Molecular and Cellular Pathobiology

SMAD2, SMAD3 and SMAD4Mutations in Colorectal Cancer

Nicholas I. Fleming1, Robert N. Jorissen1, Dmitri Mouradov1, Michael Christie1,4,Anuratha Sakthianandeswaren1, Michelle Palmieri1, Fiona Day1,4, Shan Li1, Cary Tsui1, Lara Lipton1,4,6,Jayesh Desai1,6, Ian T. Jones7, Stephen McLaughlin8, Robyn L. Ward9, Nicholas J. Hawkins10,Andrew R. Ruszkiewicz11, James Moore12, Hong-Jian Zhu5, John M. Mariadason2, Antony W. Burgess3,Dana Busam13, Qi Zhao14, Robert L. Strausberg14,15, Peter Gibbs1,4,6, and Oliver M. Sieber1,4

AbstractActivation of the canonical TGF-b signaling pathwayprovides growth inhibitory signals in the normal intestinal

epithelium. Colorectal cancers (CRCs) frequently harbor somatic mutations in the pathway members TGFBR2and SMAD4, but towhat extentmutations in SMAD2 or SMAD3 contribute to tumorigenesis is unclear. A cohort of744 primary CRCs and 36 CRC cell lines were sequenced for SMAD4, SMAD2, and SMAD3 and analyzed for allelicloss by single-nucleotide polymorphism (SNP)microarray analysis.Mutation spectrawere compared between thegenes, the pathogenicity of mutations was assessed, and relationships with clinicopathologic features wereexamined. The prevalence of SMAD4, SMAD2, and SMAD3mutations in sporadic CRCs was 8.6% (64 of 744), 3.4%(25 of 744), and 4.3% (32 of 744), respectively. A significant overrepresentation of two genetic hits was detected forSMAD4 and SMAD3, consistent with these genes acting as tumor suppressors. SMAD4mutations were associatedwith mucinous histology. Themutation spectra of SMAD2 and SMAD3were highly similar to that of SMAD4, bothin mutation type and location within the encoded proteins. In silico analyses suggested the majority of themutations were pathogenic, with most missense changes predicted to reduce protein stability or hinder SMADcomplex formation. The latter altered interface residues or disrupted the phosphorylation-regulated Ser-Ser-X-Ser motifs within SMAD2 and SMAD3. Functional analyses of selected mutations showed reductions in SMAD3transcriptional activity and SMAD2–SMAD4 complex formation. Joint biallelic hits in SMAD2 and SMAD3 wereoverrepresented andmutually exclusive to SMAD4mutation, underlining the critical roles of these three proteinswithin the TGF-b signaling pathway. Cancer Res; 73(2); 1–11. �2012 AACR.

IntroductionThe TGF-b family of cytokines are growth inhibitors of

normal epithelial cells, and loss of sensitivity to these factors

promotes tumorigenesis (1). Canonical TGF-b signaling ismediated by the TGF-b type I and type II receptors (TGFBR1,TGFBR2), which upon binding TGF-b ligands, phosphorylatethe receptor-activated SMADs (R-SMADs), SMAD2 andSMAD3, at conserved C-terminal Ser-Ser-X-Ser motifs (2, 3).The activated R-SMADs bind the common mediator SMAD(co-SMAD) SMAD4 (4–6) and the resulting complexes relocateinto the nucleus. There the proteins participate in transcrip-tional regulation of target genes in conjunction with a widerange of coregulator proteins (7).

Mutation of TGF-b pathway members is common in mul-tiple types of human malignancies including colorectal cancer(CRC; ref. 1). Germline mutations in SMAD4 cause juvenilepolyposis syndrome (JPS), an autosomal dominant predispo-sition to multiple gastrointestinal polyps and cancer (8). Insporadic CRCs, SMAD4 and TGFBR2 mutations are found inapproximately 10% and 15% of patients, respectively (9–12).TGFBR2mutations are particularly prevalent in microsatelliteunstable (MSI) tumors, with approximately 80% of such casesharboring frameshift mutations at a poly-adenosine tract inexon 4 (13).

The SMAD proteins are highly homologous and character-ized by 2 conserved regions, an N-terminal Mad homologydomain-1 (MH1) and a C-terminal Mad homology domain-2(MH2), joined by a linker domain (7). The MH1 domain

Authors' Affiliations: 1Ludwig Colon Cancer Initiative Laboratory, 2Onco-genic Transcription Laboratory, and 3Epithelial Biology Laboratory, LudwigInstitute for Cancer Research; 4Faculty of Medicine, Dentistry and HealthSciences, 5Department of Surgery, University of Melbourne; Departmentsof 6Medical Oncology and 7Colorectal Surgery, Royal Melbourne Hospital,Parkville; 8Department of Colorectal Surgery,Western Hospital, Footscray,Victoria; 9Lowy Cancer Research Centre, Prince of Wales Clinical School,10School of Medical Sciences, University of New South Wales, Sydney,New South Wales; 11Pathology Department, Institute of Medical andVeterinary Science; 12Department of Colorectal Surgery, Royal AdelaideHospital, Adelaide, South Australia, Australia; 13J. Craig Venter Institute,Rockville; 14Ludwig Collaborative Laboratory for Cancer Biology andTherapy, Department of Neurosurgery, Johns Hopkins University Schoolof Medicine, Baltimore, Maryland; and 15Ludwig Institute for CancerResearch Ltd., New York, New York

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

N.I. Fleming and R.N. Jorissen contributed equally to this work.

Corresponding Author: Oliver M. Sieber, Ludwig Colon Cancer InitiativeLaboratory, Ludwig Institute for Cancer Research, PO Box 2008, RoyalMelbourne Hospital, Parkville, VIC 3050, Australia. Phone: 61-3-93413168;Fax: 61-3-93413104; E-mail: Oliver.Sieber@ludwig.edu.au

doi: 10.1158/0008-5472.CAN-12-2706

�2012 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org OF1

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

facilitates nuclear import, DNA and transcriptional coregu-lator binding, and negatively regulates the MH2 domain (14,15). The MH2 domain is involved in SMAD protein homo-and hetero-oligomerization (4–6), cytoplasmic anchoring(16), and transcription (reviewed in ref. 7). When inactive,the SMAD4 and R-SMAD proteins reside in the cytoplasmand when activated by TGF-b signaling, it is proposed thatthe proteins form heterotrimeric transcriptional complexescontaining 1 SMAD4 and 2 R-SMAD proteins (4–6). Cancer-associated mutations in SMAD4 show a characteristic dis-tribution with respect to its domain structure. The majorityof changes cluster in the MH2 domain and often alterresidues that are close to the protein interface mediatingSMAD4 hetero-oligomerization with the R-SMADs (11, 12).Mutations in the MH1 domain have been reported to alterprotein stability, alter DNA binding, prevent nuclear trans-location, and enhance interactions with the MH2 domain(14, 15, 17).

Despite the central role of SMAD2 and SMAD3 as directmediators of TGF-b signaling and binding partners for SMAD4,evidence for mutations affecting these proteins in CRCs islimited. The mutation prevalence has been broadly estimatedat 2% to 6%, but with the exception of a recent study from TheCancer Genome Atlas (TCGA) program (18), the samplecohorts used to determine this have been small with only veryfew identified mutations (19, 20). Importantly, a series ofstudies usingmousemodels of intestinal cancer have indicatedtumor suppressor roles for Smad2 and Smad3 (21–23). Itremains unclear therefore to what extent SMAD2 and SMAD3mutations are alternative functional genetic hits in the TGF-bpathway and consequently contribute to the development ofCRCs.

