Hanyuda A et al. IRS1, Exercise and Colorectal Cancer Survival, Supplementary Materials /19
Supplementary Materials
Submission to Annals of Surgical Oncology (ASO)
Article Title
Survival Benefit of Exercise Differs by Tumor IRS1 Expression Status in Colorectal Cancer
Corresponding Author:
Shuji Ogino, MD, PhD, MS (Epidemiology)
Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School
450 Brookline Ave., Room M422, Boston, MA 02215, USA
Telephone: +1-617-632-1972; Fax: +1-617-582-8558
E-mail:
Collection of Clinical Information and Tumor Tissue in Study Cohorts
Participants have been sent follow-up questionnaires every two years to update information on potential risk factors, newly diagnosed cancers or other diseases in themselves and their first-degree relatives. A subset of the cohort participants developed colorectal cancer (CRC). Our study physicians reviewed medical records and obtained information on tumor location and stage. Paraffin-embedded tissue blocks were collected from hospitals where patients underwent resection. To avoid treatment-induced artifacts or bias, we also collected diagnostic rectal biopsy for patients who had received preoperative treatment. We included both colon and rectal cancers, considering a continuum of pathological and molecular features from rectum to proximal colon.1 Histopathological characteristics of all CRCs were confirmed by a pathologist (S.O.) blinded to other data. Tumor grade was designated as high grade (≤50% glandular area) or low grade (>50% glandular area).
Assessment of Molecular Features
Sequencing of BRAF, KRAS and PIK3CA
All tumor molecular analyses were carried out completely blinded to patient identity, clinical features and outcome data. DNA was extracted from archival paraffin-embedded colon cancer tissue.2,3 We marked tumor areas on hematoxylin-eosin stained slides, and dissected tumor tissue by a sterile needle. Genomic DNA was extracted using QIAmp DNA mini kit (Qiagen, Valencia, CA, USA). With DNA extracted from archival paraffin-embedded colon cancer tissue, polymerase chain reaction (PCR) and pyrosequencing covering BRAF codon 600, KRAS codons 12, 13, 61, 146, and PIK3CA exons 9 and 20 were conducted as previously described.3-7 All forward sequencing results were confirmed by reverse sequencing. KRAS sequencing was further validated by Pyrosequencing.3
Microsatellite Instability Analysis
MSI status was determined by analyzing variability in the length of 10 microsatellite markers (D2S123, D5S346, D17S250, BAT25, BAT26, BAT40, D18S55, D18S56, D18S67, and D18S487) from tumor DNA compared to normal DNA.8 Either forward or reverse primer for each marker8 was labeled with fluorescence, and PCR products were electrophoresed and analyzed by ABI 3730 (Applied Biosystems, Foster City, CA, USA). PCR and DNA fragment analysis for all of the markers except for D2S123, D5S346, and D17S250 was performed in duplicate.
