- Email: [email protected]
A juvenile polyposis tumor suppressor locus at 10q22 is deleted from nonepithelial cells in the lamina propria
GASTROENTEROLOGY 1997;112:1398–1403
A Juvenile Polyposis Tumor Suppressor Locus at 10q22 Is Deleted From Nonepithelial Cells in the Lamina Propria RUSSELL F. JACOBY,*,‡,§ STEVEN SCHLACK,* CAROLYN E. COLE,‡ MARLENE SKARBEK,\ CHARLES HARRIS,\ and LORRAINE F. MEISNER\ *Section of Gastroenterology, Department of Medicine, xDepartment of Cytogenetics, and ‡University of Wisconsin Comprehensive Cancer Center, University of Wisconsin, Madison; and §William S. Middleton Memorial Veteran’s Hospital, Madison, Wisconsin
Background & Aims: Juvenile polyps are characterized by an abundant lamina propria that lacks smooth muscle and may contain cystically dilated glands, with epithelium that seems normal and is nondysplastic. Rarely, an autosomal dominant inheritance pattern occurs. The aim of this study was to test the hypothesis that the genetic defect in both sporadic juvenile polyps and hereditary juvenile polyposis involves loss of function for a tumor suppressor gene. Methods: Allelic losses were detected by comparing normal DNA with tumor DNA from a series of 47 juvenile polyps from 16 patients using polymerase chain reaction amplification of microsatellite markers and fluorescent in situ hybridization (FISH). Results: Somatic deletions at 10q22 were detected in 39 of 47 juvenile polyps (83%) from 16 unrelated patients with either hereditary or sporadic juvenile polyps, and the minimum overlap localized juvenile polyposis coli to the 3-cM interval D10S219– D10S1696. Fluorescent in situ hybridization shows that the cells affected by deletion mutation reside exclusively in the lamina propria, not in the epithelium. Conclusions: The location of a novel tumor suppressor gene on chromosome 10 that is affected by deletion mutation in the majority of juvenile polyps was mapped. Unlike adenomas and carcinomas of the colonic epithelium, juvenile polyps originate in the lamina propria.
J
uvenile polyps are characterized by an abundant lamina propria that lacks smooth muscle proliferation and contains cystically dilated mucin-filled glands lined by a typically normal epithelium.1 Solitary juvenile polyps, diagnosed in approximately 1% of children, account for the majority of gastrointestinal polyps found before adulthood.2,3 Juvenile polyposis coli ( JPC) is a rare autosomal dominant syndrome (Mendelian Inheritance in Man no. 174900) characterized by numerous juvenile polyps throughout the colon and occasionally elsewhere in the gastrointestinal tract. This hereditary cancer syndrome is associated with a high risk for colon adenocarcinoma1,4 – 7 but is not linked to the adenomatous polyposis coli gene.8 To investigate the possibility that another tumor suppressor gene is responsible, we initiated a ge/ 5e1b$$0058
03-14-97 17:39:29
gasa
nome-wide scan for allele losses in sporadic and hereditary juvenile polyps. A more focused search was made possible by our discovery of a unique patient with juvenile polyposis and multiple congenital anomalies associated with a small but cytogenetically detectable deletion.9
Materials and Methods Patients and Diagnostic Criteria for Juvenile Polyps Pathognomonic criteria for juvenile polyps of the gastrointestinal tract include an abundant lamina propria that lacks smooth muscle and may contain cystically dilated glands, with epithelium that appears normal and is nondysplastic.1 Microscopic sections from all paraffin tissue blocks obtained from other hospitals were examined by our pathologist to confirm diagnoses and identify areas for microdissection and DNA extraction. Sporadic cases are defined by solitary juvenile polyps in patients without any affected relatives. Hereditary juvenile polyposis was diagnosed in patients with positive family history and/or multiple juvenile polyps. Patient classification, age at diagnosis, and polyp multiplicity are shown in Table 1.
Fluorescent In Situ Hybridization Tissue sections (5-mm) were prepared as described previously.10,11 Briefly, sections were mounted on charged slides, baked overnight at 657C, deparaffinized in d-limonene twice and 100% ethanol twice for 10 minutes each, and air-dried. Sections were then incubated in 1 mol/L sodium thiocyanate for 20 minutes at 457C, rinsed in distilled water, incubated in Hanks’ basic salt solution and 1 mg/mL protease for 2 minutes at 457C, rinsed twice in 21 SSC (11 SSC is 0.15 mol/L NaCl plus 0.015 mol/L sodium citrate) for 1 minute each, and dehydrated through an ethanol series (70%-80%90%-100%). A digoxigenin-labeled cosmid probe from 10q22 at the HK1 locus in the JPC region (Oncor, Gaithersburg, MD) was hybridized overnight at 377C in a humidified incubaAbbreviations used in this paper: FISH, fluorescent in situ hybridization; JPC, juvenile polyposis coli; LOH, loss of heterozygosity; PCR, polymerase chain reaction. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
April 1997
JUVENILE POLYPOSIS TUMOR SUPPRESSOR LOCUS 1399
Table 1. Characteristics of Patients With Juvenile Polyps Patient (n Å 16)
Age at diagnosis (yr )
Sporadic 141 143 144 Juvenile polyposis 80 81 106 117 121 122 126 127 130 131 138 139 142
Polyp multiplicity
6 66 18
1 1 1
1 4 18 6 6 2 28 6
34 ú50 3 5 ú100 2 2 ú10 3 2 23 7 ú20
1 4 5 4
alleles were quantified using a Molecular Dynamics phosphoimager and ImageQuant software (Sunnyvale, CA). Loss of heterozygosity (LOH) was scored if the signal intensity for either of the two tumor alleles was decreased by at least 0.20 compared with the corresponding normal alleles, as calculated by the formula (N1/N2-T1/T2)(N1/N2). Statistical significance was assessed as previously described using a dynamic simulation computer algorithm developed for Bayesian analysis of chromosomal deletion data.14
Results Microsatellite Map of a Germline Deletion Associated With JPC
tor after denaturation under a sealed coverslip at 727C for 6 minutes. A control cosmid probe of equal size and similar fluorescent signal properties but located on an uninvolved chromosome (21q22.3) was hybridized to a paired adjacent section from each polyp. Each slide was then washed at 707C in 11 SSC for 5 minutes. A multistage detection system was used to detect probe bound to nuclear DNA by epifluorescence microscopy after reaction with a series of three antibodies: fluorescein isothiocyanate–labeled antidigoxigenin, rabbit anti-goat, and fluorescein isothiocyanate–labeled goat antirabbit (Boehringer Mannheim, Indianapolis, IN). The mean number of fluorescent signals per nucleus was scored for 100 cells of each type (epithelial or lamina propria) from each section. The same paraffin sections had been stained first for histological identification with Giemsa and then processed for fluorescent in situ hybridization (FISH) analysis. Horizontal and vertical coordinates were recorded for randomly selected regions of the polyp to be analyzed, and Giemsa images of these areas were compared with FISH images to correlate histology with deletion data. Deletions are scored if the tumor has a mean value of õ1.6 FISH signals per nucleus, which is 2 SD less than the mean value of 1.8 observed in normal control tissues.10,11 Statistical significance was assessed using Fisher’s Exact Test to compare results with the juvenile polyposis locus probe on chromosome 10 to a randomly selected control locus on chromosome 21.
