Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Feb 1.
Published in final edited form as: Int J Cancer. 2011 Apr 27;130(3):653–664. doi: 10.1002/ijc.26047

Genetic variation in bone morphogenetic protein (BMP) and colon and rectal cancer

Martha L Slattery 1, Abbie Lundgreen, Jennifer S Herrick, Susan Kadlubar 2, Bette J Caan 3, John D Potter 4, Roger K Wolff 1
PMCID: PMC3155002  NIHMSID: NIHMS283792  PMID: 21387313

Abstract

Bone morphogenetic proteins (BMP) are part of the TGF-β-signaling pathway; genetic variation in these genes may be involved in colorectal cancer. In this study we evaluated the association between genetic variation in BMP1 (11 tagSNPs), BMP2 (5 tagSNPs), BMP4 (3 tagSNPs), BMPR1A (9 tagSNPs), BMPR1B (21 tagSNPs), BMPR2 (11 tagSNPs), and GDF10 (7 tagSNPs) with risk of colon and rectal cancer and tumor molecular phenotype. We used data from population-based case-control studies (colon cancer n=1574 cases, 1970 controls; rectal cancer n=791 cases, 999 controls). We observed that genetic variation in BMPR1A, BMPR1B, BMPR2, BMP2, and BMP4 was associated with risk of developing colon cancer, with 20 to 30% increased risk for most high-risk genotypes. A summary of high-risk genotypes showed over a twofold increase in colon cancer risk at the upper risk category (OR 2.49 95% CI 1.95, 3.18). BMPR2, BMPR1B, BMP2, and GDF10 were associated with rectal cancer. BMPR2 rs2228545 was associated with an almost twofold increased risk of rectal cancer. The risk associated with the highest category of the summary score for rectal cancer was 2.97 (95% CI 1.87, 4.72). Genes in the BMP-signaling pathway were consistently associated with CIMP+ status in combination with both KRAS-mutated and MSI tumors. BMP genes interacted statistically significantly with other genes in the TGF-β-signaling pathway, including TGFβ1, TGFβR1, Smad 3, Smad 4, and Smad 7. Our data support a role for genetic variation in BMP-related genes in the etiology of colon and rectal cancer. One possible mechanism is via the TGF-β-signaling pathway.

Keywords: bone morphogenetic protein, TGF-β, CIMP+, MSI+, genes, colon cancer, rectal cancer, GDF10

Introduction

The TGFβ-signaling pathway plays a critical role in carcinogenesis via regulation of cell growth, differentiation, and proliferation, and apoptosis 1. As members of the TGFβ-signaling pathway, bone morphogenetic proteins (BMP), may be involved in the initiation and progression of colorectal cancer. The BMP pathway has been implicated in the initiation of colorectal cancer among individuals with juvenile polyposis harboring BMPR1A receptor mutations 2 Others have shown that the BMP pathway is inactivated in the majority of sporadic colorectal cancer and may be associated with MSI+ tumors 3.

Little is known about the genetic variation in BMP genes and their associations with colon or rectal cancer. However, we know that the TGF-β-signaling pathway, of which BMP is a component, is a key regulatory pathway for colon and rectal cancer. BMPs have been shown to trigger a Smad- signaling cascade that is linked to reduced cell proliferation and cellular growth kinetics of glioblastomas 4, 5 and may play a key role in regulating tumor initiation. A recent genome-wide association study (GWAS) reported that BMP2 and BMP4 were two of the top 10 genes identified as associated wtih colon cancer6. BMP4 also has been identified as associated with colorectal cancer in the COGENT Study 7. Several studies suggest the importance of the BMP receptors, given that BMPs signal through their type I and II receptors 8. BMPR1A and BMPR1B are the two best characterized type I receptors. Substrates for these receptors include Smad proteins that play a central role in BMP signaling. Genetic variation in Smad genes has been associated with colon and rectal cancer 9, 10. GWAS have shown that Smad7 is associated with colorectal cancer6, 7. Type II BMP receptors, such as BMPR2, like type I receptors, are necessary for BMP signaling.

In this study, we examined genetic variation in BMP1, BMP2, BMP4 and their relevant receptor genes BMPR1A, BMPR1B, BMPR2, and Growth Differentiation Factor 10 (GDF10) also known as BMP3B. We evaluated associations between variants in the BMP pathway with specific tumor markers because others have shown that BMPR2 expression differs by MSI status 3. Because BMP genes are part of a larger TGF-β-signaling pathway we assessed interaction between BMP genes and other genes in that pathway, including TGFβ1, TGFβR1, Smad3, Smad4, Smad7, and NFκB1.

Methods

Two population-based study populations are included in these analyses. The first study, a population-based case-control study of colon cancer, included cases (n=1,593) and controls (n=1,994) identified between October1, 1991 and September 30, 1994 11 living in the Twin Cities Metropolitan Area, Kaiser Permanente Medical Care Program of Northern California (KPMCP) and a seven-county area of Utah. The second study, with identical data collection methods, included population-based cases with cancer of the rectosigmoid junction or rectum (n=790) and controls (n=999) who were identified between May 1997 and May 2001 in Utah and KPMCP 12. Eligible cases were between 30 and 79 years old at time of diagnosis, English speaking, mentally competent to complete the interview, no previous history of CRC, and no known (as indicated on the pathology report) familial adenomatous polyposis, ulcerative colitis, or Crohn’s disease.

Controls were matched to cases by sex and by 5-year age groups. At KPMCP, controls were randomly selected from membership lists; in Utah, controls 65 years and older were randomly selected from the Health Care Financing Administration lists and controls younger than 65 years were randomly selected from driver’s license lists. In Minnesota, controls were selected from driver’s license and state-identification lists. Study details have been previously reported 13, 14.

Interview Data Collection

Data were collected by trained and certified interviewers using laptop computers. All interviews, as previously described, were audio-taped as previously described and reviewed for quality control purposes 15. The referent period for the study was two years prior to diagnosis for cases and selection for controls. Detailed information was collected on diet, physical activity, medical history, reproductive history, family history of cancer in first-degree relatives, regular use of aspirin and non-steroidal anti-inflammatory drugs, and body size.

Tumor Marker Data

We have previously evaluated tumors for CpG island methylator phenotype (CIMP), microsatellite instability (MSI), TP53 mutations, and KRAS mutations 1619 and were therefore able to evaluate BMP-related genes in relation to tumors with specific molecular characteristics. Details of methods used to evaluate epigenetic and genetic changes have been described 1619. Given the rarity of MSI+ rectal tumors 20 we were unable to evaluate that small subset of tumors.

