Loss of Heterozygosity (LOH) at 17Q and 14Q in Human Lung Cancers

Pataer Abujiang1, Tomoharu J Mori1, Takashi Takahashi2, Fumihiko Tanaka3, Ippei Kasyu1, Shigeki Hitomi3 and Hiroshi Hiai1

1Department of Pathology and Biology of Diseases, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo- ku, Kyoto 606-8051; 2Laboratory of Ultrastructure Research, Aichi Cancer Center Research Institute, 1-1 Kanokoden, Chikusa-ku, Nagoya 464; 3Department of Thoracic Surgery, Chest Disease Research Institute, Kyoto University, Shogoin-Kawahara-cho, Sakyo-Ku, Kyoto 606, Japan

Our recent linkage study of urethane-induced pulmonary since both tumors share a number of morphological adenomas in SMXA RI strains of mouse revealed two and molecular biological properties (Malkinson et al., host resistance genes, Par1 (chromosome 11) and Par3 1992). Mutations in kras2 in mouse PA are as (chromosome 12). The map positions of Par1 and Par3 prevalent as in human lung cancers and its allelotype correspond to human 17q11-23 and 14q11-24, based on is the primary determinant of disease susceptibility synteny between mouse and human. In this study, we (Dragani et al., 1995). The involvement of multiple examined the loss of heterozygosity (LOH) in these two host genes in lung carcinogenesis has been studied homologous human chromosomal regions in 30 primary extensively in mouse models (Festing et al., 1994). lung adenocarcinoma samples with matched normal Recently, we reported two autosomal dominant DNA. Using 15 highly polymorphic markers, two resistance loci to urethane-induced PA in the SMXA commonly deleted regions were identi®ed on human recombinant inbred strain of mice, Par1 and Par3 chromosomes 14 and 17, respectively. At 17q21, nine (Abujiang et al., 1997). From the synteny map, the (53%) of 17 informative tumors showed LOH between regions of Par1 and Par3 are homologous to human D17S588 and D17S518. On the other hand, at 14q11-12, 17q11-23 and 14q11-24, respectively. Here, we analysed seven (32%) of 22 informative tumors showed LOH at human lung adenocarcinoma for LOH in homologous loci between D14S261 and D14S80. Subsequently, we regions and found two minimally deleted regions. examined 25 squamous cell carcinomas (SQ) and 24 Subsequent examination of squamous cell carcinomas small cell carcinomas (SCC). At 14q11-12, six (38%) of (SQ) and small cell carcinomas (SCC) for LOH in 16 informative SQ and ®ve (42%) of 12 informative these regions revealed that 17q contained a gene SCC showed LOH. In contrast, at 17q11-23, one (7%) aecting selectively adenocarcinomas and 14q, another of 15 informative SQ and two (14%) of 14 SCC showed gene aecting all three histological types of lung LOH. Therefore, the gene on 17q seemed to aect cancers. selectively adenocarcinomas, whereas the other gene on 14q, all three types of lung carcinomas. These Results observations indicate that a comparative genetic analysis provides a promising approach to survey genes involved Most of urethane-induced lung tumors in mice are in multifactorial process of human lung carcinogenesis. adenomas in histology. Considering the possibility that the function of the Par1 or Par3 homolog may be Keywords: pulmonary adenoma; lung cancer, LOH, somehow selective to target cells of tumorigenesis, we tumor suppressor gene ®rst selected 30 primary lung adenocarcinomas and examined their genotypes at 15 highly polymorphic microsatellite marker loci distributed in 14q11-23 and 17q11-24 regions. Figure 1 lists the marker loci in Introduction descending order from centromere to telomere, and shows their patterns and frequencies of LOH in the Lung cancer is one of most frequent malignancies in adenocarcinomas examined. Two minimal regions of industrialized countries. Lung cancer develops through deletion were identi®ed by this analysis. The ®rst a multistep process in which genetic alterations are region was bounded by D14S261 proximally and by accumulated to promote tumorigenesis. Inactivation of D14S80 distally. These marker loci have been tumor suppressor genes plays an important role in this cytogenetically mapped to chromosome 14q11 and process (Carbone et al., 1992). Loss of heterozygosity 14q21, respectively. The second region was bounded (LOH) is a frequent observation in such tumors. In by D17S588 proximally and by D17S518 distally, and human lung cancers, LOHs are found on chromosomes these marker loci have been cytogenetically mapped to 1q, 2q, 3p, 5q, 8q, 9p, 13q, 11p,and17p and chromosome 17q21 and 17q23, respectively. LOH was localizations of tumor suppressor genes have been observed most frequently at D17S588 and D14S261. con®rmed or suggested (Tsuchiya et al., 1992). Nine (53%) of 17 informative tumors showed LOH at Urethane-induced pulmonary adenoma (PA) in mice marker locus D17S588. At locus D14S261, seven (32%) is an excellent model of human lung adenocarcinoma of 22 informative tumors showed LOH. Representative cases of LOH analysis at several marker loci are shown in Figure 2. At D14S261, the patient 9 did not show LOH, since DNAs from both tumor and non-tumor Correspondence: H Hiai tissue retained two alleles, whereas the patient 22 Received 9 March 1998; revised 18 June 1998; accepted 19 June 1998 showed LOH as one allele was missing in the tumor Loss of heterozygosity in human lung cancers P Abujiang et al 3030

