Molecular Classification of Patients with Unexplained Hamartomatous and Hyperplastic Polyposis

Home , ACRV1, ACVRL1, BMPR1A, CHUK, Endoglin, Hyperplastic polyp

ORIGINAL CONTRIBUTION

Molecular Classification of Patients With Unexplained Hamartomatous and Hyperplastic Polyposis

Kevin Sweet, MS, CGC Context Significant proportions of patients with hamartomatous polyposis or with Joseph Willis, MD hyperplastic/mixed polyposis remain without specific clinical and molecular diagnosis Xiao-Ping Zhou, MD, PhD or present atypically. Assigning a syndromic diagnosis is important because it guides management, especially surveillance and prophylactic surgery. Carol Gallione, PhD Objective To systematically classify patients with unexplained hamartomatous or hy- Takeshi Sawada, MD, PhD perplastic/mixed polyposis by extensive molecular analysis in the context of central Pia Alhopuro, MD rereview of histopathology results. Sok Kean Khoo, PhD Design, Setting, and Patients Prospective, referral-based study of 49 unrelated patients from outside institutions (n=28) and at a comprehensive cancer center (n=21), Attila Patocs, MD, PhD conducted from May 2, 2002, until December 15, 2004. Germline analysis of PTEN, Cossette Martin, PhD BMPR1A, STK11 (sequence, deletion), SMAD4, and ENG (sequence), specific exon screen- Scott Bridgeman, BSc ing of BRAF, MYH, and BHD, and rereview of polyp histology results were performed. John Heinz, PhD Main Outcome Measures Molecular, clinical, and histopathological findings in patients with unexplained polyposis. Robert Pilarski, MS, CGC Results Of the 49 patients, 11 (22%) had germline mutations. Of 14 patients with Rainer Lehtonen, BSc juvenile polyposis, 2 with early-onset disease had mutations in ENG, encoding endo- Thomas W. Prior, PhD glin, previously only associated with hereditary hemorrhagic telangiectasia; 1 had hemizygous deletion encompassing PTEN and BMPR1A; and 1 had an SMAD4 mutation. Thierry Frebourg, MD, PhD One individual previously classified with Peutz-Jeghers syndrome had a PTEN dele- Bin Tean Teh, MD, PhD tion. Among 9 individuals with an unknown hamartomatous polyposis, 4 had mutations in STK11 (1), BMPR1A (2), and SMAD4 (1). Of the 23 patients with hyperplastic/ Douglas A. Marchuk, PhD mixed polyposis, 2 had PTEN mutations. Substantial discrepancies in histopathology Lauri A. Aaltonen, MD, PhD results were seen. Charis Eng, MD, PhD Conclusions Systematic molecular classification of 49 patients with unexplained hamartomatous or hyperplastic polyposis uncovered a potential novel susceptibility gene, HE INHERITED POLYPOSIS SYN- ENG, for juvenile polyposis. Importantly, given the substantial proportion of patients dromes, a heterogeneous found to have germline mutations, more extensive analysis of the known susceptibil- group of diseases, have been a ity genes is indicated. Rereview of histology results by a dedicated gastrointestinal pa- major focus of study for the thologist should be considered routinely, as organ-specific surveillance rests on defin- lastT decade and provide critical insight ing syndromic diagnosis. into the molecular pathogenesis of JAMA. 2005;294:2465-2473 www.jama.com cancer. Despite intense study, these important syndromes are still clini- lesions—hamartomas, adenomas, and develops through the adenoma- cally confusing, and proper objective hyperplastic polyps—contribute to the carcinoma sequence, initiated by ger- 1 identification is necessary for appro- various forms, although the number mline mutations of the APC gene. priate clinical management. Three pri- and distribution of polyps differ mary histopathologic subtypes of among the syndromes. The classic Author Affiliations are listed at the end of this article. paradigm is familial adenomatous pol- Corresponding Author: Charis Eng, MD, PhD, Ge- yposis (involving Ͼ100 to Ͼ1000 nomic Medicine Institute, Cleveland Clinic Founda- For editorial comment see p 2498. tion, 9500 Euclid Ave, Mailstop NE-30, Cleveland, OH adenomas), in which colorectal cancer 44195 ([email protected]).

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS

However, there is increasing evidence tion.3 Approximately 50% of all PJS tained because of distinctive mucocu- for several important categories of pol- cases are due to germline mutation in taneous lesions, benign and malignant yposis and colorectal carcinoma that the nuclear serine threonine kinase gene thyroid and breast disease, and macro- may develop from alternative routes. STK11.4 Germline mutations in the cephaly. Approximately 85% of pa- This includes the hamartomatous pol- genes of the signaling pathway of the tients with Cowden syndrome who yposis syndromes and the serrated transforming growth factor ␤ (TGF-␤) meet established diagnostic criteria have neoplasia pathway, whose morpho- superfamily (FIGURE) can lead to a mutations of the PTEN gene (Table 1). logic spectrum includes the hyperplas- range of common heritable disorders, It is widely believed that hyperplastic polyp and sessile serrated adenoma. including JPS (MIM 174900), and vas- tic polyps have no malignant poten- Given the attendant cancer risks and cular disorders, such as hereditary hem- tial. Commonly seen on colonoscopic medical management issues inherent orrhagic telangiectasia (HHT) (MIM examination (10% of patients younger in these forms of polyposis (TABLE 1), 187300) and primary pulmonary hy- than 50 years, 50% of those younger it is imperative that physicians recog- pertension (MIM 178600).5-7 For ex- than 70 years), they seldom exceed 0.5 nize that variability in histopathology ample, a combined syndrome consist- cm and are often localized to the distal and molecular etiology can hinder ing of both juvenile polyposis and colon and rectum.11,12 However, there appropriate diagnosis. hereditary HHT type 2 (MIM 175050) is evolving evidence linking the sessile The known forms of inherited ham- is known to be due to germline muta- serrated adenoma (SSA), a lesion artomatous polyposis include Peutz- tions in the SMAD4 gene.8 Mutations related to hyperplastic polyp, with Jeghers syndrome (PJS), juvenile pol- in another member of this pathway, neoplasia.13-16 SSAs tend to be multiple yposis syndrome (JPS), and Cowden BMPR1A, account for 20% to 30% of ad- and large (Ͼ1 cm), have a greater syndrome. Although collectively ac- ditional JPS cases, although appar- propensity for the proximal colon, counting for less than 1% of colorectal ently without stigmata of HHT.9 While and may comprise 15% to 20% cancer in North America,1 proper iden- other members of the TGF-␤ family of traditional hyperplastic polyps.17-20 tification of these clinically confusing would be ideal candidates for JPS, no The initiating event in the progression syndromes remains of critical impor- germline mutations in the genes from hyperplastic polyp to SSA may tance, because each syndrome carries BMPR2, ACRV1, SMAD1, SMAD2, be the activating somatic BRAF significant risks for extraintestinal ma- SMAD3, SMAD5, and SMAD7 have been (hotspot V600E) mutation.21-23 Impor- lignancy and other component fea- identified to date,10 leaving approxi- tantly, this process can be clinically tures that must be managed (Table 1). mately 50% of cases unexplained at the investigated in hyperplastic polyposis Peutz-Jeghers syndrome (Online Men- molecular level. A third hamartoma syndrome (HPS), as reappraisal of his- delian Inheritance in Man2 [MIM] syndrome, Cowden syndrome (MIM tological subtype supports the devel- 175200) causes gastrointestinal polyp- 158350), differs from both PJS and JPS opment of SSA.24-26 Interestingly, it osis, especially of the upper jejunum in that polyposis is not the defining fea- was recently shown in the serrated (78%), and mucocutaneous pigmenta- ture. Rather, most cases are ascer- adenoma mouse model that SMAD4

