Expression of DMBT1, a Candidate Tumor Suppressor Gene, Is Frequently Lost in Lung Cancer1

[CANCER RESEARCH 59, 1846–1851, April 15, 1999] Advances in Brief

Weiguo Wu, Bonnie L. Kemp, Monja L. Proctor, Adi F. Gazdar, John D. Minna, Waun Ki Hong, and Li Mao2 Departments of Thoracic/Head and Neck Medical Oncology [W. W., W. K. H., L. M.] and Pathology [B. L. K.], The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030, and Hamon Center for Therapeutic Oncology Research, The University of Texas Southwestern Medical Center, Dallas, Texas 75235 [M. L. P., A. E. G., J. D. M.]

Abstract long period of time, even with a successful, nationwide antismoking campaign. DMBT1 is a candidate tumor suppressor gene located at 10q25.3–26.1. DMBT1 was cloned through a representational differential analysis Homozygous deletion of the gene was found in a subset of medulloblas- that is used to identify potential homozygous deletions in target toma and glioblastoma multiforme; lack of expression was noted in the genomic DNA (6). The gene was localized to 10q25.3–q26.1, a region majority of these tumors. In adult tissues, DMBT1 is highly expressed only 3 in lung and small intestine tissues, indicating its important role in these with frequent LOH in many types of human cancers including lung organs. By analyzing lung cancer cell lines and primary lung tumors using cancer (7–10). Intragenic homozygous deletions of DMBT1 were reverse transcription-PCR, we found that 100% (20 of 20) of small cell found in 23–38% of medulloblastoma and glioblastoma multiforme lung cancer (SCLC) cell lines and 43% (6 of 14) of non-small cell lung cell lines and primary glioblastoma multiforme (6, 11). Interestingly, cancer (NSCLC) cell lines lacked DMBT1 expression. Furthermore, 45% loss of DMBT1 gene expression was found in 80% of these tumor cell (9 of 20) of the primary NSCLCs exhibited a markedly low level of gene lines (6). These data suggest that DMBT1 is a candidate tumor expression compared with corresponding normal lung tissues, indicating suppressor gene and may play an important role in brain tumorigen- that lack of gene expression also occurs in primary lung cancers. To esis. Deduced DMBT1 protein contains at least nine SRCR domains determine the potential mechanisms for lack of DMBT1 expression in lung that allow the gene to be classified as a new member of the SRCR cancer, we analyzed tumor cell lines for potential intragenic homozygous superfamily. Some members of this superfamily have been linked to deletions of the gene and found such homozygous deletions in 10% (4 of 40) of SCLC cell lines but in none of 14 NSCLC cell lines. Moreover, the initiation of cell proliferation and differentiation in immune system loss of expression could not be rescued by treatment with a demethylation tissue and other tissues (12–15) or have been associated with the agent (5-azacytidine) in two NSCLC cell lines lacking DMBT1 expression, polarity of epithelial cells (16). A leader sequence present in DMBT1 suggesting that de novo methylation of the promoter region of the gene is protein, together with the presence of SRCR, CUB, and ZP domains unlikely to play a role in inactivation of the gene. We then sequenced the and N-glycosylation sites, suggests that DMBT1 is likely a secreted or whole coding region of DMBT1 in 8 NSCLC cell lines that expressed membrane protein (12–17). In adult humans, DMBT1 is highly ex- DMBT1 and 20 primary NSCLCs. A potential point mutation at codon 52 pressed only in lung and small intestine, suggesting that the protein was detected in a NSCLC cell line and resulted in an amino acid change has an important role in these tissues (6). Furthermore, the gene is from serine to tryptophan. Three common polymorphisms were also located within one of the two minimally deleted regions at chromo- detected in tissues analyzed. Our data demonstrate that DMBT1 expres- some 10q identified in primary SCLCs (10), raising the possibility that sion is frequently lost in lung cancer due to gene deletion and to other not the gene may be also important in lung tumorigenesis. In this study, yet identified mechanisms, suggesting that inactivation of DMBT1 may play an important role in lung tumorigenesis. we investigated whether DMBT1 is altered in lung cancer and found an intragenic homozygous deletion of the gene in 10% of the SCLC Introduction cell lines tested and lack of the gene expression in all of the SCLC cell lines tested. Furthermore, more than 40% of NSCLC cell lines and Lung cancer is the leading cause of cancer-related death in the primary NSCLCs also lacked DMBT1 expression. The identification United States. More than 171,000 new cases are estimated to occur in of a potential point mutation in the gene in a NSCLC cell line further the United States in 1998 (1). Despite improvements in detection and supports the notion that DMBT1 may play an important role in lung treatment of the disease over the past two decades, the overall 5-year tumorigenesis. survival rate for patients with lung cancer is still Ͻ15%. It is clear that cigarette smoking is a major cause of lung cancer and that smoking Materials and Methods cessation can reduce the risk for lung cancer development. However, genetic damage caused by cigarette smoking may still exist in former Cell Lines and Tissue Specimens. All of the lung cancer cell lines have smokers (2, 3). In fact, according to recent reports, about half of all been deposited in the American Type Culture Collection (Rockville, MD). The SCLC cell lines used in this study were NCI-H60, -H69, -H82, -H128, -H146, new patients with lung cancer in the United States are former smokers -H182, -H187, -H196, -H211, -H249, -H289, -H345, -H378, -H524, -H526, (4, 5), suggesting that lung cancer incidence may remain high for a -H711, -H719, -H735, -H738, -H740, -H748, -H841, -H847, -H865, -H889, -H1045, -H1092, -H1105, -H1184, -H1238, -H1284, -H1304, -H1417, -H1450, Received 12/14/98; accepted 3/2/99. -H1514, -H1522, -H1618, -H1672, -H1688, -H1882, -H1963, -H2081, -H2107, The costs of publication of this article were defrayed in part by the payment of page -H2141, -H2171, and -H2227. The NSCLC cell lines used were NCI-H460, charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. -H157, -H226, -H1792, -H522, -H292, -H1944, -H1648, A549, -H727, S966, 1 Supported in part by American Cancer Society Grant RPG-98-054, National Cancer A427, Calu-1, and SK-Mes-1. These cell lines were grown in RPMI 1640 with Institute Grants PO1 CA74173 and Lung SPORE Grant CA 70907, and NIH Fellowship 10% fetal bovine serum. Primary NSCLC samples and corresponding normal Training Grant T32 CA 66187. W. K. H. is an American Cancer Society Clinical Research Professor. 2 To whom requests for reprints should be addressed, at Molecular Biology Labora- 3 The abbreviations used are: LOH, loss of heterozygosity; SCLC, small cell lung tory, Department of Thoracic/Head and Neck Medical Oncology, The University of Texas cancer; NSCLC, non-small cell lung cancer; RT-PCR, reverse transcription-PCR; STS, M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030. Phone: specific tag sequence; SRCR, scavenger receptor cysteine-rich; CUB, complement sub- (713) 792-6363; Fax: (713) 796-8865; E-mail: [email protected]. components Clr/Cls, Uegf, Bmp1; ZP, zona pellucida. 1846

