Summary: Researchers have identified thousands of previously unannotated, tumor-specific genetic isoforms in glioblastoma that were undetectable with conventional short-read sequencing. A subset of these novel isoforms is predicted to bind strongly to major histocompatibility complex (MHC) class I molecules, representing a newly described class of potential neoantigens. This atlas and the validated analytical framework expand the repertoire of actionable targets for next-generation personalized cancer vaccines and immunotherapies in adult neuro-oncology.
Key Facts
- Unprecedented Cellular Atlas: Built from single-cell long-read RNA sequencing of more than 210,000 individual cells from 27 glioblastoma patients.
- Overcoming Short-Read Limits: Short-read sequencing fragments transcripts and cannot reconstruct full-length isoforms at single-cell resolution; long-read methods solve this limitation.
- Tumor-Specific Isoforms: Thousands of novel isoforms were discovered that appear exclusively in glioblastoma cells and are absent from healthy tissues.
- Immunological Presentation: A subset of tumor-specific isoforms is predicted to generate peptides that bind MHC Class I molecules with high affinity, potentially acting as neoantigens recognizable by T cells.
- Translational Pipeline: The study supplies an analytical framework for clinical long-read single-cell transcriptomics to accelerate precision oncology and vaccine design.
Source: University of Hong Kong
Researchers from the Clinical Neuroscience Consortium (CNC), a collaboration between the LKS Faculty of Medicine at the University of Hong Kong and Queen Mary Hospital, together with the Hong Kong Genome Institute (HKGI), have produced the most comprehensive map to date of isoform diversity in glioblastoma, the most common and aggressive adult brain cancer.
Using advanced long-read single-cell RNA sequencing, the team detected thousands of previously unknown tumor-specific isoforms that conventional short-read technologies missed. These discoveries reveal a rich source of potential therapeutic targets, including candidates suitable for personalized cancer vaccines and other immunotherapies.
The findings were published in Nature Communications.
Revealing a previously unseen layer of tumor biology
Glioblastoma is especially difficult to treat because tumor cells within the same patient can differ dramatically in behavior, growth, and treatment response. Capturing that cellular diversity is essential for developing more effective therapies.
Each human gene can produce multiple transcript variants, or isoforms, which can alter protein sequence, localization, and function. In cancer, alternative isoforms may change how cells communicate with the immune system or enable immune evasion. Short-read single-cell RNA sequencing fragments transcripts into small pieces, making it impossible to determine how exons are combined in full-length isoforms in individual cells. Long-read single-cell sequencing overcomes this gap by reading entire transcripts at single-cell resolution.
CNC Chapter Lead (Brain Tumour) Dr Aya El Helali, Clinical Assistant Professor in Clinical Oncology at HKUMed, explained that data from over 210,000 cells across 27 patients allowed the team to build an isoform-level atlas that captures both malignant cells and the immune and stromal cells of the tumor microenvironment. The study uncovered thousands of previously unannotated isoforms, including many found only in tumor cells and not in healthy tissues.
Crucially, some of these tumor-specific isoforms are predicted to produce peptides that bind strongly to MHC class I molecules, which present peptide fragments on the cell surface for recognition by cytotoxic T cells. Such peptides qualify as candidate neoantigens and could be targeted by personalized immunotherapies.
The team also developed and validated a computational pipeline for isoform discovery using clinical long-read single-cell data, creating a practical resource for researchers working on precision oncology worldwide.
Opening the door to personalized cancer vaccines
Glioblastoma’s pronounced cellular diversity makes designing effective, broadly applicable therapies difficult. Although this discovery does not change standard clinical care immediately, it has significant implications for future precision oncology and individualized immunotherapy approaches.
Personalized anti-cancer vaccines aim to instruct a patient’s immune system to recognize tumor-specific neoantigens and destroy cancer cells while sparing healthy tissue. These vaccines require high-confidence targets unique to the patient’s tumor. The newly identified tumor-restricted isoforms expand the candidate pool for such neoantigens.
Dr Brian Chung Hon-yin, Interim CEO of the HKGI and Clinical Associate Professor at HKUMed, emphasized that many therapeutically promising targets may have been invisible to earlier technologies. Expanding the catalog of candidate neoantigens could accelerate development of next-generation, individualized immunotherapies for brain cancer.
Collaborative strengths drive future breakthroughs
The study underscores the value of close collaboration among clinicians, surgeons, oncologists, and genomics scientists within Hong Kong’s biomedical ecosystem. Professor Gilberto Leung Ka-kit, Convenor of the CNC, highlighted how this interdisciplinary effort demonstrates the region’s capacity to produce impactful translational genomics research that supports precision medicine and biotechnology development.
About the research team
Key contributors include Dr Aya El Helali (CNC Chapter Lead, Brain Tumour), Dr Brian Chung Hon-yin (Interim CEO, HKGI), Professor Gilberto Leung Ka-kit (CNC Convenor), and Dr Karrie Kiang (Research Assistant Professor), all affiliated with the School of Clinical Medicine, University of Hong Kong.
Key Questions Answered:
A: Glioblastoma shows extreme cellular and transcriptomic heterogeneity. Short-read sequencing fragments RNA and obscures how exons are combined into full-length isoforms, preventing detection of numerous altered proteins unique to cancer cells.
A: Isoforms that are unique to tumor cells can create altered protein sequences absent from healthy tissue. When processed and presented by MHC Class I molecules, these aberrant peptides can act as neoantigens that personalized vaccines can target selectively.
A: No. It does not alter current standard-of-care treatments such as surgery, radiation, or chemotherapy. However, it creates a critical target library and computational framework to support translational development of personalized immunotherapies and mRNA cancer vaccines.
Editorial Notes:
- This article was edited by an editor at Neuroscience News.
- The referenced journal paper was reviewed in full by the editorial team.
- Additional context and clarifications were added by staff to aid readability and understanding.
About this brain cancer research news
Author: Jaymee Ng
Source: University of Hong Kong
Contact: Jaymee Ng – University of Hong Kong
Image: Image credited to Neuroscience News
Original Research: Open access. “Mapping glioblastoma’s isoform diversity using long-read single-cell analysis” by Wenshu Tang et al., published in Nature Communications. DOI: 10.1038/s41467-026-72258-2
Abstract
Mapping glioblastoma’s isoform diversity using long-read single-cell analysis
Glioblastoma is an aggressive brain tumor with poor prognosis, driven in part by extensive intratumoral heterogeneity and widespread dysregulation of RNA splicing. Alternative splicing shapes cellular identity and function, contributing to tumor progression and treatment resistance.
While single-cell RNA sequencing has revealed diverse cellular states in glioblastoma, conventional short-read approaches cannot resolve full-length isoforms. In this study, single-cell long-read RNA sequencing was applied to build an isoform-level atlas of glioblastoma, capturing full-length transcripts at single-cell resolution and identifying hundreds of isoforms with differential usage across distinct tumor cell populations.
The authors developed a prioritization framework to select tumor-restricted isoforms and identified surface-intracellular target pairs across patients, highlighting opportunities for dual-specific ligand-based therapies. They discovered 6,524 isoforms absent from existing annotations, including 179 that are tumor-specific; peptides derived from these isoforms show strong predicted binding to MHC Class I molecules, suggesting potential neoantigen candidates for immunotherapy.