Abstract
Background
Pancreatic cancer is a devastating disease with a high relapse rate, even in case of resectable pancreatic cancer. Here, we aimed to identify the prognostic significance and therapeutic options of metabolic subtypes of resectable pancreatic cancer.
Method
Transcriptomic data were obtained from the TCGA-PAAD cohort in the The Cancer Genome Atlas (TCGA) data portal (n = 182). After integrative analysis of transcriptomic data in the discovery cohort, immunohistochemical (IHC) staining was performed in an independent cohort (n = 51) to validate the molecules of interest. Experimental testing for the molecules of interest was performed in vitro using pancreatic cancer cell line models AsPC1, BxPC3, MIA PaCa-2 and PANC-1.
Results
Two subtypes showing distinct gene expression patterns in the TCGA-PAAD dataset were identified. Of these, the active glucose metabolism subtype showed a significantly lower survival rate related to relapse after surgical resection. The genes SLC2A1 (GLUT1) and SLC16A3 (MCT4) were highly enriched in this subtype. The validation cohort showed a high MCT4 staining and a high relapse rate (p = 0.01). Several molecular pathways associated with aggressive tumor biology, including cell cycle regulation and Myc and mTOR downstream signaling, were highly enriched in the active glucose metabolism subtype, as well as with distinct responses to immunotherapy. MCT4 inhibition suppressed the in vitro malignant characteristics of pancreatic cancer cells and showed a synergistic effect with gemcitabine treatment.
Conclusions
From our data we conclude that MCT4 may serve as a potential therapeutic target in resectable pancreatic cancer. The precision medicine strategy for resectable pancreatic cancer should be validated in a clinical setting with a prospective study design.
Keywords: Pancreatic cancer, MCT4, Gemcitabine, Chemotherapy, Pancreatectomy
Introduction
Pancreatic cancer is a lethal disease with a dismal prognosis even in its early stages [1]. More than 60 % of patients with resectable pancreatic cancer experience recurrence after surgery [2]. There is an unmet need for strategies to predict and treat aggressive resectable pancreatic cancer. As yet, precision therapeutic strategies for pancreatic cancer are lacking. Clinical decision making for its treatment is largely determined by resectability of the tumor based on imaging [1, 2]. As yet, no targeted therapeutic agent or immunotherapy has been approved, except for conventional gemcitabine-based or combination-based chemotherapy using fluorouracil, leucovorin, irinotecan and/or oxaliplatin regimens [2]. The identification of potential therapeutic targets is considered mandatory to develop novel strategies for pancreatic cancer and to improve its clinical outcome.
Previously, we reported the prognostic significance of a metabolic phenotype of resectable pancreatic cancer that can be detected clinically using positron emission tomography (PET) scanning [3]. In addition, we identified distinct transcriptomic profiles between clinical pancreatic cancer samples with a high or low uptake of fluorodeoxyglucose (FDG) in PET scans [4]. In recent years, molecular targeted therapy has become a main option in cancer treatment [5]. Especially metabolic targets such as monocarboxylate transporter 4 (MCT4) have been reported to potentially improve therapeutic efficacies [6, 7]. MCT4 has been shown to be highly expressed in cancer cell membranes and to control glucose-lactate metabolic pathways by maintaining the intracellular pH level [8]. Several studies have shown its potential as a prognostic biomarker and a therapeutic target in various cancer types [7]. Therefore, we aimed to validate the clinical relevance of different pancreatic cancer metabolic subtypes in the The Cancer Genome Atlas (TCGA)-PAAD cohort and to identify therapeutic targets for metabolically active subtypes showing therapeutic resistance.
Materials and methods
TCGA-PAAD transcriptomic data preprocessing
To reveal the genomic features of the different metabolic phenotypes of pancreatic cancer, we obtained transcriptomic data from the The Cancer Genome Atlas (TCGA)-PAAD cohort (https://portal.gdc.cancer.gov/). Preprocessing of the transcriptomic data was conducted using log normalization and removal of duplicated genes using downward sorting of standard deviation with the remaining unique and first occurrences.
Unsupervised clustering of glucose metabolism gene panel data
A gene panel for pancreatic cancer subtypes reflecting metabolic phenotypes in PET scans was generated from the combination of several gene sets of metabolic pathways involving FDG uptake. All gene sets were collected from a molecular signature database (https://www.gsea-msigdb.org/gsea/msigdb) and the final 29 genes were selected as a gene panel for clustering of metabolic subtypes. Hierarchical clustering using uncentered correlation with centroid linkage was performed to obtain unbiased metabolic subtypes.