lung tissues were obtained from surgical resection specimens collected by the DNA polymerase (Life Technologies, Inc.). DNA was amplified for 35 cycles Department of Pathology at M. D. Anderson Cancer Center and stored at at 95°C for 30 s, 56–60°C for 60 s, and 70°C for 60 s, followed by a 5-min Ϫ80°C until the experiment. extension at 70°C in a temperature cycler (Hybaid; Omnigene) in 500-␮l DNA and RNA Extraction and cDNA Synthesis. Genomic DNA was plastic tubes. PCR products were separated on a 7% polyacrylamide-urea- isolated from SCLC and NSCLC cell lines by proteinase-K digestion (0.1 formamide gel and then exposed to X-ray film. LOH was defined as a Ͼ50% mg/ml) at 50°C overnight, followed by phenol/chloroform extraction and reduction of intensity by visual inspection in one of the two alleles as ethanol precipitation. For fresh tissues, samples were sectioned using a cryostat compared with that seen in the corresponding normal control. microtome. Selected sections were stained with H&E and reviewed for the presence of tumor cells. Only tumor samples containing about 70% or more Results tumor cells and normal lung tissues without tumor cells were used for DNA extraction. The proteinase-K digestion and DNA purification methods are the To determine the expression pattern of the DMBT1 gene in lung same as those used for cell lines. For total RNA extraction from cell lines, the cancer, we analyzed 20 SCLC cell lines, 14 NSCLC cell lines, and 20 cell pellets were quickly suspended in RNAzol B solution (Biotecx Labora- primary NSCLC tumors by RT-PCR. We used two primer sets to tories, Inc., Houston, TX) to lyse the cells. For fresh tissue samples, 20–50 mg amplify DMBT1 cDNA fragments separately at two different regions of tissue were mechanically homogenized in 1 ml of RNAzol B solution to verify expression status. The two fragments (165 and 909 bp) were ␮ (Biotecx Laboratories, Inc., Houston, TX). Five g of total RNA from each located at the 5Ј end and the 3Ј end of the coding sequences, respec- sample were then used to synthesize single-strand cDNA using SUPER- tively. Consistent results have been obtained by using the two differ- SCRIPT II RNase HϪ reverse transcriptase (Life Technologies, Grand Island, ent primer sets. No PCR product could be obtained by amplifying NY), following the manufacturer’s protocol. The synthesized cDNA was used Ј immediately for PCR amplification or stored at Ϫ20°C for further analysis. genomic DNA using the primer set for the 909-bp 3 end fragment, PCR Analysis. PCR for genomic DNA was conducted using a 20-ng indicating the presence of a large intron or introns within the cDNA genomic DNA sample as described previously (10). Primer sets used for fragment or primers designed to cross-over an intron/exon boundary. detecting deletions of DMBT1 gene were 36k, g14, g14ext, and 101n as Although the primer set for the 165-bp fragment could also be used to reported previously (6). A primer set for ␤-actin was used to amplify a 415-bp amplify genomic DNA, our negative control experiment showed that fragment as a control for DNA quality. The primer sequences for the ␤-actin genomic DNA contamination in these cDNA samples was minimal fragment were 5Ј-CTCACATCGTGCCCATCTAT-3Ј (sense) and 5Ј-GATC- (data not shown) and unlikely to change the experimental result. For Ј CTTGCGAATATCCACA-3 (antisense). For multiplex PCR analysis, one each cDNA specimen from the cell lines, a set of primers for a 543-bp primer set for a target molecule was used together with the primer set for the cDNA fragment of ␤-actin was used to verify the quality of cDNA ␤-actin fragment. To amplify cDNA, each of the PCR reactions was performed specimens. We found that DMBT1 was not expressed in any of the 20 in a 25-␮l volume containing 0.5 ␮l of reverse transcriptase reaction mixture, (100%) SCLC cell lines nor in 6 (43%) of the 14 NSCLC cell lines 3% DMSO, 1.5 mM of deoxynucleotide triphosphate, 6.7 mM of MgCl2, 16.6 ␤ ␮ (Fig. 1, A and B). Twenty primary NSCLC tumors (10 adenocarcino- mM of (NH4)2SO4,67mM of Tris, 10 mM of -mercaptoethanol, 6.7 M EDTA, 2.5 units of Taq polymerase (Life Technologies), and 0.4 ␮M each of mas and 10 squamous cell carcinomas) were also analyzed for the primers. Reactions without templates were used as negative controls to rule DMBT1 expression by semiquantitative multiplex PCR