Table 1. Incidence and Cancer Risks for Known Polyposis Syndromes Patients Meeting Clinical Criteria Population Syndrome MIM No.* Gene(s) With Mutation, % Incidence Cancer Risks Familial adenomatous polyposis 175000 APC 90 1/5000 Colorectal, duodenal, papillary thyroid, pancreatic, hepatoblastoma, CNS tumors MYH-adenomatous polyposis 608456 MYH Unknown 1/5000 Colorectal tumors, other? Hyperplastic polyposis syndrome NA NA 1/100 000 Colorectal tumors, other? Hereditary mixed polyposis syndrome 601228 CRAC1 NA NA Colorectal tumors, other? Juvenile polyposis/hemorrhagic 175050 SMAD4 20-40 1/100 000 Colorectal, gastric, duodenal, telangiectasia syndrome pancreatic tumors Juvenile polyposis syndrome 174900 BMPR1A 20-40 1/100 000 Colorectal, gastric, duodenal, pancreatic tumors Cowden syndrome 158350 PTEN 80-85 1/200 000 Breast, thyroid, uterine, melanoma, renal cell tumors Peutz-Jeghers syndrome 175200 STK11 50 1/30 000-1/100 000 Colorectal, small intestine, stomach, breast, pancreatic, sex-cord tumors Birt-Hogg Dube syndrome 135150 BHD 50-70 1/200 000 Renal tumors, other? Abbreviations: CNS, central nervous system; MIM, Mendelian Inheritance in Man; NA, not available. *From Online Mendelian Inheritance in Man.2

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS mutation results in serrated adenomas the pathological diagnosis from stan- individual also had to have tested nega- and mixed polyposis, suggesting the dard clinical pathology review from tive for mutations in SMAD4, BMPR1A, potential for involvement of the their respective hospitals of referral. or STK11, respectively. Mean age at di- TGF-␤ pathway in this alternative Only patients in group 1 (n=17) had agnosis of JPS was 9 years; the mean route to colorectal cancer.27 previous analysis of known hamarto- number of juvenile polyps was 17. Although DNA analysis often makes matous polyposis genes. Mean age at diagnosis of PJS was 33 it possible to differentiate polyposis syn- Group 1 (TABLE 2) consisted of 17 years; the mean number of PJS polyps dromes at the gene level, a significant patients with a clinical diagnosis of was 7. percentage of patients remain without either JPS (n=14) or PJS (n=3) based Group 2 (TABLE 3) comprised 9 pa- molecular diagnosis or with atypical on established clinical criteria.28,29 Each tients with an undefined hamartoma- presentation (Table 1). We therefore sought to molecularly classify 49 pa- Figure. Germline Mutations Associated With Juvenile Polyposis Syndrome (JPS) in the tients with unexplained hamartoma- Transforming Growth Factor ␤ (TGF-␤) and Bone Morphogenetic Protein (BMP) Pathways tous polyps or with hyperplastic/ mixed polyps in the context of central Signaling Pathway rereview of all histopathology results. Components TGF-β Signaling BMP Signaling Seventeen patients (34%) met clinical Basic Signaling Pathway of TGF-β Family criteria for diagnosis of a known ham- Ligand TGF-β BMP artomatous polyposis syndrome and were sequence-negative for their re- Binds spective genes prior to study entry. For Endoglin the entire cohort, extensive compre- Accessory — hensive germline sequencing and de- (ENG Gene)∗ letion analysis of PTEN, BMPR1A, and STK11, as well as sequence analysis of TGF-β Receptor BMP Receptor SMAD4 and ENG, the latter a known Receptor Type II Type II Type II Binds cause of HHT type 1, were performed. (TGFBR2 Gene) (BMPR2 Gene) Specific exon screening was performed for BRAF, MYH, and BHD. TGF-β Receptor BMP Receptor Moreover, this is the first study to evalu- Receptor Type I Type I Type IA ate germline mutation status of known (TGFBR1 Gene) (BMPR1A Gene)∗ polyposis genes in a cohort of patients with hyperplastic/mixed polyposis to SMAD2, 3 SMAD1, 5, 8 determine which, if any, have heredi- Receptor Activates Receptor SMADs (SMAD2, 3 Genes) (SMAD1, 5, 8 Genes) tary disease. Complex METHODS Binds SMAD4 SMAD4 Patients Common SMAD (SMAD4 Gene)∗ (SMAD4 Gene)∗ This was a prospective, referral-based study in which 49 unrelated patients ∗ were recruited from May 2, 2002, un- Mutation Associated With JPS til December 15, 2004, from genetics clinics at outside institutions (n=28) SMAD Complex Transcription Transcriptional and at a comprehensive cancer center Factors (n=21). All patients provided written Complex Regulation of Genes informed consent as part of a protocol Controlling Cell Cycle reviewed and approved by the respec- DNA Binding tive human subjects protection com- mittees. To be included in the study, patients had to have a minimum of 5 NUCLEUS gastrointestinal polyps on consecu- tive colonoscopic procedures, at least TGF-␤ is one of the most potent naturally occurring inhibitors of cell growth. The TGF-␤ signaling pathway is 1 of which was either a hamartoma- closely related to the BMP pathway, with which it shares a common signaling element, SMAD4. Germline tous or hyperplastic polyp. Patients mutations of the SMAD4 and BMPR1A genes have been associated with heritable JPS. Endoglin is an accessory component of the TGF-␤ receptor complex. were classified into 3 groups based on