TagSNP Selection and Genotyping

TagSNPs were selected for BMP1(rs3924229, rs1357482, rs4076873, rs7592, rs7812993, rs4872360, rs12114940, rs3924231, rs4075478, rs3857979, rs11775186), BMP2 (rs235770, rs1979855, rs7270163, rs1005464, rs3178250), BMP4 (rs17563, rs762642, rs2761887), BMPR1A (rs10887668, rs7895217, rs4934275, rs6586034, rs7088641, rs21687668, rs12765929, rs12415784, rs2883420), BMPR1B (rs7698964, rs7694043, rs7661049, rs1863652, rs9307147, rs11947569, rs13134042, rs6849425, rs4145993, rs7662504, rs12508087, rs3821968, rs6499673, rs4490463, rs10049681, rs2214395, rs2719176, rs17616243, rs17022671, rs2120834, rs3796442), BMPR2 (rs12477602, rs2350809, rs6751210, rs13430786, rs1980153, rs4303700, rs4675278, rs12621870, rs1199496, rs17199235, rs2228545), and GDF10 (rs762454, rs2853838, rs7093975, rs1198444, rs12769499, rs1902725, rs1902724) using the following parameters: LD blocks using a Caucasian LD map and an r2=0.8 defined; minor allele frequency (MAF) >0.1; range= −1500 bps from the initiation codon to +1500 bps from the termination codon; and 1 SNP/LD bin. All markers were genotyped using a multiplexed bead-array assay format based on GoldenGate chemistry (Illumina, San Diego, California). A genotyping call rate of 99.85% was attained. Blinded internal replicates represented 4.4% of the sample set; the duplicate concordance rate was 100%. A detailed summary of these SNPs is available in the online supplement. Genotyping of other genes in the candidate pathway, including NFκB1, TGFβ1, TGFβR1, Smad3, Smad4, and Smad7, which were assessed for their interactive effects with BMP genes, were genotyped on the same platform. Individuals with missing genotype data were not included in the analysis for that specific marker.

Statistical Methods

Statistical analyses were performed using SAS® version 9.2 (SAS Institute, Cary, NC). We report odds ratios (ORs) and 95% confidence intervals (95%CIs) assessed from adjusted multiple logistic regression models. TagSNP selection was based on those tagSNPs identified as being statistically significant using multiple logistic regression models adjusting for age, center, race/ethnicity, and sex. To summarize risk associated with multiple variants across the pathway we created a summary score that was based on all at-risk genotypes identified from multiple regression models for colon and rectal cancer. The score for each SNP was based on the inheritance model and its associated risk. For the co-dominant or additive model a score of zero, one, or two was assigned directly related to the number of high-risk alleles, while scores of zero or two were assigned for the dominant and recessive models. After assigning a score for each SNP, the scores were summed across SNPs to generate an individual summary score. The score variable was categorized based on the frequency distribution within the study population. P values for trend were determined by comparing a full model including the continuous score term to a score reduced model via a likelihood-ratio test.

Analysis for interaction was based on tagSNPs within each BMP gene with a Wald p value of <0.15 from the initial logistic regression analysis. These SNPs were compared to targeted candidate SNPs within genes in the proposed pathway that were previously identified as being statistically significantly associated with colon and rectal cancer at the 0.05 level or less. Genes tested for interaction were: TGFβ1 (2 SNPs for colon and rectal cancer), TGFβR1 (1 SNP for colon cancer only), Smad2 (2 SNPs for colon and 1 SNP for rectal cancer), Smad3 (4 SNPs for colon cancer and 1 SNP for rectal cancer), Smad7 (3 candidate SNPs for both colon and rectal cancer), and NFκB1 (5 SNPs for both colon and rectal cancer). BMP genes evaluated were BMP1 (1 SNP for rectal cancer), BMP2 (3 SNPs for colon cancer and 2 SNPs for rectal cancer), BMP4 (1 SNP for colon cancer), BMPR2 (2 SNPs for both colon and rectal cancer), BMPR1A (5 SNPs for colon cancer), BMPR1B (10 SNPs for colon cancer and 4 SNPs for rectal cancer), and GDF10 (1 SNP for rectal cancer).

Possible interactions between BMP genes and three hypothesized non-gene exposures associated with inflammation (i.e. recent aspirin or NSAID use), estrogen (i.e. recent estrogen use), and insulin (i.e. BMI of <25, 25–30, >30) were evaluated. We believe that inflammation, estrogen, and insulin are central to colon and rectal cancer etiology; these variables were selected as indicators of these lifestyle exposures that may interact with this candidate pathway. P values for interaction for genetic and lifestyle factors were determined using a likelihood-ratio test comparing a full model that included an interaction term with a reduced model without an interaction term.

Tumors were defined by specific somatic alterations; any TP53 mutation; any KRAS mutation; MSI+; CIMP+ defined as at least two of five markers methylated; a combination of CIMP+/KRAS-mutated; a combination of CIMP+/MSI+. As the proportion of MSI+ tumors in the rectal cases was <3% 20, we did not examine that molecular phenotype in our rectal data. Population-based controls were used to assess associations between tagSNPs in candidate genes and specific tumor molecular phenotypes using the summary score methods described above. Comparisons of cases with and without specific epigenetic and genetic changes were conducted to test for heterogeneity with specific tumor molecular phenotype. The heterogeneity p values are based on the likelihood-ratio test comparing a full model with a reduced model excluding the score term, both of which are adjusted for other tumor markers.

Adjusted multiple-comparison p values, taking into account tagSNPs within the gene, were estimated using the methods by Conneely and Boehnke21 via R version 2.11.0 (R Foundation for Statistical Computing, Vienna, Austria). Wald p values from the original models and interaction p values based on likelihood-ratio tests were used for estimates of multiple comparisons. We consider a pACT of <0.15 as being potentially important given the candidate pathway approach and the need to consider both type 1 and type 2 errors. We believe that findings at this level would merit replication.

Results

The study population is described in Table 1. The majority of participants were white non-Hispanic, male, and over 60 years of age. Table 2 describes the tagSNPs for the candidate genes carried forward into further analyses based on statistically significantly associations with colon and rectal cancer, either overall or with specific tumor markers. All tagSNPs were in HWE. Supplemental Table 1 provides a list of detailed information on all tagSNPs for these genes included on the platform.

