GENES, CHROMOSOMES & CANCER 1:240-246 (1990)

Molecular Mapping of Deletion Sites in the Short Arm of Chromosome 3 in Human Lung Cancer Hiltrud Brauch, Kalman Tory, Frederick Kotler, Adi F. Gazdar, Olive S. Pettengill, Bruce Johnson, Stephen Graziano, Tim Winton, Charles H.C.M. Buys, George D. Sorenson, Bernard J. Poiesz, John D. Minna, and Berton Zbar

Laboratory of Imrnunobiology, National Cancer Institute, Frederick Cancer Research Facility, Frederick, Maryland (H.B., K.T., F.K., B.Z.); NCI, Navy Medical Oncology Branch, Naval Hospital, Bethesda, Maryland (A.F.G., B.J.,T.W., J.D.M.): Department of Pathology, Dartrnouth Medical School, Hanover, New Hampshire (O.S.P., G.D.S.); Department of Medicine, SUNY Heatth Science Center and Veterans Administration Medical Center, Syracuse, New York (S.G., B.J.P.);Department of Human Genetics, State University of Groningen, Groningen, The Netherlands (C.H.C.M.B.)

We used I0 restriction fragment length polymorphism (RFLP) probes spanning the length of the short arm of chromosome 3 (3p) t o map deletion sites in human lung cancer. Two approaches were used. I) When a patient’s tumor and normal tissue were available, loci with allelic heterozygosity in the normal tissue were tested for loss of alleles at 3p. 2) When the corresponding normal tissue was not available, the frequency of heterozygosity at each locus in a panel of tumors was compared t o the corresponding published frequencies in nontumor tissue of healthy individuals or patients with lung cancer. In 14 small cell lung carcinomas (SCLC) with normal D N A for comparison, allele loss was found at all heterozygous loci, with one exception at a locus near the 3p centromere (D3S4). In the total of 53 SCLCs, which included tumors without paired normal tissue, frequency of heterozygosity was significantly reduced in all 10 3p loci. Three loci, DNF15S2, R A F I , and D3Sl8, were homozygous in all tumors in the SCLC panel. These loci, which are in regions 3p2 I and 3p25, may thus be involved in the origin or evolution of SCLC. We also investigated 24 non-SCLC tumors. In this panel, frequency of heterozygosity was significantly reduced at seven of the 10 loci tested. Comparison of the results shows that the pattern of allele loss on 3p is different in SCLC and non-SCLC, suggesting a difference in pathogenesis at the genetic level.

INTRODUCTION

A deletion of the short arm of chromosome 3 (3p) is one of the most frequent genetic changes in small cell lung carcinoma (SCLC) (Whang-Peng et al., 1982; de Leij et al., 1985; Zech et al., 1985; Brauch et al., 1987; Kok et al., 1987; Naylor et al., 1987; Yokota et al., 1987; Dobrovic et al., 1988; Leduc et al., 1989). This observation suggests that a gene located on 3p plays a role in the origin or evolution of this tumor. One tumor suppressor gene has been located on 3p. T h e von HippelLindau disease (VHL) gene, a gene that predisposes to renal cell carcinomas, hemangioblastomas, and pheochromocytomas, is linked to RAFl, a gene located at 3p24-25 (Seizinger et al., 1988; Tory et al., 1989). It is not known whether the VHL gene plays a role in lung cancer. One approach to determining 3p loci involved in the pathogenesis of SCLC is to determine the 3p region(s) deleted in all SCLC tumors: This region would be the region most likely to harbor a regulating gene that may be involved in SCLC. Previous cytogenetic studies put the region that is deleted in all SCLC at 3p14-23 (Whang-Peng et al., 1982), 3p21-22 (Kok et al., 1987), and 3p23-24 (Ibsen et al., 1987). We used restriction fragment 0 1990 WILEY-LISS, INC.