MSI-high was defined as the presence of instability in ≥30% of the markers and MSS as no instability of all markers. MSI-low (<30% unstable markers) tumors were grouped with microsatellite stable (MSS) tumors (no unstable markers) because we have previously demonstrated that these two groups show similar features.8
Methylation Analyses for CpG Islands and LINE-1
We quantified DNA methylation in eight CpG island methylator phenotype (CIMP)-specific promoters (CACNA1G, CDKN2A, CRABP1, IGF2, MLH1, NEUROG1, RUNX3, and SOCS1).9,10 Bisulfite treatment on genomic DNA and subsequent real-time PCR (MethyLight) were validated and performed as previously described.11 CIMP-high was defined as the presence of 6 or more methylated promoters, CIMP-low as 1 to 5 methylated promoters, and CIMP-negative as the absence (0/8) of methylated promoters, according to the previously established criteria.12 Methylation levels of long interspersed nucleotide element-1 (LINE-1) were quantified by validated bisulfite DNA treatment, polymerase chain reaction, and pyrosequencing assay as previously described.10,11
Immunohistochemical Staining of IRS1 and IRS2
For IRS1 and IRS2 immunostaining, TMA sections were deparaffinized overnight with xylene, followed by immersion in 100% ethanol for 15 min. For antigen retrieval, slides were then immersed in citrate buffer at pH 6 (BioGenex, Fremont, CA, USA) and heated in a pressure cooker for 30 min at 95°C. After cooling at ambient temperature for 45 min and rinsing with Tris-buffered saline (TBS, pH 7), specimens were treated with dual endogenous enzyme block (Dako, Carpinteria, CA, USA) for 30 min, followed by 10% fetal bovine serum (Life Technologies, Carlsbad, CA, USA) in TBS for 30 min. Samples were incubated with IRS1 antibody (rabbit 06-248, Millipore, Billerica, MA, USA; 1:200 dilution) or IRS2 antibody (rabbit 06-506, Millipore; 1:500) at 4oC for 16 h, then washed thoroughly in TBS, incubated with anti-rabbit IgG (Vector Laboratories, Burlingame, CA, USA) for 30 min, rinsed again, and treated with streptavidin-peroxidase (ABC kit, Vector Laboratories) according to the manufacturer’s instructions. Slides were then treated with liquid DAB Plus substrate chromogen (Dako) for 2-3 min for color development, washed with distilled water, and counterstained with hematoxylin for 3 min. Finally, specimens were dehydrated with graded alcohol and xylene, and cover glasses were placed.
Statistical Analysis
All statistical analyses were conducted using SAS software (version 9.3, SAS Institute, Cary, NC, USA). All P-values were two-sided. Our primary hypothesis testing was assessment of the interaction between physical activity and tumor IRS1 expression level in CRC-specific mortality analysis in the combined cohort. All other analyses and hypothesis testing in this study were secondary analyses, and therefore, results were interpreted cautiously. In particular, we were aware of multiple testing inherent in the subgroup analyses of prognostic associations of physical activity in strata of IRS1 (or IRS2) status. To test differences in the frequency distribution of categorical data, the chi-square test was performed. One-way analysis of variance (ANOVA) was used to compare mean age and mean LINE-1 methylation level. For the primary endpoint CRC-specific mortality, participants were censored at the time of death if death was not due to CRC. Multivariable Cox proportional hazards regression models were used to control for potential confounders, and were stratified by stage and sex to limit the number of variables in multivariable models. A statistical interaction was assessed by a likelihood ratio test, using the cross-product of post-diagnosis physical activity (ordinal variable of four categories: Q1, Q2, Q3 and Q4) and IRS1 status (three ordinal categories of negative, low, and high level) as the interaction term. Pinteraction value was calculated by comparing the model with the interaction term to the model without the interaction term.
We used an initial model including the interaction term, post-diagnosis physical activity, tumor IRS1 status, and possible covariates: sex, age at diagnosis (continuous, increase by 10 years), year of diagnosis (continuous), post-diagnosis body mass index (<30 vs. ≥30kg/m2), family history of CRC in first degree relatives (absent vs. present), tumor location (caecum, ascending to transverse, splenic flexure to sigmoid, rectum), tumor grade (low vs. high), MSI status (low/MSS vs. high), CIMP (negative vs. low vs. high), LINE-1 methylation (continuous), BRAF, KRAS, PIK3CA mutations, TP53 expression and CTNNB1 nuclear expression. A backward elimination was performed with P=0.05 as a threshold to limit the number of variables in the final model and avoid overfitting. The covariates selected by this procedure were "year of CRC diagnosis (continuous)" for CRC-specific survival analysis, and "age at diagnosis (continuous)" and "tumor location (caecum vs. ascending to transverse vs. splenic flexure to sigmoid vs. rectum)" for overall survival analysis. These covariates were consistently included in the all other multivariable analysis models.
In addition to regression models for the interaction term, in our secondary analysis, we calculated survival hazard ratio (HR) for high post-diagnosis physical activity (vs. low activity) in each stratum of IRS1 expression level (negative, low, or high). We divided post-diagnosis physical activity into two categories (Q1-Q2 as low level and Q3-Q4 as high level).