In this study, we determined the prevalence, spectra, andLOH status of SMAD4, SMAD2, and SMAD3 somatic mutationsin 744 sporadic and 36 CRC cell lines. The distribution andnature of the SMAD2 and SMAD3 mutations were comparedwith those of SMAD4. The pathogenicity of missense muta-tions was estimated in silico using homology and crystalstructure information, and for selected cases, was tested invitro using assays for transcriptional function and heteromericcomplex formation. Relationships with patient clinical char-acteristics were examined. Our results reveal that themutationspectra of SMAD2 and SMAD3 are highly similar to that ofSMAD4 and suggest that joint inactivation of SMAD2 andSMAD3 constitutes a novel mode of TGF-b pathway inactiva-tion in CRCs.

Materials and MethodsPatients

Fresh-frozen tumor and normal specimens were analyzedfrom 744 patients with sporadic CRCs treated at St Vincent'sHospital, Sydney (134 patients), the Royal Melbourne andWestern Hospitals, Melbourne (369 patients), and the RoyalAdelaide Hospital, Adelaide (241 patients), Australia. Informedconsent was given by all participants according to local ethicsregulations. Information on patient clinical characteristics wasretrieved from hospital databases. The primary cancers com-

prised 70 stage I, 228 stage II, 347 stage III, and 99 stage IV cases.Three hundred and thirteen cancers were from the proximalcolon, 241 from the distal colon, and 189 from the rectum;tumor location data were unavailable for one patient. Themedian age at cancer diagnosis was 69.2 years and ranged from25 to 99 years; 414 patients were male and 330 were female(Table 1).

CRC cell linesA total of 36 colon cancer cell lines were studied: CACO2,

COLO201, COLO320, COLO741, DLD1, Gp5d, HCA7, HCC2998,HCT116, HDC135, HDC143, HDC57, HDCC114, HT29, HT55,KM12, LIM1215, LIM1863, LIM1899, LIM2099, LIM2405,LIM2550, LIM2551, LOVO, LS174T, RKO, RW2982, SKCO1,SW1116, SW1222, SW403, SW48, SW480, SW837, SW948, andT84. The cell lines were authenticated in 2010 by short tandemrepeat (STR) analysis. Cells were cultured with Dulbecco'smodified Eagle's medium (DMEM) and 10% FBS at 37�C and5% CO2. Literature references for the lines are provided inSupplementary Table S1.

Mutation detectionFor cancer samples, hematoxylin and eosin–stained sec-

tions were reviewed by an anatomical pathologist and macro-dissected in areas comprising greater than 60% neoplasticcells. Genomic DNA was extracted using the AllPrep DNA/RNA Mini Kit (Qiagen). Coding regions and exon–intronboundaries of the entire SMAD4, SMAD2, and SMAD3 geneswere amplified using the PCR, and direct DNA sequencingwas conducted using BigDye Terminator v3.1 Ready ReactionMix (Applied Biosystems). Primer sequences are availablefrom the authors. Reaction products were run on 3730xl DNAAnalyzers (Applied Biosystems) using Biomek FX robots(Beckman-Coulter) and Pixsys 4200 nanoliter liquid handlingsystems (Cartesian Technologies). Mutations detected wereconfirmed as somatic by bidirectional resequencing of newPCR products from tumor and matched normal DNA or cellline DNA.

LOH and DNA copy number analysisLOH and DNA copy number status at the SMAD4, SMAD2,

and SMAD3 loci was determined from single-nucleotide poly-morphism (SNP) analyses (Human610-Quad BeadChip arrays,Illumina) for cell lines, tumor, and matched normal DNAsamples using OncoSNP software as described previously(24). SNP call rates for normal samples were greater than98% (median, 99.8%; range, 98.3%–99.9%); the median call ratefor tumor and cell line samples was 97.3% (range: 83.7%–99.8%)and 95.1% (range: 87.9%–99.8%), respectively. Correct pairingof tumor and normal samples was verified from SNP genotypematching. For the cell lines, where normal reference DNA wasunavailable, LOH was assigned as present for regions wherecontiguous SNP homozygosity extended for more than 2 Mb.For assignment of CIN status, copy number for individualautosomes was estimated by calculating the mode of absoluteDNA copy number states across SNPs. Samples for whichmorethan 3 autosomes showed deviations from 2nwere classified asCIN positive.

Fleming et al.

Cancer Res; 73(2) January 15, 2013 Cancer ResearchOF2

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

Expression constructs and site-directed mutagenesisExpression constructs for N-terminal HA-tagged SMAD2,

untagged SMAD3 and N-terminal FLAG-tagged SMAD4 weregenerated in pcDNA3.1þ (Invitrogen) by PCR subcloningfrom the Mammalian Genome Collection clones MGC34440,MGC60396, and MGC8602, respectively (Dana-Farber/Har-vard Cancer Center DNA Resource Core, Boston, MA). C-

terminal HA-tagged TGFBR1 was generated from a previ-ously published FLAG-tagged version (25). Site-directedmutagenesis was conducted using the QuikChange II XL kit(Agilent Technologies), and the following missense muta-tions were produced: S276L, D300N, P305L, T413N, D450E,and S647P for SMAD2; D258N, R268C, P336S, R373Q, D408Y,and S425C for SMAD3; and T204D for HA-TGFBR1. The

Table 1. Characteristics of 744 CRC patients and associations with SMAD4, SMAD2, or SMAD3mutationstatus

SMAD4 SMAD2 SMAD3

Characteristic All patients Wild-type (%) Mutant (%) P Wild-type (%) Mutant (%) P Wild-type (%) Mutant (%) P

Total N 744 680 (91.4) 64 (8.6) 719 (96.6) 25 (3.4) 712 (95.7) 32 (4.3)Age, yMean � SD 69.1 � 11.3 69.3 � 11.5 67.5 � 9.3 0.113 69.1 � 11.4 68.5 � 9.5 0.719 69.2 � 11.3 67.6 � 11.1 0.377Range 25.0–99.2 25.0–99.2 45.7–85.9 25.0–99.2 48.7–83.8 25.0–99.2 46.6–87.1

GenderMale 414 384 (92.8) 30 (7.2) 0.149 400 (96.6) 14 (3.4) 1 402 (97.1) 12 (2.9) 0.045a

Female 330 296 (89.7) 34 (10.3) 319 (96.7) 11 (3.3) 310 (93.9) 20 (6.1)Tumor locationRight colon 313 282 (90.1) 31 (9.9) 0.274 300 (95.8) 13 (4.2) 0.376 295 (94.2) 18 (5.8) 0.172Left colon 241 219 (90.9) 22 (9.1) 236 (97.9) 5 (2.1) 235 (97.5) 6 (2.5)Rectum 189 178 (94.2) 11 (5.8) 182 (96.3) 7 (3.7) 181 (95.8) 8 (4.2)Not available 1 1 0 1 0 1 0

AJCC stageI 70 66 (94.3) 4 (5.7) 0.384 68 (97.1) 2 (2.9) 0.373 69 (98.6) 1 (1.4) 0.360II 228 206 (90.4) 22 (9.6) 218 (95.6) 10 (4.4) 218 (95.6) 10 (4.4)III 347 321 (92.5) 26 (7.5) 339 (97.7) 8 (2.3) 328 (94.5) 19 (5.5)IV 99 87 (87.9) 12 (12.1) 94 (94.9) 5 (5.1) 97 (98) 2 (2)

T stage1 21 21 (100) 0 (0) 0.388 19 (90.5) 2 (9.5) 0.303 19 (90.5) 2 (9.5) 0.2032 91 85 (93.4) 6 (6.6) 87 (95.6) 4 (4.4) 88 (96.7) 3 (3.3)3 516 471 (91.3) 45 (8.7) 500 (96.9) 16 (3.1) 497 (96.3) 19 (3.7)4 116 103 (88.8) 13 (11.2) 113 (97.4) 3 (2.6) 108 (93.1) 8 (6.9)

N stage0 326 296 (90.8) 30 (9.2) 0.732 312 (95.7) 14 (4.3) 0.379 314 (96.3) 12 (3.7) 0.6941 255 236 (92.5) 19 (7.5) 247 (96.9) 8 (3.1) 242 (94.9) 13 (5.1)2 163 148 (90.8) 15 (9.2) 160 (98.2) 3 (1.8) 156 (95.7) 7 (4.3)