Cytogenetic studies of a patient with juvenile polyposis and multiple congenital abnormalities of the head, extremities, and abdomen previously showed that the only chromosomal defect was an interstitial deletion of 10q22.3–q24.1.9 This cytogenetically evident abnormality affected only one homologue of chromosome 10 in the patient and was not observed in either of the parents who were unaffected by juvenile polyposis.9 We verified paternity by genotyping microsatellites at multi-
Microsatellite Allele Loss Studies Polymerase chain reaction (PCR) amplification was performed using normal genomic DNA extracted from peripheral blood lymphocytes compared with juvenile polyp DNA extracted from paraffin-embedded, formalin-fixed tissue sections. Microsatellite loci12 on chromosome 10 were amplified with the incorporation of [a-32P]adenosine triphosphate to label PCR products, which were then separated by electrophoresis on 6% polyacrylamide gels.13 Heterozygous microsatellite
/ 5e1b$$0058
03-14-97 17:39:29
gasa
Figure 1. Deletions associated with somatic LOH in a series of sporadic and familial juvenile polyps and a germline deletion in a patient with multiple congenital anomalies including juvenile polyps. Solid lines indicate deleted markers, dotted lines noninformative markers, and arrowheads the proximal and distal retained markers. D Pt, patient deletion.
WBS-Gastro
1400 JACOBY ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
ple loci (data not shown), confirming that the child’s deletion was a de novo event. The genetic map of this deletion was then defined by comparing the patient’s genotype to parental genotypes for 41 microsatellite markers on chromosome 10 that were informative in this family. The patient did not inherit any paternal alleles for 5 contiguous informative markers (D10S573, D10S541, D10S579, D10S564, and D10S571) spanning an interval of approximately 17 cM (Figure 1, right). Inheritance of alleles from both parents was shown for 29 markers that were proximal to D10S219 and for 7 markers that were distal to D10S603. Genetic markers mapping within the deleted interval were then used for allele loss studies in our series of sporadic and hereditary juvenile polyps from 16 unrelated patients (Figure 1, left). Somatic Deletions at Chr10q22 Occur in Most Juvenile Polyps We observed LOH in the majority of juvenile polyps for microsatellite markers on chromosome 10,
particularly in the interval D10S583–D10S569 (Figure 2). The most frequently affected genetic marker, D10S219, was deleted from 39 of 47 (83%) of juvenile polyps. The probability of deletion diminished as genetic map distance from D10S219 increased (Figure 3). Maps indicating the contiguous extent of deletion for each individual affected polyp were constructed from LOH data for the most informative markers and indicated that the minimum region of overlap involves the 3-cM interval D10S219–D10S1696 (Figure 1). Consistent with the hypothesized inherited tumor suppressor gene mutation in patients with multiple juvenile polyps, different tumors from the same polyposis patient always had somatic deletions affecting the same allele (Figure 2). FISH Identifies Cells Affected by Deletion Juvenile polyps are often removed at the time of colonoscopic examination by snare electrosurgical excision without resecting much adjacent normal tissue. FISH was useful to detect LOH when tissues of this type
Figure 2. Deletion of genetic markers on chromosome 10 in juvenile polyps. Allele loss for PCR-amplified microsatellites was quantified as described in Materials and Methods for juvenile polyp DNA compared with normal DNA. LOH is indicated by bold large typeface and dark grey shading for significant allele loss ratios ú0.30, or large typeface with medium grey shading if between 0.20 and 0.29, or light grey if minimally significant (between 0.17 and 0.19). ‘‘NI’’ indicates that the marker was not informative (alleles homozygous or not easily quantitated separately because similar size caused overlapping of the allele ladders on gels). Blank spaces indicate data that were unavailable because of poor amplification of paraffin section DNA, which occurs frequently if the microsatellite is ú200 base pairs. ‘‘S’’ or ‘‘L’’ indicates which of the microsatellite alleles (shorter or longer) showed LOH as decreased relative intensity of that allele.
/ 5e1b$$0058
03-14-97 17:39:29
gasa
WBS-Gastro
April 1997
JUVENILE POLYPOSIS TUMOR SUPPRESSOR LOCUS 1401
Deletion mutations exclusively affected cells in the lamina propria in 29 of 30 juvenile polyps (Figure 4A). Epithelial cells, smooth muscle cells, fibroblasts, and polymorphonuclear leukocytes did not seem to be involved because FISH shows that their nuclei are all disomic (normal). Light microscopy showed that the lamina propria cells affected by deletion appeared to have round nuclei and scant cytoplasm typical of lymphocytes and macrophages or histiocytes.