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS tous polyposis, due either to discrep- hyperplastic polyp (Ͼ1 cm) among yps were recorded from the accompa- ancies in the pathology report(s) or to other polyps and a mean age at diag- nying colonoscopic requisition sheet or lack of supportive clinical features to nosis of 43 years. The remainder had from the gross description on the pa- establish a diagnosis. Mean age at di- multiple hyperplastic polyps and ad- thology report. The histopathology agnosis of first polyp was 36 years. enomas, with a mean age at diagnosis slides were centrally reviewed by a Group 3 (TABLE 4) consisted of 23 of 52 years. Initial pathology reports single gastrointestinal pathologist patients with a combination of hyper- identified SSA in only 3 members of (J.W.) without knowledge of the mo- plastic polyps and adenomas. Eleven pa- group 3. lecular findings. tients met criteria for the diagnosis of Medical records were requested and hyperplastic polyposis syndrome.30 hematoxylin and eosin–stained slides Molecular Genetic Analyses Mean age at diagnosis was 47 years. An obtained. History of colorectal or other Mutation analysis was carried out on additional 5 patients had a single large cancer was noted. Size and site of pol- coded samples in a blinded fashion. Ge-

Table 2. Group 1: Patient Demographics, Characteristics, and Initial and Final Histopathologic Classification No. of Polyps by Histological Examination Age at First Cancer Diagnosis Family History Central Pathology Patient Polyp, y JP PJ HP TA MHAP (Age, y) of CRC Review With JPS A213 −JP B* 1.5 Ͼ50 −JP C515 −JP D* 313 −JP E 4 17 1 − JP F644 −JP G55 ϩ JP H* 31 33 2 ϩ JP I137 −JP J1330 ϩ JP K 19 5 Bilateral pheochromocytoma (15, 18) − JP L753 ϩ JP M1212 −JP N* 55 −JP With PJS O* 59 3 3 3 1 Ovary (64) − JP, HP, and TA P11 11 ϩ JP Q28 5 − Fundic gland polyps Abbreviations: CRC, colorectal cancer; HP, hyperplastic polyps; JP, juvenile polyps; JPS, juvenile polyposis syndrome; MHAP, mixed hyperplastic polyp/adenomas; PJ, Peutz- Jeghers polyps; PJS, Peutz-Jeghers syndrome; TA, tubular adenomas. *Found to have germline mutation (see Table 5).

Table 3. Group 2: Patient Demographics, Characteristics, and Initial and Final Histopathologic Classification No. of Polyps by Histologic Examination Age at First Cancer Diagnosis Family History Central Pathology Patient Polyp, y HA JP PJ HP TA MHAJP (Age, y) of CRC Review A 40 1 1 3 3 Ovarian teratoma (15) − Inflammatory polyps, HP, and TA B 69 2 3 − JP and HP C* 33 Ͼ5 −JP D 50 5 5 − HP and TA E3613 ϩ Fundic gland polyps F* 17 Ͼ70 3 3 ϩ JP and TA G* 36 11 3 7 4 CRC (36) ϩ JP and TA H* 8 Ͼ20 7 6 ϩ JP and TA I 35 1 4 MFH (33), PTC (35) − HP and TA Abbreviations: CRC, colorectal cancer; HA, hamartomatous polyps; HP, hyperplastic polyps; JP, juvenile polyps; MHAJP, mixed hyperplastic/adenoma/juvenile polyps; MFH, malignant fibrous histiocytoma; PJ, Peutz-Jeghers polyps; PTC, papillary thyroid cancer; TA, tubular adenomas. *Found to have germline mutation (see Table 5).

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS nomic DNA was extracted from periph- RESULTS 11 of ENG. The K513R mutation in- eral blood leukocytes, and the entire Group 1 volves a residue within the zona pellu- coding sequence, the exon-intron In group 1, comprising previously mu- cida domain that is conserved from Gal- boundaries, and the flanking se- tation-negative individuals with a clini- lus gallus to Mus musculus. Neither quences of PTEN, BMPR1A, SMAD4, cal diagnosis of JPS or PJS, 5 of 17 (29%) mutation was found in 105 North Ameri- STK11, and ENG were analyzed for mu- were found to have germline mutations can controls. Both patients presented tations using polymerase chain reac- (Table 2 and TABLE 5). Central pathol- with juvenile polyposis of unusually early tion–based denaturing gradient gel elec- ogy review confirmed the diagnosis of onset (age 3 and 5 years, respectively), trophoresis and direct sequencing.8,9,31-33 multiple juvenile polyps in all 14 pa- with no abnormalities noted in skin pig- PTEN and BMPR1A deletion analyses tients with JPS. Among the 14 patients mentation or features consistent with a were carried out.34 Genomic rearrange- with JPS who tested mutation-negative diagnosis of HHT (Table 5). A third pa- ments of STK11 were investigated us- in the 2 known genes, SMAD4 and tient with JPS (1B) was found to have a ing quantitative multiplex polymerase BMPR1A, 4 occult germline mutations 1.2-Mb deletion encompassing BMPR1A chain reaction of short fluorescence.35 were found. Specifically, 2 patients were and PTEN. This patient presented at an Targeted screening of BRAF (exons 11 found to have germline mutations in early age (18 months) with multiple ju- and 15), MYH (exons 7 and 13), and ENG, encoding endoglin, previously as- venile polyps (Table 5). This early polyp BHD (exons 7, 9, 11, 12, and 13), in- sociated only with HHT (Table 5). Pa- presentation clearly distinguished these cluding intron-exon boundaries, was tient 1N had a 1711C→T mutation of 3 patients (mean age, 3 years; range, 18 performed using direct sequencing or exon 12, causing the amino acid change months to 5 years) from the remaining denaturing high-performance liquid R571C. A second patient (1D) had the patients with JPS in group 1 (mean age, chromatography.36 1538 A→G (K513R) mutation in exon 11 years; range, 2-31 years).