Lymphovascular invasionAbsent 289 260 (90) 29 (10) 0.338 279 (96.5) 10 (3.5) 0.812 274 (94.8) 15 (5.2) 0.526Present 206 191 (92.7) 15 (7.3) 198 (96.1) 8 (3.9) 198 (96.1) 8 (3.9)Not available 249 229 20 242 7 240 9

DifferentiationWell/moderate 546 498 (91.2) 48 (8.8) 0.754 526 (96.3) 20 (3.7) 0.476 529 (96.9) 17 (3.1) 0.042a

Poor 171 158 (92.4) 13 (7.6) 167 (97.7) 4 (2.3) 159 (93) 12 (7)Not available 27 24 3 26 1 24 3

MucinousNo 578 540 (93.4) 38 (6.6) <0.001a 562 (97.2) 16 (2.8) 0.082 554 (95.8) 24 (4.2) 1Yes 157 133 (84.7) 24 (15.3) 148 (94.3) 9 (5.7) 151 (96.2) 6 (3.8)Not available 9 7 2 9 0 7 2

MSIStable 646 588 (91.0) 58 (9) 0.441 626 (96.9) 20 (3.1) 0.359 620 (96.0) 26 (4.0) 0.296Unstable 98 92 (93.9) 6 (6.1) 93 (94.9) 5 (5.1) 92 (93.9) 6 (6.1)

CINNegative 167 150 (89.8) 17 (10.2) 0.759 155 (92.8) 12 (7.2) 0.004a 156 (93.4) 11 (6.6) 0.116Positive 464 421 (90.7) 43 (9.3) 455 (98.1) 9 (1.9) 448 (96.6) 16 (3.4)Not available 113 109 4 109 4 108 5

aP < 0.05; comparisons were made with the Fisher exact and Kruskal–Wallis tests.

Co- and R-SMAD Mutations in Colorectal Cancer

www.aacrjournals.org Cancer Res; 73(2) January 15, 2013 OF3

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

primer sequences used are provided in SupplementaryTable S2.

SMAD3 transcription luciferase reporter assayExpression constructs for constitutively active TGFBR1 (ca-

TGFBR1; substitution T204D in ref. 26), either wild-type ormutated SMAD3, a luciferase reporter for SMAD3 mediatedtranscription (pCAGA12; ref. 27) and an expression constructfor Renilla luciferase to indicate transfection efficiency (28)were cotransfected into HEK293T cells. After 24 hours, lucif-erase expressionwas analyzed by use of the Dual-Glo luciferasereporter assay (Promega) on a LumiStar Galaxy luminometer(BMG Labtech).

CoimmunoprecipitationExpression constructs for ca-TGFBR1, FLAG-SMAD4, and

either wild-type or mutated HA-SMAD2 were cotransfectedinto HEK293T cells. After 24 hours, lysates were prepared withlysis buffer containing 1% Triton-X 100 (Sigma-Aldrich) andincubated with anti-FLAG M2 Affinity Gel beads (Sigma-Aldrich). The beads were washed 3 times with lysis buffer,and bound proteins were eluted by boiling in nondenaturingsample loading buffer. The proteins were separated by PAGE,transferred to nitrocellulosemembranes, and immunoblottingwas conducted using antibodies for FLAG (M2; Sigma-Aldrich)and SMAD2 (D43B4; Cell Signaling Technology).

Pathogenicity predictionFour in silico algorithms were applied to predict the path-

ogenicity of missense mutations in SMAD4, SMAD2, andSMAD3: SIFT-Blink (29), PolyPhen-2 (30), MAPP (31), and I-Mutant-3.0 (ref. 32; Supplementary Methods). The 3-dimen-sional structure information used for these predictions issummarized in Supplementary Table S3. Pathogenicity pre-dictions were conducted for putative splice site mutationsusing MaxEntScan (ref. 33; Supplementary Methods).

Statistical analysisStatistical analyses were conducted using the R statistical

computing software (34). Differences between groups wereassessed using the Fisher exact test for categorical variablesand the Kruskal–Wallis test for continuous variables. Whenmultiple continuous variables were compared with a controlvariable, ANOVA was conducted followed by Dunnett post hoccomparison analysis. All statistical analyses were 2-sided andconsidered significant when P < 0.05.

ResultsPrevalence of SMAD4, SMAD2, and SMAD3 mutations inCRC

Primary CRCs from 744 patients and 36 CRC cell lines werescreened for somatic mutations in the entire coding regionsand exon–intron boundaries of SMAD4, SMAD2, and SMAD3 bydirect DNA sequencing (Supplementary Table S4). LOH wasdetermined from SNP array data for 631 of the primary cancersand all of the cell lines. For the primary cancers, the prevalenceof somatic truncating, frameshift, missense, and splice site

mutations in SMAD4, SMAD2, and SMAD3was 8.6% (64 of 744),3.4% (25 of 744), and 4.3% (32 of 744), respectively, and thecombined prevalence was 14.8% (110 of 744). Ten furtherchanges were silent mutations that were excluded from sub-sequent analyses. LOH at SMAD4, SMAD2, and SMAD3 wasdetected in 54.0% (341 of 631), 53.4% (337 of 631), and 27.1%(171 of 631) of primary cases, respectively, with the loss atSMAD4 and SMAD2 being highly correlated because of theirclose proximity on chromosome 18q (Pearson correlationcoefficient ¼ 0.93). The CRC cell lines showed higher frequen-cies of SMAD4 and SMAD2mutations than the primary cancersof 22.2% (8 of 36) and 13.9% (5 of 36), respectively, but themutation frequency for SMAD3 (5.6%, 2 of 36) was similar. LOHat the SMAD4, SMAD2, and SMAD3 loci was present in 50.0% (18of 36), 50.0% (18 of 36), and 16.7% (6 of 36) of the cell lines,consistent with the primary cancers.

Spectrum of somatic SMAD4 mutationsFor SMAD4, the most frequent types of somatic mutation

detected in primary cancers and cell lines were missensemutation (66.7%, 52 of 78), followed by nonsense (19.2%, 15of 78), splice site (5.1%, 4 of 78), frameshift (3.8%, 3 of 78), andin-frame insertion/deletion mutations (3.8%, 3 of 78; Supple-mentary Fig. S1A). In one case, a nucleotide substitutionaltered the stop codon, presumably resulting in a proteinextension. Sixty-one SMAD4 mutations were unique and 9were recurrent changes, with 68.6% (48 of 70) to our knowledgerepresenting novel somatic mutations not previously reportedin CRCs (Supplementary Table S4). The nonsense and trun-cating mutations occurred throughout the gene, but 78.8% (41of 52) of the missensemutations clustered in theMH2 domain,which represents only 41.5% of the coding sequence (P¼ 0.023,Fisher exact test; Fig. 1A). In line with previous observations(10–12), a hotspot region formissensemutations was apparentand we could refine its definition to include the 6 codonsspanning from Asp351 to Pro356 and the nearby Arg361(Figs. 1–3). Missense changes at these residues accountedfor 43.9% (18 of 41) of the MH2 domain missense changesdetected. In addition, 2 of the 3 detected in-frame insertion/deletion mutations also mapped to these codons (D351del andP356delinsQK).

When the SMAD4 missense mutations were mapped ontothe crystal structures of SMAD4:SMAD2:SMAD2 and SMAD4:SMAD3:SMAD3MH2 domain heterotrimers (4–6), 46.3% (19 of41) of the mutations occurred within the conserved R-SMADbinding surface. Furthermore, the identified mutation hotspotregion (Asp351-Pro356, Arg361) mapped to a defined proteinloop (L1 loop) directly involved in binding the R-SMADs(ref. 4; Fig. 3). Notably, the L1 loop is conserved across the 3SMADproteins and is used in the formation of both homo- andhetero-oligomeric complexes. At the base of the loop, a saltbridge connects Asp351 to Arg361, which in turn forms afurther salt bridge with another conserved aspartate residuein the neighboring subunit of the trimeric SMAD complex (4).In SMAD4, the opposing aspartate (Asp537; that would bebound by the L1 loop of a neighboring SMADprotein) aswell asan associated leucine residue (Leu540) were also altered bymissense mutations in our cases (Fig. 3A).