in which out the possibility of contamination. Thermal cycling was performed in a ␤-actin was used as an internal control for both cDNA quality and temperature cycler (Hybaid; Omnigene, Woodbridge, NJ) in 500-␮l plastic efficiency of PCR amplification. DMBT1 expression was greatly tubes for one initial cycle of denaturation at 95°C for 2 min, followed by 30 reduced in nine (45%) tumors (four adenocarcinomas and five squa- cycles of denaturation at 95°C for 30 s, annealing at 59°C for 45 s, extension mous cell carcinomas) compared with DMBT1 expression in paired at 70°C for 1 min, and a final elongation step at 70°C for 5 min. The PCR normal lung tissues (Fig. 1C), suggesting that gene expression was ␮ products were then separated in a 2% agarose gel containing 0.5 g/ml lost in these primary tumors. ethidium bromide. Primers used for cDNA amplification are as follows: To determine whether intragenic homozygous deletions of the gene DMBT-1 A/2AS,5Ј-GCAGCAGAAATATACCACCC-3Ј (sense) and 5Ј- CCACCCACCTGTAGATAG-3Ј (antisense); DMBT-1 4S/4AS,5Ј-CAG- contributed to such loss of gene expression, we analyzed genomic GAGCTATCTCCAATC-3Ј (sense) and 5Ј-ACACCAAGAGGAACATCC-3Ј DNA from 40 SCLC cell lines including 14 cell lines used for gene (antisense). expression analysis and 14 NSCLC cell lines that had all been used for Sequence Analysis. After PCR, we excised the bands we were interested in gene expression analysis. Three intragenic STS markers (36k, g14, from a 1.5% agarose gel and purified them using Qiaquick resin columns and 101n) flanking an ϳ65-kb fragment of DMBT1 (6) were used to (Qiagen, Valencia, CA). Between 10 and 15 ng of DNA per sample were used screen these tumor cell lines. A primer set for the ␤-actin genomic for each direct sequencing reaction containing a primer labeled with DNA fragment was used as an internal control in each PCR reaction. ␥ 33 [ - P]ATP and amplified by PCR for 35 cycles using the AmpliCycle Homozygous deletion was detected in 4 (10%) SCLC cell lines sequencing kit (Perkin-Elmer, Branchburg, NJ) according to the manufactur- (H711, H889, H1450, and H1963) at g14 (Fig. 2A). Because g14 is er’s protocol. Each amplified product (3 ␮l) was run on a 6% long-range gel located between 36k and 101n, which were not homozygously deleted (FMC BioProducts, Rockland, ME) and exposed to film. RT-PCR and sequencing primers for DMBT1 cDNA are as follows: DMBT-1 S1,5Ј-GCTA- in these four tumor cell lines, the homozygous deletions should be GGTACCTATAAATGTC-3Ј; DMBT1 AS1,5Ј-TCTTCACTATGGCCA- smaller than 65 kb. No homozygous deletion was found in any of the CAG-3Ј; DMBT-1 S2,5Ј-TTGTCCTGGATGATGTGC-3Ј; DMBT-1 AS2,5Ј- 14 NSCLC cell lines tested. To rule out the possibility that the lack of AACACACTTAGCA-TTGTTGG-3; DMBT-1 S3,5Ј-ATTGTGGTGGCT- PCR amplification was due to a rare nucleotide polymorphism within TCTTATTC-3Ј; DMBT-1 AS3,5Ј-ACCAG-GTTCGGCGTTATC-3Ј; the primers used for g14, we tested another primer set, g14ext, which DMBT-1 S4,5Ј-CAGGAGCTATCTCCAATC-3Ј; DMBT-1 AS4,5Ј-ACAC- amplifies a genomic DNA fragment close to g14. DNA also could not CAAGAGGAACATCC-3Ј; DMBT-1 AF,5Ј-GCAGCAGAAATATAC- be amplified by g14ext in the four cell lines also showing homozygous Ј Ј Ј CACCC-3 ; and DMBT-1 BR,5-TCCTGAACCCTGGCCAAA-3 . deletion at g14 (Fig. 2B), confirming the existence of intragenic LOH Analysis. Fifty ng of DNA were used in each PCR amplification. The homozygous deletions in DMBT1 in these tumor cell lines. microsatellite markers used were D10S209 and D10S587 (Research Genetics, De novo methylation of the promoter regions has been shown to Huntsville, AL). For PCR amplification, one of the primers for each marker was end-labeled with [␥-32P]ATP (4500 Ci/mmol; ICN Biomedicals, Costa play an important role in inactivation of certain tumor suppressor Mesa, CA) and T4 DNA polynucleotide kinase (New England Biolabs, Bev- genes, such as p16 and VHL, in multiple tumor types (18–20). erly, MA). PCR reactions were carried out in a 12.5-␮l volume containing 3% Because the promoter sequence of DMBT1 has not been identified, it ␮ ␮ DMSO, 200 M of deoxynucleotide triphosphate, 1.5 mM MgCl2, 0.4 M of is unclear whether its promoter sequences contain CpG sites that PCR primers including 0.01 ␮M [␥-32P] labeled primer, and 0.5 unit of Taq might be the targets for methylation. Nevertheless, we treated two 1847