Discussion
Figure 3. Proportion of juvenile polyps with LOH for microsatellite markers on Chr 10q in the juvenile polyposis/Cowden critical region. D10S196 is near the centromere.
were available from the pathology archives, but PCR allele loss studies could not be performed because normal DNA was not available from the decreased patient or preserved tissue. For FISH analysis of juvenile polyps, we obtained a cosmid clone for the gene HK1 that maps (Human Genome Database) between D10S201 and D10S564 near the minimum region of overlap of germline and somatic JPC deletions. Using this probe that appears to be closely linked to the JPC locus, somatic deletions were shown by FISH in 30 of 39 (77%) of juvenile polyps (Figure 4A). FISH results agreed with LOH data obtained by PCR analysis of microsatellite alleles in 16 of 19 cases in which both techniques could be used on the same polyp specimens (80-1, 80-2, 803, 80-5, 117-2, 117-3, 121-1, 126-2, 131-1, 138-1, 139-1, 139-2, 139-3, 139-4, 139-5, and 139-6; compare Figure 2 and Figure 4A). The JPC locus is apparently separated from these markers by a small genetic distance because 2 of 19 polyps (0.10) deleted only the FISH and 1 of 19 (0.05) deleted only the PCR markers. The high rate of deletion observed at chromosome 10q22 probably has a functional basis. Although common in colon carcinomas, aneuploidy is not a characteristic of benign juvenile polyps. As expected, deletions were not observed at a control locus on another chromosome (chromosome 21), indicating that nonspecific aneuploidy is unlikely to account for the deletions we observed at the putative JPC locus on Chr10q (Figure 4B). FISH analysis is particularly useful because it can identify mutations in individual cells in tissue sections from tumors composed of multiple cell types, e.g., juvenile polyps. Cellular architecture was recorded after Giemsa staining, and then FISH was used to detect deletion mutations affecting identified cells in the same section. / 5e1b$$0058
03-14-97 17:39:29
gasa
We show that a locus mapping to the long arm of chromosome 10 is deleted from the majority of both sporadic and hereditary juvenile polyps and that congenital deletion of the same region is associated with a syn-
Figure 4. Disomy or monosomy (LOH) assessed by FISH at the (A ) JPC locus (10q) compared with (B ) a control locus (21q). The mean number of FISH probes bound per nucleus of epithelial cells (,) or cells in the lamina propria ((), was determined for 39 juvenile polyps from 14 unrelated patients. Disomic nuclei have a mean value of Ç1.8 { 0.1, õ2.0, because approximately 10% of nuclei in thin sections are invariably not intact. We considered cells to have LOH (arrowheads) if the mean FISH signals per nucleus were decreased by 2 SD compared with normal. The criterion level of 1.6 is indicated by a dotted line. The 10q lamina propria deletions are statistically significant at a level of P õ 0.00000000001 by Fisher’s Exact Test compared with the control locus. Patients with juvenile polyps are identified (as in Table 1 and Figure 2) by numbers ranging from 80 to 144; the numbers after the dash indicate up to 7 different juvenile polyps from the same patient. Multiple sections from one polyp, number 117-1, are indicated by the letters a–e.
WBS-Gastro
1402 JACOBY ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
Figure 5. The microscopic view on the left is an H&E-stained section of a juvenile polyp (no. 106-3) showing the typical features of increased cellularity in the lamina propria and nondysplastic epithelium. FISH was used to identify cell types affected by deletion mutations at the JP1 locus on chromosome 10, by counting the number of signals bound to 100 cell nuclei of each type, in a series of 39 juvenile polyps (see Figure 3). The diagram on the right summarizes these FISH results, which show that epithelial cells are unaffected by the deletion mutations (always have two FISH signals indicating normal disomy at the JP1 locus). Vascular endothelial cells, polymorphonuclear lymphocytes, and smooth muscle cells are also normal (two FISH signals). The majority of cells in the lamina propria have deletion mutations affecting the JP1 locus (only one FISH signal).
drome of juvenile polyposis. The high frequency of LOH in juvenile polyps (83%), the clonal expansion of cells in the lamina propria of polyps affected by deletion, and the increased risk for malignancy in this syndrome suggest that this locus functions as a tumor suppressor. Cowden disease, associated with a variety of tumors including phenotypically similar intestinal hamartomas, has recently been linked to the same markers that we independently identified in this study.15 The possible allelic relationship between these syndromes should be clarified by molecular identification of the tumor suppressor gene(s) in this region.15 None of the known genes mapped to the 10q22–24 region are strong candidates: CD39, the apoptosis-signaling receptor FAS/APT1, retinol binding protein RBP4, interferon-induced protein IFI56, zinc finger protein ZNF32, homeobox T-cell lymphoma protein HOX11, and the homeotic gene PAX2. The FAS gene might be considered a plausible candidate because defective apoptosis has been implicated in colonic polyps,16 but the only known mutation affecting this gene in the lpr mouse has a phenotype of autoimmune disease similar to lupus erythematosis.17,18 Frequent LOH on 10q occurs in other tumor types, including prostate and endometrial adenocarcinoma, but the affected regions have not been mapped with enough detail to define their relationship to the juvenile polyposis/Cowden JPC locus.19,20 The relatively low rate of LOH / 5e1b$$0058
03-14-97 17:39:29
gasa
previously observed in adenocarcinomas of the colon21,22 is consistent with the clinical impression that only a minority of these tumors arise from juvenile polyps in the general population. However, we suspect that mutations inactivating the JPC locus would be observed more frequently in colon carcinomas selected from patients with hereditary juvenile polyposis. Juvenile polyps can develop foci of adenomatous change and, in some cases, progress to severe dysplasia or adenocarcinoma.1,5 – 7,23 – 25 Identification of the JPC gene should provide important insights into a unique alternative pathway of neoplastic development distinct from that previously described for gastrointestinal tumors, not originating in the epithelium. The identity of the neoplastic progenitor in the lamina propria is unknown, but morphologically the mutated cells have characteristics of a macrophage (histiocyte) or lymphocyte. Lymphocytes and histiocytes are both normal residents of the lamina propria and may regulate immunologic responses to mucosal injury or infection.26 Further studies will be necessary to thoroughly characterize the specific cell lineage(s) involved. One of the pathognomonic histological features of juvenile polyps is hypercellularity of the lamina propria,1 which appears to be now explained by the increased abundance of cells with JPC locus deletions in the lamina propria detected by FISH (Figure 5). The epithelial changes, including WBS-Gastro
April 1997
JUVENILE POLYPOSIS TUMOR SUPPRESSOR LOCUS 1403
dilated mucin-filled glands and the risk for adenomatous degeneration, could be caused by altered interactions with the mutated cells in the lamina propria because the epithelial cells do not share the clonal deletion mutation. Further studies should clarify how relationships between the lamina propria and the epithelium are perturbed during the development of neoplasia in juvenile polyps. Molecular diagnosis of juvenile polyps can now be performed as we described using paraffin sections readily available from most pathology archives for detection of JPC locus deletions at 10q22 by FISH. Another clinical implication of this study is that DNA-based predictive testing in informative juvenile polyposis families can now be used to indicate who would more likely benefit from endoscopic surveillance for early detection of gastrointestinal neoplasia.