Fleming et al.

Cancer Res; 73(2) January 15, 2013 Cancer ResearchOF4

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

A further 3 SMAD4 MH2 domain missense mutations(A406T, K428T, R515T) altered residues that are separate fromthe primary mutation hotspot region and are involved inbinding the C-terminal phosphorylated Ser-Ser-X-Ser motifsof SMAD2 and SMAD3 (refs. 4–6; Fig. 3B). Of the remaining 16MH2 missense mutations, 13 were predicted to affect proteinstability by in silico pathogenicity prediction and 3 weredeemed benign or ambiguous (Supplementary Table S4).A total of 19 SMAD4 mutations were outside the MH2

domain, with 14 mapping within the MH1 domain and 5occurring in the linker domain. Nine of the MH1 mutationswere missense changes, including 4 examples of A118V, 2changes altering Arg100 (R100G, R100T), and the singlechanges K45N, G65R, and L98F. All of these except K45N werepredicted to alter protein stability by our in silico analyses(Supplementary Table S4). K45N and L98F occurred near thebound DNA interface (Fig. 2), but the pathogenicity of K45Nmay be due to disruption of a conserved nuclear localization

signal (NLS; ref. 35) rather than an alteration of DNA binding.Hence, we found limited evidence to suggest that MH1 domainmutationswere selected to target the DNA-binding interface ofSMAD4. Finally, the 2missense mutations in the SMAD4 linkerdomain, L172M and T197I, were predicted to be benign in ouranalyses. However, phosphorylation of the SMAD4 linkerdomain has been implicated in the regulation of nuclearimport/export (36) and it is possible that T197I removes aregulatory phosphorylation site.

Spectra of somatic SMAD2 and SMAD3 mutationsThe somatic mutation spectra of SMAD2 and SMAD3 were

highly similar to the mutation spectrum of SMAD4 both inmutation type and location. The most common changes weremissense mutations (23 of 32, 71.9% and 23 of 36, 63.9%,respectively), then nonsense mutations (6 of 32, 18.8% and 5of 36, 13.9%) and then other mutation types (3 of 32, 9.4% and 8of 36, 22.2%; Supplementary Fig. S1B and S1C). As observed inSMAD4, the majority of the mutations in both R-SMADsmapped to the protein MH2 domains and were overrepresent-ed relative to the number of amino acids present in thedifferent regions of the proteins (20 of 32, 62.5%, P ¼ 0.026and 24 of 36, 66.7%, P ¼ 0.038, respectively, Fisher exacttest; Fig. 1B and C). Overall for SMAD2 and SMAD3, 28 and34 mutations were unique and 3 and 2 were recurrent changes,with 77.4% (24 of 31) and 88.9% (32 of 36) to our knowledgerepresenting novel somatic mutations not previously reportedin CRCs (Supplementary Table S4).

Within the MH2 domains of SMAD2 and SMAD3, thedistribution of missense mutations revealed further similari-ties to SMAD4. 58.5% (10 of 17) of SMAD2 and 33.3% (5 of 15) ofSMAD3 MH2 domain missense changes clustered at residuesdirectly involved in oligomerization or at amino acids sup-porting those residues (Figs. 1–3). Notably, 35.0% (7 of 20) ofSMAD2 and 12.5% (3 of 24) of SMAD3 mutations occurred atresidues homologous to those of the principal mutation hot-spot region in SMAD4 (Asp351-Pro356, Arg361; Figs. 1–3A).The Ser-Ser-X-Ser motif of SMAD2 was directly altered by 2missense mutations (S464L, S467P), a recurring nonsensemutation (S464X, 2 cases), and probably by a detected stopcodon change predicted to extend the protein as seen forSMAD4 (Figs. 1 and 3B). For SMAD3, the Ser-Ser-X-Ser motifwas compromised by a missense mutation (S425C), an in-frame insertion (S422_S423insRN), and a frameshift mutation(S423VfsX13; Figs. 1 and 3B). As also seen for SMAD4, themajority of the remaining missense mutations in the MH2domains of SMAD2 and SMAD3were buried within the proteinstructure and predicted to affect protein stability by in silicoanalysis with 4 notable exceptions: The mutation A354T inSMAD2 was homologous to the A406T change identified inSMAD4 and is therefore likely to compromise SMAD oligo-merization in an analogous manner. The change S276L inSMAD2, which occurred in 3 cases, is partially surface exposedand was previously identified in an N-ethyl-N-nitrosoureamutagenesis screen where it caused a hypomorphic develop-mental phenotype in mice by an unknown mechanism (37).The wild-type residue of E228K in SMAD3 is solvent-exposed,but of unknown function. Finally, Ser266 in SMAD3, altered by

Figure 1. Somatic mutation spectra for SMAD4 (A), SMAD2 (B), andSMAD3 (C) for 744 primary CRCs and 36 CRC cell lines. Blue line,missense mutation; red line, nonsense mutation; blue triangle, in-frameinsertion/deletion; red triangle, out-of-frame insertion/deletion; bluesquare, combined deletion and insertion; black triangle, splice sitemutation (arrow direction indicates whether it occurs before or after theexon); black star, start/stop site mutation. Height units indicate count ofmutations.

Co- and R-SMAD Mutations in Colorectal Cancer

www.aacrjournals.org Cancer Res; 73(2) January 15, 2013 OF5

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

S266L, has been previously implicated in binding of the tran-scriptional coregulators Ski and Sno (38).

Within the MH1 domain of SMAD2, 2 of the 3 detectedmissense mutations were predicted to affect protein stabilityby in silico analysis and the third, A22V, was deemed benign.Of the 6 SMAD3 MH1 domain missense mutations, only one,K43R, had a solvent exposed wild-type residue and it waspredicted to disrupt the conserved NLS (35) similar to theK45N change detected in SMAD4. Of the 5 remainingchanges, 4 were predicted to affect protein stability, includ-ing R93N that altered the homologous residue to Arg100 inSMAD4. There was no evidence that MH1 mutations inSMAD2 and SMAD3 had altered DNA binding. Finally, forthe linker domains, all 5 observed missense changes in theproteins were predicted to be benign, again mirroring theresults for SMAD4.

SMAD3 and SMAD4 tend to acquire two genetic hitsTo determine whether the SMAD4, SMAD2 and SMAD3 genes

had a tendency to acquire 2 genetic hits in line with theirproposed tumor suppressor roles, mutation data were integrat-edwith LOHdata for the primary tumors. As SMAD2 is localizedwithin approximately 3.1 Mb of SMAD4 on chromosome 18q,analyses for SMAD2 were restricted to SMAD4 wild-type casesand analyses for SMAD4 were restricted to SMAD2 wild-typecases. For SMAD4, there was a significant overrepresentation ofcases with 2 genetic hits in the form of either 2 mutations or 1

mutation plus LOH (P < 0.001, Fisher exact test; Table 2). Asimilar overrepresentationwas observed for SMAD3 (P¼ 0.001),but in the case of SMAD2, there was no evidence for an excess of

Figure 2. Distribution of missense mutations within the MH1 and MH2 domains of SMAD4, SMAD2, and SMAD3. Structures for MH1 domains are shown forcomplexeswithDNA. Structures forMH2domains are shown for heterotrimeric complexes of SMAD4with theR-SMADs.Orange,mutation at interface; cyan,mutation adjacent to interface; red, mutation at serine in R-SMAD Ser-Ser-X-Ser motif; blue, amino acid deletion; green, other mutation (mainly internal); andpink, zinc ion. The SMAD2 MH1 domain structure is based on the SMAD3 MH1 domain structure and represents the DNA binding spliceoform.

