Spatial transcriptomics of retinoblastoma: a visual window on intra-patient heterogeneity.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BACKGROUND: Retinoblastoma is the most common intraocular malignant tumor in childhood. Although current treatments offer a high survival rate, treatment toxicity, tumor relapse, and treatment resistance require a deeper understanding of the disease mechanisms to develop adapted therapies. Microscopically, this tumor is characterized by different states of differentiation and proliferation, ranging from poorly differentiated to well-differentiated retinoblastoma. Retinocytoma, on the other hand, is a benign non-proliferative tumor. Recent next-generation multi-omics analyses classified retinoblastoma tumors in subtypes 1 or 2, subtype 2 presenting a later age of onset, more genetic alterations, and higher metastatic potential. In parallel, several single-cell transcriptomics studies demonstrated intratumoral heterogeneity. However, mapping the different cell populations directly on the tumor and comparing histological features and molecular subtypes remains an unmet need. METHODS: Spatial transcriptomics was used to characterize a primary enucleated case with two histologically distinct areas. The whole transcriptomic profile of sixteen regions of interest, covering the two differentiation patterns of the tumor and the non-tumoral retina, was analyzed. RESULTS: The clustering of the regions of interest based on whole transcriptome data correlated with the histological description: cluster 1 (6 regions of interest) corresponded to highly differentiated areas and cluster 2 (6 regions of interest) to poorly differentiated areas. They showed enrichment for phototransduction and proliferation respectively, confirmed by immunohistochemistry for markers of these pathways. The publicly available molecular signatures of the two retinoblastoma subtypes categorized our regions of interest into two groups, which correlated perfectly with histological observation and transcriptomic profiles. Further analysis of the expression of specific senescence markers in the well-differentiated area did not support the diagnosis of retinocytoma as it did not confirm the expected up-regulation seen in this tumor type. CONCLUSIONS: This study demonstrates a strong correlation between histological observation and molecular profiling, representing the first mapping of retinoblastoma composed of adjacent molecular subtypes 1 and 2. It also highlights the diagnostic ambiguity between retinocytoma and well-differentiated non-proliferative subtype 1 retinoblastoma and emphasizes the need for specific biomarkers to differentiate these two tumor types.