Table 1.

Description of Study Population

Colon Rectal
Control Case Control Case
n (%) n (%) n (%) n (%)
Center Utah 378 (19.33) 249 (16.01) 365 (38.06) 274 (36.34)
Kaiser, CA 787 (40.24) 744 (47.85) 594 (61.94) 480 (63.66)
Minnesota 791 (40.44) 562 (36.14) NA NA
Age 30–39 40 (2.04) 23 (1.48) 21 (2.19) 19 (2.52)
40–49 128 (6.54) 102 (6.56) 101 (10.53) 96 (12.73)
50–59 326 (16.67) 290 (18.65) 243 (25.34) 196 (25.99)
60–69 673 (34.41) 538 (34.60) 329 (34.31) 250 (33.16)
70–79 789 (40.34) 602 (38.71) 265 (27.63) 193 (25.60)
Race/Ethnicity NHW 1828 (93.46) 1428 (91.83) 824 (85.92) 625 (82.89)
Hispanics 75 (3.83) 59 (3.79) 63 (6.57) 61 (8.09)
Black 53 (2.71) 68 (4.37) 43 (4.48) 29 (3.85)
Asian 29 (3.02) 39 (5.17)
Gender Male 1047 (53.53) 870 (55.95) 541 (56.41) 451 (59.81)
Female 909 (46.47) 685 (44.05) 418 (43.59) 303 (40.19)

Table 2.

Descriptive table of tagSNPs associated with colon and rectal cancer.

Symbol Location SNP Major/Minor Allele NHW MAF1 Hisp AA FDR2 HWE
BMP2 20p12 rs235770 C/T 0.38 0.33 1.00 1.00
rs1979855 T/C 0.17 0.11 0.16 0.69
rs3178250 T/C 0.2 0.18 0.23 0.97
BMP4 14q22-q23 rs17563 C/T 0.44 0.40* 0.21* 0.75
BMPR1A 10q22.3 rs6586034 T/G 0.44 0.49 0.23* 0.91
rs7088641 T/C 0.31 0.35 0.07 0.67
rs2883420 T/C 0.39 0.43 0.22* 0.96
BMPR1B 4q22-q24 rs7694043 C/T 0.35 0.26 0.12 1.00
rs1863652 C/T 0.35 0.41 0.23 0.91
rs9307147 A/G 0.45 0.38 0.24 1.00
rs11947569 T/C 0.21 0.19 0.22 1.00
rs13134042 G/A 0.21 0.23 0.08 0.74
rs6849425 C/T 0.21 0.23 0.19 0.96
rs7662504 A/C 0.39 0.48* 0.48 1.00
rs4490463 A/G 0.42 0.34 0.45* 1.00
rs2719176 C/G 0.38 0.4 0.49* 1.00
rs17616243 C/T 0.15 0.18 0.07 1.00
rs2120834 G/C 0.39 0.36 0.47 0.75
BMPR2 2q33-q34 rs6751210 A/G 0.49 0.47 0.44* 0.85
rs17199235 A/G 0.11 0.04 0.04 0.67
rs2228545 G/A 0.03 <.01 0.02 1.00
GDF10 10q11.22 rs762454 A/G 0.33 0.29 0.34 1.00
1

Minor Allele Frequency (MAF) based on control population;

*

Indicates major/minor allele differs for Hispanic (Hisp) or African American (AA) from non-Hispanic white (NHW)

2

FDR (HWE) = False Discovery Rate Hardy Weinberg Equilibrium test.

Associations between tagSNPs, and risk of colon and rectal cancer are shown in Table 3. For several genes, more than one tagSNP was independently associated with colon cancer; BMPR1B rs13134042, rs2120834, rs17616243, rs2719176, and rs1863652 were all associated with colon cancer with ORs of similar magnitudes of risk. Seven SNPs had pACT of <0.15 for colon cancer and three SNPs, BMP2 rs12979855, rs3178250, and BMPR1A rs2883420 had pACT values of <0.05. For rectal cancer seven independent tagSNPs were associated with disease risk, representing four genes, BMP2, BMPR1B, BMPR2, and GDF10. Of these, BMPR2 rs17199235 had an adjusted pACT of <0.05 and BMPR2 rs228545 and GDF10 rs762454 had pACT values of <0.15. For both colon and rectal cancer, the summary score across tagSNPs showed a significant linear trend of increasing risk associated with increasing number of higher risk genotypes.

Table 3.