Fig. 1. Expression of DMBT1 in lung cancer cell lines and primary NSCLC tissues. A, examples of DMBT1 expression in SCLC cell lines. No DMBT1 cDNA fragment could be detected. B, expression of DMBT1 in NSCLC cell lines. The gene expression was not detectable in six cell lines (H460, H727, H522, A427, H226, and H1648). C, examples of DMBT1 expression in primary NSCLC. DMBT1 expression was significantly reduced in tumors (11 and 15) using a semiquantitative RT-PCR. D, loca- tions of primer sets used to determine DMBT1 expression.

NSCLC cell lines lacking expression of DMBT1 with 0.5 ␮M and 1.0 also present. One substitution is located at codon 42. This codon was ␮M concentrations of 5-azacytidine for 3 and 5 days, respectively. originally determined to be glutamine, derived from a coding se- DMBT1 transcription could not be rescued by such treatment, indi- quence CAA (GenBank Accession Number AJ000342). However, cating methylation of the promoter region of the gene was unlikely to based on our sequencing analysis of DMBT1 cDNA from 50 individ- be the mechanism by which DMBT1 expression came to be lacking. uals, the coding for this codon appears to be ACA encoding a To determine whether there are mutations in DMBT1 that play a threonine residue. A nucleotide substitution changed the codon from role in its inactivation, we sequenced the coding region of DMBT1 in ACA to CCA, resulting in the threonine to proline substitution that eight NSCLC cell lines expressing the gene. Four nucleotide substi- was observed in four cell lines (Table 1). Another substitution was tutions were identified in these cell lines compared with the published identified at codon 54 in the same four cell lines (Table 1). The coding DMBT1 cDNA sequence (GenBank accession number AJ000342). sequence for this codon was changed from TTG to TCG, resulting in One substitution (nucleotide 5227 G3 A) was identified at the 3Ј end an amino acid change from a leucine to a serine. Finally, a substitution of the gene in these lines (Calu-1, H1792, and SK-MES-1) without a was found at codon 52 in one cell line (Calu-1). The coding sequence change in the encoded amino acid. The other three nucleotide substi- TCG was changed to TGG, resulting in an amino acid substitution tutions were found at the 5Ј end of the gene with amino acid changes from a serine to a tryptophan (Fig. 3).