12.
13.
14.
15.
16.
17.
18.
References 1. Jass JR, Williams CB, Bussey HJR, Morson BC. Juvenile polyposis: a precancerous condition. Histopathology 1988;13:619– 630. 2. Heiss KF, Schaffner D, Ricketts RR, Winn K. Malignant risk in juvenile polyposis coli: increasing documentation in the pediatric age group. J Pediatr Surg 1993;28:1188–1193. 3. Sturniolo GC, Montino MC, Dall’Igna F, D’Inca R, Missineo A, Cecchetto A, Previtera C, Riddell RH. Familial juvenile poyposis coli: results of endoscopic treatment and surveillance in two sisters. Gastrointest Endosc 1993;39:561–565. 4. Horrilleno EG, Eckert C, Ackerman LV. Polyps of the rectum and colon in children. Cancer 1957;10:1210–1220. 5. Stemper TJ, Kent TH, Summers RW. Juvenile polyposis and gastrointestinal carcinoma: a study of a kindred. Ann Intern Med 1975;83:639–646. 6. Jarvinen H, Franssila KO. Familial juvenile polyposis coli: increased risk of colorectal cancer. Gut 1984;25:792–800. 7. Giardiello FM, Hamilton SR, Kern SE, Johan G, Offerhaus A, Green PA, Celano P, Krush AJ, Booker SV. Colorectal neoplasia in juvenile polyposis or juvenile polyps. Arch Dis Childhood 1991;66: 971–975. 8. Leggett BA, Thomas LR, Knight N, Healey S, Chenevix-Trench G, Searle J. Exclusion of APC and MCC as the gene defect in one family with familial juvenile polyposis. Gastroenterology 1993; 105:1313–1316. 9. Jacoby RF, Schlack S, Sekhon G, Laxova R. Del(10)(q22.3q24.1) associated with hereditary juvenile polyposis. Am J Med Sci (in press). 10. Han K, Lee W, Harris CP, Simsimian RC, Lee K, Kang C, Meisner LF. Comparison of chromosome aberrations in leiomyoma and leiomyosarcoma using FISH on archival tissues. Cancer Genet Cytogenet 1994;74:19–24. 11. Lee W, Han K, Harris CP, Shim S, Kim S, Meisner LF. Use of
/ 5e1b$$0058
03-14-97 17:39:29
gasa
19.
20. 21. 22. 23. 24.
25.
26.
FISH to detect chromosomal translocations and deletions. Am J Pathol 1993;143:15–19. Dib C, Faure S, Fizames C, et al. A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature 1996;380:152–154. Jacoby RF, Marshall DJ, Schlack S, Kailas S, Harms B, Love R. Genetic instability associated with adenoma to carcinoma progression in hereditary non-polyposis colon cancer. Gastroenterology 1994;109:73–82. Newton MA, Wu S, Reznikoff CA. Assessing the significance of chromosome-loss data: where are suppressor genes for bladder cancer? Stat Med 1994;13:839–858. Nelen MR, Padberg GW, Peeters EAJ, et al. Localization of the gene for Cowden disease to chromosome 10q22-23. Nat Genet 1996;13:114–116. Que FG, and Gores GJ. Cell death by apoptosis: basic concepts and disease relevance for the gastroenterologist. Gastroenterology 1996;110:1238–1243. Cheng J, Liu C, Koopman WJ, Mountz JD. Characterization of human Fas gene exon/intron organization and promoter region. J Immunol 1995;154:1239–1245. Cheng J, Zhou T, Liu C, et al. Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 1994; 263:1759–1762. Peiffer SL, Herzog TJ, Tribune DJ, Mutch DG, Gersell DJ, and Goodfellow RJ. Allelic loss of sequences from the long arm of chromosome 10 and replication errors in endometrial cancers. Cancer Res 1996;55:1922–1926. Ittmann M. Allelic loss on chromosome 10 in prostate adenocarcinoma. Cancer Res 1996;56:2143–2147. Reichmann A, Martin P, Levin B. Chromosomal banding patterns in human large bowel cancer. Int J Cancer 1981;28:431–440. Vogelstein B, Fearon ER, Kern SE, et al. Allelotype of colorectal carcinomas. Science 1989;244:207–211. Kaschula RO. Mixed juvenile adenomatous and intermediate polyposis coli. Dis Colon Rectum 1971;14:368–374. Goodman ZD, Yardley JH, Milligan FD. Pathogenesis of colonic polyps in multiple juvenile polyposis. Cancer 1979;43:1906– 1913. Lipper S, Kahn LB, Sandler RS, et al. Multiple juvenile polyposis: a study of the pathogenesis of juvenile polyps and their relationships to the colonic adenomas. Hum Pathol 1981;12:804–813. Strober W, Brown WR. The mucosal immune system. In: Samter M, ed. Immunological diseases. Boston: Little, Brown, 1988:24.
Received December 3, 1996. Accepted January 14, 1997. Address requests for reprints to: Russell F. Jacoby, M.D., H6/516 Clinical Science Center, 600 Highland Avenue, Madison, Wisconsin 53792. Fax: (608) 262-7641. e-mail: [email protected]. Supported partially by grants P30 CA14520 and U01 CA59352 from the National Institutes of Health and a Merit Review grant from the Department of Veteran’s Affairs (to R.F.J.). A preliminary report of these findings was presented at the plenary session of the American Gastroenterological Association annual meeting, May 20, 1996, in San Francisco and recorded on CD-ROM. The authors thank Dr. Marta Voytovich and Dr. Jill Madsen for review of histopathological diagnoses and David J. Marshall and Greg Kuhlman for technical assistance.