Associations between BMP genes and colon and rectal cancer

Colon Cancer Controls Cases OR (95% CI) P value2 pACT3
BMP2 rs19798554 0.00056 0.0027
TT 1381 1014 1.00
TC/CC 575 541 1.29 (1.11,1.48)
rs3178250 0.0094 0.0359
TT 1256 931 1.00
TC/CC 700 624 1.20 (1.05,1.38)
rs235770 0.0223 0.063
CC 796 691 1.00
CT 884 680 0.91 (0.78,1.05)
TT 276 184 0.78 (0.63,0.97)
BMP4 rs17563 0.0315 0.0791
CC/CT 1567 1191 1.00
TT 387 364 1.20 (1.02,1.41)
BMPR1A rs2883420 0.0065 0.0492
TT 685 584 1.00
TC/CC 1271 971 0.87 (0.76,1.00)
rs70886415 0.0098 0.0666
TT 904 793 1.00
TC/CC 1052 762 0.83 (0.73,0.95)
rs6586034 0.0487 0.2592
TT 555 497 1.00
TG/GG 1388 1051 0.82 (0.71,0.95)
BMPR1B rs9307147 0.0028 0.0538
AA 604 548 1.00
AG 954 740 0.86 (0.74,1.00)
GG 398 267 0.75 (0.62,0.91)
rs17616243 0.0094 0.1561
CC 1427 1077 1.00
CT/TT 529 478 1.22 (1.05,1.41)
rs7662504 0.0169 0.2497
AA 706 605 1.00
AC 931 729 0.90 (0.78,1.04)
CC 318 221 0.78 (0.64,0.96)
rs4490463 0.0375 0.4134
AA 646 562 1.00
AG 964 746 0.89 (0.76,1.03)
GG 345 246 0.82 (0.67,1.00)
rs2719176 0.0439 0.428
CC/CG 1673 1291 1.00
GG 283 264 1.21 (1.01,1.45)
rs13134042 0.0400 0.4213
GG/GA 1856 1498 1.00
AA 99 57 0.70 (0.50,0.98)
rs1863652 0.0299 0.3634
CC/CT 1711 1398 1.00
TT 243 157 0.79 (0.64,0.98)
rs2120834 0.0189 0.2647
GG/GC 1648 1354 1.00
CC 307 201 0.79 (0.65,0.96)
BMPR2 rs6751210 0.0366 0.2816
AA/AG 1501 1145 1.00
GG 455 410 1.18 (1.01,1.38)
Summary Score
(1 – 11) 414 194 1.00
(12 – 14) 483 368 1.60 (1.28,1.99)
(15 – 17) 477 400 1.76 (1.42,2.19)
(18 – 20) 339 312 1.94 (1.54,2.44)
(21 – 30) 243 281 2.49 (1.95,3.18)
P Trend <.0001
Rectal Cancer
BMP2 rs3178250 0.0403 0.1738
TT/TC 926 711 1.00
CC 33 42 1.63 (1.02,2.60)
BMPR1B rs7694043 0.02607 0.3656
CC 413 367 1.00
CT/TT 546 387 0.82 (0.67,1.00)
rs6849425 0.042 0.4945
CC/CT 902 725 1.00
TT 57 29 0.62 (0.39,0.98)
rs11947569 0.0459 0.5057
TT/TC 924 710 1.00
CC 35 44 1.68 (1.06,2.65)
BMPR2 rs17199235 0.0016 0.0162
AA 798 596 1.00
AG/GG 161 158 1.35 (1.05,1.73)
rs2228545 0.0174 0.1349
GG 917 695 1.00
GA/AA 42 59 1.93 (1.28,2.91)
GDF10 rs762454 0.0218 0.1068
AA/AG 862 651 1.00
GG 97 102 1.42 (1.05,1.91)
Summary Score
(0 – 2) 402 244 1.00
(3 – 4) 394 324 1.32 (1.07,1.65)
(5 – 6) 131 129 1.61 (1.20,2.16)
(7 – 11) 32 57 2.97 (1.87,4.72)
P Trend <.0001
1

Odds Ratio (OR) and 95% Confidence Interval (CI) from multiple logistic regression analysis adjusting for age, sex, center, and race/ethnicity

2

Wald p value

3

pACT from methods of Conneely and Boehnke21

4

BMP2 rs1979855 and rs3178250 r2=0.59

5

BMPR1A rs7088641 and rs6586034 r2 = 0.58

Assessment of interaction between BMP-related genes and other genes in the candidate pathway that were hypothesized as interacting with BMP genes, showed several statistically significant interactions. For colon cancer BMP2 interacted with NFκB1, Smad3, TGFβ1, Smad2 and Smad7; BMP4 interacted statistically with NFκB1 and Smad3, BMPR1B with NFκB1, Smad2, Smad7, Smad3, Smad4, and TGFβ1; BMPR1A with Smad7 and TGFβ1, and BMPR2 interacted significantly with Smad3, Smad7, and TGFβ1. For rectal cancer BMP1interacted with NFκB1, Smad7, and TGFβ1; BMP2 interacted statistically significantly with TGFβ1; BMPR1B interacted with Smad7 and TGFβR1; BMPR2 interacted with NFκB1 and TGFβ1; and GDF10 interacted with NFκB1, Smad2, and TGFβ1 (Table 4). Of the 357 SNPs evaluated (21BMP SNPs with Wald p <0.15 and 17 gene pathway SNPS) for interaction in colon cancer, 62 had a p value of <0.05, of which 37 had a pACT of <0.15 and 11 had pACT values of <0.05. For rectal cancer, we tested 120 SNP interactions (10 BMP SNPS with 12 gene pathway SNPS), of which 23 were significant at the 0.05 level; after adjustment of these SNPs for multiple comparisons, 19 had a pACT at the 0.15 level and 7 had a pACT at the 0.05 level.

Table 4.

