Intestinal Cancer in Patients with a Germline Mutation in the Down-Regulated in Adenoma (DRA) Gene
Total Page:16
File Type:pdf, Size:1020Kb
Oncogene (1998) 16, 681 ± 684 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 SHORT REPORT Intestinal cancer in patients with a germline mutation in the down-regulated in adenoma (DRA) gene Akseli Hemminki1, Pia HoÈ glund1, Eero Pukkala2, Reijo Salovaara1,3, Heikki JaÈ rvinen4, Reijo Norio5 and Lauri A Aaltonen1 Departments of 1Medical Genetics, and 3Pathology, Haartman Institute, PO Box 21 (Haartmaninkatu 3), 00014 University of Helsinki, Finland; 2Finnish Cancer Registry, Liisankatu 21 B, 00171 Helsinki, Finland; 4Second Department of Surgery, Helsinki University Central Hospital, Haartmaninkatu 4, 00250 Helsinki, Finland; 5Department of Medical Genetics, the Family Federation of Finland, PO Box 849 (Kalevankatu 16), 00101 Helsinki, Finland A recent study has revealed that germline mutations of and in dedierentiated states of mucosa, such as in the down-regulated in adenoma (DRA) gene are a likely embryonic tissues or in neoplasia, DRA expression is cause of a recessive intestinal absorption defect, not detected (Schweinfest et al., 1993; Silberg et al., congenital chloride diarrhea. This ®nding was in 1995; HoÈ glund et al., 1996). Although the association accordance with previous works showing that DRA between lost DRA expression and neoplasia is encodes a sodium independent transporter for sulfate apparent, it is not known if this loss has a role in and oxalate. Although DRA was originally reported as a tumorigenesis, or if it is merely a consequence of the candidate tumor suppressor, these studies have ques- dedierentiated status of the epithelium. tioned the relevance of DRA in cancer. To evaluate Congenital chloride diarrhea (CLD) is an autosomal whether further studies on the role of DRA in recessive disorder characterized by voluminous watery tumorigenesis are still of interest, we examined whether stools containing a high concentration of chloride individuals carrying germline DRA mutations have an (McKusick, 1994). Other features include hydramnios excess of intestinal cancer. Cancer status of 229 and premature birth. Heterozygous carriers are members of 36 Finnish congenital chloride diarrhea symptomless. The majority of Finnish CLD patients families (44 homozygous patients, 70 heterozygous originate from the Eastern part of Finland (see map in parents, and 115 grandparents at 50% risk of being a HoÈ glund et al., 1995). The population of this area is DRA mutation carrier) was checked at the Finnish approximately 1.35 million people. Strong evidence was Cancer Registry and the risk of intestinal cancer was recently presented indicating that CLD is caused by found slightly elevated (standardized incidence ratio 3.4, germline mutations in the DRA gene. A three base pair 95% con®dence interval 1.4 ± 7.0, P50.05). While this in-frame deletion (DV317) and a C307W change were result does not unambiguously demonstrate an increased found in all Finnish CLD chromosomes studied intestinal cancer risk in DRA mutation carriers, it should (n=64). However, the C307W seemed like a poly- promote further studies to determine the possible role of morphism because of its high population frequency. DRA in cancer. Furthermore, it was not disease-causing when present without DV317 (HoÈ glund et al., 1996). Keywords: DRA; CLD; cancer; neoplasia; tumor The association between digestive tract cancer and suppressor homozygous defects in another anion transporter, the cystic ®brosis transmembrane conductance regulator gene, CFTR, has been recently documented (Neglia et al., 1995). Some of the molecular mechanisms that Four years ago, the down-regulated in adenoma might associate gastrointestinal cancer predisposition (DRA) gene was cloned (Schweinfest et al., 1993) and and defects in CFTR have been recently revealed consecutively mapped to 7q22-q31.1 (Taguchi et al., (Abraham et al., 1996). To determine whether the 1994). DRA has a great deal of structural similarity germline defect in DRA has relevance in intestinal with the diastrophic dysplasia sulphate transporter cancer, we utilized two unique data sets that were gene, which led investigators to presume a functional available to us. First, all identi®ed patients (n=44), similarity between these two genes (HaÈ stbacka et al., parents (n=70) and grandparents (n=115) of Finnish 1994). Structural similarity with a rat liver sulfate CLD families were checked for cancer at the Finnish transporter led other researchers to study the function Cancer Registry. Second, 728 patients with colorectal of DRA, and the gene was shown to encode a sodium neoplasia, 485 of them from the high CLD frequency independent transporter for sulfate and oxalate (Silberg area, were studied for the presence of the DV317 et al., 1995). The expression of DRA is limited to the mutation. 