Table 4. Group 3: Patient Demographics, Characteristics, and Initial and Final Histopathologic Classification No. of Polyps by Histologic Examination Family Family Age at First Cancer Diagnosis History History Central Pathology Patient Polyp, y HP TA SSA (Age, y) of HPS of CRC Review A 41 31 3 Sigmoid (41) − ϩ HP, TA, and SSA B 57 74 8 − − HP, TA, and SSA C 19 52 3 Cecum (19) ϩ − HP, TA, and SSA D68Ͼ30 5 4 Sigmoid (68) − ϩ HP, TA, and SSA Cecum (68) Ascending colon (69) Transverse colon (70) E623(2Ͼ1.0 cm) 2 − ϩ HP, TA, and SSA F509(2Ͼ1.0 cm) 1 1 − ϩ HP, TA, and SSA G45Ͼ30 2 Splenic flexure (45) − ϩ HP and TA H* 16 Ͼ10 Ͼ10 ϩϩHP and TA I54Ͼ30 2 − − HP, TA, and SSA J6220(2Ͼ1.0 cm) 9 − − HP, TA, and SSA K4022- ϩϩHP, TA, and SSA With a single HP (Ͼ1 cm) L 23 10 2 Cecum (26) − ϩ HP, TA, and SSA M38123 −ϩ HP and TA N 58 14 3 Cecum (60) − ϩ HP, TA, and SSA O43121 −ϩ HP and TA P 53 13 34 − − HP, TA, and SSA With multiple HP Q521017 −ϩ HP and TA R 58 8 8 CLL (54) − − HP, TA, and SSA S4735 −ϩ HP and TA T 59 6 11 GIST (60) − ϩ HP and TA U* 63 5 9 NSCL (63), breast (64) − − HP and TA V 37 2 3 Seminoma (35) − ϩ HP and TA W4526 −ϩ HP and TA Abbreviations: CLL, chronic lymphocytic leukemia; CRC, colorectal cancer; GIST, gastrointestinal stromal tumor; HP, hyperplastic polyps; HPS, hyperplastic polyposis syndrome; NCSL, non–small cell lung cancer; SSA, sessile serrated adenoma; TA, tubular adenomas. *Found to have germline mutation (see Table 5).

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS

A fourth patient with JPS (1H) was (Tables 3 and 5). Patient 2C had a 29– (c.1409_10insCCCT) in exon 10 of found to have a 1524T→A (W508R) base pair (bp) deletion of exon 1 of SMAD4. Notably, these 2 individuals mutation in exon 11 of SMAD4. This STK11 and an IVS7-32 (A→T) poly- (2G and 2H) had previous negative mo- mutation was not found in 100 nor- morphism of SMAD4. Central pathol- lecular workup for the attenuated form mal controls. This mutation had been ogy review based on 5 lesions was con- of familial adenomatous polyposis prior missed on previous analysis by an out- sistent with the diagnosis of juvenile to study entry. Central pathology re- side research laboratory but was con- polyps. Thus, unexpectedly, a germ- view noted a higher degree of juvenile firmed on reanalysis. line STK11 mutation, previously asso- polyposis with marked predisposition Among the 3 individuals present- ciated only with PJS, has been found in to adenomatous transformation com- ing with the diagnosis of PJS, patient an individual with juvenile polyps. pared with the other study cases. 1O was found to have a hemizygous de- Patient 2F was found to carry a ger- letion involving the PTEN promoter re- mline IVS3ϩ 5G→C mutation of Group 3 gion and exon 1. Interestingly, all 3 BMPR1A. RNA extraction and reverse Of 23 patients with a combination of individuals lacked characteristic mu- transcriptase polymerase chain reac- hyperplastic and adenomatous pol- cocutaneous pigmentation and had tion revealed that this mutation led to yps, 2 (9%) were found to have germ- been diagnosed solely on polyp histol- a splicing alteration, resulting in a 97-bp line PTEN mutation (Tables 4 and 5). ogy results. Indeed, central rereview of intronic insertion between the coding Review of the clinical phenotype re- the polyp results did not confirm ini- sequences of exons 3 and 4. Central pa- vealed features reminiscent of, but not tial clinical pathology findings for any thology review noted juvenile polyps, diagnostic for, Cowden syndrome of the patients with PJS (Table 2). some with adenomatous change, from (Table 5).37 Patient 3H, with a 612insC each colonoscopic procedure. mutation in exon 6 of the PTEN gene, Group 2 A second splice site mutation, presented with a 20-year history of pol- Of the 9 individuals with undefined IVS1ϩ1(G→T) of BMPR1A, was found yposis and a significant family his- hamartomatous polyps, 4 (44%) were in patient 2G. A fourth patient in this tory. Patient 3U had the IVS3- found to have germline mutations group (2H) harbored a 4-bp insertion 3~7delCTTTT mutation of PTEN,a