Associations between BMP genes and candidate genes in hypothesized pathway

BMP Gene SNP (high-risk genotype) Pathway Gene SNP (high- risk genotype) Interaction p value1 pACT2
Colon Cancer
BMP2 rs1979855 (TC/CC) NFκB1 rs3821958 (GG) 0.0339 0.1618
SMAD3 rs12901071 (AA) 0.0065 0.0309
TGFβ1 rs1800469 (GG) 0.0080 0.0357
rs4803455 (AA) 0.0406 0.0907
rs235770 (CC) SMAD2 rs1787199 (TT) 0.0327 0.0978
SMAD3 rs2414937 (GG/GC) 0.0015 0.0082
rs7163381 (GG/GA) 0.0325 0.1226
SMAD7 rs12953717 (TT) 0.0335 0.1397
rs4939827 (TT) 0.0098 0.0509
TGFβ1 rs1800469 (GG) 0.0093 0.0363
rs4803455 (AA) 0.0103 0.0343
rs3178250 (TC/CC) SMAD2 rs4940086 (CC) 0.0307 0.0984
SMAD3 rs12901071 (AA) 0.0126 0.0553
BMP4 rs17563 (TT) rs12901071 (AA) 0.0024 0.0051
NFκB1 rs230510 (AA) 0.0052 0.0136
rs3821958 (GG) 0.0219 0.0409
BMPR1A rs2168730 (GG) SMAD7 rs4939827 (TT) 0.0015 0.0153
rs2883420 (TT) rs4464148 (CC) 0.0478 0.2620
rs7895217 (AA) rs12953717 (TT) 0.0204 0.1377
rs4464148 (CC) 0.0108 0.0816
TGFβ1 rs1800469 (GG) 0.0214 0.1339
BMPR1B rs13134042 (GG/GA) NFκB1 rs4648110 (TT/TA) 0.0226 0.3525
SMAD2 rs1787199 (TT) 0.0140 0.1347
SMAD7 rs12953717 (TT) 0.0058 0.0995
rs4939827 (TT) 0.0033 0.0627
rs1863652 (CC/CT) NFκB1 rs11722146 (AA) 0.0449 0.5433
rs230510 (AA) 0.0193 0.3187
rs3821958 (GG) 0.0228 0.3526
SMAD3 rs1498506 (AA) 0.0364 0.4523
SMAD7 rs12953717 (TT) 0.0192 0.2498
rs4464148 (CC) 0.0469 0.4390
rs2120834 (GG/GC) SMAD2 rs1787199 (TT) 0.0404 0.3096
TGFβ1 rs1800469 (GG) 0.0084 0.1128
rs4803455 (AA) 0.0070 0.1023
rs2719176 (GG) NFκB1 rs230510 (AA) 0.0032 0.0776
SMAD3 rs1498506 (AA) 0.0293 0.3923
rs3821968 (TT) NFκB1 rs11722146 (AA) 0.0407 0.5180
rs4490463 (AA) rs11722146 (AA) 0.0062 0.1322
rs230510 (AA) 0.0045 0.1016
rs3821958 (GG) 0.0089 0.1775
SMAD2 rs1787199 (TT) 0.0497 0.3550
rs4940086 (CC) 0.0292 0.2431
SMAD3 rs12901071 (AA) 0.0232 0.3338
rs1498506 (AA) 0.0045 0.0872
TGFβ1 rs1800469 (GG) 0.0230 0.2477
rs4699673 (GG) SMAD7 rs12953717 (TT) 0.0194 0.2467
rs4939827 (TT) 0.0096 0.1501
rs7662504 (AA) NFκB1 rs4648110 (TT/TA) 0.0201 0.3245
SMAD2 rs1787199 (TT) 0.0001 0.0011
rs4940086 (CC) 0.0010 0.0129
SMAD3 rs12901071 (AA) 0.0494 0.5470
SMAD7 rs12953717 (TT) 0.0125 0.1796
rs4464148 (CC) 0.0384 0.4035
rs4939827 (TT) 0.0044 0.0808
rs9307147 (AA) SMAD2 rs4940086 (CC) 0.0118 0.1186
SMAD3 rs1498506 (AA) 0.0106 0.1809
SMAD7 rs4464148 (CC) 0.0206 0.2555
BMPR2 rs1980153 (AA) SMAD3 rs1498506 (AA) 0.0422 0.1452
SMAD7 rs12953717 (TT) 0.0183 0.0761
rs4464148 (CC) 0.0332 0.0979
rs4939827 (TT) 0.0261 0.0956
rs6751210 (GG) TGFβ1 rs4803455 (AA) 0.0172 0.0593
Rectal Cancer
BMP1 rs3924229 (CC) NFκB1 rs11722146 (AA) 0.0461 0.1251
rs3821958 (GG) 0.0076 0.0289
rs4648110 (AA) 0.0418 0.1249
SMAD7 rs12953717 (CC) 0.0238 0.0617
rs4464148 (TT) 0.0320 0.0624
TGFβ1 rs1800469 (GG) 0.0241 0.0598
BMP2 rs1979855 (CC) rs4803455 (AA) 0.0290 0.0561
rs3178250 (CC) rs4803455 (AA) 0.0077 0.0270
BMPR1B rs11947569 (CC) SMAD7 rs12953717 (CC) 0.0185 0.1263
rs4939827 (TT) 0.0257 0.1580
rs13134042 (GA/AA) NFκB1 rs230510 (AA) 0.0404 0.3058
rs4648110 (AA) 0.0196 0.1703
SMAD7 rs12953717 (CC) 0.0356 0.1907
rs6849425 (CC/CT) NFκB1 rs3821958 (GG) 0.0135 0.1300
SMAD7 rs4464148 (TT) 0.0053 0.0463
rs4939827 (TT) 0.0166 0.1224
BMPR2 rs17199235 (AG/GG) TGFβ1 rs1800469 (GG) 0.0017 0.0063
rs2228545 (GA/AA) NFκB1 rs230510 (AA) 0.0213 0.0960
rs3821958 (GG) 0.0067 0.0360
TGFβ1 rs1800469 (GG) 0.0344 0.0903
GDF10 rs762454 (GG) NFκB1 rs4648110 (AA) 0.0492 0.1216
SMAD2 rs1792689 (CC) 0.0206 0.0206
TGFβ1 rs4803455 (AA) 0.0086 0.0165
1

likelihood ratio p value

2

pACT from methods of Conneely and Boehnke 21

We evaluated the combined effects of the BMP genes with various molecularly defined colon and rectal tumor phenotypes (Table 5). The colon tumor phenotypes most influenced by the BMP genes were CIMP+, MSI+, KRAS-mutated, and combinations of these epigenetic and genetic molecular changes. KRAS-mutated tumors were not associated with BMP-related genes for rectal cancer, although TP53-mutated tumors were. Risk summary scores showed increasing risk with increasing number of at-risk genotypes for both colon and rectal cancer. The p value for heterogeneity indicates that the majority of associations were unique to the specific tumor molecular phenotype. The magnitude of the associations with tumor markers was slightly stronger for rectal tumors than for colon tumors.

Table 5.

Associations between colon and rectal tumor molecular phenotype and BMP genes

Score1 N
OR (95%CI)2 OR (95% CI)2 Heterogeneity
Controls Case+ Cases− Case+ vs Controls CIMP+ CASE+ vs. CASE− p value3
Colon Cancer
(0 – 3) 488 43 179 1.00 1.00
(4 – 6) 968 137 368 1.57 (1.10,2.26) 1.46 (0.94,2.26) 0.0869
(7 – 8) 425 75 162 2.04 (1.37,3.04) 2.07 (1.25,3.42) 0.0040
(9 – 10) 75 17 29 2.60 (1.40,4.82) 1.99 (0.89,4.44) 0.1000
KRAS
(0 – 4) 492 52 183 1.00 1.00
(5 – 6) 561 101 203 1.63 (1.14,2.33) 1.70 (1.09,2.64) 0.0178
(7 – 8) 441 92 162 1.94 (1.34,2.79) 2.09 (1.33,3.30) 0.0013
(9 – 12) 462 103 195 2.05 (1.44,2.94) 1.70 (1.08,2.68) 0.0208
MSI+
(0 – 6) 451 23 218 1.00 1.00
(7 – 8) 323 24 156 1.43 (0.79,2.58) 1.92 (0.81,4.58) 0.1382
(9 – 11) 567 52 313 1.80 (1.08,2.98) 1.34 (0.68,2.66) 0.3992
(12 – 14) 435 49 219 2.15 (1.28,3.60) 2.47 (1.25,4.88) 0.0078
(15 – 18) 180 37 84 3.98 (2.29,6.89) 4.27 (1.86,9.80) 0.0004
CIMP+& KRAS
(0 – 2) 1073 27 243 1.00 1.00
(4 – 4) 704 31 171 1.71 (1.01,2.90) 1.71 (0.96,3.04) 0.0680
(6 – 6) 179 16 36 3.58 (1.88,6.80) 3.80 (1.73,8.38) 0.0013
CIMP+ & MSI+
(0 – 6) 592 15 196 1.00 1.00
(7 – 8) 518 24 172 1.85 (0.96,3.57) 2.02 (0.94,4.32) 0.0660
(9 – 11) 631 43 244 2.70 (1.48,4.93) 1.89 (0.94,3.80) 0.0699
(12 – 16) 215 26 71 5.12 (2.64,9.93) 4.11 (1.85,9.14) 0.0004
Rectal Cancer
CIMP+
(0 – 3) 345 6 152 1.00 1.00
(4 – 5) 327 19 163 3.46 (1.36,8.79) 2.95 (1.13,7.69) 0.0185
(6 – 7) 219 23 114 6.30 (2.51,15.80) 5.00 (1.93,12.95) 0.0003
(8 – 11) 68 11 43 9.71 (3.45,27.32) 11.31 (3.05,41.95) <.0001
TP53
(0 – 0) 226 55 75 1.00 1.00
(1 – 2) 315 63 77 0.77 (0.52,1.16) 1.13 (0.67,1.89) 0.6562
(3 – 4) 266 79 72 1.18 (0.80,1.74) 1.70 (1.03,2.82) 0.0376
(5 – 6) 113 55 39 1.93 (1.24,2.99) 2.08 (1.15,3.76) 0.0145
(7 – 12) 39 25 18 2.58 (1.44,4.63) 1.73 (0.81,3.70) 0.1562
CIMP+& KRAS
(0 – 5) 663 5 234 1.00 1.00
(6 – 8) 227 8 70 4.72 (1.53,14.59) 5.20 (1.59,16.97) 0.0057
(9 – 14) 69 8 33 15.81 (5.01,49.89) 12.93 (3.80,43.97) <.0001
1