Validation cohort
A total of 51 patients who underwent left-sided pancreatic cancer resection from 2010 to 2012 consecutively were enrolled in a retrospective cohort for the validation of MCT4 expression with prognostic significance after approval by the Institutional Review Board (Severance Hospital, Seoul, Korea, IRB No. 4-2015-0573). Clinical parameters, especially information on overall survival (OS) and recurrence-free survival (RFS), as well as formalin fixed paraffin-embedded (FFPE) tissue blocks of resected pancreatic cancers were obtained and analyzed.
Immunohistochemical validation
To validate results from the TCGA-PAAD cohort using transcriptomic data, immunohistochemistry (IHC) was performed in the independent retrospective cohort. Tissue sections of 4 μm thickness were deparaffinized, rehydrated and washed two times in a buffer. To reduce nonspecific background staining due to endogenous peroxidase, the slides were incubated in hydrogen peroxide for 10 min and washed four times in a buffer. Next, a primary antibody directed against monocarboxylate transporter 4 (MCT4, 1:50, Santa Cruz, Santa Cruz, CA, USA) was applied. The slides were subsequently incubated with a primary antibody enhancer for 20 min at room temperature and washed four times in a buffer. Then, horseradish peroxidase polymer was applied to the slides, incubated for 30 min at room temperature and washed four times in a buffer. Finally, the slides were stained with hematoxylin chromogen, washed four times in deionized water, and counterstained.
Gene set enrichment analysis
Single-sample gene set enrichment analysis (ssGSEA) was performed using TCGA-PAAD transcriptome data from the web server of GenePattern (https://www.genepattern.org/). The hallmark pathway and REACTOME pathways were used to calculate pathway scores for each sample and for comparison with metabolic subtypes and MCT4 expression.
Clinical response to gemcitabine treatment and prediction of immunotherapy response
Clinical information on cancer drugs for each TCGA patient was retrieved [5] and analyzed with respect to the metabolic subtype using unsupervised clustering of transcriptomic data. Prediction of immunotherapy response was performed using the tumor immune dysfunction and exclusion (TIDE) algorithm based on the estimation of immune exclusion and dysfunction deconvoluted from the transcriptome data [9].
Pancreatic cancer and ductal epithelial cell lines
Human pancreatic cancer cell lines AsPC1, BxPC3, MIA PaCa-2 and PANC-1 and a human pancreatic ductal epithelium (HPDE) cell line were purchased from the ATCC and tagged with green fluorescent protein (GFP) using a lentiviral transduction system. All cell lines were incubated and maintained in a humidified CO2 chamber at 37 °C with 5 % CO2 and grown in RPMI-1640 medium (ThermoFisher, Waltham, MA, USA) supplemented with 10 % fetal bovine serum (FBS), 100 units/ml penicillin and 100 µg/ml streptomycin. GFP was cloned into a pLenti CMV/TO Puro empty vector purchased from Addgene (Cambridge, MA, USA). Lentiviruses were prepared by co-transfection with packaging vectors pMD2G, pMDLg/pRRE and pRSV-Rev, and pLenti CMV/TO Puro-GFP into HEK 293T cells. GFP-tagged pancreatic cancer cell lines were established by lentiviral infection using a transduction technique, after which selection was performed for 1-2 weeks using 1 µg/ml puromycin.
Western blot analysis
Pancreatic cancer cells were processed using a Pro-Prep protein extraction kit (iNtRON Biotechnology, Seongnam, Korea). Equal amounts of protein were loaded on sodium dodecyl sulfate-polyacrylamide gels, electrophoresed and transferred to nitrocellulose membranes (Invitrogen). The resulting blots were processed with 5 % nonfat dry milk at room temperature and next incubated with primary antibodies directed against MCT4 followed by incubation with peroxidase-labeled secondary antibodies. Blot images were generated using an ImageQuant LAS 4000 biomolecular imager (GE Healthcare Life Sciences). Quantification of the protein bands was conducted using Multi-Gauge version 3.0 software.
In vitro cell proliferation assay using RNA interference and gemcitabine treatment
RNA interference for MCT4 was conducted using si-RNA (Silencer® Select s17416, s17417, Negative Control No.1 siRNA, Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. Transfection of siRNA into pancreatic cancer cells was performed using Lipofectamine RNAiMAX (Life Technologies, USA). For the proliferation assay, fluorescence intensity was measured daily after the transfection procedure using a fluorescence microplate reader (Varioskan Flash Multimode Reader, Thermo Scientific, USA) with bottom optic readings. The excitation/emission wavelengths were 488/507 nm for GFP. Gemcitabine hydrochloride was purchased (G6423, Sigma Aldrich, USA) and dissolved in dimethylsulfoxide (DMSO) to prepare stock solutions. Next, pancreatic cancer cells were treated with gemcitabine hydrochloride for 24 h at concentrations ranging from 0.01 to 10 µmol/L (0.01, 0.1, 1 and 10 µmol/L).