Table 5. Mutation Results and Clinical Features Age at Final Pathology Group/Patient Mutation Diagnosis, y Result Polyp Location Extracolonic Features Group 1 B Hemizygous deletion of BMPR1A 1.5 JP Duodenum Macrocephaly (OFC, 53.5 cm); frontal and PTEN and colon bossing; depressed nasal bridge; high arched palate; broad thumbs and toes; ASD with VSD with absence of portal vein; cytogenetic studies of 46,XX,t(2;10)(q31;p15). D ENG, K513R (1538 A→G) 3 JP Pancolonic None reported H SMAD4, W508R (1524T→A) 31 JP and TA Pancolonic None reported N ENG, R571C (1711C→T) 5 JP Distal descending Macrocephaly (OFC, 54.5 cm); frontal and cecum bossing; down-slanting, palpebral fissures; elongated philtrum; high arched palate O Hemizygous deletion of PTEN 59 JP, HP, and TA Gastric and Mucinous ovarian adenocarcinoma promoter and exon 1 pancolonic (64 y); deep vein thrombosis* Group 2 C STK11, 191-219 del 29 of exon 1; 33 JP Pancolonic None reported SMAD4, IVS7-32 (A→T) F BMPR1A, IVS3 ϩ 5(G→C) 17 JP and TA Pancolonic None reported G BMPR1A, IVS1 ϩ 1(G→T) 36 JP and TA Pancolonic Hepatic flexure CRC (36 y)* H SMAD4 c.1409_10insCCCT 8 JP and TA Pancolonic None reported Group 3 H PTEN, 612insC 16 HP and TA Rectosigmoid Macrocephaly (OFC 64.4 cm); single lipoma; keratoderma palmare punctata U PTEN, IVS3-3 ~ 7delCTTTT 63 HP and TA Pancolonic Macrocephaly (OFC 58.0 cm); non–small cell lung cancer (62 y); invasive ductal carcinoma (63 y)* Abbreviations: ASD, atrial septal defect; CRC, colorectal cancer; HP, hyperplastic polyps; JP, juvenile polyps; OFC, occipital-frontal circumference; PJ, Peutz-Jeghers polyps; SSA, sessile serrated adenoma; TA, tubular adenomas; VSD, ventricular septal defect. *Numbers in parentheses indicate age at diagnosis.

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS history of non–small cell lung cancer It has been predicted that a propor- tempting to speculate that the mis- and invasive ductal carcinoma of the tion of patients with polyposis would sense mutations in this region of the breast, and 5 hyperplastic polyps and share clinical and molecular features, protein may be gain-of-function mu- 9 adenomas identified since age 63 given the crosstalk and interaction that tations. This has been seen in achon- years. Neither PTEN mutation– occurs between many of the known droplasia41 and in the NEU/erbB2 re- positive patient had SSA on central pa- genes. As an example, the TGF-␤ sig- ceptor,42 in which mutations in the thology review. naling pathway is an important mecha- transmembrane domains are activated Central review of pathology reports re- nism for the pathogenesis of many heri- due to altered dimerization properties vealed the highest frequency of discrep- table diseases, including juvenile of the proteins. ancy in group 3. Initial pathology re- polyposis, HHT, and primary pulmo- Currently there is great interest fo- ports identified only 3 of 23 patients nary hypertension (Figure). Patients cused on angiogenesis and its poten- (13%) with SSA. Central pathology re- with mutations in ALK1 need careful tial role in solid-tumor development, view noted that 13 of 23 patients (57%) assessment and management for com- as seen with colorectal carcinoma. had at least 1 SSA. Of the 11 patients with plications not only of HHT but also of It is known that angiogenesis pro- HPS, 9 had at least 1 SSA, with the ma- primary pulmonary hypertension. Simi- motes the development of small adeno- jority (78%) found in the proximal co- larly, mutations in the SMAD4 gene can matous polyps, with recent evidence lon. Seven colon carcinomas occurred in cause a combined syndrome of juve- showing alterations in the microvas- 4 patients with HPS (mean age at diag- nile polyposis and HHT.8 Other candi- culature at even earlier stages of colon nosis, 54 years; range, 19-70 years). In date genes of the TGF-␤ family have carcinogenesis.43-45 Endoglin is a ho- the remaining cohort of patients in group been studied in patients with juvenile modimeric transmembrane glycopro- 3, 4 had at least 1 SSA. Of these, 2 (aged polyposis, but no mutations have been tein predominantly expressed in vas- 26 and 60 years, respectively) were di- identified.10 Endoglin, a protein prod- cular endothelial cells but also present agnosed with cecal carcinoma. uct of the ENG gene previously asso- in several nonendothelial tissues.6,46 No- ciated with only HHT, acts as a cor- tably, elevated levels of endoglin ex- COMMENT eceptor and accessory protein in this pression have been detected on hu- Our systematic histopathologic and signaling pathway. Previously, 1 small man microvascular endothelium and on molecular evaluation of 49 unrelated series of patients (n=7) with features vascular endothelial cells in tissues and patients with unexplained hamartoma- of both HHT and JPS were studied for tumors undergoing active angiogen- tous polyposis or mixed polyps revealed either ENG, ACVRL1,orSMAD4 mu- esis.46-48 In fact, endoglin serves as a that 11 (22%) had germline mutations. tations; each was found to have a ger- useful prognostic marker, as the mi- Of these, we characterize a new geneti- mline SMAD4 mutation.8 Therefore, this crovessel density increases during the cally defined form of juvenile polyposis represents the first report of germline progressive stages of colorectal carci- in 2 patients with germline ENG muta- ENG mutation as a cause of juvenile nogenesis.6,46,49,50 Gain-of-function mu- tions previously only found in individu- polyposis. Interestingly, the 2 patients tations might serve to further up- als with HHT. Seven additional patients with ENG mutation in our series pre- regulate endoglin expression in the had germline mutation of known pol- sented with juvenile polyposis of un- vasculature network, which, along with yposis genes, and 2 others were found usually early onset (Table 5). Neither modifier genes and environment, con- to have a PTEN mutation, although their patient has stigmata of HHT, although tribute to polyp development and car- clinical history was not diagnostic for both are young. This suggests either that cinogenesis. Recently, Lebrin et al51 im- Cowden syndrome (Table 5). Colorec- a subset of patients with JPS and ENG plicated endoglin as a modulator of the tal polyps do occur at increased fre- mutations will have polyposis with- balance between the TGF-␤/ACVRL1 quency in patients with Cowden syn- out HHT, as may be the case with and TGF-B/ALK-5 signaling pathways drome, sometimes as early as age 5 years, BMPR1A mutation–positive families, or in endothelial cells (Figure). Thus, im- and are typically found distal to the that HHT may have age-related pen- balance between signaling pathways hepatic flexure.38,39 Usually these are etrance. Penetrance of HHT symp- might be a clue to understanding the hamartomatous, although adenomas and toms by age 8 years is approximately biological role of endoglin in the con- hyperplastic and inflammatory polyps 25% (D.A.M., unpublished data). The text of JPS and HHT. have been reported.38,39 It remains to be advent of clinical disease might also be Hyperplastic polyposis syndrome is seen what role the PTEN gene plays in dependent on the site and type of mu- not a single entity, as histological and the development of colorectal polyps. tation within a particular gene. Inter- molecular analysis has shown a mixed Overall, we reclassified 6 patients (1B, estingly, the vast majority of HHT type picture. This condition is character- 1O, 2C, 2F, 2G, 2H), not only on the 1–specific missense mutations occur be- ized by multiple or large hyperplastic basis of this more extensive molecular tween exons 1 and 9, compared with polyps. Recent studies suggest that the analysis but also by reevaluation of polyp the novel JPS-related ENG mutations SSA may be a component of HPS and, histology results. that occur in exons 11 and 12.40 It is as such, may be associated with the in-