length polymorphisms (RFLPs) to test for loss of alleles at loci on 3p in 14 paired SCLC samples. T e n 3p loci distributed over the length of the short arm of chromosome 3 were tested against a panel of 52 SCLC cell lines and one primary SCLC tumor. T h e results suggest that 3p loci in the regions 3p21 and 3p24-25 are consistently deleted or homozygous in SCLC. There is controversy about loss of heterozygosity at 3p loci in non-SCLC. Loss of heterozygosity at the DNF15S2 locus was reported for all lung carcinomas (Kok et al., 1987). We reported earlier that loss of heterozygosity was found in about 25% of primary non-SCLC (Brauch et al., 1987). We have now studied 24 non-SCLC cell lines and found significant reduction in frequency of heterozygosity a t some but not all 3p loci tested. Although both SCLC and non-SCLC showed significant reduction in frequency of heterozygosity at 3p loci, the frequency of heterozygosity in the SCLC panel was significantly lower than in the non-SCLC Received September 19, 1989; accepted October 20, 1989. Address reprint requests to Dr. Hiltrud Brauch, Cellular Immunity Section, Laboratory of Immunobiology, Bldg. 560, Rm. 12-34, National Cancer Institute, Frederick Cancer Research Facility, Frederick, MD 21701.

LOSS OF 3p LOCI I N LUNG CANCER

panel, suggesting that different loci are responsible for the development of SCLC and non-SCLC. MATERIALS AND METHODS

Tumors and normal tissues were obtained from patients treated at the NCI-Navy Medical Oncology Branch (Bethesda, Maryland), Dartmouth Medical School (Hanover, New Hampshire), SUNY Health Science Center (Syracuse, New York), and Johns Hopkins School of Medicine (Baltimore, Maryland). T h e methods to establish the cell lines and their characterization as SCLC and non-SCLC have been reported (Gazdar et al., 1980; Pettengill et al., 1980; Graziano et al., 1987). Two SCLC cell lines were obtained from the Department of Human Genetics, State University Groningen (Groningen, T h e Netherlands) (Kok et al., 1987). Detailed information about the cell lines is available upon request. DNA polymorphisms were compared in tumor and normal tissues of patients with small cell carcinoma of the lung. Therefore, it was essential to obtain paired samples, that is, tumor and normal tissue, from each patient. We obtained 16 paired samples from patients with SCLC. T o make use of SCLC and non-SCLC cell lines, for which paired normal tissue was not available, we determined the frequency of heterozygosity a t loci on 3p in 37 unpaired SCLC and 24 unpaired non-SCLC cell lines. DNA extraction, digestion with restricrion enzymes, Southern blotting, hybridization, and evaluation of the results were performed as described previously (Brauch et al., 1987). We analyzed lung carcinoma samples with 10 recombinant DNA probes to loci on 3p. D3S4, D3S30, DNF15S2, D3S32, and D3S2 are members of a linkage group on 3p (Leppert et al., 1987); D3S4 is detected by the probe B67 (Mode et al., 1984), D3S30 by pYNZ86.1 (Nakamura et al., 1987), DNFlSSZ ( 3 ~ 2 1 by ) pH3H2 (Carritt et al., 1986), D3S32 by pEFD145.1 (Fujimoto et al., 1988), and D3S2 ( 3 ~ 2 1 by ) p12-32 (Naylor et al., 1984; Barker et al., 1986). D3S18 and D3S17 are members of a second linkage group on 3p; D3S18 is detected by CRI-L162; D3S17 is detected by CRI-L892 (Donis-Keller e t al., 1987). THRB (EMU, 3p22-24.1) is detected by pBH302 and pheA4 (Rider et al., 1987; Gareau et al., 1988; Drabkin et al., 1988); RAFI ( 3 ~ 2 5is) detected by p627 (Bonner et al., 1984). D3S3 ( 3 ~ 1 4 )is detected by pMSI-37 (Shih and Weinberger, 1982). Restriction enzymes used were MspI for B67, pYNZ86.1, pMSI-37, and p12-32, Tap1 for B67; pEFD145.1, p627, CRI-L162 and CRI-L892, Hind111 for pH3H2 and pBH302.