SNPs included in summary score

Colon CIMP+: BMP4 (rs17563 and (rs762642), BMPR1A (rs12765929), BMPR2 (rs12477602 and rs4303700);

Rectal CIMP+: BMP1 (rs4076873), BMP2 (rs235770), BMPR1B (rs11947569, rs3821968 and rs7662504), GDF10 (rs762454);

Colon KRAS: BMPR1B (rs17616243, rs2719176, rs4490463, rs7698964, rs9307147), GDF10 (rs2853838);

Rectal TP53: BMP2 (rs1979855, rs235770), BMPR1B (rs11947569), BMPR2 (rs17199235, rs2228545, rs4675278);

Colon MSI: BMP1 (rs3857979, rs4075478), BMP4 (rs17563), BMPR1A (rs6586034), BMPR1B (rs12508087, rs4490463, rs6849425), BMPR2 (rs12621870, rs4675278);

Colon CIMP+ & KRAS: BMP1 (rs3924231), BMPR2 (rs13430786), GDF10 (rs7093975;

Rectal CIMP+ & KRAS: BMP1 (rs4076873), BMP2 (rs235770), BMPR1B (rs1863652, rs2214395, rs3821968), BMPR2 (rs1199496), GDF10 (rs1902724);

Colon CIMP+ & MSI: BMP1 (rs13257482, rs3857979), BMP2 (rs1005464, rs3178250), BMP4 (rs17563), BMPR1B (rs12508087, rs4490463), BMPR2 (rs4675278);.

2

Odds Ratios (OR) and 95% Confidence Intervals (CI) estimated from multiple logistic regression models adjusting for age, center, ethnicity/race, and sex

3

p for heterogeneity based on regression models comparing cases with and without tumor molecular phenotype

There were few statistically significant interactions between BMP genes and obesity and recent use of aspirin/NSAIDS or estrogen status (data not shown in table). BMP2 rs235770 interacted statistically significantly with BMI; those with the TT genotype had a greater risk of rectal cancer if they had a BMI of 30 or more (OR 2.08 95% CI 1.13,3.82 compared to OR of 0.73 95% CI 0.44,1.24 for normal weight and TT genotype; p interaction 0.0098; pACT 0.02). BMPR1B rs9307147 interacted statistically significantly with aspirin/NSAIDs; having the GG genotype reduced colon cancer risk among those without recent use (OR 0.63 95% CI 0.49,0.80 while the GG genotype among aspirin/NSAID users was 1.0; p interaction 0.0288; pACT 0.22). No other meaningful interactions were detected.

Discussion

This study highlights the potential importance of the BMP genes in colon and rectal carcinogenesis. Both independently and compositely, these genes are associated with cancer risk. Our findings corroborate the hypothesis that type I and type II receptors of BMP genes play a significant role in disease risk. Given the interaction with many other genes within the TGF-β-signaling pathway, it is probable that at least part of their influence in disease risk is through this signaling pathway and that the pathway may operate through CIMP-related mechanisms in combination with KRAS-mutated tumors and MSI+ tumors.

Loss of BMP signaling has been shown to be highly prevalent in sporadic colon cancers 22. BMP acts as a tumor suppressor that is involved in apoptosis; disturbances in BMP signaling could lead to tumorigenesis 23. BMP also is a member of the TGF-β superfamily that plays a critical role in colorectal cancer. BMP signaling is mediated by its receptors and their downstream molecules such as Smad. Approximately 50% of individuals with juvenile polyposis carry germline mutations in either BMPR1A or Smad4 genes 24. Thus, there is a clear biologically plausible role for BMP genes in the etiology of colorectal cancer.

An important consideration when determining risk associated with genes hypothesized as being a component of a candidate pathway is how they work together as well as independently. It is generally unknown if having one or multiple SNPs have similar effects on risk. For instance, does the risk increase with the number of high-risk genotypes or do multiple high-risk genotypes have a minimal effect beyond any individual high-risk genotype in the candidate pathway? For both colon and rectal cancer, it appears that having multiple high-risk genotypes increases the risk of cancer. The summary risk appeared to have slightly greater effect for rectal cancer than for colon cancer. Our findings illustrate the importance of assessing multiple candidate genes together to obtain a better understanding of their relevance to the overall pathway.

In addition to evaluating how BMP genes work together, we evaluated how these work as part of the TGF-β-signaling pathway. We observed statistically significant interactions with several genes within this pathway, including TGFβ1, TGFβR1, Smad3, Smad4, and Smad7. The statistically significant interaction observed between BMP-related genes and other genes within the TGF-β-signaling pathway supports the concept that multiple components in the pathway influence disease risk, not just isolated genes or SNPs. Additionally, the combined effects of variation in genes within the TGF-β-signaling pathway on colon and rectal cancer risk provides additional support for the importance of this pathway in colon and rectal cancer.