Fig. 2. Homozygous deletion of DMBT1 in SCLC cell lines. A, homozygous deletion was detected in four SCLC cell lines (14, H711; 20, H889; 29, H1450; and 35, H1963) using a g14 primer set for an intragenic STS. B, the homozygous deletion was confirmed in these cell lines by using another primer set (g14ext) for STS adjacent to the genomic sequence of g14. 1848

Table 1 Summary of DMBT1 alteration in NSCLC cell lines and primary tumors Histological DMBT1 Polymorphism LOH Samples typea expression (with amino acid change) Mutation (D10S209 and D10S587) NSCLC cell lines H460 Large cell Ϫ NEb NE H157 SCC ϩ Codons 42, 54 No H226 SCC Ϫ NE NE H292 NSCLC ϩ Codons 42, 54 No H522 AC Ϫ NE NE H727 NSCLC Ϫ NE NE H1648 AC Ϫ NE NE H1792 AC ϩ No H1944 AC ϩ No A427 NSCLC Ϫ NE NE A549 NSCLC ϩ Codons 42, 54 No S966 NSCLC ϩ No SK-MES-1 SCC ϩ No Calu-1 NSCLC ϩ Codons 42, 54 Yes (Codon 52) Paired NSCLC primary tumor 1 AC (M) No changec Codons 42, 54 No Ϫ 2 AC (M) Reduced4 No ϩ 3AC(W5) No change No Ϫ 4AC(P5) Reduced Codons 42, 54 No ϩ 5 AC (M) No change No Ϫ 6 AC (P) No change No Ϫ 7 AC (P) No change No Ϫ 8 AC (P) Reduced No ϩ 9 AC (M) Reduced Codons 42, 54 No ϩ 10 AC (W) No change No Ϫ 11 SCC (M) Reduced No ϩ 12 SCC (P) No change Codons 42, 54 No ϩ 13 SCC (M) No change No ϩ 14 SCC (P) No change No Ϫ 15 SCC (P) Reduced No Ϫ 16 SCC (M) No change Codons 42, 54 No Ϫ 17 SCC (P) Reduced Codons 42, 54 No ϩ 18 SCC (M) Reduced No ϩ 19 SCC (M) No change Codons 42, 54 No Ϫ 20 SCC (M) Reduced Codons 42, 54 No ϩ Normal lung tissues 1 No 2 No 3 No 4 No 5 No 6 No 7 Codons 42, 54 No 8 No 9 No 10 No 11 No 12 Codons 42, 54 No 13 No 14 No 15 No 16 No 17 Codons 42, 54 No 18 Codons 42, 54 No 19 Codons 42, 54 No 20 Codons 42, 54 No 21 No 22 No a AC, adenocarcinoma; SCC, squamous cell carcinoma; W, well differentiated; M, moderately differentiated; P, poorly differentiated. b NE, Not examined. c The expression level of DMBT1 in tumor tissues was compared with corresponding normal tissues.

