EP536 - Spatial Exploration of the Cancer-associated Fibroblast Landscape in Early-Onset Colorectal Cancer

Purpose/Background: The incidence of colorectal cancer in average-risk adults younger than 50 years, known as early-onset colorectal cancer (EO-CRC), has been rapidly increasing, making it a public health issue. The onco-immunogenic factors governing the tumor microenvironment (TME) in EO-CRC are unknown. There is evidence that tumors with a high proportion of stroma and the presence of cancer-associated fibroblasts (CAFs) are associated with worse prognosis and therapeutic efficacy. Understanding these TME dynamics is essential to improving prognostic tools and treatments for EO-CRC. The non-immune, fibroblast stroma, contributes to the pathogenesis of CRC. However, whether differential expression of cancer-associated fibroblasts exists between EO- and AO-CRC is not known. Recent findings suggest that EO-CRC is more fibrotic than AO-CRC. Therefore, we hypothesize that EO-CRC tumors will have an increased fibrotic or chronically inflamed CAF landscape than AO-CRC patients.

Methods/Interventions: Surgical specimens preserved in FFPE from patients with sporadic EO-CRC or AO-CRC were subjected to In Situ spatial transcriptomic testing using 10x Genomics' Xenium platform. Briefly, 5 µm sections were taken from the FFPE tissue blocks. The Xenium In Situ for FFPE - Tissue Preparation Guide (CG000578, Rev C) was followed for the Xenium workflow after tissue quality check was achieved. A pre-designed, off-the-shelf Human Colorectal Gene Set of 325 genes from 10x genomics and a personalized 50-gene set to subclassify CAFs into general fibroblasts, inflammatory CAF (iCAF), myofibroplastic CAF (myCAF), and antigen-presenting CAFs (apCAF) were used for in-situ transcriptome mapping. The CAF subclassification gene list was extracted from a Pan-Cancer CAF transcriptomic atlas. Student's t test was performed to compare differences between EO-CRC and AO-CRC. Statistical significance was defined as p < 0.05.

Results/Outcomes: Sixteen samples were selected to characterize the tumor microenvironment and CAF representation of a diverse cohort of patients, including 5 non-Hispanic White, 7 Hispanic, and 4 Black patients. Cellular characterization showed no differences in cell representation of the TME between EO-CRC and AO-CRC. CAFs subtype mapping identified a higher proportion of myCAFs in EO-CRC compared to AO-CRC (p= 0.049). No differences were found in other CAF subtypes. Cell-to-cell interaction analysis showed higher proportion of myCAFs to be in direct contact, defined as distance of < 50um, with cancer cells in EO-CRC than in AO CRC (p=0.028).

Conclusion/Discussion: A deeper understanding of the underlying relationships and interactions in the tumor microenvironment in EO-CRC is critical for improving prognostic and therapeutic tools. Despite our small sample size, we identified some differences in the CAF landscape between EO-CRC and AO-CRC patients. Our next steps include extracting morphological features from our spatial transcriptomic data set to train an AI-based digital pathology tool for cellular identification on H&E images and validating our spatial transcriptomic findings using immunostaining on a larger sample.

Cancer-associated Fibroblast Subtypes in EO-CRC and AO-CRC