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS creased risk for neoplasia.13-16,24-26 It re- of Medicine (Dr Eng), Cleveland; Department of Mo- 5. Newman JH, Wheeler L, Lane KB, Loyd E, Gaddi- lecular Genetics and Microbiology, Duke University pati R, Phillips JA III, Loyd JE. Mutation in the gene mains unknown whether SSAs arise Medical Center, Durham, NC (Drs Gallione and Mar- for bone morphogenetic protein receptor II as a cause from traditional hyperplastic polyps or chuk); Biomedicum Helsinki, Department of Medical of primary pulmonary hypertension in a large kindred. Genetics and Molecular and Cancer Biology Re- N Engl J Med. 2001;345:319-324. are a unique polyp of themselves, as well search Program, University of Helsinki, Helsinki, Fin- 6. Waite KA, Eng C. From developmental disorder to as which of a number of molecular path- land (Drs Alhopuro and Aaltonen and Mr Lehtonen); heritable cancer: it’s all in the BMP/TGF-␤ family. Nat ways may be involved. In our series of Cancer Genetics Laboratory, Van Andel Research In- Rev Genet. 2003;4:763-773. stitute, Grand Rapids, Mich (Drs Khoo and Teh); De- 7. Blobe G, Schiemann W, Lodish H. Role of trans- patients with HPS, most polyps were hy- partment of Genetics, Faculty of Medicine University forming growth factor ␤ in human disease. N Engl J perplastic polyps, indistinguishable from Hospital of Rouen, Rouen, France (Drs Martin and Fre- Med. 2000;342:1350-1358. bourg); and Cancer Research UK Human Cancer Ge- 8. Gallione CJ, Repetto GM, Legius E, et al. A com- common sporadic hyperplastic polyps netics Research Group, University of Cambridge, Cam- bined syndrome of juvenile polyposis and hereditary apart from the number and, in some bridge, England (Dr Eng). haemorrhagic telangiectasia associated with muta- cases, the large size and proximal loca- Author Contributions: Dr Eng had full access to all of tions in MADH4 (SMAD4). Lancet. 2004;363: the data in the study and takes responsibility for the integ- 852-859. tion. The finding that most (81%) of our rity of the data and the accuracy of the data analysis. 9. Zhou XP, Woodford-Richens K, Lehtonen R, et al. patients with HPS had at least 1 SSA, Study concept and design: Sweet, Aaltonen, Eng. Germline mutations in BMPR1A/ALK3 cause a sub- Acquisition of data: Sweet, Willis, Gallione, Sawada, set of cases of juvenile polyposis syndrome and of while 4 of 11 (36%) had a previous di- Alhopuro, Khoo, Martin, Bridgeman, Heinz, Pilarski, Cowden and Bannayan-Riley-Ruvalcaba syndromes. agnosis of colorectal cancer, supports the Prior, Frebourg, Teh, Marchuk, Aaltonen, Eng. Am J Hum Genet. 2001;69:704-711. 19,24,26 Analysis and interpretation of data: Sweet, Zhou, 10. Howe JR, Sayed MG, Ahmed AF. The preva- results of previous work. Gallione, Alhopuro, Patocs, Martin, Bridgeman, lence of MADH4 and BMPR1A mutations in juvenile Our overall observations in this pro- Lehtonen, Prior, Frebourg, Marchuk, Aaltonen, Eng. polyposis and absence of MBPR2, BMPR1B, spectively collected series illustrate the Drafting of the manuscript: Sweet, Willis, Alhopuro, and ACVR1 mutations. J Med Genet. 2004;41: Aaltonen, Eng. 484-491. importance of defining molecular diag- Critical revision of the manuscript for important in- 11. Imperiale TF, Wagner DR, Lin CY, Larkin GN, nosis to influence surveillance and medi- tellectual content: Sweet, Zhou, Gallione, Sawada, Rogge JD, Ransohoff DF. Results of screening colo- Alhopuro, Khoo, Patocs, Martin, Bridgeman, Heinz, noscopy among persons 40 to 49 years of age. N Engl cal intervention strategies for the pa- Pilarski, Lehtonen, Prior, Frebourg, Teh, Marchuk, J Med. 2002;346:1781-1785. tient and to allow for determination of Aaltonen, Eng. 12. DiSario J, Foutch PG, Mai H, Pardy K, Manne R. Statistical analysis: Sweet, Eng. Prevalence and malignant potential of colorectal pol- disease status for at-risk family mem- Obtained funding: Aaltonen, Eng. yps in asymptomatic, average-risk men. Am J bers. This is because each specific syn- Administrative, technical, or material support: Willis, Gastroenterol. 1991;86:941-945. Zhou, Alhopuro, Patocs, Martin, Heinz, Pilarski, drome carries different organ-specific 13. Hawkins NJ, Ward RL. Sporadic colorectal can- Lehtonen, Aaltonen, Eng. cers with microsatellite instability and their possible ori- Study supervision: Frebourg, Aaltonen, Eng. risks of neoplasia. Many of our study pa- gin in hyperplastic polyps and serrated adenomas. Financial Disclosures: None reported. J Natl Cancer Inst. 2001;93:1307-1313. tients, previously diagnosed and man- Funding/Support: This study was partially funded by 14. Wynter CV, Walsh MD, Higuchi T, Leggett BA, the American Cancer Society (grant RSG02-151-01- aged as having a particular form of pol- Young J, Jass JR. Methylation patterns define two types CCE to Dr Eng). Dr Eng is a recipient of the Doris Duke of hyperplastic polyp associated with colorectal cancer. yposis, were reclassified based on Distinguished Clinical Scientist Award. Dr Aaltonen re- Gut. 2004;53:573-580. extended molecular analyses and re- ceives grants from the Academy of Finland (44870, 15. Hawkins NJ, Bariol C, Ward R. The serrated neo- Finnish Center of Excellence Program 2000–2005), view of histopathology results. Because plasia pathway. Pathology. 