To determine whether the three nucleotide substitutions that cause of the specimens from the 42 unrelated individuals tested and seven amino acid changes represent polymorphisms, we analyzed 20 pri- cell lines studied. mary NSCLC tumors and their corresponding normal lung tissues and We further performed microsatellite analysis to determine po- 22 normal lung tissues from patients with cancers other than lung tential LOH at the DMBT1 region in the 20 primary NSCLC tumors cancer. We found that the threonine3proline substitution at codon 42 in which DMBT1 expression had been analyzed using two micro- was present in 8 (40%) of the 20 NSCLC tumors and in their satellite markers flanking the gene. LOH at the locus was found in corresponding normal lung tissues (Table 1). This substitution was 10 (50%) of the 20 tumors. Interestingly, 8 (80%) of the 10 tumors also found in 6 (27%) of the 22 normal lung tissues from patients with with LOH at the locus had reduced DMBT1 gene expression, other types of malignancies. The leucine3serine substitution at codon whereas the other two tumors with LOH had normal amounts of 54 was found in the same tumor and normal tissue specimens exhib- DMBT1 expression (Table 1). Statistical analysis showed that iting the codon 42 substitution. Thus, we think that these two substi- tumors with LOH at DMBT1 locus have a significantly higher rate tutions represent frequent, linked polymorphisms. However, the of lack of DMBT1 expression than those without LOH at the locus serine3tryptophan substitution at codon 52 was not observed in any (P ϭ 0.005 by Fisher’s exact test). 1849

domains, and one ZP domain (15). A unique domain with high homology was also found between every SRCR domain in CRP- ductin; it is known as CRP. The ␣ form of the protein contains a short transmembrane domain and a cytoplasmic domain, whereas the ␤ form lacks these domains (15). In fetal lung tissue, there are three DMBT1 mRNA species of 8.0, 7.5, and 6.0 kb. The sequence reported previously was from the smallest transcript (6). Therefore, it is possible that the larger transcripts of DMBT1 contain more coding sequences for a transmembrane and a cytoplasmic domain as demon- Fig. 3. A potential point mutation was identified in Calu-1 cells. The sequence of codon ␣ 52 was changed from TCG to TGG, resulting in an amino acid substitution from serine to strated in the mouse CRP-ductin form. Alternatively, more SRCR tryptophan. Arrow, site where the nucleotide substitution was located. domains are present in the larger form of DMBT1. Nevertheless, the lack of gene expression was detected using primers for both the 3Ј end and the 5Ј end of the DMBT1 cDNA sequence in this study, indicating Discussion that none of these potential isoforms were expressed in these lung tumors. DMBT1 was identified by using a representative difference analy- Previous studies suggest that some of the SRCR superfamily mem- sis, a method for identifying homozygous deletions in target tissues bers play a role in cell differentiation (14, 15). Thus, it is possible that (21), to subtract DNA between normal tissues and medulloblastoma the lack of DMBT1 expression results in the loss of epithelial cell cell lines (6). The gene was mapped to chromosome 10q25.3–26.1, a differentiation. However, we did not observe clear association be- region frequently lost in many tumor types and that spans less than 65 tween loss of DMBT1 expression and tumor differentiation status or kb (6). In the original report, intragenic homozygous deletions were squamous versus adenocarcinoma histology (Table 1). found in three of the nine brain tumor cell lines and 22% (7/32) of the We have identified three amino acid substitutions in DMBT1 at a primary glioblastomas, indicating that the gene may be disrupted by narrowed NH2-terminal region in lung cancer cell lines and primary such deletions in these tumors. A subsequent report confirmed the tumors. The region contains about 70 amino acids and has no homol- observation by finding intragenic homozygous deletion of DMBT1 in ogy to any known protein. Although these substitutions represent 38% (8/21) of primary glioblastomas (11). However, the lack of nonconservative amino acid changes, two of the substitutions were DMBT1 expression was also observed in glioblastoma cell lines found in both primary NSCLC tumors and in the