241

Loci tested were chosen to represent 3p loci distributed over the entire length of 3p. Their sequential order is based on linkage analysis with probes to loci D3S4, D3S30, DNF15S2, D3S32, and D3S2 (Leppert et al., 1987) and linkage analysis with probes to D3S18 and D3S17 (DonisKeller et al., 1987). It is also based on results of in situ hybridization (Shih and Weinberger, 1982; Bonner et al., 1984; Rider et al., 1987; Dobrovic et al., 1988; Drabkin et al., 1988), tests with cytogenetically defined somatic cell hybrids (Brauch et al., 1989), and studies of a derivative chromosome 3 for regional localization of loci in 3p21 (Drabkin et al., 1989). We evaluated 16 paired samples at the above 3p loci. T h e data of 14 paired samples are presented. It was not possible to evaluate NCI-HI28 for allele loss. T h e constitutional DNA of this patient was homozygous a t all 3p loci tested. It was also not possible to evaluate DMSll4 for allele loss because the densitometry results were consistent with trisomy of chromosome 3. T h e observed frequencies of heterozygosity were compared to their expected frequencies by using the normal approximation to the binomial distribution, corrected for continuity (Armitage, 1971). Frequencies of heterozygosity at 3p loci were compared among the panels of SCLC and non-SCLC samples according to the Fisher exact test (Steel and Torrie, 1980). RESULTS Paired SCLC Samples and Corresponding Normal Tissue

DNA from paired tumor and normal tissue of 16 SCLC patients was analyzed. Normal DNA from 14 patients was heterozygous at one or more 3p loci. In each of those 14 patients, tumor DNA showed loss of heterozygosity at one or more of the 10 loci tested. As is shown in Figure 1, with one exception (D3S4), all 10 loci in all 14 SCLC samples were homozygous, reflecting either homozygosity in both normal and tumor DNA (hatched boxes) or loss of heterozygosity in tumor DNA (open boxes). There was widespread allele loss in tumor DNA over the region of 3p represented by the tested loci. T h e one exception was retention of heterozygosity by tumor DMS153 at locus D3S4, thought to be located close to the centromere. This pattern of allele loss was consistent with a terminal deletion of 3p. Unpaired SCLC Lines

T o supplement the information obtained from the paired SCLC lines, we tested 37 unpaired

242

BRAUCH ET AL.

Figure I . Loss of alleles on the short arm of chromosome 3 (3p) in SCLC. RFLPs of I 0 3p loci were compared between tumor and normal D N A in 14 patients with SCLC. The loci listed from left t o right indicate the order from centromere to telomere. Open boxes are loci with heterozygous genotype in the normal tissue and loss of one allele in the tumor. Hatched boxes are homozygous loci in normal and tumor tissue. The black box is a locus with heterozygous genotype in the normal tissue and retained heterozygosity in the tumor. Boxes with Xs are loci not tested because of an insufficient amount of DNA. A t some loci. results correspond to those reported earlier: D3S2, D3S3, and DNF I552 (Brauch et al., 1987); THRB (Leduc et al., 1989); and D3S32 (Sithanandam et al., 1989).

SCLC samples at the 10 3p loci. For these unpaired SCLC samples, matching normal DNA was not available, and loss of heterozygosity could not be determined directly. Therefore, we compared the frequencies of heterozygosity in SCLC to the frequencies of heterozygosity expected in the normal population. T h e normal frequencies of heterozygosity for each locus were taken from the literature and were found not to be statistically different from the frequencies of heterozygosity obtained from the normal tissues of a SCLC population and a non-SCLC population, except for the D3S3 locus, where the normal frequency of heterozygosity was significantly higher for both groups than the value in the literature. Using the former for comparison, the reduction in frequency of heterozygosity in SCLC patients and in non-SCLC patients at the D3S3 locus becomes significant. T h e results show that frequencies of heterozygosity in SCLC are significantly reduced at all loci shown (Fig. 2A). T h e results at the loci DNF-

15S2, D3S32, RAFl, and D3S17 were particularly striking: Oh3 SCLC samples were heterozygous at the DNF15S2 locus (expected 25/53); 1/50 SCLC samples were heterozygous at the D3S32 locus (expected 24/50); 0/46 SCLC samples were heterozygous at the RAFl locus (expected 16/46); 2/43 SCLC samples were heterozygous at the D3S17 locus (expected 31/43). Among the 37 unpaired SCLC cell lines, six were heterozygous at one or more loci on 3p (Fig. 3A). Heterozygosity occurred in almost all the loci for which loss of alleles was generally found in the other SCLC cell lines (Fig. 1). However, in three loci (DNF15S2, RAFl, and D3S18), heterozygosity was not demonstrated not only among these six cell lines but in all 53 SCLCs tested. Non-SCLC Lines

In our previous study we found that 25% of primary non-SCLCs showed loss of heterozygosity at

LOSS OF 3p LOCl IN LUNG CANCER

243

A: SCLC 46

52

53

D3S30

D3S3

53

50

52

53

43

46

45

THRB

RAFl

D3Sl8

D3S17

23

19

20

18

70 60 $ 5r

In

50

0

0)

E7 2 40

al c. 0

I

6 30 5r

u

E

=

0)

U

20

?!