Table 2. Relationship between somaticmutations and allelic loss status at the SMAD4,SMAD2, and SMAD3 loci in 631 primary CRCs

GeneNumber ofmutations No loss LOH P

SMAD4a 0 264 (47.9) 287 (52.1) <0.001b

1 11 (20.4) 43 (79.6)2 4 (80.0) 1 (20.0)

SMAD2a 0 265 (48.1) 286 (51.9) 0.9041 9 (47.4) 10 (52.6)2 1 (100.0) 0 (0.0)

SMAD3 0 448 (74.2) 156 (25.8) <0.001b

1 11 (42.3) 15 (57.7)3 1 (100.0) 0 (0.0)

aGiven the close proximity of SMAD2 and SMAD4 on chro-mosome 18q, analyses for SMAD2 were restricted toSMAD4 wild-type cases and analyses for SMAD4 wererestricted to SMAD2 wild-type cases.bP< 0.05; comparisonsweremadewith the Fisher exact test.

Fleming et al.

Cancer Res; 73(2) January 15, 2013 Cancer ResearchOF6

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

caseswith 2 hits (P¼ 0.904). These relationships remainedwhenthe cell line datawere included (Supplementary Table S5). Therewas no association between presence of SMAD4, SMAD2, orSMAD3mutation and locus-specific DNA copy number (P > 0.05for all comparisons, Supplementary Table S6).

Relationships between SMAD4, SMAD2, and SMAD3 genemutation statusOverall, 1.5% (12 of 780) of cases showed mutations in more

than one of the Co- or R-SMAD genes: 1 case had jointmutations in SMAD4 and SMAD2, 4 cases showed joint muta-tion of SMAD4 and SMAD3, and 7 cases, including 6 primaryCRCs, had mutations in SMAD2 and SMAD3. Within theprimary CRC cohort, cooccurrence of SMAD4 and R-SMADgene mutations was consistent with random association (P ¼0.794, Fisher exact test), but combined SMAD2 and SMAD3mutationwas significantly overrepresented (P< 0.001; Table 3).These relationships remained when cell line data were includ-ed (Supplementary Table S7). Strikingly, the 6 primary CRCswithmutations in both R-SMADgenes had biallelic genetic hitsfor both, and furthermore, all the somatic mutations involvedmapped to the MH2 domains and were either nonsense orframeshift changes or missense mutations clearly predicted tobe pathogenic (Supplementary Table S8). Finally, in all 6 cases,SMAD4 was wild-type, suggesting that joint inactivation of theR-SMAD genes represents an alternative to SMAD4 bialleliccompromise in CRCs.

Pathogenicity validation for novel SMAD2 and SMAD3mutationsTo test the predicted pathogenicity of the detected mis-

sense mutations in SMAD2 and SMAD3, selected changes

were introduced experimentally into the correspondingproteins by site-directed mutagenesis. For SMAD2, theseincluded D300N, P305L, and D450E at the oligomerization

Figure 3. Mutation hotspot regionsin SMAD4 compared withhomologous regions in SMAD2 andSMAD3. A, SMAD4 hotspot regionspanning Asp351-Pro356 andincluding Arg361, which forms a keyinteraction interface with SMAD2 orSMAD3. B, SMAD4 hotspot regioncomprising residues Ala406, Arg515,and Lys428 mediating interactionwith the C-terminal Ser-Ser-X-Sermotif of SMAD2 or SMAD3. Purple,SMAD2 or SMAD3; yellow, SMAD4.Frequencies of reoccurring changesindicated in parentheses.

Table 3. Relationships between somaticmutations in SMAD4, SMAD2, and SMAD3 for744 primary CRCs

SMAD4

Wild-type Mutant P

SMAD2 Wild-type 656 (91.2) 63 (8.8) 0.715Mutant 24 (96) 1 (4)

SMAD4

Wild-type Mutant PSMAD3 Wild-type 652 (91.6) 60 (8.4) 0.346

Mutant 28 (87.5) 4 (12.5)

SMAD4

Wild-type Mutant PSMAD2 and/orSMAD3

Wild-type 634 (91.5) 59 (8.5) 0.794

Mutant 46 (90.2) 5 (9.8)

SMAD3

Wild-type Mutant PSMAD2 Wild-type 693 (96.4) 26 (3.6) <0.001a

Mutant 19 (76.0) 6 (24.0)

aP < 0.05; comparisonsweremadewith the Fisher exact test.

Co- and R-SMAD Mutations in Colorectal Cancer

www.aacrjournals.org Cancer Res; 73(2) January 15, 2013 OF7

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

interface (Fig. 3A), S467P within the Ser-Ser-X-Ser motif (Fig.3B), the partially exposed S276L reported to cause a hypo-morphic development phenotype in mice (37), and T413Npredicted to alter protein stability. For SMAD3, we selectedD258N, R268C, and D408Y at the oligomerization interface(Fig. 3A), S425C targeting the Ser-Ser-X-Ser motif (Fig. 3B),P336S that alters a proline facilitating protein backbonebending, and the partially exposed R373Q that was detectedtwice in this study and also previously in a CRC cell line,SNU-769A (39).

SMAD3 directly binds DNA and regulates gene transcrip-tion (27). To test whether the observed mutations in SMAD3altered this activity, expression constructs for the mutant

SMAD3 proteins or the wild-type protein were analyzed in awell-established model of TGF-b signaling where HEK293Tcells are activated by a constitutively active version ofTGFBR1 (T204D; ca-TGFBR1; ref. 26). SMAD3 transcription-al activity was monitored using a specific luciferase reporterpCAGA12 based on the promoter of plasminogen activatorinhibitor-type 1 (PAI1/SERPINE1; ref. 27). Transfection withwild-type SMAD3 or ca-TGFBR1 alone increased the lumi-nescent signal 4-fold compared with cells transfected withvector control. Cotransfection with wild-type SMAD3 andca-TGFBR1 increased the signal to 10-fold (Fig. 4A). Whenthe various mutant SMAD3 constructs were cotransfectedwith ca-TGFBR1, the reporter signals were significantly

Figure 4. Functional impact of novelsomatic mutations in SMAD2 andSMAD3. A, SMAD-binding motifmediated transcription by wild-type and mutant SMAD3 proteins.Luciferase assays were conductedin HEK293T cells 24 hours aftertransfection with 5 ng of thepCAGA12 firefly luciferase reporter(27) and 90 ng of the indicatedSMAD3 and ca-TGFBR1expression constructs. About 5 ngof a Renilla luciferase expressionconstruct was included to indicatetransfection efficiency. The 2luciferase signals were quantitatedand mean signal ratios of 3experiments were generated. Errorbars, SEM. Indications ofsignificance, �, P < 0.001, weretested by ANOVA followed byDunnett post hoc analysis andrelate to comparisons with thewild-type SMAD3/ca-TGFBR1cells (bar 5). B, co-IP of wild-typeand mutant SMAD2 proteins withFLAG-SMAD4. Representative of 3experiments. Indications ofsignificance, �, P < 0.01, weretested by ANOVA followed byDunnett post hoc analysis andrelate to comparisons with theFLAG-SMAD4/wild-type SMAD2/ca-TGFBR1 pulldown (lane 4).

Fleming et al.

Cancer Res; 73(2) January 15, 2013 Cancer ResearchOF8

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

lower than the wild-type SMAD3/ca-TGFBR1 signal (P <0.001 for all comparisons).Full-length SMAD2 is not able to directly bind DNA due to

inclusion of an alternatively spliced exon not present inthe highly homologous SMAD3 sequence (40), but it still actsas a transcriptional activator through complex formationwith SMAD4 (5). As SMAD2 reporter assays remain poorlydefined, we tested selected mutations for altered complexformation with SMAD4 in coimmunoprecipitation experi-ments. HEK293T cells were transfected with constructs encod-ing N-terminal FLAG-tagged SMAD4, wild-type or mutantSMAD2, and ca-TGFBR1. Lower amounts of the mutantSMAD2 proteins were pulled down by FLAG-SMAD4 thanwild-type SMAD2 (P < 0.01, for all comparisons), suggestingthat the selected mutations all reduce the ability of SMAD2 toform a complex with SMAD4 (Fig. 4B, Supplementary Fig. S2).