Figure 1 Deletion mapping of chromosomes 14 and 17. Fifteen microsatellite markers spanning 14q11-23 and 17q11-24, were used to study 30 human lung adenocarcinomas for evidence of allelic loss. Patient numbers are shown above the map. Percentage of informative cases with LOH (%LOH) is shown in the right column

DNA. Similar LOH was observed in the patients 3 and 10 at D17S588 and D17S855, respectively. To examine whether LOHs in these regions are speci®c to adenocarcinomas or observed also in other types of lung cancers, we further studied 25 cases of SQ and 24 cases of SCC. Figure 3A,B schematically shows the patterns of LOH and their frequencies in these tumors. Comparing with lung adenocarcinomas, LOHs at 17q21 were less frequent in SQ and SCC as summarized in Table 1. Therefore, the LOH on 17q seemed to be associated with adenocarcinomas (P50.05). On the other hand, LOHs at 14q11-12 in SQ (30%) and SCC (42%) were as frequent as in adenocarcinomas (32%) (Table 1), indicating the LOHs on 14q were prevalent to all types of lung cancers. To further characterize the role of the putative tumor suppressor gene in this region in lung cancer, we examined the possible associations of LOHs on 14q and 17q with clinicopathological parameters including sex, degree of dierentiation, smoking history, TNM stage and lymph node metastasis. However, no positive association was found (Table 2).

Discussion

Comparative genetic analysis is one of promising approaches to characterize complicated diseases in human. It is increasingly evident that there is an extensive syntenic conservation of genomic organiza- tion among species, and comparative genetic maps have been used to predict the locations of genes in other species. The present study aimed to extend the genetic information obtained in chemical-induced lung tumors in the mouse to survey genetic changes in human lung cancers. Analysis of 79 cases of human Figure 2 Representative data of detection of loss of hetero- lung cancer revealed frequent LOHs in two distinct zygosity. PCR products for the microsatellite locus from normal chromosomal regions. One of locus at 17q21 seem to lung (N) and tumor DNA were compared. Patient numbers are shown at left. The upper boxed ®gure indicates the length of the play signi®cant role in lung adenocarcinomas. On the fragment, and the lower boxed ®gure shows the peak area other hand, another locus at 14q11-12 seemed to play Loss of heterozygosity in human lung cancers PAbujianget al 3031 signi®cant role in all types of lung cancers. These have been reported in a variety of malignancies results suggest the existence of putative tumor including breast (Cropp et al., 1993), ovary suppressor genes in these regions. LOHs in 17q21 (Wertheim et al., 1996) and prostate cancers (Gao et

Figure 3 Deletion map of (A) 25 squamous cell carcinomas and (B) 24 small cell carcinomas. Patient numbers are shown above the map

Table 1 Frequency of LOH on chromosomes 14 and 17 in lung cancer Human Allelic loss/heterozygosity (% LOH) chromosome Marker locusa Adenocarcinoma Squamous (SQ) Small (SCC) 17q21 D17S588 9/17 (53%) 1/15 (7%) 2/14 (14%) 14q11±12 D14S261 7/22 (32%) 3/10 (30%) 5/12 (42%) aThe markers with the highest frequency of LOH on 17q and 14q