Statistics
Statistical comparisons of the continuous variables of the subtypes were conducted using Student’s t-test and categorical variables using the chi-square test. Survival analysis was performed using the log-rank test with Kaplan-Meier estimates. A p-value < 0.05 was considered statistically significant.
Results
Identification of prognostic subtypes from unsupervised clustering of gene sets for glucose metabolism
Unsupervised clustering of transcriptomic data from the TCGA-PAAD cohort using the metabolic gene panel revealed two distinct metabolic subtypes (Fig. 1A). Dimension reduction using t-SNE plot showed a differently clustered distribution of metabolic subtypes consistent with hierarchical clustering results (Fig. 1B). The two metabolic subtypes showed significantly different clinical outcomes, especially recurrence after surgical resection (p = 0.033, Fig. 1C).
Fig. 1.
Metabolic subtypes deduced from the TCGA-PAAD cohort using a metabolic gene panel of transcriptome data.(A) Hierarchical clustering matrix from gene expression data of 29 metabolic genes. Gene expression was normalized to the Z score from the standardization method. (B) t-SNE plot from dimension reduction of metabolic gene expression profiles. Two-dimensional distribution was plotted using two t-SNE components. (C) Survival analysis using Kaplan-Meier plots for overall and recurrence-free survival. Red line indicates the patient group with cluster A and blue line for cluster B. P-values were estimated using the log-rank test for comparisons between the two metabolic clusters
MCT4 serves as a potential biomarker for clinically relevant subtypes
In order to examine the expression of each metabolic subtype, several molecules were identified for each metabolic cluster (Fig. 2A). We observed a significantly higher expression of MCT4 (SLC16A3), a cell membrane protein for lactate transport, in metabolic cluster A than in metabolic cluster B and normal pancreas (Fig. 2B). Protein expression of MCT4, as confirmed by IHC staining using an independent validation cohort, showed a broad and intense expression in metabolic cluster A (Fig. 2C and D). The prognostic relevance of MCT4 expression was found to be significant in resected pancreatic cancers of the validation cohort (OS, p = 0.03; RFS, p = 0.01; Fig. 2E).
Fig. 2.
MCT4 as a representative molecule for metabolic cluster A. (A) Volcano plot for statistical testing to identify metabolic cluster-specific molecules. Fold ratios and p-values were used to plot each molecule specific to the metabolic subtypes. (B) Box plot representing SLC16A3 (MCT4) expression in metabolic clusters A, B and normal pancreas. P-values were calculated using ANOVA. (C) Representative images for each staining intensity (grade, and 1, 2, 3, respectively) from IHC staining of MCT4. (D) Pie charts showing the proportion of MCT4 IHC staining intensity and area. (E) Survival analysis using Kaplan-Meier plotting for overall and recurrence-free survival. Red line indicates the patient group with high MCT4 staining intensity (> grade 3) and blue line that with low MCT4 staining intensity (grade 1 or 2). P-values were estimated using the log-rank test for comparisons between the two metabolic clusters
Gene set enrichment analysis for metabolic subtypes and association with MCT4 expression
Several cancer-specific pathways in terms of the cell cycle regulation, apoptosis and TGF-β signaling were found to be highly enriched in metabolic cluster A, representing an aggressive cancer biology (Fig. 3A). Furthermore, MCT4 was highly expressed in metabolic cluster A and was significantly correlated with the mitogen-activated protein kinase (MAPK) pathway, RAS signaling, and epithelial-mesenchymal transition (EMT) features associated with transforming growth factor β (TGF-β) (Fig. 3B).
Fig. 3.
Single sample gene set enrichment analysis for metabolic subtypes and MCT4 expression. (A) Bar plot for the top 10 pathways specific to each metabolic subtype. The Hallmark gene set was used to estimate pathway correlations. The color in each box represents p-value and length for log ratio (cluster A compared with cluster B). (B) Heatmap showing the pattern of molecular pathways according to MCT4 expression using the REACTOME gene set database. The gene set enrichment score was normalized by the Z score from the standardization method
Therapeutic resistance of the metabolic subtype showing high MCT4 expression
The patients with inadequate response to gemcitabine treatment were found to be significantly associated with metabolic cluster A, whereas about 80 % of the responders were assigned to cluster B (Fig. 4A). In silico prediction of immunotherapy showed a significant correlation between MCT4 expression and the TIDE score (Fig. 4B). In fact, it was noted that most patients with metabolic cluster A were predicted to be non-responders to immunotherapy for pancreatic cancer (Fig. 4C).