2002;34:548-555. Finnish Cancer Society, Sigrid Juselius Foundation, and 16. Jass J, Whitehall V, Young J, Leggett B. Emerg- phenotypic features can be shared by a European Commission (GLG2-CT-2001-01861). ing concepts in colorectal neoplasia. Gastroenterology. number of hamartoma syndromes, a mo- Role of the Sponsors: The funding agencies had no role in design and conduct of the study; the collec- 2002;123:862-876. lecular analysis of a battery of the known tion, management, analysis, and interpretation of the 17. Jass J. Serrated adenoma of the colorectum: susceptibility genes (PTEN, BMPR1A, data; or the preparation, review, or approval of the a lesion with teeth. Am J Pathol. 2003;162:705-708. manuscript. 18. Bariol C, Hawkins N, Turner J, Meagher A, Wil- SMAD4, and STK11) should be consid- Acknowledgment: We thank the genetic counselors, liams DB, Ward R. Histopathological and clinical evalu- ered for patients who have suggestive physicians, and families who have so kindly contrib- ation of serrated adenomas of the colon and rectum. Mod Pathol. 2003;16:417-423. clinical and histopathologic features but uted to this study; in particular, Barbara Kunz, MS, CGC, and Joy Larsen Haidle, MS, CGC, Hubert Hum- 19. Snover DC, Jass JR, Fenoglio-Preiser C, Batts KP. who are negative for mutations in the phrey Institute, Minneapolis, Minn; Katherine Hunt, Serrated polyps of the large intestine: a morphologic most obvious genes. Lastly, given the sig- MS, CGC, University of Texas, San Antonio; John Quil- and molecular view of an evolving concept. Am J Clin lin, MS, CGC, Virginia Commonwealth University, Rich- Pathol. 2005;124:380-391. nificant discrepancies seen in histology mond; Shonee Lesh, MS, CGC, Kaiser Permanente, 20. Torlakovic E, Skovlund E, Snover DC, Torlakovic reports, rereview by a dedicated gastro- Sacramento, Calif; Dawn Allain, MS, CGC, Wauke- G, Nesland J. Morphologic reappraisal of serrated sha Memorial Hospital, Waukesha, Wis; and Aimee colorectal polyps. Am J Surg Pathol. 2003;27: intestinal pathologist should be consid- Fiorito, MS, CGC. 65-81. ered, since that may guide accurate se- 21. Yang S, Farraye F, Mack C, Posnik O, O’Brien M. BRAF and KRAS mutations in hyperplastic polyps and lection of which gene(s) to begin testing. REFERENCES serrated adenomas of the colorectum: relationship to 1. Nagy R, Sweet K, Eng C. Highly penetrant heredi- histology and CpG island methylation status. Am J Surg Author Affiliations: Clinical Cancer Genetics Pro- tary cancer syndromes. Oncogene. 2004;23: Pathol. 2004;28:1452-1459. gram, Human Cancer Genetics Program, Compre- 6445-6470. 22. Kambara T, Simms LA, Whitehall VL, et al. BRAF hensive Cancer Center (Drs Zhou, Sawada, Patocs, and 2. National Center for Biotechnology Information. On- mutation is associated with DNA methylation in ser- Eng and Mssrs Sweet and Pilarski), Division of Hu- line Mendelian Inheritance in Man. Available at: rated polyps and cancers of the colorectum. Gut. 2004; man Genetics, Department of Internal Medicine (Dr http://www/ncbi.nlm.nih.gov/entrez/query.fcgi 53:1137-1144. Eng and Mssrs Sweet and Pilarski), and Department ?db-OMIM. Accessed October 6, 2005. 23. Chan T, Zhao W, Project CG, Leung S, Yuen S. of Pathology (Drs Prior and Heinz and Mr Bridge- 3. Jeghers H, Katz K. Generalized intestinal polyp- BRAF and KRAS mutations in colorectal hyperplastic man), The Ohio State University, Columbus; Ge- osis and melanin spots of the oral mucosa, lips and polyps and serrated adenomas. Cancer Res. 2003;63: nomic Medicine Institute, Lerner Research Institute, digits: a syndrome of diagnostic significance. N Engl 4878-4881. Cleveland Clinic Foundation, Cleveland, Ohio (Drs Pa- J Med. 1949;241:993-1005. 24. Leggett BA, Devereaux B, Biden K, Searle J, Young tocs and Eng); Department of Pathology, University 4. Hemminki A, Markie D, Tomlinson I, et al. A serine/ J, Jass J. Hyperplastic polyposis: association with Hospitals Health System (Dr Willis) and Department threonine kinase gene defective in Peutz-Jeghers colorectal cancer. Am J Surg Pathol. 2001;25: of Genetics, Case Western Reserve University School syndrome. Nature. 1998;391:184-187. 177-184.