For IRS2 in our secondary analysis, we started from initial model of interaction analysis with cross-product of post-diagnosis physical activity (4 ordinal categories from Q1 to Q4) and IRS2 status (3 ordinal categories from negative to high expression). Among the covariates described above [sex, age at diagnosis (continuous, increase by 10 years), year of diagnosis (continuous), post-diagnosis body mass index (<30 vs. ≥30kg/m2), family history of CRC in first degree relatives (absent vs. present), tumor location (caecum, ascending to transverse, splenic flexure to sigmoid, rectum), tumor grade (low vs. high), MSI status (low/MSS vs. high), CIMP (negative vs. low vs. high), LINE-1 methylation (continuous), BRAF, KRAS, PIK3CA mutations, TP53 expression and CTNNB1 nuclear expression)], we selected the following covariates by backward elimination with P=0.05 as a threshold; "year of diagnosis of CRC" and "MSI" as final covariates for CRC-specific survival analyses, and "year of CRC diagnosis", "age at diagnosis" and "tumor location" as final covariates for overall survival analyses.
Combined categories of disease stage (I, II, III, missing) and sex were used as a stratifying variable using the “strata” option in the SAS “proc phreg” command to minimize residual confounding and overfitting.
To handle missing data in multivariable survival analyses, we incorporated cases with missing information into the majority category of the given covariate. Proportion of missing values for each variable was as follows: tumor location, 0.5%; tumor grade, 0.3%; MSI status, 0.8%; CIMP status, 1.1%; LINE-1 methylation, 2.5%; BRAF mutation, 1.3%; KRAS mutation, 0.3%; PIK3CA mutation, 8.6%; TP53 expression, 1.1%; nuclear CTNNB1 expression, 4.6%. Excluding cases with missing information in any of the covariates did not substantially alter results.
The proportionality of hazards assumption was evaluated using a time-dependent variable, which was cross-product of IRS1 variable and survival time (all P-values >0.20).
References for Supplementary Materials
1. Yamauchi M, Morikawa T, Kuchiba a., et al. Assessment of colorectal cancer molecular features along bowel subsites challenges the conception of distinct dichotomy of proximal versus distal colorectum. Gut. 2012;61(6):847-854. doi:10.1136/gutjnl-2011-300865.
2. Ogino S, Brahmandam M, Kawasaki T, Kirkner GJ, Loda M, Fuchs CS. Combined analysis of COX-2 and p53 expressions reveals synergistic inverse correlations with microsatellite instability and CpG island methylator phenotype in colorectal cancer. Neoplasia. 2006;8(6):458-464. doi:10.1593/neo.06247.
3. Ogino S, Kawasaki T, Brahmandam M, et al. Sensitive sequencing method for KRAS mutation detection by Pyrosequencing. J Mol Diagn. 2005;7(3):413-421. doi:10.1016/S1525-1578(10)60571-5.
4. Imamura Y, Lochhead P, Yamauchi M, et al. Analyses of clinicopathological, molecular, and prognostic associations of KRAS codon 61 and codon 146 mutations in colorectal cancer: cohort study and literature review. Mol Cancer. 2014;13(1):135. doi:10.1186/1476-4598-13-135.
5. Liao X, Morikawa T, Lochhead P, et al. Prognostic role of PIK3CA mutation in colorectal cancer: Cohort study and literature review. Clin Cancer Res. 2012;18(8):2257-2268. doi:10.1158/1078-0432.CCR-11-2410.
6. Nosho K, Kawasaki T, Ohnishi M, et al. PIK3CA mutation in colorectal cancer: relationship with genetic and epigenetic alterations. Neoplasia. 2008;10(6):534-541. doi:10.1593/neo.08336.
7. Ogino S, Meyehardt J a., Cantor M, et al. Molecular alterations in tumors and response to combination chemotherapy with gefitinib for advanced colorectal cancer. Clin Cancer Res. 2005;11(18):6650-6656. doi:10.1158/1078-0432.CCR-05-0738.