corresponding without homozygous deletions (6), suggesting the existence of alter- normal lung tissues in 40% of cases analyzed, indicating they are native mechanisms for inactivation of DMBT1 in these tumors. Cy- likely to be examples of polymorphism. These polymorphisms are togenetic analysis has shown a complex rearrangement of this region probably also common in patients with other types of cancers (27%). in a glioblastoma cell line, and fluorescence in situ hybridization We are presently unable to examine the frequency of these nucleotide analysis suggested that one of the translocation breakpoints lies in the changes in a normal population because DMBT1 is not expressed in vicinity of the gene (11). Although the functions of DMBT1 as a tumor peripheral lymphocytes, and the genomic sequence of DMBT1 is not suppressor are presently unknown, the genetic evidence supports the yet available. Therefore, we cannot exclude the possibility that these notion that the gene is important in brain tumors. changes may have functional consequences in tumorigenic processes The study presented in this report is the first demonstrating that and may be enriched in patients with cancer as a germ-line alteration DMBT1 is also altered frequently in lung cancer. The abnormalities predisposing to cancer development. The other amino acid substitu- exist not only in tumor cell lines but also in primary tumors. We found tion at codon 52 was found in only one NSCLC cell line but not in any a similar rate (ϳ45%) of loss of DMBT1 expression in both NSCLC of the other tumor cell lines, primary tumors, and normal tissues from cell lines and primary NSCLC specimens. The discordance between the frequency of homozygous deletions (10% in SCLC and none in 49 unrelated individuals, suggesting this substitution may be a somatic NSCLC) and the lack of DMBT1 expression (100% in SCLC and 43% mutation. However, this hypothesis requires confirmation by func- in NSCLC) in lung cancer cell lines suggests that other, yet to be tional analysis in future studies. Thus, it will be interesting to see how identified mechanisms play a role in the inactivation of the gene. replacing wild-type DMBT1 in our tumor cell lines with homozygous However, it is unlikely that methylation in the promoter region of deletions compares with a DMBT1 gene with the codon 42 and 54 DMBT1 gene plays an important role because the gene is not GC-rich substitutions. in sequence content, and a demethylation treatment could not restore In summary, we have demonstrated that expression of DMBT1 is DMBT1 expression. However, we cannot exclude the possibility that frequently lost in lung cancer, particularly in SCLC. We found that small deletions exist and are located within DMBT1 but away from intragenic homozygous deletions of DMBT1 exist in a fraction of regions we examined. It is interesting that most of the primary SCLC cell lines and that LOH status at the DMBT1 region was NSCLC tumors exhibiting lack of DMBT1 expression also lost one of significantly associated with reduced expression of DMBT1 in pri- the DMBT1 alleles (LOH), whereas LOH was rare in tumors express- mary NSCLC tumors. We also identified three nucleotide substitu- ing normal amounts of DMBT1 (P ϭ 0.005; Table 1), suggesting that tions in DMBT1 resulting in nonconservative amino acid changes, LOH plays an important role in inactivating DMBT1. Further studies including one potential point mutation. These data support the hy- are required to study subtle alterations within the gene and to identify pothesis that DMBT1 is a tumor suppressor gene and that it plays an molecules that control the gene’s expression. important role in lung tumorigenesis. DMBT1 is considered a member of the SRCR superfamily because at least nine SRCR domains are present in its predicted protein Acknowledgments sequence. It also contains two CUB domains and one ZP domain. It is interesting that a unique domain with 20–23 amino acid residues We thank Julie Starr for critical editorial review of the manuscript. (known as SID, for SRCR interspersed domain) is located between References every SRCR domain (6). In the mouse, the structure of CRP-ductin is 1. Landis, S. H., Murray, T., Bolden, S., and Wingo, P. A. Cancer statistics. CA Cancer very similar to that of DMBT1, with eight SRCR domains, 5 CUB J. Clin., 48: 6–29, 1998. 1850