Table 2 Correlation between allelic loss and clinicopathological features Cases with Cases without Cases with Cases without 17q LOH 17q LOH P value 14 q LOH 14q LOH P value Sex Male 14 22 16 20 Female 3 8 0.48 5 20 0.08 Dierentiation Well 3 4 5 9 Mod, poorly 8 16 0.64 9 12 0.67 Smoking history Non-smokers 6 5 7 8 Smokers 8 23 0.08 14 22 0.6 TNM I 7 9 7 9 II, III 5 14 0.27 11 15 0.92 Metastasis Without 6 15 6 10 With 5 8 0.54 8 11 0.78 Loss of heterozygosity in human lung cancers P Abujiang et al 3032 al., 1995a). The tumor suppressor genes prohibitin suggest the presence of shared genetic steps in lung (Cliby et al., 1993a), BRCA1 (Gao et al., 1995b) and carcinogenesis across species. NME1 (Niederacher et al., 1997) are mapped in this region. A large panel of human cancers of the breast, ovary, liver, and lung have been examined for somatic Materials and methods mutations in the prohibitin gene. Although mutations are observed in a few sporadic breast cancers, none is Samples and DNA extraction. identi®ed in any other cancers (Sato et al., 1993). Recent positional cloning of the BRCA1 from that Lung tumors and corresponding nontumorous tissue were region and its mutation screening support that the obtained at surgery from 79 patients. Types of tumors were BRCA1 is responsible for inherited breast (Jandrig et con®rmed histologically. Tissues were frozen immediately andstoredat7808C. DNA was extracted from frozen al., 1996) and ovarian cancer (Merajver et al., 1995). tissue according to the methods described previously It would be interesting to screen for BRCA1 (Abujiang et al., 1996). mutations in tumor samples which demonstrate LOH at the BRCA1 locus. This will provide evidence to either prove or disprove the hypothesis that the Fluorescent microsatellite analysis. BRCA1 gene is indeed involved in lung cancer. We chose 15 microsatellite markers along 14q and 17q arms. Fong et al. (1995) also showed 45% of LOH in lung Map positions of these loci were based on the Human adenocarcinomas and 29% of LOH in SQ overall 17q. Genetic Map (Cold Spring Harbor Laboratory Press 1993, They used Southern blot analysis of RFLPs (Fong et Stephen J O'Brien) and CEPH/CHLC databases. The al., 1995), so that it is dicult to compare their results oligonucleotides were labeled with 6-FAM ¯uorescent dye, and the TAMRA dye was reserved for the size standard. with ours. Whether or not these observations direct PCR for ¯uorescent markers was carried out in a volume of the deletion of the same critical locus will require 9 ml and included 106PCR buer, 50 ng of DNA, 250 mM further study. of each dNTP, 0.5 units of Dynazyme Taq polymerase and Cancers of colorectal tract (Chang et al., 1995), 10 pmol of each primer. The thermal pro®le was as follows: kidney (Thrash-Bingham et al., 1995), endometrium 948C for 2 min; 35 cycles of 948Cfor30s,538Cfor40s, (Chang et al., 1995), bladder (Chang et al., 1995), and 728C for 1 min; and a ®nal extension of 728Cfor ovary (Cliby et al., 1993b) and neuroblastoma 10 min. One ml of products were added to 2.5 mlof (Takayama et al., 1992) are solid cancers associated formamide, 0.5 ml of blue dextran and 0.5 mlofTAMRA with a signi®cant degree of 14q LOH. The proximal 500 size standard (Applied Biosystems), loaded on 4% end of the region de®ned in our study overlaps with the polyacrylamide 6-M urea gels, and run for 2 h in a 377A Automated Sequencer (Applied Biosystems). The data were region de®ned in the ovarian carcinoma study. collected automatically and analysed using GeneScan Assuming that the deletion observed at 14q11-12 in software (Applied Biosystems); ®nally, the Genotyper bladder, ovarian and lung cancers are driven by the software (Applied Biosystems) and Exel software were same gene, the identi®cation and characterization of used for allele scoring and assessment of LOH. In the case this tumor suppressor locus would be expected to of constitutional heterozygotes, two alleles are detected in represent a substantial advance in the ®eld of cancer normal tissue, and if one is absent in the tumor the result is molecular genetics. Although these regions are not LOH. The program checks for LOH by calculating the ratio small enough to plausibly indicate a positional cloning of the constitutional alleles (Canxian et al., 1996). eort, it is possible to narrow down the minimal regions of deletion through the identi®cation of Statistical analysis. additional microsatellite markers to saturate these The X2 test was performed to compare LOH frequencies in regions and to perform LOH analysis of a larger respect of histopathological type, stage, and grade. P50.05 series of tumors with them. We concentrated on lung was considered to indicate statistical signi®cance. cancers, because the primary aim of this study was to seek common genetic events among species in carcinogenesis of homologous type tumors. In murine urethane-induced PA, four resistance genes Abbreviations Par, pulmonary adenoma resistant; Pas, pulmonary (Par1-4) have been identi®ed (Abujiang et al., 1997; adenoma susceptible; PA, pulmonary adenoma; RI, Manenti et al., 1997). Par2 is mapped on chromosome recombinant inbred; LOH, loss of heterozygosity; SQ, 18 close to the putative tumor suppressor gene DCC. squamous cell carcinoma; SCC, Small cell carcinoma. Although we did not examine the Par2 homologous region, frequent LOH of DCC in human lung cancers Acknowledgements has been reported (Zhang et al., 1995). Par4 was We thank Drs Hiromi Wada (Department of Thoracic recently mapped to chromosome 6 (Manenti et al., Surgery, Chest Disease Research Institute, Kyoto Univer- 1997), but its syntenic region has not been examined. To sity) for valuable discussions. This study was supported by date, therefore, 3 of 4 mouse PA resistance genes have grants from the Ministry of Education, Culture, and syntenic human chromosomal regions bearing hot spots Science, from the Ministry of Health and Welfare, Japan for LOH in lung adenocarcinomas. These observations and from the Japanese Owner's Association.

