Abstract 2798: Cancer stem cells profiling of the colorectal cancer patient-derived xenograft model and corresponding in vitro organoid model by multiplex IHC

Abstract

Abstract Background. Organoids are derived from stem cells from patient tissue, with self-capability as well as self-organization [1] and have been reported to mimic the structural and functional features of their in vivo counterpart[2]. Patient-derived xenografts (PDXs) are considered predictive in vivo cancer model for preclinical research and we have built and characterized a large library of PDX biobank. We recently have also established a large biobank of PDX-derived organoids (PDXOs) from our PDX biobank (>200) using the Hubrecht Organoid Technology (HUB), with evidence of “biological equivalence” to their parental PDXs, sharing similar profiles of genomics and pathology, as well as pharmacology. PDX and the derived PDXO thus constitute matched in vivo and in vitro model pairs, with great potential pharmaceutical applications. This report attempted to verify and characterize the CSC components of both models by multiplex IHC. Method. Five CRC PDX/PDXO pairs (CR1250, CR2258, CR3310,CR5048, CR11372) were subjected to histopathology (HE staining) and CSCs (Multiplex IHC staining) analysis. TSA-based multiplex IHC staining in Leica Bond Rx automatic platform was performed using a 4-plex CSC panel: DAPI (Sigma, #9542), CD133 (Cell Signaling, #86781), CD44 (abcam, ab51037), SOX2 (Cell Signaling, #14962) and basal cell marker pan-CK (Cell Signaling, #4545), followed by whole slide scan using Vectra® Polaris™ automated quantitative pathology imaging system (PerkinElmer) and quantified by HALO image analysis (Indica labs). Results. H&E staining of the five CRC-PDXOs showed typical adenocarcinoma histology structure, e.g. recognizable adenocarcinoma cells in tubular or vesicular glands, similar to those of the matched PDXs. The presence of the basal cells marker (pan-CK) was observed predominantly on the cell surface with broader pattern in corresponding matched models. Similarly, CD44, a known dominant marker for CRC CSC was observed broadly on the surface of tumor cells in both models. Moreover, pan-CK and CD44 showed universal co-localization, thus suggesting that pan-CK+CD44+ cells become the primary cellular composition of CRC PDXOs and PDXs. Interestingly, CD133, another marker of CSC, displayed distinct staining patterns between PDXOs and PDXs: predominantly outside of tumor glandular in PDXs (periglandular-like) vs. intraglandular in PDXOs (intraglandular-like). SOX2, a key transcription factor for CSCs and vital to CSC development, was also broadly present in all PDXOs and PDXs with similar pattern. Conclusion. Multiplex IHC could be a powerful tool to investigate CSC component in tumors and tumor models. Our preliminary data seem to show that most cells, in both PDX and PDXO models, resemble basal cell layer where pan-CK+/CD44+ CSCs also co-localized, consistent with hypothesis of both being CSC-driven disease models.