Accepted Manuscript Pancreatic Cancer Subtypes: Beyond Lumping and Splitting Andrew J. Aguirre PII: S0016-5085(18)35213-2 DOI: https://doi.org/10.1053/j.gastro.2018.11.004 Reference: YGAST 62235 To appear in: Gastroenterology Please cite this article as: Aguirre AJ, Pancreatic Cancer Subtypes: Beyond Lumping and Splitting, Gastroenterology (2018), doi: https://doi.org/10.1053/j.gastro.2018.11.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
<DOC>Editorials <AT>Pancreatic Cancer Subtypes: Beyond Lumping and Splitting <AU>Andrew J. Aguirre <AFN>Department of Medical Oncology Dana Farber Cancer Institute and Harvard Medical School Boston, Massachusetts and Institute of MIT and Harvard Cambridge, Massachusetts Acknowledgments <ARTFN>Funding AJA is supported by funding from the Pancreatic Cancer Action Network, the Lustgarten Foundation, the Dana-Farber Cancer Institute Hale Center for Pancreatic Cancer Research, the Isenberg Fund for Innovation in pancreatic cancer, the Doris Duke Charitable Foundation, and the National Cancer Institute K08 CA218420-01, P50CA127003, and U01 CA224146. <ARTFN>Conflicts of interest The author discloses the following: AJA has consulted for Oncorus, Inc. <COR>Reprint requests Address requests for reprints to: Andrew J. Aguirre, MD, PhD, E- mail: Andrew_Aguirre@dfci.harvard.edu <TXBX>See Stratification of pancreatic ductal adenocarcinomas based on tumor and microenvironment features, by Puleo F, Nicolle R, Blum Y, et al, on page 000. <BEGIN ARTICLE>Pancreatic ductal adenocarcinoma (PDAC) is the most common form of pancreatic cancer and has often been viewed as a disease with a poor prognosis and few
clinically useful molecular subtypes. However, over the last decade multiple groups have characterized the complex molecular landscape of PDAC to reveal several distinct classes of disease. In this issue of Gastroenterology, Puleo et al 1 report the results of a large, multicenter study examining gene expression programs in resected PDAC specimens and define 5 transcriptomic subtypes. Their robust stratification system corroborates key features of previous studies, while establishing an integrated classification that defines subtypes by gene expression features from both tumor and stromal compartments. Recent studies have defined numerous molecular subtypes of PDAC. Genome sequencing has elucidated classes of tumors based on mutations in key signaling pathways, including KRAS, DNA damage repair, transforming growth factor-beta, chromatin remodeling, and others. 2-7 Several studies have examined gene expression using microarrays or RNA sequencing and have identified distinct RNA signatures of PDAC. 2,3,8-10 Collisson et al 9 originally described 3 PDAC gene expression subtypes, including classical, quasimesenchymal, and exocrine-like subtypes. Moffitt et al 8 later used a virtual microdissection approach on data from primary and metastatic PDAC tumors to digitally separate tumor, stromal, and normal cell gene expression. They identified 2 tumor-specific signatures, including a classical subtype and a basal-like gene expression program that resembled the basal subtype of breast and bladder cancers. Moreover, they also identified 2 stroma-specific gene expression signatures and demonstrated that combinations of the stroma and tumor-specific subtypes represent distinct biology with different prognostic implications. The International Cancer Genome Consortium subsequently defined 4 major gene expression subtypes of PDAC, including a pancreatic progenitor class that showed strong overlap with the Collisson and Moffitt classical subtypes, as well as a squamous classification that closely resembles the Moffitt basal-like subtype. 2
Additionally, the International Cancer Genome Consortium study also described an aberrantly differentiated endocrine exocrine subtype and a novel immunogenic subtype that was associated with evidence of a significant immune infiltration. The Cancer Genome Atlas project subsequently provided further data suggesting that the aberrantly differentiated endocrine exocrine/exocrine-like subtype and the immunogenic subtype likely represented gene expression from non-neoplastic cells. 3 Beyond these messenger RNA subtypes, noncoding RNAs as well as protein expression have also been used to molecularly stratify PDAC specimens. 3 More recently, Tuveson et al 11 performed drug sensitivity profiling of PDAC organoid models and elucidated novel functional subtypes that were used to define gene expression signatures that predict chemotherapy sensitivity. Puleo et al 1 now report an impressive analysis of gene expression data from 309 resected PDAC tumors that were consecutively collected across 4 academic centers. They rediscovered the classical/pancreatic progenitor and basal-like/squamous subtypes in higher cellularity tumors and used unsupervised consensus clustering of all tumors in the cohort to define 5 distinct subtypes comprised of both tumor-specific and microenvironment-derived signatures. Two of these classes showed low stromal signals, including pure classical and pure basal-like subtypes. The other 3 classes were defined by the impact of high stromal content, including the immune classical, desmoplastic, and stroma activated subtypes. Although the Puleo classification has similarities with the Moffitt classification, the authors argue that their 5-subtype stratification incorporates more information about immune cell and proinflammatory signals and propose that tumor and microenvironment expression features must be integrated to define optimal clinically prognostic subtypes of PDAC. The demonstration of robust gene expression signatures from a large cohort of formalin-fixed paraffin embedded specimens (standard in clinical practice) is a
