B. Szeglin1, C. Wu9,11, I. Wasserman2, S. Uppada3, X. Chen6, K. Ganesh8, A. Elghouayel7,11, J. Shia5, A. Barlas10, P. B. Paty11, M. R. Weiser11, J. G. Guillem11, G. M. Nash11, K. Manova-Todorova10, P. Dhawan3, R. Beauchamp4, N. E. Kemeny8, J. Garcia-Aguilar11, C. L. Sawyers9, J. Smith9,11 1Albert Einstein College Of Medicine,Bronx, NY, USA 2Icahn School Of Medicine At Mount Sinai,New York, NY, USA 3University Of Nebraska Medical Center,Department Of Biochemistry And Molecular Biology,Omaha, NE, USA 4Vanderbilt University Medical Center,Section Of Surgical Sciences,Nashville, TN, USA 5Memorial Sloan-Kettering Cancer Center,Department Of Pathology,New York, NY, USA 6University Of Miami Miller School Of Medicine,Department Of Bioinformatics And Biostatistics,Miami, FL, USA 7College Of William And Mary,Williamsburg, VA, USA 8Memorial Sloan-Kettering Cancer Center,Department Of Medical Oncology,New York, NY, USA 9Memorial Sloan-Kettering Cancer Center,Human Oncology And Pathogenesis Program,New York, NY, USA 10Memorial Sloan-Kettering Cancer Center,Department Of Molecular Cytology,New York, NY, USA 11Memorial Sloan-Kettering Cancer Center,Colorectal Service, Department Of Surgery,New York, NY, USA
Introduction:
Loss of SMAD4, the central node of the TGF-β superfamily, occurs in 10-20% of colorectal cancer (CRC) cases. SMAD4 loss in the context of activated Wnt signaling may play a role in disease progression and resistance to standard 5-fluorouracil (5-FU) based chemotherapy, but the underlying mechanisms are poorly understood. Development of relevant CRC models to better study SMAD4 biology and associated chemoresistance is needed.
Methods:
Fresh CRC specimens were obtained at time of resection. Tumors were dissociated to individual cells and seeded within a Matrigel matrix in our 3D tumoroid cell culture model. SMAD4 mutant versus SMAD4-wildtype (wt) tumoroids were injected subcutaneously into immunocompromised mice. The mice were treated with systemic 5-FU and tumors weighed at necropsy. In addition, CRISPR/Cas9 was used to knockdown (kd) SMAD4 in patient-derived tumoroids ex vivo. SMAD4 expression was restored in SMAD4 mutant SW480 CRC cells. The SMAD4-kd tumoroids, SMAD4-wt tumoroids, and CRC cells with restored SMAD4 expression were treated with 5-FU or FOLFIRI (5-FU, leucovorin, and irinotecan) to determine dose-response differences. In a parallel, exploratory analysis, microarray expression data from 250 CRC patients was used to generate a SMAD4 signature (FDR < 10-7). The Illumina BaseSpace Correlation Engine was used to correlate this signature with compounds that could be used in synergy with 5-FU based chemotherapy in the context of SMAD4 loss.
Results:
Engrafted SMAD4-deficient tumors did not respond to 5-FU treatment, while SMAD4-retained tumors demonstrated decreased tumor weight compared to vehicle (p < 0.02). SMAD4-kd tumoroids treated with either 5-FU or FOLFIRI ex vivo were significantly more resistant to treatment than SMAD4-wt tumoroids (p < 0.01). Conversely, restoration of SMAD4 expression in CRC cells mutant for SMAD4 was associated with significant response to 5-FU based therapy (p < 0.01). Finally, the SMAD4 signature implicated 3-3-diindolylmethane, a putative Wnt pathway inhibitor, as a lead candidate for use in the context of SMAD4 deficiency and 5-FU based chemotherapy.
Conclusion:
We demonstrate that loss of SMAD4 is associated with chemoresistance to 5-FU and FOLFIRI treatment in in vivo and ex vivo CRC tumoroid models, thereby establishing relevant biological systems to study patient-specific resistance mechanisms. Furthermore, in silico analysis of a SMAD4 gene expression signature reveals 3-3-diindolylmethane as a possible therapeutic compound to target activated Wnt signaling in the context of SMAD4 loss in CRC patients undergoing 5-FU based chemotherapy.