Fig. 4.
Therapeutic resistance specific to metabolic subtypes and MCT4 expression. (A) Sankey diagram showing associations between metabolic subtypes and gemcitabine response in the TCGA-PAAD cohort. Each percentage indicates the proportion of the patients belonging to each group. (B) Scatter plot for tumor immune dysfunction exclusion (TIDE) score and MCT4 expression. The correlation coefficient and p-value were estimated using the Pearson method. The red line represents a trend line between MCT4 expression and TIDE score. (C) Waterfall plot representing the TIDE score for the prediction of immunotherapy. A TIDE score < 0 was assigned to the responder
MCT4 inhibition as a potential strategy for the subtype showing high glucose metabolism
Relatively high expression of MCT4 protein was observed in pancreatic cancer cell lines MIA PaCa-2 and PANC-1 (Fig. 5A, B). After siRNA-mediated MCT4 silencing we observed a marked growth inhibition in MCT4 high-expressing cell lines compared to that in the controls (Fig. 5C). Gemcitabine treatment in combination with siRNA-mediated MCT4 silencing revealed a statistically significant reduction in viability of MCT4 high-expressing cells compared to gemcitabine mono-treatment at all concentrations ranging from 0.01 to 10 µmol/L. (Fig. 5D).
Fig. 5.
In vitro therapeutic potential of MCT4 using pancreatic cancer cell line models. (A) Western blot analysis of MCT4 and control glyceraldehyde 3-phosphate dehydrogenase (GAPDH) in pancreatic cancer cell lines (MIA PaCa-2, HPDE, BxPC3, AsPC1, PANC-1). (B) Bar plot showing relative intensities (MCT4/GAPDH). The standard error and mean were calculated from the results of triplicate experiments. (C) Line chart indicating an inhibitory effect of RNA interference of MCT4 using siRNA in PANC-1 and MIA PaCa-2 cells. The standard error and mean were calculated from the results of triplicate experiments (* p < 0.05, ** p < 0.01, *** p < 0.001). (D) Bar plot comparing the inhibitory effect of RNA interference of MCT4 using siRNA combined with gemcitabine in PANC-1 and MIA PaCa-2 cells. The standard error and mean were calculated from the results of triplicate experiments (* p < 0.05, ** p < 0.01)
Discussion
Even in early clinical stages in which margin-negative (R0) resection is discerned from curative resection, cancer relapse frequently occurs with local and systemic metastasis, especially in pancreatic cancer [10]. There is no consensus guideline for resectable pancreatic cancer to improve clinical outcomes by selecting therapeutic strategies for high-risk patients [8, 11]. Therefore, precision medicine, including the prediction of a potentially systemic molecular subtype with clinically available biomarkers, is mandatory in managing resectable pancreatic cancer [12, 13]. Despite the failure to find specific subtypes with oncological relevance to the treatment of pancreatic cancer [14], we tried to classify pancreatic cancers in order to individualize their treatment options [4, 15–19]. Previously, we identified precise strategies for resectable pancreatic cancer regarding clinically available biomarkers, such as metabolic phenotypes in PET scans [3]. We found that high glucose (FDG) uptake of tumor tissue in PET scans correlated with lymph node metastasis and early recurrence after curative resection, as well as with distinct molecular traits in resectable pancreatic cancers [4]. Therefore, we here set out to validate the prognostic significance of metabolic subtypes and to identify potential therapeutic targets in the aggressive subset of resectable pancreatic cancers showing active metabolism. We identified two transcriptomic subtypes using the gene set involving glucose metabolism, showing prognostic significance for cancer relapse in the TCGA-PAAD cohort. Two cell surface molecules, SLC2A1 (GLUT1) and SLC14A3 (MCT4), showing significantly elevated expression levels in the metabolic subtype with a poorer recurrence-free survival, were identified as potential biomarkers. Considering off-target effects, MCT4 was selected as potential biomarker for the diagnosis and treatment of the metabolically active subtype showing an aggressive behavior in resectable pancreatic cancer [6, 7]. In addition, we validated this result in an independent cohort using IHC. More than 60 % of the resected tumors showed a high staining intensity for MCT4 and a high relapse rate in this cohort, even after R0 resection.