2. Strauss, G. M., Gleason, R., and Sugarbaker, D. J. Screening for lung cancer 190 12. Starling, G. C., Whitney, G. S., Siadak, A. W., Llewellyn, M. B., Bowen, M. A., Farr, re-examined. A reinterpretation of the Mayo Lung Project randomized trial on lung A. G., and Aruffo, A. A. Characterization of mouse CD6 with novel monoclonal cancer screening. Chest, 1034 (Suppl. 4): 3375–3415, 1993. antibodies which enhance the allogeneic mixed leukocyte reaction. Eur. J. Immunol., 3. Tong, L., Spitz, M. R., Fueger, J. J., and Amos, C. I. Lung carcinoma in former 26: 738–746, 1996. smokers. Cancer (Phila.), 78: 1004–1010, 1996. 13. Ullrich, A., Sures, I., D’Egidio, M., Jallal, B., Powell, T. J., Herbst, R., Dreps, 4. Mao, L., Lee, J. S., Kurie, J. M., Fan, Y. H., Lippman, S. M., Lee, J. J., Ro, J. Y., A., Azam, M., Rubinstein, M., and Natoli, C. The secreted tumor-associated Broxson, A., Yu, R., Morice, R. C., Kemp, B. L., Khuri, F. R., Walsh, G. L., antigen 90K is a potent immune stimulator. J. Biol. Chem., 269: 18401–18407, Hittelman, W. N., and Hong, W. K. Clonal genetic alterations in the lungs of current 1994. and former smokers. J. Natl. Cancer Inst., 89: 857–862, 1997. 14. Law, S. K., Micklem, K. J., Shaw, J. M., Zhang, X. P., Dong, Y., Willis, A. C., and 5. Wistuba, I. I., Lam, S., Behrens, C., Virmani, A. K., Fong, K. M., LeRiche, J., Samet, J. M., Srivastava, S., Minna, J. D., and Gazdar, A. F. Molecular damage in the Mason, D. Y. A new macrophage differentiation antigen which is a member of the bronchial epithelium of current and former smokers. J. Natl. Cancer Inst., 89: scavenger receptor superfamily. Eur. J. Immunol., 23: 2320–2325, 1993. 1366–1373, 1997. 15. Cheng, H., Byerknes, M., and Chen, H. CRP-ductin: A gene expressed in intestinal 6. Mollenhauer, J., Wiemann, S., Scheurlen, W., Korn, B., Hayashi, Y., Wilgenbus, crypts and in pancreatic and hepatic ducts. Anat. Rec., 244: 327–343, 1996. K. K., von Deimling, A., and Poustka, A. DMBT1, a new member of the SRCR 16. Al-Awqati, Q. Plasticity in epithelial polarity of renal intercalated cells: targeting of ϩ superfamily, on chromosome 10q25.3–26.1 is deleted in malignant brain tumors. Nat. the H -ATPase and band 3. Am. J. Physiol., 270: 1571–1580, 1996. Genet., 17: 32–39, 1997. 17. Resnick, D., Pearson, A., and Krieger, M. The SRCR superfamily: a family reminis- 7. Rasheed, B. K. A., McLendon, R. E., Friedman, H. S., Friedman, A. H., Fuchs, H. E., cent of the lg superfamily. Trends Biochem. Sci., 19: 5–8, 1994. Bigner, D. D., and Bigner, S. H. Chromosome 10 deletion mapping in human 18. Merlo, A., Herman, J. G., Mao, L., Lee, D. J., Burger, P. C., Baylin, S. B., and gliomas: a common deletion region in 10q25. Oncogene, 10: 2243–2246, 1995. Sidransky, D. 5Ј CpG island methylation is associated with transcriptional silencing 8. Peiffer, S. L., Herzog, T. J., Tribune, D. J., Mutch, D. G., Gersell, D. J., and of the tumor suppressor p16/CDKN2/MTS1 in human cancers. Nat. Med., 1: 686– Goodfellow, P. J. Allelic loss of sequences from the long arm of chromosome 10 and 692, 1995. replication errors in endometrial cancers. Cancer Res., 55: 1922–1926, 1995. 19. Herman, J. G., Merlo, A., Mao, L., Lapidus, R. G., Issa, J. J., Davison, N. E., 9. Gray, I. C., Phillips, S. M., Lee, S. J., Neoptolemos, J. P., Weissenbach, J., and Spurr, Sidransky, D., and Baylin, S. Inactivation of CDKN2/p16/MTS1 gene is frequently N. K. Loss of the chromosomal region 10q23–25 in prostate cancer. Cancer Res., 55: associated with aberrant DNA methylation in all common human cancers. Cancer 4800–4803, 1995. Res., 55: 4525–4530, 1995. 10. Kim, S. K., Ro, J. Y., Kemp, B. L., Lee, J. S., Kwon, T. J., Hong, W. K., and Mao, 20. Herman, J. G., Latif, F., Weng, Y., Lerman, M. I., Zbar, B., Liu, S., Samid, D., Duan, L. Identification of two distinct tumor-suppressor loci on the long arm of chromosome 10 in small cell lung cancer. Oncogene, 17: 1749–1753, 1998. D. S., Gnarra, J. R., Linehan, W. M., and Baylin, S. Silencing of the VHL tumor- 11. Somerville, R. P. T., Shoshan, Y., Eng, C., Barnett, G., Miller, D., and Cowell, J. K. suppressor gene by DNA methylation in renal carcinoma. Proc. Natl. Acad. Sci. USA, Molecular analysis of two putative tumor suppressor genes, PTEN and DMBT, which 91: 9700–9704, 1994. have been implicated in glioblastoma multiforme disease progression. Oncogene, 17: 21. Lisitsyn, N., Lisitsyn, N., and Wigler, M. Cloning the differences between two 1755–1757, 1998. complex genomes. Science (Washington DC), 259: 946–951, 1993.

1851

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 1999 American Association for Cancer Research. Expression of DMBT1, a Candidate Tumor Suppressor Gene, Is Frequently Lost in Lung Cancer

Weiguo Wu, Bonnie L. Kemp, Monja L. Proctor, et al.

Cancer Res 1999;59:1846-1851.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/59/8/1846

Cited articles This article cites 21 articles, 6 of which you can access for free at: http://cancerres.aacrjournals.org/content/59/8/1846.full#ref-list-1

Citing articles This article has been cited by 13 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/59/8/1846.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/59/8/1846. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.