WBS-Gastro
A Juvenile Polyposis Tumor Suppressor Locus at 10q22 Is Deleted From Nonepithelial Cells in the Lamina Propria RUSSELL F. JACOBY,*,‡,§ STEVEN SCHLACK,* CAROLYN E. COLE,‡ MARLENE SKARBEK,\ CHARLES HARRIS,\ and LORRAINE F. MEISNER\ *Section of Gastroenterology, Department of Medicine, xDepartment of Cytogenetics, and ‡University of Wisconsin Comprehensive Cancer Center, University of Wisconsin, Madison; and §William S. Middleton Memorial Veteran’s Hospital, Madison, Wisconsin
Background & Aims: Juvenile polyps are characterized by an abundant lamina propria that lacks smooth muscle and may contain cystically dilated glands, with epithelium that seems normal and is nondysplastic. Rarely, an autosomal dominant inheritance pattern occurs. The aim of this study was to test the hypothesis that the genetic defect in both sporadic juvenile polyps and hereditary juvenile polyposis involves loss of function for a tumor suppressor gene. Methods: Allelic losses were detected by comparing normal DNA with tumor DNA from a series of 47 juvenile polyps from 16 patients using polymerase chain reaction amplification of microsatellite markers and fluorescent in situ hybridization (FISH). Results: Somatic deletions at 10q22 were detected in 39 of 47 juvenile polyps (83%) from 16 unrelated patients with either hereditary or sporadic juvenile polyps, and the minimum overlap localized juvenile polyposis coli to the 3-cM interval D10S219– D10S1696. Fluorescent in situ hybridization shows that the cells affected by deletion mutation reside exclusively in the lamina propria, not in the epithelium. Conclusions: The location of a novel tumor suppressor gene on chromosome 10 that is affected by deletion mutation in the majority of juvenile polyps was mapped. Unlike adenomas and carcinomas of the colonic epithelium, juvenile polyps originate in the lamina propria.
J
uvenile polyps are characterized by an abundant lamina propria that lacks smooth muscle proliferation and contains cystically dilated mucin-filled glands lined by a typically normal epithelium.1 Solitary juvenile polyps, diagnosed in approximately 1% of children, account for the majority of gastrointestinal polyps found before adulthood.2,3 Juvenile polyposis coli ( JPC) is a rare autosomal dominant syndrome (Mendelian Inheritance in Man no. 174900) characterized by numerous juvenile polyps throughout the colon and occasionally elsewhere in the gastrointestinal tract. This hereditary cancer syndrome is associated with a high risk for colon adenocarcinoma1,4 – 7 but is not linked to the adenomatous polyposis coli gene.8 To investigate the possibility that another tumor suppressor gene is responsible, we initiated a ge/ 5e1b$$0058
03-14-97 17:39:29
gasa
nome-wide scan for allele losses in sporadic and hereditary juvenile polyps. A more focused search was made possible by our discovery of a unique patient with juvenile polyposis and multiple congenital anomalies associated with a small but cytogenetically detectable deletion.9
Materials and Methods Patients and Diagnostic Criteria for Juvenile Polyps Pathognomonic criteria for juvenile polyps of the gastrointestinal tract include an abundant lamina propria that lacks smooth muscle and may contain cystically dilated glands, with epithelium that appears normal and is nondysplastic.1 Microscopic sections from all paraffin tissue blocks obtained from other hospitals were examined by our pathologist to confirm diagnoses and identify areas for microdissection and DNA extraction. Sporadic cases are defined by solitary juvenile polyps in patients without any affected relatives. Hereditary juvenile polyposis was diagnosed in patients with positive family history and/or multiple juvenile polyps. Patient classification, age at diagnosis, and polyp multiplicity are shown in Table 1.
Fluorescent In Situ Hybridization Tissue sections (5-mm) were prepared as described previously.10,11 Briefly, sections were mounted on charged slides, baked overnight at 657C, deparaffinized in d-limonene twice and 100% ethanol twice for 10 minutes each, and air-dried. Sections were then incubated in 1 mol/L sodium thiocyanate for 20 minutes at 457C, rinsed in distilled water, incubated in Hanks’ basic salt solution and 1 mg/mL protease for 2 minutes at 457C, rinsed twice in 21 SSC (11 SSC is 0.15 mol/L NaCl plus 0.015 mol/L sodium citrate) for 1 minute each, and dehydrated through an ethanol series (70%-80%90%-100%). A digoxigenin-labeled cosmid probe from 10q22 at the HK1 locus in the JPC region (Oncor, Gaithersburg, MD) was hybridized overnight at 377C in a humidified incubaAbbreviations used in this paper: FISH, fluorescent in situ hybridization; JPC, juvenile polyposis coli; LOH, loss of heterozygosity; PCR, polymerase chain reaction. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
April 1997
JUVENILE POLYPOSIS TUMOR SUPPRESSOR LOCUS 1399
Table 1. Characteristics of Patients With Juvenile Polyps Patient (n Å 16)
Age at diagnosis (yr )
Sporadic 141 143 144 Juvenile polyposis 80 81 106 117 121 122 126 127 130 131 138 139 142
Polyp multiplicity
6 66 18
1 1 1
1 4 18 6 6 2 28 6
34 ú50 3 5 ú100 2 2 ú10 3 2 23 7 ú20
1 4 5 4
alleles were quantified using a Molecular Dynamics phosphoimager and ImageQuant software (Sunnyvale, CA). Loss of heterozygosity (LOH) was scored if the signal intensity for either of the two tumor alleles was decreased by at least 0.20 compared with the corresponding normal alleles, as calculated by the formula (N1/N2-T1/T2)(N1/N2). Statistical significance was assessed as previously described using a dynamic simulation computer algorithm developed for Bayesian analysis of chromosomal deletion data.14
Results Microsatellite Map of a Germline Deletion Associated With JPC
tor after denaturation under a sealed coverslip at 727C for 6 minutes. A control cosmid probe of equal size and similar fluorescent signal properties but located on an uninvolved chromosome (21q22.3) was hybridized to a paired adjacent section from each polyp. Each slide was then washed at 707C in 11 SSC for 5 minutes. A multistage detection system was used to detect probe bound to nuclear DNA by epifluorescence microscopy after reaction with a series of three antibodies: fluorescein isothiocyanate–labeled antidigoxigenin, rabbit anti-goat, and fluorescein isothiocyanate–labeled goat antirabbit (Boehringer Mannheim, Indianapolis, IN). The mean number of fluorescent signals per nucleus was scored for 100 cells of each type (epithelial or lamina propria) from each section. The same paraffin sections had been stained first for histological identification with Giemsa and then processed for fluorescent in situ hybridization (FISH) analysis. Horizontal and vertical coordinates were recorded for randomly selected regions of the polyp to be analyzed, and Giemsa images of these areas were compared with FISH images to correlate histology with deletion data. Deletions are scored if the tumor has a mean value of õ1.6 FISH signals per nucleus, which is 2 SD less than the mean value of 1.8 observed in normal control tissues.10,11 Statistical significance was assessed using Fisher’s Exact Test to compare results with the juvenile polyposis locus probe on chromosome 10 to a randomly selected control locus on chromosome 21.