novel and important contribution to the literature. Additionally, the authors provide an impressive online histogenomics application allowing the correlation of histology with transcriptomic signatures. An important limitation of the study is its focus only on primary resected tumors rather than metastatic disease, which may harbor distinct tumor and stromal features. With several well-done studies now proposing multiple gene expression classifications of PDAC, how does the field make sense of these subtypes to benefit patient care? These classifications are more similar than different, and several common themes have emerged. The classical/pancreatic progenitor and the basal-like/squamous neoplastic subtypes have been validated across multiple studies in primary 1-3,8-10 and metastatic samples. 8,12,13 Basallike/squamous tumors harbor a significantly worse prognosis than classical/progenitor tumors. 1-3,8-10,13 These basal-like/squamous tumors also display more frequent TP53 mutations, 2,3,8 a higher pathologic grade, 1,13 and a poorer response to modern chemotherapy regimens. 13 Puleo et al 1 now also present immunohistochemistry data showing that nuclear GLI1 expression may be an important biomarker for this basal-like subtype. The classical/pancreatic progenitor subtype has been firmly associated with higher levels of GATA6 expression, 3,8,9,13 and Puleo et al 1 also report a novel and intriguing association of KRAS G12R mutations with this subtype. Moreover, the presence of an activated or desmoplastic type stroma confers a worse prognosis within the classical/pancreatic progenitor tumors. 1,8 We must now translate these gene expression classifications into clinical practice through prospective biopsy-based clinical trials (Figure 1). The future of precision medicine in patients with PDAC will rest on comprehensive molecular and functional characterization of a patient s tumor to define integrated subtypes of disease and predictive models that guide improved patient
care. Given the aggressive nature of PDAC, molecular profiling must be done early in a patient s disease course to inform real-time treatment decisions. Furthermore, development of rapid clinicready assays for surrogate biomarkers of transcriptional subtypes (eg, GATA6 in classical type tumors or GLI1 in basal-like/squamous tumors) may enable timely identification of these subtypes to inform first-line therapy selection. Repeat characterization of on-treatment and postprogression biopsy specimens and/or monitoring of circulating blood biomarkers will inform our understanding of how these disease subtypes evolve in response to treatment. Emerging single cell RNA sequencing data from human PDAC specimens promise to add even more complexity to PDAC gene expression stratifications. As treatment options for PDAC patients expand to include more highly active chemotherapy regimens, targeted agents, and immunotherapy strategies that enable better differentiation of responders and nonresponders, increased granularity of molecular subtyping will undoubtedly hold greater meaning. References 1. Puleo F, Nicolle R, Blum Y, et al. Stratification of pancreatic ductal adenocarcinomas based on tumor and microenvironment features. Gastroenterology 2018;000:000-000. 2. Bailey P, Chang DK, Nones K, et al. Genomic analyses identify molecular subtypes of pancreatic cancer. Nature 2016;531:47-52. 3. Cancer Genome Atlas Research Network. Integrated genomic characterization of pancreatic ductal adenocarcinoma. Cancer Cell 2017;32:185-203 e13. 4. Notta F, Chan-Seng-Yue M, Lemire M, et al. A renewed model of pancreatic cancer evolution based on genomic rearrangement patterns. Nature 2016;538:378-382. 5. Witkiewicz AK, McMillan EA, Balaji U, et al. Whole-exome sequencing of pancreatic cancer defines genetic diversity and therapeutic targets. Nat Commun 2015;6:6744.
6. Waddell N, Pajic M, Patch AM, et al. Whole genomes redefine the mutational landscape of pancreatic cancer. Nature 2015;518:495-501. 7. Biankin AV, Waddell N, Kassahn KS, et al. Pancreatic cancer genomes reveal aberrations in axon guidance pathway genes. Nature 2012;491:399-405. 8. Moffitt RA, Marayati R, Flate EL, et al. Virtual microdissection identifies distinct tumorand stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat Genet 2015;47:1168-78. 9. Collisson EA, Sadanandam A, Olson P, et al. Subtypes of pancreatic ductal adenocarcinoma and their differing responses to therapy. Nat Med 2011;17:500-503. 10. Birnbaum DJ, Finetti P, Birnbaum D, et al. Validation and comparison of the molecular classifications of pancreatic carcinomas. Mol Cancer 2017;16:168. 11. Tiriac H, Belleau P, Engle DD, et al. Organoid profiling identifies common responders to chemotherapy in pancreatic cancer. Cancer Discov 2018;8:1112-1129. 12. Aguirre AJ, Nowak JA, Camarda ND, et al. Real-time genomic characterization of advanced pancreatic cancer to enable precision medicine. Cancer Discov 2018;8:1096-1111. 13. Aung KL, Fischer SE, Denroche RE, et al. Genomics-driven precision medicine for advanced pancreatic cancer: early results from the COMPASS trial. Clin Cancer Res 2018;15:1344-1354. Figure 1. Pancreatic cancer subtyping to inform precision medicine. The integration of clinical and pathologic features with genomic, transcriptomic, and functional data from biopsy specimens can be used to define clinically relevant subtypes and predictive models that guide improved patient care.