Upregulation of MCT4 during tumorigenesis and its prognostic value have been noted in many human cancers [20] but, as yet, studies on pancreatic cancer have been scarce. Recently, Baek et al. [6] reported that MCT4 expression may serve as an important determinant of cancer prognosis and glycolytic metabolism. They successfully demonstrated that MCT4 depletion could induce a compensatory metabolism, leading to retardation of tumor growth and vulnerability to metabolic stress. Our current integrative molecular analysis revealed a distinct underlying tumor biology in cluster A in terms of hallmark pathways for cell cycle regulation, TGF-β and mTOR signaling and cancer metabolism. We found that MCT4 expression was positively correlated with aggressive cancer-specific pathways. In addition, we found that this aggressive phenotype in metabolic cluster A was related to significant resistance to chemotherapy using gemcitabine and to immunotherapy. As yet, R0 resection is known to be the most effective monotherapy for pancreatic cancer, but the R0 resection rate is less than 20 %, and more than half of the patients experience tumor recurrence even after curative resection [21]. Therefore, chemotherapy should be the main therapeutic option for treating pancreatic cancer. Considering that desmoplastic stroma and genetic alterations of pancreatic cancer cells can be associated with resistance to chemotherapy and immunotherapy [22], we believe that our present observations suggest a potential therapeutic application of MCT4 in overcoming treatment resistance in resected pancreatic cancer in the near future. It is well-known that immunotherapy can be effective in a very limited subset of pancreatic cancer patients. Our observations and in silico predictions suggest that metabolically active tumors with a high MCT4 expression are more resistant to immunotherapy as well as to gemcitabine-based chemotherapy. We collected in vitro evidence indicating that MCT4 may be a potential therapeutic target in an aggressive subset of resectable pancreatic cancers. Molecular inhibition of MCT4 using siRNA resulted in a significant suppression of pancreatic cancer cell growth and a synergistic effect of gemcitabine treatment compared to controls. Although MCT4-targeting anti-cancer drugs are currently not available, our results revealed a potential therapeutic strategy using a combination of metabolic targeting and current chemotherapy. Considering the clinically relevant side effects of gemcitabine [23, 24], these observations suggest a potential way of reducing the gemcitabine dose in order to achieve a patient-comfort chemotherapy as well as to improve the quality of life during anticancer treatment [25]. Preclinical studies of MCT4 inhibition are warranted to confirm its efficacy in vivo. Robust clinical validation using multiple clinical cohorts with a prospective study design will also be mandatory to apply this novel therapeutic target in a clinical setting in resectable pancreatic cancer. Finally, clinical biomarkers to visualize a tumor’s glucose metabolism after PET scanning may be used as molecular traits for resectable pancreatic cancers in independent cohorts.
In conclusion, we found that metabolic subtypes for glucose are significantly correlated with an aggressive pancreatic cancer tumor biology and a dismal prognosis. Integrative analysis revealed that MCT4 may serve as a potential therapeutic target for the aggressive subset of resectable pancreatic cancers. Robust clinical validation and preclinical testing are mandatory to confirm this novel therapeutic option.
Funding
This study was supported by a faculty research grant from the Yonsei University College of Medicine (6-2015-0053).
Data availability
Transcriptome data used in this study are available via the The Cancer Genome Atlas (TCGA)-PAAD cohort (https://portal.gdc.cancer.gov/).