Cytogenetic studies of a patient with juvenile polyposis and multiple congenital abnormalities of the head, extremities, and abdomen previously showed that the only chromosomal defect was an interstitial deletion of 10q22.3–q24.1.9 This cytogenetically evident abnormality affected only one homologue of chromosome 10 in the patient and was not observed in either of the parents who were unaffected by juvenile polyposis.9 We verified paternity by genotyping microsatellites at multi-
Microsatellite Allele Loss Studies Polymerase chain reaction (PCR) amplification was performed using normal genomic DNA extracted from peripheral blood lymphocytes compared with juvenile polyp DNA extracted from paraffin-embedded, formalin-fixed tissue sections. Microsatellite loci12 on chromosome 10 were amplified with the incorporation of [a-32P]adenosine triphosphate to label PCR products, which were then separated by electrophoresis on 6% polyacrylamide gels.13 Heterozygous microsatellite
/ 5e1b$$0058
03-14-97 17:39:29
gasa
Figure 1. Deletions associated with somatic LOH in a series of sporadic and familial juvenile polyps and a germline deletion in a patient with multiple congenital anomalies including juvenile polyps. Solid lines indicate deleted markers, dotted lines noninformative markers, and arrowheads the proximal and distal retained markers. D Pt, patient deletion.
WBS-Gastro
1400 JACOBY ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
ple loci (data not shown), confirming that the child’s deletion was a de novo event. The genetic map of this deletion was then defined by comparing the patient’s genotype to parental genotypes for 41 microsatellite markers on chromosome 10 that were informative in this family. The patient did not inherit any paternal alleles for 5 contiguous informative markers (D10S573, D10S541, D10S579, D10S564, and D10S571) spanning an interval of approximately 17 cM (Figure 1, right). Inheritance of alleles from both parents was shown for 29 markers that were proximal to D10S219 and for 7 markers that were distal to D10S603. Genetic markers mapping within the deleted interval were then used for allele loss studies in our series of sporadic and hereditary juvenile polyps from 16 unrelated patients (Figure 1, left). Somatic Deletions at Chr10q22 Occur in Most Juvenile Polyps We observed LOH in the majority of juvenile polyps for microsatellite markers on chromosome 10,
particularly in the interval D10S583–D10S569 (Figure 2). The most frequently affected genetic marker, D10S219, was deleted from 39 of 47 (83%) of juvenile polyps. The probability of deletion diminished as genetic map distance from D10S219 increased (Figure 3). Maps indicating the contiguous extent of deletion for each individual affected polyp were constructed from LOH data for the most informative markers and indicated that the minimum region of overlap involves the 3-cM interval D10S219–D10S1696 (Figure 1). Consistent with the hypothesized inherited tumor suppressor gene mutation in patients with multiple juvenile polyps, different tumors from the same polyposis patient always had somatic deletions affecting the same allele (Figure 2). FISH Identifies Cells Affected by Deletion Juvenile polyps are often removed at the time of colonoscopic examination by snare electrosurgical excision without resecting much adjacent normal tissue. FISH was useful to detect LOH when tissues of this type
Figure 2. Deletion of genetic markers on chromosome 10 in juvenile polyps. Allele loss for PCR-amplified microsatellites was quantified as described in Materials and Methods for juvenile polyp DNA compared with normal DNA. LOH is indicated by bold large typeface and dark grey shading for significant allele loss ratios ú0.30, or large typeface with medium grey shading if between 0.20 and 0.29, or light grey if minimally significant (between 0.17 and 0.19). ‘‘NI’’ indicates that the marker was not informative (alleles homozygous or not easily quantitated separately because similar size caused overlapping of the allele ladders on gels). Blank spaces indicate data that were unavailable because of poor amplification of paraffin section DNA, which occurs frequently if the microsatellite is ú200 base pairs. ‘‘S’’ or ‘‘L’’ indicates which of the microsatellite alleles (shorter or longer) showed LOH as decreased relative intensity of that allele.
/ 5e1b$$0058
03-14-97 17:39:29
gasa
WBS-Gastro
April 1997
JUVENILE POLYPOSIS TUMOR SUPPRESSOR LOCUS 1401
Deletion mutations exclusively affected cells in the lamina propria in 29 of 30 juvenile polyps (Figure 4A). Epithelial cells, smooth muscle cells, fibroblasts, and polymorphonuclear leukocytes did not seem to be involved because FISH shows that their nuclei are all disomic (normal). Light microscopy showed that the lamina propria cells affected by deletion appeared to have round nuclei and scant cytoplasm typical of lymphocytes and macrophages or histiocytes.