Others have reported that BMPR2 is associated with MSI+ tumors 25. Our data suggest that in addition to associations with MSI, BMP-related genes are associated with CIMP+ tumors. Statistically significant associations were observed for CIMP+ tumors in combination with both MSI+ tumors and KRAS- mutated tumors. Our previous report on polymorphism in TGFβ1, TGFβR1, and Smad genes 26 also suggested that CIMP+ tumors were highly associated with these genes. These data add to the evidence that the TGF-β-signaling pathway is important in the etiology of CIMP+ tumors.

This study was hypothesis driven, assessing candidate genes along a biologically defined candidate pathway. The genes were selected because of their biologic function and potential importance in the regulation of the TGF-β-signaling pathway. Because little is known about these genes, including which SNPs are functional, we used a tagSNP approach to characterize genetic variation within the gene that may influence disease risk. The Cogent GWAS Study reported significant associations for BMP4 rs44442357. A small subset of our data have this tagSNP available and we observed a non-significant risk estimate of 1.18 (95% CI 0.88,1.57) for this SNP, which is comparable to the significant risk estimate of 1.12 (95% CI 1.07–1.18) reported for rs4444235 in the Cogent Study. Although we identified several BMP4 SNPs that were associated with colon cancer they had low D′ values compared to this previously reported SNP; BMP4 rs762642 with a D′ of 0.447 was the only BMP SNP with a value of greater than 0.08 compared to rs444235. TagSNPs, although not necessarily functional, serve as an indication that variation in a relevant gene contributes to disease risk. The identification of functional SNPs in linkage with the tagSNPs is outside the scope of this report, but identification of functional SNPs within these genes could potentially contribute to both improved risk assessment and the development of targeted therapies

Our analysis plan included many comparisons that were necessary to consolidate the data into a more coherent picture of how BMP-related genes are associated with colon and rectal cancer. To address how tagSNPs operated together we calculated summary scores across high-risk genotypes as defined from our initial analysis. Given our limited information on these genes prior to our analysis such selection seems justified. We used the pACT to give an indication of the potential importance of statistically significant individual tagSNPs considering the comparisons being made. We report Wald and likelihood ratio p values that were used to adjust for multiple comparisons. The pACT and other methods should be viewed as an indication of risk of false positive results taking into account the comparisons being made. However, it is important for other studies to replicate these results and conduct experiments to test the functionality of potentially important SNPs and genes, thus we considered a pACT of <0.15 as potentially meaningful for replication purposes and to avoid type 2 errors. These results need confirmation in other large studies of colon and rectal cancer, particularly the question of whether variation in BMP-related genes do, indeed, act in concert to cumulatively elevate risk and if there is a similar pattern of cumulative risk with other members of the TGF-β-signaling pathway

Few studies have examined BMP-related genes and risk of colon and rectal cancer despite the biologic plausibility for an association. Their importance is potentially highlighted by GWAS that have identified both BMP2 and BMP4 among top 10 hits with colon cancer. Here we report that in addition to confirming the role of BMP2 and BMP4 in colon and rectal cancer etiology, we show that other BMP genes also contribute to both colon and rectal cancer risk. Our data support the role of BMP genes as an important component of the TGF-β-signaling pathway and further suggest that this pathway may act to elevate risk of CIMP+ colorectal cancer.

Supplementary Material

Supp Table S1

Acknowledgments

This study was funded by NCI grants CA48998 and CA61757. This research also was supported by the Utah Cancer Registry, which is funded by Contract #N01-PC-67000 from the National Cancer Institute, with additional support from the State of Utah Department of Health, the Northern California Cancer Registry, and the Sacramento Tumor Registry. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official view of the National Cancer Institute. We would like to acknowledge the contributions of Sandra Edwards, Roger Edwards, Leslie Palmer, Donna Schaffer, Dr. Kristin Anderson and Judy Morse for data management and collection.