Declarations
Conflict of interest
The authors declare that they have no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.F. Bray, J. Ferlay, I. Soerjomataram, R.L. Siegel, L.A. Torre, A. Jemal, Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 68, 394–424 (2018) [DOI] [PubMed] [Google Scholar]
- 2.A. Oba, C. Croce, P. Hosokawa, C. Meguid, R.J. Torphy, M.H. Al-Musawi, S. Ahrendt, A. Gleisner, R.D. Schulick, and M. Del Chiaro, Prognosis based definition of resectability in pancreatic cancer: a road map to new guidelines. Ann. Surg. (2020). 10.1097/SLA.0000000000003859 Online ahead of print. [DOI] [PubMed] [Google Scholar]
- 3.C.M. Kang, S.H. Lee, H.K. Hwang, M. Yun, W.J. Lee, Preoperative volume-based PET parameter, MTV2.5, as a potential surrogate marker for tumor biology and recurrence in resected pancreatic cancer. Medicine 95, 1–8 (2016) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.S.H. Lee, H.K. Hwang, W.J. Lee, M. Yun, C.M. Kang, Preoperative metabolic tumor volume2.5 associated with early systemic metastasis in resected pancreatic cancer: a transcriptome-wide analysis. Gut Liver 13, 356–365 (2019) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Z. Ding, S. Zu, J. Gu, Evaluating the molecule-based prediction of clinical drug responses in cancer. Bioinformatics 32, 2891–2895 (2016) [DOI] [PubMed] [Google Scholar]
- 6.G.H. Baek, Y.F. Tse, Z. Hu, D. Cox, N. Buboltz, P. McCue, C.J. Yeo, M.A. White, R.J. DeBerardinis, E.S. Knudsen, A.K. Witkiewicz, MCT4 defines a glycolytic subtype of pancreatic cancer with poor prognosis and unique metabolic dependencies. Cell Rep. 9, 2233–2249 (2014) [DOI] [PubMed] [Google Scholar]
- 7.A. Javaeed, S.K. Ghauri, MCT4 has a potential to be used as a prognostic biomarker - a systematic review and meta-analysis. Oncol. Rev. 13, 88–96 (2019) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.T. Seufferlein, J. Mayerle, Pancreatic cancer in 2015: Precision medicine in pancreatic cancer-fact or fiction? Nat. Rev. Gastroenterol. Hepatol. 13, 74–75 (2016) [DOI] [PubMed] [Google Scholar]
- 9.P. Jiang, S. Gu, D. Pan, J. Fu, A. Sahu, X. Hu, Z. Li, N. Traugh, X. Bu, B. Li, J. Liu, G.J. Freeman, M.A. Brown, K.W. Wucherpfennig, X.S. Liu, Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat. Med. 24, 1550–1558 (2018) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.J.P. Neoptolemos, J. Kleeff, P. Michl, E. Costello, W. Greenhalf, D.H. Palmer, Therapeutic developments in pancreatic cancer: current and future perspectives. Nat. Rev. Gastroenterol. Hepatol. 15, 333–348 (2018) [DOI] [PubMed] [Google Scholar]
- 11.S. Lawrence, V. Adsay, G. Petersen, D. Klimstra, N. Bardeesy, M.D.M. Leiserson, R. Bowlby, K. Kasaian, I. Birol, K.L. Mungall, S. Sadeghi, J.N. Weinstein, P.T. Spellman,Y. Liu, L.T. Amundadottir, J. Tepper, A.D. Singhi, R. Dhir, D. Paul, T. Smyrk, L. Zhang,P. Kim, J. Bowen, J. Frick, J.M. Gastier-Foster, M. Gerken, K. Lau, K.M. Leraas, T.M. Lichtenberg, N.C. Ramirez, J. Renkel, M. Sherman, L. Wise, P. Yena, E. Zmuda, J. Shih, A. Ally, M. Balasundaram, R. Carlsen, A. Chu, E. Chuah, A. Clarke, N. Dhalla, R.A.Holt, S.J.M. Jones, D. Lee, Y. Ma, M.A. Marra, M. Mayo, R.A. Moore, A.J. Mungall, J.E. Schein, P. Sipahimalani, A. Tam, N. Thiessen, K. Tse, T. Wong, D. Brooks, J.T. Auman, S. Balu, T. Bodenheimer, D.N. Hayes, A.P. Hoyle, S.R. Jefferys, C.D. Jones, S. Meng, P.A. Mieczkowski, L.E. Mose, C.M. Perou, A.H. Perou, J. Roach, Y. Shi, J.V. Simons, T. Skelly, M.G. Soloway, D. Tan, U. Veluvolu, J.S. Parker, M.D. Wilkerson, A. Korkut, Y. Senbabaoglu,P. Burch, R. McWilliams, K. Chaffee, A. Oberg, W. Zhang, M.C. Gingras, D.A. Wheeler, L. Xi, M. Albert, J. Bartlett, H. Sekhon, Y. Stephen, Z. Howard, M. Judy, A. Breggia, R.T. Shroff, S. Chudamani, J. Liu, L. Lolla, R. Naresh, T. Pihl, Q. Sun, Y. Wan, Y. Wu, S. Jennifer, K. Roggin, K.F. Becker, M. Behera, J. Bennett, L. Boice, E. Burks, C.G. Carlotti