416 control individuals were also analysed mature mucosal cells of the small and large intestine for this mutation. (Schweinfest et al., 1993; Silberg et al., 1995; HoÈ glund As a result of the centralized health care system of et al., 1996) and to the prostate (HoÈ glund et al., 1996), Finland, diagnostic hallmarks, and well-studied genealogy of the disease (Norio et al., 1971) the ascertainment of CLD is believed to be almost Correspondence: LA Aaltonen complete. Altogether 36 CLD families were known Received 16 April 1997; revised 1 September 1997; accepted 1 to us. In 27 families the three base pair deletion September 1997 DV317 has been demonstrated to cause the disease Mutation in DRA and intestinal cancer AHemminkiet al 682 (HoÈ glund et al., 1996), from the rest of the families no Table 1 Cancer in CLD familes DNA samples were available. In all families, the Number patients were diagnosed by characteristic clinical of cases Average features, and by demonstrating high fecal chloride (parents/ age at content. Previous genealogical studies have identi®ed Cancer location Cancer histology grandparents)diagnosis parents and grandparents of the patients (HoÈ glund et Intestine 7a (3/4) 58 al., 1995; Norio et al., 1971). Since the disease is Rectum adenocarcinoma 3 (2/1) 47 Sigmoid colon adenocarcinoma 2 (0/2) 69 recessive, all parents are obligate heterozygous Ascending colon adenocarcinoma 1 (0/1) 72 carriers, and 50% of the grandparents carry one Ileum adenocarcinoma 1 (1/0) 53 defective allele. One mother of a CLD patient was Lung 8 (1/7) 67 exluded from this study because paternal uniparental microcellular carcinoma 4b (0/4) 69 adenocarcinoma 2 (1/1) 60 isodisomy of chromosome 7 caused the disease in the unknown 2 (0/2) 70 aected child (HoÈ glund et al., 1994). Data was Other 18 (1/17) 68 available from 44 patients, 70 parents and 115 Stomach adenocarcinoma 2 (0/2) 74 grandparents. One parent, and 33 grandparents were Prostrate adenocarcinoma 2 (0/2) 70 lost during follow up, usually because they had died Endometrium adenocarcinoma 1 (0/1) 62 Ovary cystadenocarcinoma 1 (1/0) 42 before personal identi®cation numbers necessary for Breast ductal carcinoma 1 (0/1) 79 computer linking were introduced in 1967. The Kidney renal cell carcinoma 1 (0/1) 66 Finnish Cancer Registry has legal status, is popula- Thyroid papillary carcinoma 1 (0/1) 73 tion based, and has almost 100% coverage since 1953 Eye (choroid epithelium) melanoma 1 (0/1) 57 (KylloÈ nen et al., 1987; Teppo et al., 1994). Follow-up Bladder unknown 1 (0/1) 75 for cancer started at the birth of the ®rst CLD patient Skin basocellular carcinoma 7 (0/7) 69 in each family (for the grandparents follow-up started at the birth of the CLD patient's parent), or on the Total 33 (5/28) 66 1st of January 1967, whichever was later, and ended Results from the Cancer Registry follow-up: cancers found in CLD at death or 31st of December 1993, whichever patients' parents and grandparents. aNumber of cases compared with occurred ®rst. The numbers of observed cases and the age and sex adjusted expected value was statistically signi®cant with the Mantel-Haenzel chi-square test, P50.05. bNumber of cases person-years at risk were counted, by 5 year age compared with the age and sex adjusted expected value was groups, separately for the three calendar periods statistically signi®cant with the Mantel-Haenzel chi-square test, (1967 ± 1975, 1976 ± 1984 and 1985 ± 1993). The ex- P50.01 pected numbers of cases for total cancer and for speci®c cancer types were calculated by multiplying the number of person-years in each age group by the In the unselected patients with colorectal neoplasia, corresponding average cancer incidence in the whole DV317 was not found in the low frequency areas Finnish population during the period of observation. (South-Central Finland) in 243 patients (223 carcino- To calculate the standard incidence ratio, the observed mas, 20 adenomas). In the high CLD frequency area 6/ number of cases (see Table 1) was divided by the 485 (1.2%) patients, 2/198 (1.0%) with adenoma and expected number. The statistical signi®cance was 4/287 (1.4%) with carcinoma, were heterozygous for tested by the Mantel ± Haenzel chi-square test, on DV317 (Figure 1). Four (1.0%) of the 416 controls the presumption that the number of observed cases displayed the mutation (dierence not statistically followed a Poisson distribution. signi®cant). The CLD patients were followed for 697 person In three out of the four unselected patients with years, parents 1250 person years, and grandparents DV317, carcinoma DNA was also available. Loss of 2430 person years. At the end of follow-up, there were the wild type allele could not be detected in any of no cases of cancer reported to the Cancer Registry these tumors (data not shown). The same was true with among the CLD patients (homozygous for the genetic four paran embedded tumors from the CLD families defect), all 32 years old or younger. Two of the parents displaying heterozygosity for DV317. (obligate heterozygous carriers of the defect) had The study focused on occurrence of intestinal altogether three primary intestinal cancers (P50.01, cancer, but incidentally as many as eight cases of standardized incidence ratio (SIR) 20.0, 95% con- lung cancer were detected during the cancer follow-up ®dence interval (CI) 4.1 ± 58.5).