Relationship between SMAD4, SMAD2, and SMAD3mutations and patient clinical characteristicsMutations in SMAD4, SMAD2, and SMAD3were analyzed for

associationwith clinicopathologic features (Table 1), includingage at diagnosis, gender, tumor location, American JointCommittee on Cancer (AJCC) stage, tumor (T) stage, node(N) stage, lymphovascular invasion, differentiation, mucinoushistology, MSI, and CIN status. SMAD4 mutations were asso-ciated with mucinous histology (P < 0.001, Fisher exact test),but no further correlations were apparent. Mutation of SMAD2was associated with CIN negative status (P ¼ 0.004), whereasmutation of SMAD3 was associated with female gender (P ¼0.045) andpoor differentiation (P¼ 0.042).When the combinedoccurrence of mutation in the 3 genes was evaluated, therelationships with CIN negative status (P¼ 0.047) and mucin-ous histology (P ¼ 0.005) remained significant, but no furtherrelationships were detected.

DiscussionThis study presents the first substantial characterization of

somaticmutations in SMAD2 and SMAD3 in sporadic CRCs andis the most comprehensive survey to date of the somaticmutation spectrum of SMAD4. Our data show that somaticmutations in SMAD2 and SMAD3 make an important contri-bution to themutational load in the TGF-b pathway, occurringin 3.4% and 4.3% of primary cancers, respectively. The R-SMADgenes have similar mutation spectra to that of SMAD4, withpathogenic mutations acting through analogous mechanisms.Furthermore, joint biallelic inactivation of the R-SMAD genesappears to represent an alternative mechanism to SMAD4inactivation for TGF-b pathway compromise.The prevalence and spectrum of detected SMAD4mutations

in primary cancers and cell lines was consistent with previoussmaller studies (10–12) and the recent TCGA study (ref. 18;Supplementary Fig. S3A). The majority of the detected changeswere missense mutations and these tended to cluster at MH2residues involved in the SMAD complex interface. Here, wedefined a mutation hotspot region (Asp351-Pro356, Arg361)mapping to the L1 loop of SMAD4 and a second group oftargeted residues (Ala406, Arg515, and Lys428) involved inbinding the Ser-Ser-X-Ser motifs of the R-SMAD proteins.

Importantly, functional studies of the L1 residues D351 andR361, and of Arg515 within the second set, have previouslyshown that these amino acids are important for complexformation (5, 15, 41). Within the MH1 domain of SMAD4, therewas no evidence for clustering of missense mutations along theDNA-binding interface, and instead most changes were pre-dicted to affect protein stability. For the changes at Gly65 andArg100, this is supported byprevious functional studies showingreduced stability for proteins containing relevant site-directedchanges (15, 17, 42). Residue A118 showed recurrent mutation(A118V) and pathogenicity of this alteration is probable as it hasbeen detected in pancreatic cancer (12, 43). The main conse-quence of SMAD4mutations inCRCs therefore appears to be theprevention of SMAD4 homo- and/or hetero-oligomerizationwith the R-SMAD proteins or the reduction of SMAD4 stability.

Previous studies have only identified a small number ofSMAD2 and SMAD3 mutations in primary CRCs, leaving openthe question as to whether these genes contribute to colorectaltumorigenesis (19, 20). Our identification of 32 SMAD2 and 36SMAD3mutations reveal the somaticmutation spectra for thesegenes in CRCs. Both spectra showed a high similarity to themutation spectrum of SMAD4, strongly suggesting that thedetected changes were similarly pathogenic. Similar trendsare apparent in the recent TCGA study (ref. 18; Supplemen-tary Fig. S3B and S3C). The majority of SMAD2 and SMAD3changes were missense mutations clustering in the MH2domain and, in particular, occurred at residues involved inSMAD complex formation. Mutations were observed atresidues homologous to the mutation hotspot region in theL1 loop of SMAD4 and within the Ser-Ser-X-Ser motifscritical for R-SMAD activation and subsequent binding toSMAD4. Functional analyses of selected interface, buried,and Ser-Ser-X-Ser motif mutations in SMAD2 and SMAD3confirmed their pathogenicity, with SMAD3 changes reduc-ing activity in luciferase reporter assays and SMAD2 changesreducing the amount of the SMAD2 protein pulled downwith SMAD4 in coimmunoprecipitation. It is likely that theselosses of function reflect a combination of effects on proteinstability, complex formation, and/or DNA binding, which arelikely to be interdependent. Our results are in agreementwith previous functional studies that showed that changes atD258 and D408 in SMAD3, and alterations within the Ser-Ser-X-Ser motifs of both proteins, reduced complex forma-tion with SMAD4 (2, 3, 5).

For the MH1 domains of SMAD2 and SMAD3, the missensemutations identified were primarily predicted to affect proteinstability rather than alter protein–protein interactions, andsimilarly to the situation in SMAD4, there was no evidence tosuggest targeting of DNA binding function. For the linkerdomains, all 5 relevant missense changes in SMAD2 andSMAD3 were predicted to be benign mirroring the results forSMAD4. A number of reports have implicated inappropriatephosphorylation of serine/threonine residues within the R-SMAD linker domains in the pathology of CRCs (44, 45), but wefound only one change (T184A) in the linker domain of SMAD2that could represent a relevant change. Taken together, ourdata suggest that SMAD4, SMAD2, and SMAD3 are mutated ina similar manner with homologous MH1 and MH2 domain

Co- and R-SMAD Mutations in Colorectal Cancer

www.aacrjournals.org Cancer Res; 73(2) January 15, 2013 OF9

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

changes acting through analogous mechanisms to prevent theformation of activated complexes.

Somatic SMAD4 mutations have been reported to be morecommon in advanced stages of CRCs (11, 46), and LOH at theSMAD4 locus has been associated with poor prognosis (10). Inour cohort of primary CRCs, presence of SMAD4 mutationshowed no relationship to AJCC stage, T stage, N stage, orlymphovascular invasion, and the only significant associationidentified was with tumor mucinous histology. Mucinous his-tology has been suggested to be a poor prognostic factor in somestudies (47, 48) and it may be in this context that SMAD4 LOH isrelevant to prognosis (10). Consistent with our findings, chro-mosome 18q LOH has also been previously associated withmucinous histology in CRCs (49). The power to detect clinico-pathologic associations for SMAD2 and SMAD3was limited, butSMAD2 mutations may be more prevalent in CIN negativetumors and SMAD3mutations may be more frequent in tumorsfrom females and or those with poorer differentiation.

A significant overrepresentation of 2 genetic hits wasobserved for SMAD4 and SMAD3, consistent with these genesacting as tumor suppressors. This tendency was not detectedfor SMAD2, although interpretation of LOH at this locus isconfounded by its close proximity to SMAD4. It is also remainspossible that missense mutations within the SMAD proteinsconfer dominant-negative effects in the context of SMADcomplexes (50).

Despite the unclear overall relationship between SMAD2mutation and LOH, we found strong evidence for selection ofbiallelic mutation in both SMAD2 and SMAD3. Markedly, all ofthe mutations in these double-mutant cases were either trun-cating mutations or MH2 missense changes of unambiguouspathogenicity, and all of these tumors were SMAD4 wild-type.Taken together, these data strongly suggest that dual biallelicmutation of the R-SMADs is an alternative pathogenic mech-anism to biallelic SMAD4 mutation. However, our results donot exclude the possibility that single-gene SMAD2 or SMAD3compromise has significant functional effects.