Junior, J. Chabot, D. Pretti da Cunha Tirapelli, J. Sebastião dos Santos, M. Dubina, J. Eschbacher, M. Huang, L. Huelsenbeck-Dill, R. Jenkins, A. Karpov, R. Kemp, V. Lyadov, S. Maithel, G. Manikhas, E. Montgomery, H. Noushmehr, A. Osunkoya,T. Owonikoko, O. Paklina, O. Potapova, S. Ramalingam, W.K. Rathmell, K. Rieger-Christ, C. Saller, G. Setdikova, A. Shabunin, G. Sica, T. Su, T. Sullivan, P. Swanson, K. Tarvin, M. Tavobilov, L.B. Thorne, S. Urbanski, O. Voronina, T. Wang, D. Crain, E. Curley, J. Gardner, D. Mallery, S. Morris, J. Paulauskis, R. Penny, C. Shelton, T. Shelton, K.P. Janssen, O. Bathe, N. Bahary, J. Slotta-Huspenina, A. Johns, H. Hibshoosh, R.F. Hwang, A. Sepulveda, A. Radenbaugh, S.B. Baylin, M. Berrios, M.S. Bootwalla, A. Holbrook, P.H. Lai, D.T. Maglinte, S. Mahurkar, T.J. Triche, D.J. Van Den Berg, D.J. Weisenberger, L. Chin, R. Kucherlapati, M. Kucherlapati, A. Pantazi, P. Park, G. Saksena, D. Voet, P. Lin, S. Frazer, T. Defreitas, S. Meier, L. Chin, S.Y. Kwon, Y.H. Kim, S.J. Park, S.S. Han, S.H. Kim, H. Kim, E. Furth, M. Tempero, C. Sander, A. Biankin, D. Chang, P. Bailey, A. Gill, J. Kench, S. Grimmond, A. Johns, A.P. Cancer Genome Initiative (APGI, R. Postier, R. Zuna, H. Sicotte, J.A. Demchok, M.L. Ferguson, C.M. Hutter, K.R. Mills Shaw, M. Sheth, H.J. Sofia, R. Tarnuzzer, Z. Wang, L. Yang, J. (julia) Zhang, I. Felau, and J.C. Zenklusen, Integrated genomic characterization of pancreatic ductal adenocarcinoma. Cancer Cell 32, 185-203 (2017) [DOI] [PMC free article] [PubMed]
- 12.S.B. Dreyer, M. Pinese, N.B. Jamieson, C.J. Scarlett, E.K. Colvin, M. Pajic, A.L. Johns, J.L. Humphris, J. Wu, M.J. Cowley, A. Chou, A.M. Nagrial, L. Chantrill, V.T. Chin, M.D. Jones, K. Moran-Jones, C.R. Carter, E.J. Dickson, J.S. Samra, N.D. Merrett, A.J. Gill, J.G. Kench, F. Duthie, D.K. Miller, S. Cooke, D. Aust, T. Knösel, P. Rümmele, R. Grützmann, C. Pilarsky, N.Q. Nguyen, E.A. Musgrove, P.J. Bailey, C.J. McKay, A.V. Biankin, D.K. Chang, Australian Pancreatic Cancer Genome Initiative, and Glasgow Precision Oncology Laboratory, precision oncology in surgery: patient selection for operable pancreatic cancer. Ann. Surg. 272, 366–376 (2020) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.S.B. Dreyer, M. Pinese, N.B. Jamieson, C.J. Scarlett, E.K. Colvin, M. Pajic, A.L. Johns, J.L. Humphris, J. Wu, M.J. Cowley, A. Chou, A.M. Nagrial, L. Chantrill, V.T. Chin, M.D. Jones, K. Moran-Jones, C.R. Carter, E.J. Dickson, J.S. Samra, N.D. Merrett, A.J. Gill, J.G. Kench, F. Duthie, D.K. Miller, S. Cooke, D. Aust, T. Knösel, P. Rümmele, R. Grützmann, C. Pilarsky, N.Q. Nguyen, E.A. Musgrove, P.J. Bailey, C.J. McKay, A.V. Biankin, D.K. Chang, Australian Pancreatic Cancer Genome Initiative, and Glasgow Precision Oncology Laboratory, precision oncology in surgery: patient selection for operable pancreatic cancer. Ann. Surg. 1, 366–376 (2018) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.M.A. Tempero, D. Klimstra, J. Berlin, T. Hollingsworth, P. Kim, N. Merchant, M. Moore, D. Pleskow, A. Wang-Gillam, A.M. Lowy, Changing the way we do business: recommendations to accelerate biomarker development in pancreatic cancer. Clin. Cancer Res. 19, 538–540 (2013) [DOI] [PubMed] [Google Scholar]