References

Abujiang P, Nishimura M, Kamoto T, Ichioka K, Sato M Canxian F, Salovaara R, Hemminki A, Kristo P, Chadwick and Hiai H. (1997). Cancer Res., 57, 2904 ± 2908. RB, Aaitonen LA and Chapelle AL. (1996). Cancer Res., Abujiang P, Kamoto T, Lu LM, Yamada Y and Hiai H. 56, 3331 ± 3337. (1996). Cancer Res., 56, 3716 ± 3720. Loss of heterozygosity in human lung cancers PAbujianget al 3033 Carbone DP and Minna JD. (1992). Adv. Intern. Med., 37, Malkinson AM. (1992). Cancer Res., 52, 2670 ± 2676. 153 ± 171. Manenti G, Gariboldi M, Fiorino A, Zanesi N, Pierotti MA Chang WY, Cairns P, Schoenberg MP, Polascik TJ and and Dragani TA. (1997). Cancer Res., 57, 4164 ± 4166. Sidransky D. (1995). Cancer Res., 55, 3246 ± 3249. Merajver SD, Pham TM, Cadu RF, Chen M, Poy EL, Cliby W, Ritland S, Hartmann L, Dodson M, Halling KC, Cooney KA, Weber BL, Collins FS, Johnston C and Keeney G, Podratz KC and Jenkins RB. (1993b). Cancer. Frank TS. (1995). Nat. Genet., 9, 439 ± 443. Res., 53, 2393 ± 2398. Niederacher D, Picard F, Roeyen C, An HX, Bender HG and Cliby W, Sarkar G, Ritland SR, Hartmann L, Podratz KC Beckmann MW. (1997). Genes Chromosomes Cancer, 18, and Jenkins RB. (1993a). Gynecol. Oncol., 50, 34 ± 37. 181 ± 192. Cropp CS, Champeme MH, Lidereau R and Callahan R. Sato T, Sakamoto T, Takita K, Saito H, Okui K and (1993). Cancer Res., 53, 5617 ± 5619. Nakamura Y. (1993). Genomics, 17, 762 ± 764. Dragani TA., Manenti G and Pierotti MA. (1995a). Adv. Takayama H, Suzuki T, Mugishima H, Fujisawa T, Ookuni Cancer Res., 67, 83 ± 112. M, Schwab M, Ghring M, Nakamura Y, Sugimura T and Festing MF, Yang A and Malkinson AM. (1994). Genet Res., Terada M. (1992). Oncogene, 7, 1185 ± 1189. 64, 99 ± 106. Thrash-Bingham CA, Greenberg RE, Howard S, Bruzel A, FongKM,KidaY,ZimmermanPV,IkenagaMandSmith Bremer M, Goll A, Salazar H, Freed JJ and Tartof KD. PJ. (1995a). Cancer Res., 55, 4268 ± 4272. (1995). Proc. Natl. Acad. Sci. USA., 92, 2854 ± 2858. GaoX,ZacharekA,SalkowaskiA,GrignonDJ,SakrW, Tsuchiya E, Nakamura Y, Weng SY, Nakagawa K, Tsuchiya Porter AT and Honn KV. (1995a). Cancer Res., 55, 1002 ± S, Sugano H and Kitagawa T. (1992). Cancer Res., 52, 1005. 2478 ± 2481. GaoX,ZacharekA,GrignonDJ,SakrW,PowellIJ,Porter Wertheim I, Tangir J, Muto MG, Welch WR, Berkowitz RS, AT and Honn KV. (1995b). Oncogene, 11, 1241 ± 1247. Chen WY and Mok SC. (1996). Oncogene, 12, 2147 ± 2153. Jandrig B, Grade K, Seitz S, Waindzoch B, Muller M, Zhang J, Ding F and Wang X. (1995). Chung Hua I Hsueh Bender E, Nothnagel A, Rohde K, Schlag PM, Kath R, Tsa Chih. (China Med.J.), 75, 211 ± 213. Hoken K and Scherneck S. (1996). Int. J Cancer., 68, 188 ± 192.