8. Ogino S, Brahmandam M, Cantor M, et al. Distinct molecular features of colorectal carcinoma with signet ring cell component and colorectal carcinoma with mucinous component. Mod Pathol. 2006;19(1):59-68. doi:10.1038/modpathol.3800482.
9. Hinoue T, Weisenberger DJ, Lange CPE, et al. Abstract LB-173: Genome-scale analysis of aberrant DNA methylation in colorectal cancer. Cancer Res. 2011;71(8 Supplement):LB - 173 - LB - 173. doi:10.1158/1538-7445.AM2011-LB-173.
10. Ogino S, Kawasaki T, Nosho K, et al. LINE-1 hypomethylation is inversely associated with microsatellite instability and CpG island methylator phenotype in colorectal cancer. Int J Cancer. 2008;122(12):2767-2773. doi:10.1002/ijc.23470.
11. Ogino S, Kawasaki T, Brahmandam M, et al. Precision and performance characteristics of bisulfite conversion and real-time PCR (MethyLight) for quantitative DNA methylation analysis. J Mol Diagn. 2006;8(2):209-217. doi:10.2353/jmoldx.2006.050135.
12. Ogino S, Nosho K, Kirkner GJ, et al. CpG island methylator phenotype, microsatellite instability, BRAF mutation and clinical outcome in colon cancer. Gut. 2009;58(1):90-96. doi:10.1136/gut.2008.155473.
Hanyuda A et al. IRS1, Exercise and Colorectal Cancer Survival, Supplementary Materials /19
Supplementary Tables
Table S1. Clinical, pathologic and molecular characteristics in colorectal cancer cases of male population according to post-diagnosis physical activity quartile
Clinical, pathologic or molecular characteristics / Total No. / Post-diagnosis physical activity quartile ofmale population* / P†
Q1 / Q2 / Q3 / Q4
All cases / 179 / 43 / 46 / 45 / 45
Age, years (mean ± SD) / 69.1 ± 8.4 / 71.2 ± 8.9 / 69.3 ± 9.8 / 67.9 ± 7.7 / 68.1 ± 6.7 / 0.25‡
Year of diagnosis / 0.11
Prior to 1996 / 90 (50%) / 18 (42%) / 22 (48%) / 24 (53%) / 26 (58%)
1996 to 2008 / 89 (50%) / 25 (58%) / 24 (52%) / 21 (47%) / 19 (42%)
Body mass index, kg/m2 / 0.77
<30 / 163 (91%) / 38 (88%) / 42 (91%) / 43 (96%) / 40 (89%)
≥30 / 16 (8.9%) / 5 (12%) / 4 (8.7%) / 2 (4.4%) / 5 (11%)
Family history of colorectal cancer in first degree relatives / 0.80
Absent / 141 (79%) / 34 (79%) / 34 (74%) / 38 (84%) / 35 (78%)
Present / 38 (21%) / 9 (21%) / 12 (26%) / 7 (16%) / 10 (22%)
Tumor location / 0.56
Caecum / 45 (25%) / 10 (23%) / 14 (30%) / 10 (23%) / 11 (24%)
Ascending and transverse colon / 55 (31%) / 15 (35%) / 16 (35%) / 11 (25%) / 13 (29%)
Splenic flexure to sigmoid colon / 36 (20%) / 7 (16%) / 8 (17%) / 11 (25%) / 10 (22%)
Rectum / 42 (24%) / 11 (26%) / 8 (17%) / 12 (27%) / 11 (24%)