References

  • 1.Elliott RL, Blobe GC. Role of transforming growth factor Beta in human cancer. J Clin Oncol. 2005;23:2078–93. doi: 10.1200/JCO.2005.02.047. [DOI] [PubMed] [Google Scholar]
  • 2.Hardwick JC, Kodach LL, Offerhaus GJ, van den Brink GR. Bone morphogenetic protein signalling in colorectal cancer. Nat Rev Cancer. 2008;8:806–12. doi: 10.1038/nrc2467. [DOI] [PubMed] [Google Scholar]
  • 3.Kodach LL, Wiercinska E, de Miranda NF, Bleuming SA, Musler AR, Peppelenbosch MP, Dekker E, van den Brink GR, van Noesel CJ, Morreau H, Hommes DW, Ten Dijke P, et al. The bone morphogenetic protein pathway is inactivated in the majority of sporadic colorectal cancers. Gastroenterology. 2008;134:1332–41. doi: 10.1053/j.gastro.2008.02.059. [DOI] [PubMed] [Google Scholar]
  • 4.Piccirillo SG, Vescovi AL. Bone morphogenetic proteins regulate tumorigenicity in human glioblastoma stem cells. Ernst Schering Foundation symposium proceedings; 2006. pp. 59–81. [DOI] [PubMed] [Google Scholar]
  • 5.Piccirillo SG, Reynolds BA, Zanetti N, Lamorte G, Binda E, Broggi G, Brem H, Olivi A, Dimeco F, Vescovi AL. Bone morphogenetic proteins inhibit the tumorigenic potential of human brain tumour-initiating cells. Nature. 2006;444:761–5. doi: 10.1038/nature05349. [DOI] [PubMed] [Google Scholar]
  • 6.Chen LS, Hutter CM, Potter JD, Liu Y, Prentice RL, Peters U, Hsu L. Insights into colon cancer etiology via a regularized approach to gene set analysis of GWAS data. Am J Hum Genet. 86:860–71. doi: 10.1016/j.ajhg.2010.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Houlston RS, Webb E, Broderick P, Pittman AM, Di Bernardo MC, Lubbe S, Chandler I, Vijayakrishnan J, Sullivan K, Penegar S, Carvajal-Carmona L, Howarth K, et al. Meta-analysis of genome-wide association data identifies four new susceptibility loci for colorectal cancer. Nat Genet. 2008;40:1426–35. doi: 10.1038/ng.262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Chen D, Zhao M, Mundy GR. Bone morphogenetic proteins. Growth Factors. 2004;22:233–41. doi: 10.1080/08977190412331279890. [DOI] [PubMed] [Google Scholar]
  • 9.Slattery ML, Herrick J, Lundgreen A, Wolff RK. Genetic variation in the TGF{beta}-signaling pathway and colon and rectal cancer risk. Cancer Epidemiol Biomarkers Prev. doi: 10.1158/1055-9965.EPI-10-0843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Slattery ML, Herrick J, Curtin K, Samowitz W, Wolff RK, Caan BJ, Duggan D, Potter JD, Peters U. Increased risk of colon cancer associated with a genetic polymorphism of SMAD7. Cancer Res. 70:1479–85. doi: 10.1158/0008-5472.CAN-08-1792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Slattery ML, Potter JD, Duncan DM, Berry TD. Dietary fats and colon cancer: assessment of risk associated with specific fatty acids. International journal of cancer. 1997;73:670–7. doi: 10.1002/(sici)1097-0215(19971127)73:5<670::aid-ijc10>3.0.co;2-a. [DOI] [PubMed] [Google Scholar]
  • 12.Slattery ML, Caan BJ, Benson J, Murtaugh M. Energy balance and rectal cancer: an evaluation of energy intake, energy expenditure, and body mass index. Nutrition and cancer. 2003;46:166–71. doi: 10.1207/S15327914NC4602_09. [DOI] [PubMed] [Google Scholar]
  • 13.Slattery ML, Potter J, Caan B, Edwards S, Coates A, Ma KN, Berry TD. Energy balance and colon cancer--beyond physical activity. Cancer Res. 1997;57:75–80. [PubMed] [Google Scholar]
  • 14.Slattery ML, Edwards S, Curtin K, Ma K, Edwards R, Holubkov R, Schaffer D. Physical activity and colorectal cancer. Am J Epidemiol. 2003;158:214–24. doi: 10.1093/aje/kwg134. [DOI] [PubMed] [Google Scholar]
  • 15.Edwards S, Slattery ML, Mori M, Berry TD, Caan BJ, Palmer P, Potter JD. Objective system for interviewer performance evaluation for use in epidemiologic studies. Am J Epidemiol. 1994;140:1020–8. doi: 10.1093/oxfordjournals.aje.a117192. [DOI] [PubMed] [Google Scholar]
  • 16.Samowitz WS, Curtin K, Ma KN, Edwards S, Schaffer D, Leppert MF, Slattery ML. Prognostic significance of p53 mutations in colon cancer at the population level. Int J Cancer. 2002;99:597–602. doi: 10.1002/ijc.10405. [DOI] [PubMed] [Google Scholar]
  • 17.Slattery ML, Curtin K, Anderson K, Ma KN, Ballard L, Edwards S, Schaffer D, Potter J, Leppert M, Samowitz WS. Associations between cigarette smoking, lifestyle factors, and microsatellite instability in colon tumors. J Natl Cancer Inst. 2000;92:1831–6. doi: 10.1093/jnci/92.22.1831. [DOI] [PubMed] [Google Scholar]
  • 18.Samowitz WS, Curtin K, Schaffer D, Robertson M, Leppert M, Slattery ML. Relationship of Ki-ras mutations in colon cancers to tumor location, stage, and survival: a population-based study. Cancer Epidemiol Biomarkers Prev. 2000;9:1193–7. [PubMed] [Google Scholar]
  • 19.Slattery ML, Curtin K, Sweeney C, Levin TR, Potter J, Wolff RK, Albertsen H, Samowitz WS. Diet and lifestyle factor associations with CpG island methylator phenotype and BRAF mutations in colon cancer. Int J Cancer. 2007;120:656–63. doi: 10.1002/ijc.22342. [DOI] [PubMed] [Google Scholar]
  • 20.Slattery ML, Curtin K, Wolff RK, Boucher KM, Sweeney C, Edwards S, Caan BJ, Samowitz W. A comparison of colon and rectal somatic DNA alterations. Dis Colon Rectum. 2009;52:1304–11. doi: 10.1007/DCR.0b013e3181a0e5df. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Conneely KN, Boehnke M. So Many Correlated Tests, So Little Time! Rapid Adjustment of P Values for Multiple Correlated Tests. Am J Hum Genet. 2007;81:1158–68. doi: 10.1086/522036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Kodach LL, Wiercinska E, de Miranda NFCC, Bleuming SA, Musler AR, Peppelenbosch MP, Dekker E, van den Brink GR, van Noesel CJM, Morreau H, Hommes DW, ten Dijke P, et al. The Bone Morphogenetic Protein Pathway Is Inactivated in the Majority of Sporadic Colorectal Cancers. Gastroenterology. 2008;134:1332–41.e3. doi: 10.1053/j.gastro.2008.02.059. [DOI] [PubMed] [Google Scholar]
  • 23.Hardwick JC, Van Den Brink GR, Bleuming SA, Ballester I, Van Den Brande JM, Keller JJ, Offerhaus GJ, Van Deventer SJ, Peppelenbosch MP. Bone morphogenetic protein 2 is expressed by, and acts upon, mature epithelial cells in the colon. Gastroenterology. 2004;126:111–21. doi: 10.1053/j.gastro.2003.10.067. [DOI] [PubMed] [Google Scholar]
  • 24.Howe JR, Bair JL, Sayed MG, Anderson ME, Mitros FA, Petersen GM, Velculescu VE, Traverso G, Vogelstein B. Germline mutations of the gene encoding bone morphogenetic protein receptor 1A in juvenile polyposis. Nat Genet. 2001;28:184–7. doi: 10.1038/88919. [DOI] [PubMed] [Google Scholar]
  • 25.Kodach LL, Bleuming SA, Musler AR, Peppelenbosch MP, Hommes DW, van den Brink GR, van Noesel CJ, Offerhaus GJ, Hardwick JC. The bone morphogenetic protein pathway is active in human colon adenomas and inactivated in colorectal cancer. Cancer. 2008;112:300–6. doi: 10.1002/cncr.23160. [DOI] [PubMed] [Google Scholar]
  • 26.Slattery MLHJ, Lundgreen A, Wolff RK. Genetic variation in the TGF-Beta-signaling pathway and colon and rectal cancer risk. Cancer Epidemiology Biomarkers and Prevention. 2011 doi: 10.1158/1055-9965.EPI-10-0843. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp Table S1

RESOURCES