- 15.H.K. Hwang, S.H. Lee, H.I. Kim, S.H. Kim, J. Choi, C.M. Kang, W.J. Lee, Yonsei criteria, a potential linkage to intratumoral Foxp3 + / CD8 + ratio for the prediction of oncologic outcomes in resected left-sided pancreatic cancer. Yonsei Med. J. 61, 291–300 (2020) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.S.Y. Rho, M. Yun, C.M. Kang, S.H. Lee, H.K. Hwang, W.J. Lee, Different biological behaviors in left-sided pancreatic cancer according to Yonsei criteria: proposal of a modified Yonsei criteria score. Pancreatology 18, 990–995 (2018) [DOI] [PubMed] [Google Scholar]
- 17.J.U. Chong, S.H. Kim, H.K. Hwang, C.M. Kang, W.J. Lee, Yonsei criteria: a clinical reflection of stage I left-sided pancreatic cancer. Oncotarget 8, 110830–110836 (2017) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.S.H. Lee, H.K. Hwang, C.M. Kang, W.J. Lee, The Yonsei criteria as a clinically detectable parameter for excellent prognosis in resected left-sided pancreatic cancer: outcomes of a propensity score-matched analysis. Surg. Endosc. 31, 4656 (2017) [DOI] [PubMed]
- 19.J.U. Chong, H.K. Hwang, J.H. Lee, M. Yun, C.M. Kang, W.J. Lee, Clinically determined type of 18F-fluoro-2-deoxyglucose uptake as an alternative prognostic marker in resectable pancreatic cancer. PLoS One 12, 1–15 (2017) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.V.L. Payen, E. Mina, V.F. Van Hée, P.E. Porporato, P. Sonveaux, Monocarboxylate transporters in cancer. Mol. Metab. 33, 48–66 (2020) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.K.J. Labori, M.H. Katz, C.W. Tzeng, B.A. Bjørnbeth, M. Cvancarova, B. Edwin, E.H. Kure, T.J. Eide, S. Dueland, T. Buanes, I.P. Glsadhaug, Impact of early disease progression and surgical complications on adjuvant chemotherapy completion rates and survival in patients undergoing the surgery first approach for resectable pancreatic ductal adenocarcinoma - a population-based cohort study. Acta Oncol. 55, 265–277 (2016) [DOI] [PubMed] [Google Scholar]
- 22.S. Yu, C. Zhang, K.P.K. Xie, Therapeutic resistance of pancreatic cancer: roadmap to its reversal. Biochim. Biophys. Acta (BBA) - Rev. Cancer 1875, 188461 (2020) [DOI] [PubMed] [Google Scholar]
- 23.M. Sinn, M. Bahra, T. Liersch, K. Gellert, H. Messmann, W. Bechstein, D. Waldschmidt, L. Jacobasch, M. Wilhelm, B.M. Rau, R. Grützmann, A. Weinmann, G. Maschmeyer, U. Pelzer, J.M. Stieler, J.K. Striefler, M. Ghadimi, S. Bischoff, B. Dörken, H. Oettle, H. Riess, CONKO-005: adjuvant chemotherapy with gemcitabine plus erlotinib versus gemcitabine alone in patients after r0 resection of pancreatic cancer: a multicenter randomized phase III trial. J. Clin. Oncol. 35, 3330–3337 (2017) [DOI] [PubMed] [Google Scholar]
- 24.K. Uesaka, N. Boku, A. Fukutomi, Y. Okamura, M. Konishi, I. Matsumoto, Y. Kaneoka, Y. Shimizu, S. Nakamori, H. Sakamoto, S. Morinaga, O. Kainuma, K. Imai, N. Sata, S. Hishinuma, H. Ojima, R. Yamaguchi, S. Hirano, T. Sudo, Y. Ohashi, Adjuvant chemotherapy of S-1 versus gemcitabine for resected pancreatic cancer: a phase 3, open-label, randomised, non-inferiority trial (JASPAC 01). Lancet 388, 248–257 (2016) [DOI] [PubMed] [Google Scholar]
- 25.Y. Hagiwara, Y. Ohashi, K. Uesaka, N. Boku, A. Fukutomi, Y. Okamura, M. Konishi, I. Matsumoto, Y. Kaneoka, Y. Shimizu, S. Nakamori, H. Sakamoto, S. Morinaga, O. Kainuma, K. Imai, N. Sata, S. Hishinuma, H. Ojima, R. Yamaguchi, S. Hirano, T. Sudo, Health-related quality of life of adjuvant chemotherapy with S-1 versus gemcitabine for resected pancreatic cancer: Results from a randomised phase III trial (JASPAC 01). Eur. J. Cancer 93, 79–88 (2018) [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Transcriptome data used in this study are available via the The Cancer Genome Atlas (TCGA)-PAAD cohort (https://portal.gdc.cancer.gov/).