Downloaded from jama.ama-assn.org by guest on December 21, 2011 MOLECULAR CLASSIFICATION OF PATIENTS WITH UNEXPLAINED POLYPOSIS

25. Jass JR, Iino H, Ruszkiewicz A, et al. Neoplastic promoter mutations and deletions in Cowden/ 43. Wali RK, Roy HK, Kim YL, et al. Increased micro- progression occurs through mutator pathways in hy- Bannayan-Riley-Ruvalcaba syndrome result in aber- vascular blood content is an early event in colon perplastic polyposis of the colorectum. Gut. 2000;47: rant PTEN protein and dysregulation of the phosphoi- carcinogenesis. Gut. 2005;54:654-660. 43-49. nositol-3-kinase/Akt pathway. Am J Hum Genet. 2003; 44. Aotake T, Lu CD, Chiba Y, Muraoka R, Tanigawa 26. Rashid A, Houlihan P, Booker S, Petersen G, 73:404-411. N. Changes in angiogenesis and tumor cell apoptosis Giardiello FM, Hamilton S. Phenotypic and molecu- 35. Le Meur N, Martin C, Saugier-Veber P, et al. Com- during colorectal carcinogenesis. Clin Cancer Res. 1999; lar characteristics of hyperplastic polyposis. plete germline deletion of the STK11 gene in a family 5:135-142. Gastroenterology. 2000;119:323-332. with Peutz-Jeghers syndrome. Eur J Hum Genet. 2004; 45. Shpitz B, Gochberg S, Neufeld D, et al. Angio- 27. Hohenstein P, Molenaar L, Elsinga J, et al. Ser- 5:415-418. genic switch in earliest stages of human colonic rated adenomas and mixed polyposis caused by a splice 36. Al-Tassan N, Chmiel NH, Maynard J, et al. Inher- tumorigenesis. Anticancer Res. 2003;23:5153- acceptor deletion in the mouse Smad4 gene. Genes ited variants of MYH associated with somatic G:C→T:A 5157. Chromosomes Cancer. 2002;36:273-282. mutations in colorectal tumors. Nat Genet. 2002;30: 46. Fonsatti E, Maio M. Highlights on endoglin (CD105): 28. Jass J, Williams CB, Bussey HJR, Morson B. Juve- 227-232. from basic findings towards clinical applications in hu- nile polyposis—a precancerous condition. Histo- 37. Eng C. Will the real Cowden syndrome please stand man cancer. J Transl Med. 2004;2:18. pathology. 1988;13:619-630. up: revised diagnostic criteria. J Med Genet. 2000;37: 47. Wong SH, Hamel L, Chevalier S, Philip A. Endog- 29. Aaltonen LA, Jarvinen H, Gruber SB. Peutz- 828-830. lin expression on human microvascular endothelial cells Jeghers syndrome. In: Hamilton S, Aaltonen LA, eds. 38. Lashner BA, Riddell RH, Winans CS. Ganglioneu- association with betaglycan and formation of higher or- Pathology and Genetics of Tumours of the Digestive romatosis of the colon and extensive glycogenic acan- der complexes with TGF-␤ signalling receptors. Eur J System. Lyon, France: IARC Press; 2000:74-76. thosis in Cowden’s disease. Dig Dis Sci. 1986;31: Biochem. 2000;267:5550-5560. 30. Jass J, Burt R. Hyperplastic polyposis. In: Hamil- 213-216. 48. Miller DW, Graulich W, Karges B, et al. Elevated ton S, Aaltonen LA, eds. Pathology and Genetics of 39. Weber HC, Marsh D, Lubensky I, Lin A, Eng C. Germ- expression of endoglin, a component of the TGF-␤- Tumours of the Digestive System. Lyon, France: IARC line PTEN/MMAC1/TEP1 mutations and association with receptor complex, correlates with proliferation of Press; 2000:136-138. gastrointestinal manifestiations in Cowden disease. tumor endothelial cells. Int J Cancer. 1999;81: 31. Marsh DJ, Coulon V, Lunetta KL, et al. Mutation Gastroenterology. 1998;114S:G2902. 568-572. spectrum and genotype-phenotype analyses in 40. Lux A, Gallione CJ, Marchuk DA. Expression analy- 49. Saad RS, Liu Y, Nathan G, Celebrezze J, Medich D, Cowden disease and Bannayan-Zonana syndrome, two sis of endoglin missense and truncation mutations: in- Silverman JF. Endoglin (CD105) and vascular endothe- hamartoma syndromes with germline PTEN mutation. sights into protein structure and disease mechanisms. lial growth factor as prognostic markers in colorectal Hum Mol Genet. 1998;7:507-515. Hum Mol Genet. 2000;22:745-755. cancer. Mod Pathol. 2003;17:197-203. 32. Mutter GL, Lin MC, Fitzgerald JT, et al. Altered 41. Naski MC, Wang Q, Xu J, Ornitz DM. Graded ac- 50. Akagi K, Ikeda Y, Sumiyoshi Y, et al. Estimation of PTEN expression as a diagnostic marker for the ear- tivation of fibroblast growth factor receptor 3 by mu- angiogenesis with anti-CD105 immunostaining in the liest endometrial precancers. J Natl Cancer Inst. 2000; tations causing achondroplasia and thanatophoric process of colorectal cancer development. Surgery. 2002; 92:924-931. dysplasia. Nat Genet. 1996;2:233-237. 131(1 suppl):S109-S113. 33. Roth S, Sistonen P, Hemminki A, et al. SMAD genes 42. Bargmann CI, Hung MC, Weinberg RA. Multiple 51. Lebrin F, Goumans MJ, Jonker L, et al. Endoglin in juvenile polyposis. Genes Chromosomes Cancer. independent activations of the neu oncogene by a point promotes endothelial cell proliferation and TGF-␤/ 1999;26:54-61. mutation altering the transmembrane domain of p185. ALK1 signal transduction. EMBO J. 2004;23: 34. Zhou XP, Waite KA, Pilarski R, et al. Germline PTEN Cell. 1986;45:649-657. 4018-4028.

Downloaded from jama.ama-assn.org by guest on December 21, 2011