In conclusion, we have shown a combined prevalence ofSMAD4, SMAD2, and SMAD3 mutation of 14.8% in primarysporadic CRCs. SMAD4 mutations were the most commonalterations, with missense changes predicted to disruptcomplex formation and/or protein stability and in particulartargeting the L1 loop of the protein. Our data suggest thatSMAD2 and SMAD3 mutations are bona fide contributors tothe mutation burden in CRCs. The mutation spectra of the

R-SMAD genes mirrored that of SMAD4, with homologousMH1 and MH2 domain changes shown to act throughanalogous pathogenic mechanisms. Biallelic SMAD4 muta-tion and joint biallelic SMAD2 and SMAD3 mutation appearto be functional alternatives for TGF-b pathway inactivationin CRCs.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception anddesign:N.I. Fleming,M. Palmieri, L. Lipton, J. Desai, R.L.Ward,H.-J. Zhu, A.W. Burgess, R.L. Strausberg, P. Gibbs, O.M. SieberDevelopment ofmethodology:N.I. Fleming, A. Sakthianandeswaren, F. Day, C.Tsui, H.-J. Zhu, P. GibbsAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): N.I. Fleming, M. Christie, A. Sakthianandeswaren, M.Palmieri, F. Day, C. Tsui, J. Desai, I.T. Jones, S. McLaughlin, R.L. Ward, N.J.Hawkins, A.R. Ruszkiewicz, J. Moore, J.M. Mariadason, D. Busam, R.L. Strausberg,P. GibbsAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis):N.I. Fleming, R.N. Jorissen, D. Mouradov, M. Christie,C. Tsui, H.-J. Zhu, J.M. Mariadason, Q. Zhao, R.L. Strausberg, O.M. SieberWriting, review, and/or revision of the manuscript: N.I. Fleming, R.N.Jorissen, M. Christie, A. Sakthianandeswaren, L. Lipton, J. Desai, R.L. Ward, A.R. Ruszkiewicz, H.-J. Zhu, J.M. Mariadason, A.W. Burgess, P. Gibbs, O.M. SieberAdministrative, technical, or material support (i.e., reporting or orga-nizingdata, constructingdatabases):N.I. Fleming, A. Sakthianandeswaren, F.Day, S. Li, I.T. Jones, N.J. Hawkins, A.R. Ruszkiewicz, D. BusamStudy supervision: O.M. Sieber

AcknowledgmentsThis work was conducted under the auspices of the Hilton Ludwig Cancer

Metastasis Initiative. The authors thank the patients for participating in thisstudy, the Victorian Cancer BioBank and Biogrid Australia for the provision ofspecimens and access to the clinical data, respectively, and Prof.Manfred Schwabat the DKFZ for the provision of cell lines.

Grant SupportThe study was supported by the CSIRO Preventative Health Flagship through

an Australian Cancer Grid Project Grant (L. Lipton, P. Gibbs, O.M. Sieber), theNHMRC through a Project Grant (Application ID 489418; L. Lipton, P. Gibbs, O.M.Sieber, R.L. Ward), a Program Grant (Application ID 487922; A.W. Burgess), theHilton Ludwig Cancer Metastasis Initiative (L. Lipton, P. Gibbs, O.M. Sieber), theVictorian Government through a Victorian Cancer Agency Translation CancerResearch Grant (L. Lipton, P. Gibbs, O.M. Sieber), and the Operational Infra-structure Support Program. M. Christie and F. Day are supported by the CancerCouncil Victoria through Postgraduate Cancer Research Scholarships and L.Lipton by the CSIRO Preventative Health Flagship through a Clinical ResearcherFellowship.

The costs of publication of this article were defrayed in part by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 13, 2012; revised September 27, 2012; accepted October 17, 2012;published OnlineFirst November 8, 2012.

References1. Massague J. TGFb in cancer. Cell 2008;134:215–30.2. AbdollahS,Macias-SilvaM, Tsukazaki T,Hayashi H, AttisanoL,Wrana

JL. TbRI phosphorylation of Smad2 on Ser465 and Ser467 is requiredfor Smad2-Smad4 complex formation and signaling. J Biol Chem1997;272:27678–85.

3. Souchelnytskyi S, Tamaki K, Engstrom U, Wernstedt C, ten Dijke P,Heldin CH. Phosphorylation of Ser465 and Ser467 in the C terminus ofSmad2mediates interactionwith Smad4 and is required for transform-ing growth factor-beta signaling. J Biol Chem 1997;272:28107–15.

4. Shi Y, Hata A, Lo RS, Massague J, Pavletich NP. A structural basis formutational inactivation of the tumour suppressor Smad4. Nature1997;388:87–93.

5. Chacko BM, Qin B, Correia JJ, Lam SS, de Caestecker MP, Lin K. TheL3 loop and C-terminal phosphorylation jointly define Smad proteintrimerization. Nat Struct Biol 2001;8:248–53.

6. Chacko BM, Qin BY, Tiwari A, Shi G, Lam S, Hayward LJ, et al.Structural basis of heteromeric smad protein assembly in TGF-bsignaling. Mol Cell 2004;15:813–23.

7. Brown KA, Pietenpol JA, Moses HL. A tale of two proteins: differentialroles and regulation of Smad2 and Smad3 in TGF-b signaling. J CellBiochem 2007;101:9–33.

8. Houlston R, Bevan S, Williams A, Young J, Dunlop M, Rozen P, et al.Mutations in DPC4 (SMAD4) cause juvenile polyposis syndrome, butonly account for a minority of cases. HumMol Genet 1998;7:1907–12.

Fleming et al.

Cancer Res; 73(2) January 15, 2013 Cancer ResearchOF10

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

9. MarkowitzS,Wang J,Myeroff L, ParsonsR,Sun L, Lutterbaugh J, et al.Inactivation of the type II TGF-b receptor in colon cancer cells withmicrosatellite instability. Science 1995;268:1336–8.

10. Koyama M, Ito M, Nagai H, Emi M, Moriyama Y. Inactivation of bothalleles of the DPC4/SMAD4 gene in advanced colorectal cancers:identification of seven novel somatic mutations in tumors from Jap-anese patients. Mutat Res 1999;406:71–7.

11. Miyaki M, Iijima T, Konishi M, Sakai K, Ishii A, Yasuno M, et al. Higherfrequency of Smad4 gene mutation in human colorectal cancer withdistant metastasis. Oncogene 1999;18:3098–103.

12. Iacobuzio-Donahue CA, Song J, Parmiagiani G, Yeo CJ, Hruban RH,Kern SE. Missense mutations of MADH4: characterization of themutational hot spot and functional consequences in human tumors.Clin Cancer Res 2004;10:1597–604.

13. Iacopetta BJ, Welch J, Soong R, House AK, Zhou XP, Hamelin R.Mutation of the transforming growth factor-b type II receptor gene inright-sided colorectal cancer: relationship to clinicopathological fea-tures and genetic alterations. J Pathol 1998;184:390–5.

14. Hata A, Lo RS, Wotton D, Lagna G, Massague J. Mutations increasingautoinhibition inactivate tumour suppressors Smad2 and Smad4.Nature 1997;388:82–7.

15. Kuang C, Chen Y. Tumor-derived C-terminal mutations of Smad4 withdecreased DNA binding activity and enhanced intramolecular inter-action. Oncogene 2004;23:1021–9.

16. WuG, Chen YG, Ozdamar B, Gyuricza CA, Chong PA, Wrana JL, et al.Structural basis of Smad2 recognition by theSmad anchor for receptoractivation. Science 2000;287:92–7.

17. Moren A, Itoh S, Moustakas A, Dijke P, Heldin CH. Functional con-sequences of tumorigenic missense mutations in the amino-terminaldomain of Smad4. Oncogene 2000;19:4396–404.

18. TCGA. Comprehensive molecular characterization of human colonand rectal cancer. Nature 2012;487:330–7.

19. SjoblomT, Jones S,Wood LD, Parsons DW, Lin J, Barber TD, et al. Theconsensuscoding sequencesof humanbreast and colorectal cancers.Science 2006;314:268–74.