Discussion
Figure 3. Proportion of juvenile polyps with LOH for microsatellite markers on Chr 10q in the juvenile polyposis/Cowden critical region. D10S196 is near the centromere.
were available from the pathology archives, but PCR allele loss studies could not be performed because normal DNA was not available from the decreased patient or preserved tissue. For FISH analysis of juvenile polyps, we obtained a cosmid clone for the gene HK1 that maps (Human Genome Database) between D10S201 and D10S564 near the minimum region of overlap of germline and somatic JPC deletions. Using this probe that appears to be closely linked to the JPC locus, somatic deletions were shown by FISH in 30 of 39 (77%) of juvenile polyps (Figure 4A). FISH results agreed with LOH data obtained by PCR analysis of microsatellite alleles in 16 of 19 cases in which both techniques could be used on the same polyp specimens (80-1, 80-2, 803, 80-5, 117-2, 117-3, 121-1, 126-2, 131-1, 138-1, 139-1, 139-2, 139-3, 139-4, 139-5, and 139-6; compare Figure 2 and Figure 4A). The JPC locus is apparently separated from these markers by a small genetic distance because 2 of 19 polyps (0.10) deleted only the FISH and 1 of 19 (0.05) deleted only the PCR markers. The high rate of deletion observed at chromosome 10q22 probably has a functional basis. Although common in colon carcinomas, aneuploidy is not a characteristic of benign juvenile polyps. As expected, deletions were not observed at a control locus on another chromosome (chromosome 21), indicating that nonspecific aneuploidy is unlikely to account for the deletions we observed at the putative JPC locus on Chr10q (Figure 4B). FISH analysis is particularly useful because it can identify mutations in individual cells in tissue sections from tumors composed of multiple cell types, e.g., juvenile polyps. Cellular architecture was recorded after Giemsa staining, and then FISH was used to detect deletion mutations affecting identified cells in the same section. / 5e1b$$0058
03-14-97 17:39:29
gasa
We show that a locus mapping to the long arm of chromosome 10 is deleted from the majority of both sporadic and hereditary juvenile polyps and that congenital deletion of the same region is associated with a syn-
Figure 4. Disomy or monosomy (LOH) assessed by FISH at the (A ) JPC locus (10q) compared with (B ) a control locus (21q). The mean number of FISH probes bound per nucleus of epithelial cells (,) or cells in the lamina propria ((), was determined for 39 juvenile polyps from 14 unrelated patients. Disomic nuclei have a mean value of Ç1.8 { 0.1, õ2.0, because approximately 10% of nuclei in thin sections are invariably not intact. We considered cells to have LOH (arrowheads) if the mean FISH signals per nucleus were decreased by 2 SD compared with normal. The criterion level of 1.6 is indicated by a dotted line. The 10q lamina propria deletions are statistically significant at a level of P õ 0.00000000001 by Fisher’s Exact Test compared with the control locus. Patients with juvenile polyps are identified (as in Table 1 and Figure 2) by numbers ranging from 80 to 144; the numbers after the dash indicate up to 7 different juvenile polyps from the same patient. Multiple sections from one polyp, number 117-1, are indicated by the letters a–e.
WBS-Gastro
1402 JACOBY ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
Figure 5. The microscopic view on the left is an H&E-stained section of a juvenile polyp (no. 106-3) showing the typical features of increased cellularity in the lamina propria and nondysplastic epithelium. FISH was used to identify cell types affected by deletion mutations at the JP1 locus on chromosome 10, by counting the number of signals bound to 100 cell nuclei of each type, in a series of 39 juvenile polyps (see Figure 3). The diagram on the right summarizes these FISH results, which show that epithelial cells are unaffected by the deletion mutations (always have two FISH signals indicating normal disomy at the JP1 locus). Vascular endothelial cells, polymorphonuclear lymphocytes, and smooth muscle cells are also normal (two FISH signals). The majority of cells in the lamina propria have deletion mutations affecting the JP1 locus (only one FISH signal).
drome of juvenile polyposis. The high frequency of LOH in juvenile polyps (83%), the clonal expansion of cells in the lamina propria of polyps affected by deletion, and the increased risk for malignancy in this syndrome suggest that this locus functions as a tumor suppressor. Cowden disease, associated with a variety of tumors including phenotypically similar intestinal hamartomas, has recently been linked to the same markers that we independently identified in this study.15 The possible allelic relationship between these syndromes should be clarified by molecular identification of the tumor suppressor gene(s) in this region.15 None of the known genes mapped to the 10q22–24 region are strong candidates: CD39, the apoptosis-signaling receptor FAS/APT1, retinol binding protein RBP4, interferon-induced protein IFI56, zinc finger protein ZNF32, homeobox T-cell lymphoma protein HOX11, and the homeotic gene PAX2. The FAS gene might be considered a plausible candidate because defective apoptosis has been implicated in colonic polyps,16 but the only known mutation affecting this gene in the lpr mouse has a phenotype of autoimmune disease similar to lupus erythematosis.17,18 Frequent LOH on 10q occurs in other tumor types, including prostate and endometrial adenocarcinoma, but the affected regions have not been mapped with enough detail to define their relationship to the juvenile polyposis/Cowden JPC locus.19,20 The relatively low rate of LOH / 5e1b$$0058
03-14-97 17:39:29
gasa
previously observed in adenocarcinomas of the colon21,22 is consistent with the clinical impression that only a minority of these tumors arise from juvenile polyps in the general population. However, we suspect that mutations inactivating the JPC locus would be observed more frequently in colon carcinomas selected from patients with hereditary juvenile polyposis. Juvenile polyps can develop foci of adenomatous change and, in some cases, progress to severe dysplasia or adenocarcinoma.1,5 – 7,23 – 25 Identification of the JPC gene should provide important insights into a unique alternative pathway of neoplastic development distinct from that previously described for gastrointestinal tumors, not originating in the epithelium. The identity of the neoplastic progenitor in the lamina propria is unknown, but morphologically the mutated cells have characteristics of a macrophage (histiocyte) or lymphocyte. Lymphocytes and histiocytes are both normal residents of the lamina propria and may regulate immunologic responses to mucosal injury or infection.26 Further studies will be necessary to thoroughly characterize the specific cell lineage(s) involved. One of the pathognomonic histological features of juvenile polyps is hypercellularity of the lamina propria,1 which appears to be now explained by the increased abundance of cells with JPC locus deletions in the lamina propria detected by FISH (Figure 5). The epithelial changes, including WBS-Gastro
April 1997
JUVENILE POLYPOSIS TUMOR SUPPRESSOR LOCUS 1403
dilated mucin-filled glands and the risk for adenomatous degeneration, could be caused by altered interactions with the mutated cells in the lamina propria because the epithelial cells do not share the clonal deletion mutation. Further studies should clarify how relationships between the lamina propria and the epithelium are perturbed during the development of neoplasia in juvenile polyps. Molecular diagnosis of juvenile polyps can now be performed as we described using paraffin sections readily available from most pathology archives for detection of JPC locus deletions at 10q22 by FISH. Another clinical implication of this study is that DNA-based predictive testing in informative juvenile polyposis families can now be used to indicate who would more likely benefit from endoscopic surveillance for early detection of gastrointestinal neoplasia.
