why" for these errors. In certain brain cancers (IDH1-mutant oligodendrogliomas), the loss of specific chromosomal regions (1p/19q) leads to a shortage of splicing factors like SF3A3 and SNRPD2. This shortage directly causes the "messy" splicing that creates the neoantigens.
Key Results
- High-Confidence Targets: Out of 789 identified neojunctions, 32 were confirmed at both the RNA (transcriptomic) and protein (proteomic) levels.
- Functional Killing: T-cells engineered with TCRs targeting RPL22 and GNAS neojunctions successfully recognized and killed multiple glioblastoma and melanoma cell lines that naturally expressed these targets.
- Biological Validation: Using siRNA to knock down splicing factors in "normal" cancer cells successfully induced the expression of these neoantigens, proving the causal link between splicing machinery failure and antigen production.
Limitations and Future Work
While the abstract highlights a major breakthrough, a few areas remain for future exploration:
- HLA Restriction: The current high-confidence targets are focused on the HLA-A*02:01 allele. While common, this means the specific therapy would only work for patients with that specific genetic background.
- Clinical Translation: The study primarily uses in vitro (lab-based) models and cell lines. Moving these "off-the-shelf" TCRs into human clinical trials will be the next major hurdle to prove safety and efficacy in living patients.
- Breadth of Cancers: While 10 cancer types were screened, the deep-dive validation focused heavily on glioma and melanoma; further work is needed to see if the GNAS or RPL22 targets are equally effective in other solid tumors.
Summary updated