20. Eppert K, Scherer SW, Ozcelik H, Pirone R, Hoodless P, Kim H, et al.MADR2 maps to 18q21 and encodes a TGFb-regulated MAD-relatedprotein that is functionally mutated in colorectal carcinoma. Cell1996;86:543–52.

21. Hamamoto T, Beppu H, Okada H, Kawabata M, Kitamura T, MiyazonoK, et al. Compound disruption of smad2 accelerates malignant pro-gression of intestinal tumors in apc knockout mice. Cancer Res2002;62:5955–61.

22. Zhu Y, Richardson JA, Parada LF, Graff JM. Smad3 mutant micedevelop metastatic colorectal cancer. Cell 1998;94:703–14.

23. Sodir NM, Chen X, Park R, Nickel AE, Conti PS, Moats R, et al. Smad3deficiency promotes tumorigenesis in the distal colon of ApcMin/þ

mice. Cancer Res 2006;66:8430–8.24. Yau C, Mouradov D, Jorissen RN, Colella S, Mirza G, Steers G, et al. A

statistical approach for detecting genomic aberrations in heteroge-neous tumor samples from single nucleotide polymorphism genotyp-ing data. Genome Biol 2010;11:R92.

25. Zhu HJ, Sizeland AM. Extracellular domain of the transforming growthfactor-b receptor negatively regulates ligand-independent receptoractivation. J Biol Chem 1999;274:29220–7.

26. Wieser R, Wrana JL, Massague J. GS domain mutations that consti-tutively activate TbR-I, the downstream signaling component in theTGF-b receptor complex. EMBO J 1995;14:2199–208.

27. Dennler S, Itoh S, Vivien D, ten Dijke P, Huet S, Gauthier JM. Directbinding of Smad3 and Smad4 to critical TGFb-inducible elements inthe promoter of human plasminogen activator inhibitor-type 1 gene.EMBO J 1998;17:3091–100.

28. Lorenz WW, McCann RO, Longiaru M, Cormier MJ. Isolation andexpression of a cDNA encoding Renilla reniformis luciferase. ProcNatl Acad Sci U S A 1991;88:4438–42.

29. Ng PC, Henikoff S. Predicting deleterious amino acid substitutions.Genome Res 2001;11:863–74.

30. Adzhubei IA, Schmidt S, Peshkin L, RamenskyVE,GerasimovaA,BorkP, et al. A method and server for predicting damaging missensemutations. Nat Methods 2010;7:248–9.

31. Stone EA, Sidow A. Physicochemical constraint violation bymissensesubstitutions mediates impairment of protein function and diseaseseverity. Genome Res 2005;15:978–86.

32. Capriotti E, Fariselli P, Rossi I, Casadio R. A three-state prediction ofsingle point mutations on protein stability changes. BMC Bioinformat-ics 2008;9 Suppl 2:S6.

33. Yeo G, Burge CB. Maximum entropy modeling of short sequencemotifs with applications to RNA splicing signals. J Comput Biol2004;11:377–94.

34. Ihaka R, Gentleman R. R: a language for data analysis and graphics.J Comput Graph Statist 1996;5:299–314.

35. Xiao Z, Liu X, Henis YI, Lodish HF. A distinct nuclear localization signalin the N terminus of Smad 3 determines its ligand-induced nucleartranslocation. Proc Natl Acad Sci U S A 2000;97:7853–8.

36. Matsuura I, Denissova NG, Wang G, He D, Long J, Liu F. Cyclin-dependent kinases regulate the antiproliferative function of Smads.Nature 2004;430:226–31.

37. Vivian JL, Chen Y, Yee D, Schneider E,Magnuson T. An allelic series ofmutations in Smad2 andSmad4 identified in a genotype-based screenof N-ethyl-N- nitrosourea-mutagenized mouse embryonic stem cells.Proc Natl Acad Sci U S A 2002;99:15542–7.

38. MizuideM,Hara T, Furuya T, TakedaM, Kusanagi K, InadaY, et al. Twoshort segments of Smad3 are important for specific interaction ofSmad3 with c-Ski and SnoN. J Biol Chem 2003;278:531–6.

39. Ku JL, Park SH, Yoon KA, Shin YK, Kim KH, Choi JS, et al. Geneticalterations of the TGF-b signaling pathway in colorectal cancer celllines: a novel mutation in Smad3 associated with the inactivation ofTGF-b-induced transcriptional activation. Cancer Lett 2007;247:283–92.

40. Dennler S, Huet S, Gauthier JM. A short amino-acid sequence in MH1domain is responsible for functional differences between Smad2 andSmad3. Oncogene 1999;18:1643–8.

41. De Bosscher K, Hill CS, Nicolas FJ. Molecular and functional con-sequences of Smad4 C-terminal missense mutations in colorectaltumour cells. Biochem J 2004;379:209–16.

42. Xu J, Attisano L. Mutations in the tumor suppressors Smad2 andSmad4 inactivate transforming growth factor beta signaling by target-ing Smads to the ubiquitin-proteasome pathway. Proc Natl Acad SciU S A 2000;97:4820–5.

43. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, et al.Core signaling pathways in human pancreatic cancers revealed byglobal genomic analyses. Science 2008;321:1801–6.

44. Yamagata H, Matsuzaki K, Mori S, Yoshida K, Tahashi Y, Furukawa F,et al. Acceleration of Smad2 andSmad3 phosphorylation via c-JunNH(2)-terminal kinase during human colorectal carcinogenesis. CancerRes 2005;65:157–65.

45. Matsuzaki K, Kitano C, Murata M, Sekimoto G, Yoshida K, UemuraY, et al. Smad2 and Smad3 phosphorylated at both linker andCOOH-terminal regions transmit malignant TGF-b signal inlater stages of human colorectal cancer. Cancer Res 2009;69:5321–30.

46. Mamot C, Mild G, Reuter J, Laffer U, Metzger U, Terracciano L, et al.Infrequent mutation of the tumour-suppressor gene Smad4 in early-stage colorectal cancer. Br J Cancer 2003;88:420–3.

47. Farhat MH, Barada KA, Tawil AN, Itani DM, Hatoum HA, ShamseddineAI. Effect of mucin production on survival in colorectal cancer: a case-control study. World J Gastroenterol 2008;14:6981–5.

48. Catalano V, Loupakis F, Graziano F, Torresi U, Bisonni R, Mari D,et al. Mucinous histology predicts for poor response rate and overallsurvival of patients with colorectal cancer and treated with first-lineoxaliplatin- and/or irinotecan-based chemotherapy. Br J Cancer2009;100:881–7.

49. Ogino S, Brahmandam M, Cantor M, Namgyal C, Kawasaki T, KirknerG, et al. Distinctmolecular features of colorectal carcinomawith signetring cell component and colorectal carcinoma with mucinous com-ponent. Mod Pathol 2006;19:59–68.

50. Hoodless PA, Tsukazaki T, Nishimatsu S, Attisano L, Wrana JL,Thomsen GH. Dominant-negative Smad2 mutants inhibit activin/Vg1signaling and disrupt axis formation in Xenopus. Dev Biol 1999;207:364–79.

Co- and R-SMAD Mutations in Colorectal Cancer

www.aacrjournals.org Cancer Res; 73(2) January 15, 2013 OF11

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706

Published OnlineFirst November 8, 2012.Cancer Res   Nicholas I. Fleming, Robert N. Jorissen, Dmitri Mouradov, et al.  

Mutations in Colorectal CancerSMAD4 and SMAD3, SMAD2

  Updated version

  10.1158/0008-5472.CAN-12-2706doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2012/11/08/0008-5472.CAN-12-2706.DC1

Access the most recent supplemental material at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/early/2013/01/10/0008-5472.CAN-12-2706To request permission to re-use all or part of this article, use this link

Research. on March 27, 2021. © 2012 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 8, 2012; DOI: 10.1158/0008-5472.CAN-12-2706