12.
13.
14.
15.
16.
17.
18.
References 1. Jass JR, Williams CB, Bussey HJR, Morson BC. Juvenile polyposis: a precancerous condition. Histopathology 1988;13:619– 630. 2. Heiss KF, Schaffner D, Ricketts RR, Winn K. Malignant risk in juvenile polyposis coli: increasing documentation in the pediatric age group. J Pediatr Surg 1993;28:1188–1193. 3. Sturniolo GC, Montino MC, Dall’Igna F, D’Inca R, Missineo A, Cecchetto A, Previtera C, Riddell RH. Familial juvenile poyposis coli: results of endoscopic treatment and surveillance in two sisters. Gastrointest Endosc 1993;39:561–565. 4. Horrilleno EG, Eckert C, Ackerman LV. Polyps of the rectum and colon in children. Cancer 1957;10:1210–1220. 5. Stemper TJ, Kent TH, Summers RW. Juvenile polyposis and gastrointestinal carcinoma: a study of a kindred. Ann Intern Med 1975;83:639–646. 6. Jarvinen H, Franssila KO. Familial juvenile polyposis coli: increased risk of colorectal cancer. Gut 1984;25:792–800. 7. Giardiello FM, Hamilton SR, Kern SE, Johan G, Offerhaus A, Green PA, Celano P, Krush AJ, Booker SV. Colorectal neoplasia in juvenile polyposis or juvenile polyps. Arch Dis Childhood 1991;66: 971–975. 8. Leggett BA, Thomas LR, Knight N, Healey S, Chenevix-Trench G, Searle J. Exclusion of APC and MCC as the gene defect in one family with familial juvenile polyposis. Gastroenterology 1993; 105:1313–1316. 9. Jacoby RF, Schlack S, Sekhon G, Laxova R. Del(10)(q22.3q24.1) associated with hereditary juvenile polyposis. Am J Med Sci (in press). 10. Han K, Lee W, Harris CP, Simsimian RC, Lee K, Kang C, Meisner LF. Comparison of chromosome aberrations in leiomyoma and leiomyosarcoma using FISH on archival tissues. Cancer Genet Cytogenet 1994;74:19–24. 11. Lee W, Han K, Harris CP, Shim S, Kim S, Meisner LF. Use of
/ 5e1b$$0058
03-14-97 17:39:29
gasa
19.
20. 21. 22. 23. 24.
25.
26.
FISH to detect chromosomal translocations and deletions. Am J Pathol 1993;143:15–19. Dib C, Faure S, Fizames C, et al. A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature 1996;380:152–154. Jacoby RF, Marshall DJ, Schlack S, Kailas S, Harms B, Love R. Genetic instability associated with adenoma to carcinoma progression in hereditary non-polyposis colon cancer. Gastroenterology 1994;109:73–82. Newton MA, Wu S, Reznikoff CA. Assessing the significance of chromosome-loss data: where are suppressor genes for bladder cancer? Stat Med 1994;13:839–858. Nelen MR, Padberg GW, Peeters EAJ, et al. Localization of the gene for Cowden disease to chromosome 10q22-23. Nat Genet 1996;13:114–116. Que FG, and Gores GJ. Cell death by apoptosis: basic concepts and disease relevance for the gastroenterologist. Gastroenterology 1996;110:1238–1243. Cheng J, Liu C, Koopman WJ, Mountz JD. Characterization of human Fas gene exon/intron organization and promoter region. J Immunol 1995;154:1239–1245. Cheng J, Zhou T, Liu C, et al. Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 1994; 263:1759–1762. Peiffer SL, Herzog TJ, Tribune DJ, Mutch DG, Gersell DJ, and Goodfellow RJ. Allelic loss of sequences from the long arm of chromosome 10 and replication errors in endometrial cancers. Cancer Res 1996;55:1922–1926. Ittmann M. Allelic loss on chromosome 10 in prostate adenocarcinoma. Cancer Res 1996;56:2143–2147. Reichmann A, Martin P, Levin B. Chromosomal banding patterns in human large bowel cancer. Int J Cancer 1981;28:431–440. Vogelstein B, Fearon ER, Kern SE, et al. Allelotype of colorectal carcinomas. Science 1989;244:207–211. Kaschula RO. Mixed juvenile adenomatous and intermediate polyposis coli. Dis Colon Rectum 1971;14:368–374. Goodman ZD, Yardley JH, Milligan FD. Pathogenesis of colonic polyps in multiple juvenile polyposis. Cancer 1979;43:1906– 1913. Lipper S, Kahn LB, Sandler RS, et al. Multiple juvenile polyposis: a study of the pathogenesis of juvenile polyps and their relationships to the colonic adenomas. Hum Pathol 1981;12:804–813. Strober W, Brown WR. The mucosal immune system. In: Samter M, ed. Immunological diseases. Boston: Little, Brown, 1988:24.
Received December 3, 1996. Accepted January 14, 1997. Address requests for reprints to: Russell F. Jacoby, M.D., H6/516 Clinical Science Center, 600 Highland Avenue, Madison, Wisconsin 53792. Fax: (608) 262-7641. e-mail: [email protected]. Supported partially by grants P30 CA14520 and U01 CA59352 from the National Institutes of Health and a Merit Review grant from the Department of Veteran’s Affairs (to R.F.J.). A preliminary report of these findings was presented at the plenary session of the American Gastroenterological Association annual meeting, May 20, 1996, in San Francisco and recorded on CD-ROM. The authors thank Dr. Marta Voytovich and Dr. Jill Madsen for review of histopathological diagnoses and David J. Marshall and Greg Kuhlman for technical assistance.
WBS-Gastro