Cancer stage / 0.54
I / 50 (30%) / 16 (41%) / 11 (27%) / 9 (21%) / 16 (37%)
II / 62 (38%) / 10 (26%) / 13 (32%) / 20 (47%) / 19 (44%)
III / 52 (32%) / 13 (33%) / 17 (41%) / 14 (33%) / 8 (19%)
Tumor grade / 0.51
Low / 167 (94%) / 42 (98%) / 43 (93%) / 39 (89%) / 43 (96%)
High / 11 (6.2%) / 1 (2.3%) / 3 (6.5%) / 5 (11%) / 2 (4.4%)
MSI status / 0.91
MSI-low/MSS / 154 (87%) / 39 (91%) / 37 (82%) / 38 (84%) / 40 (89%)
MSI-high / 24 (13%) / 4 (9.3%) / 8 (18%) / 7 (16%) / 5 (11%)
MLH1 promoter hypermethylation / 0.67
Negative / 161 (90%) / 39 (93%) / 40 (87%) / 39 (87%) / 43 (96%)
Positive / 17 (9.6%) / 3 (7.1%) / 6 (13%) / 6 (13%) / 2 (4.4%)
CIMP status / 0.46
CIMP-negative / 77 (43%) / 18 (43%) / 19 (41%) / 18 (40%) / 22 (49%)
CIMP-low / 81 (46%) / 18 (43%) / 23 (50%) / 20 (44%) / 20 (44%)
CIMP-high / 20 (11%) / 6 (14%) / 4 (8.7%) / 7 (16%) / 3 (6.7%)
LINE-1 methylation, %
(mean ± SD) / 60.4±10.4 / 58.8±10.8 / 58.1±10.6 / 62.4±10.9 / 62.2±8.8 / 0.098‡
BRAF mutation / 0.25
Negative / 171 (96%) / 41 (98%) / 45 (98%) / 43 (96%) / 42 (93%)
Positive / 7 (3.9%) / 1 (2.4%) / 1 (2.2%) / 2 (4.4%) / 3 (6.7%)
KRAS mutation / 0.074
Negative / 94 (53%) / 25 (58%) / 27 (59%) / 24 (53%) / 18 (40%)
Positive / 85 (47%) / 18 (42%) / 19 (41%) / 21 (47%) / 27 (60%)
PIK3CA mutation / 0.15
Negative / 136 (80%) / 35 (85%) / 38 (88%) / 30 (70%) / 33 (79%)
Positive / 33 (20%) / 6 (15%) / 5 (12%) / 13 (30%) / 9 (21%)
TP53 expression / 0.60
Negative / 100 (56%) / 25 (58%) / 21 (46%) / 28 (64%) / 25 (58%)
Positive / 78 (44%) / 18 (42%) / 25 (54%) / 16 (36%) / 19 (42%)
CTNNB1 (β-catenin)
expression (nuclear) / 0.54
Negative / 75 (45%) / 19 (49%) / 19 (41%) / 21 (51%) / 16 (38%)
Positive / 93 (55%) / 20 (51%) / 27 (59%) / 20 (49%) / 26 (62%)
PTGS2 (COX-2) expression / 0.028
Negative / 70 (39%) / 20 (47%) / 21 (46%) / 18 (40%) / 11 (24%)
Positive / 109 (61%) / 23 (53%) / 25 (54%) / 27 (60%) / 34 (76%)
IRS1 expression / 0.26
Negative / 15 (7.0%) / 3 (7.0%) / 4 (8.7%) / 5 (11%) / 3 (6.7%)
Low / 106 (63%) / 27 (63%) / 33 (72%) / 21 (47%) / 25 (56%)
High / 58 (32%) / 13 (30%) / 9 (20%) / 19 (42%) / 17 (38%)
IRS2 expression / 0.60
Negative / 12 (7.1%) / 2 (5.1%) / 4 (8.9%) / 5 (11%) / 1 (2.5%)
Low / 101 (60%) / 23 (59%) / 28 (62%) / 28 (64%) / 22 (55%)
High / 55 (33%) / 14 (36%) / 13 (29%) / 11 (25%) / 17 (43%)
CIMP, CpG island methylator phenotype; LINE-1, long interspersed nucleotide element 1; METS, metabolic equivalent task score; MSI, microsatellite instability; MSS, microsatellite stable; SD, standard deviation