LL

10

0

1 D3S

I1 LII D3S2

D3S32

VF 15S2

n

D

Chromosome 3p Loci

6: non-SCLC 19

70

s 2ln

60

0 0

c

2 50 al al

c1

I

6

40

5r 0 C

$

30

2 u. 20 1c

C

I! 03S4

24

16

22

24

D3S2

D3S32

24

7

DNF

15S2

Chromosome 3p Loci

Figure 2. Frequencies of heterozygosity at loci of the short arm of chromosome 3 (3p) in SCLC (A) and non-SCLC (6). Black bars represent the frequencies of heterozygosity at 3p loci derived from a normal population. White bars give the frequencies of heterozygosity found in the rumors. The numbers of tumors tested at a particular locus are given above the columns. SCLC frequencies of heterozygosity were significantly reduced at D3S4 (P = 0.001 5), D3S30. DNFl5S2. D3S32. D3S2, THRB, RAFI, and D3S17 (P< O.OOOl), D3S18 (P = 0.001 I),and D3S3 (P = 0.009). Non-SCLC frequencies of heterozygosity were significantly reduced at D3S30, D3S32, D3S2, THRB, and D3S17 (P < O.OOI), 0353 (P = 0.022). and DNF15S2 (P = 0.02). Frequencies of heterozygosity were not significantly reduced at D3S4 (P = 0.47), RAFl (P = 0.076), and D3S18 (P = 0.43).

244

BRAUCH ET A L

A: SCLC

B: non-SCLC

Figure 3. Chromosome 3 loci in unpaired tumors that showed heterozygosity at one or more of the tested loci on 3p; A SCLC and B non-SCLC. RFLPs were tested at I 0 chromosome 3p loci. Hatched boxes indicate homozygous loci. Black boxes indicate heterozygous loci. Boxes with X are loci not tested because of an insufficient amount of DNA. A SCLC Of 37 unpaired SCLC samples. 6 were heterozygous at one or more loci. There was no SCLC tumor heterozygous at the loci DNF 1552. RAFl and D3S 18. B Non-SCLC: O f 24 unpaired non-SCLC samples 8 samples were heterozygous at multiple loci and 3 samples were heterozygous at one 3p locus. The loci DNF I5S2 and D3S I 7 were most frequently found heterozygous. Loci with the lowest number of sampler found heterozygous were D3S3. D3S2, D3S32 and THRB.

loci on 3p (Brauch et al., 1987). Other investigators have stated that all lung carcinomas show a deletion at 3p (Kok et al., 1987). We used non-SCLC cell lines because these lines would not have the normal tissue contamination that may obscure allele loss. With the exception of loci D3S4, RAFZ, and D3S18, there was significant reduction in frequency of heterozygosity at all 3p loci (Fig. ZB). T h e number of non-SCLC tumors with retained heterozygosity was higher than in the SCLC group. In eight non-SCLC cell lines, we found retained

heterozygosity at multiple loci, and, in an additional three, a single heterozygous locus was found (Fig. 3B). T h e most frequently occurring heterozygous loci were DNF15S2 (6/24, expected 12/24), D3S17 (6/18, expected 13/18), D3S4 (5/19, expected 5/19), D3S30 (4/24, expected 13/24), RAFl (3/19, expected 7/19), and D3S18 (3/20, expected 4/20). Loci with the highest reduction in frequency of heterozygosity were D3S2 (2/24, expected 11/ 22), D3S32 (2/20, expected 10/22), and THRB (ERBA2) (2123, expected 10/23). Comparison of

LOSS O F 3p LOCI IN LUNG CANCER

the frequencies of heterozygosity between the SCLC and non-SCLC panels showed that the groups had a different pattern of 3p loss at the following loci: D3S4 (P = 0.047), D3S30 (P = 0.036), DNF15S2 (P< 0.001), RAFZ (P = 0.026), and D3S17 (P = 0.008). DISCUSSION

We used RFLP to define the 3p region lost in SCLC. A panel of 53 SCLC samples was tested, with 10 polymorphic 3p probes distributed throughout the length of 3p. We found that all SCLC samples were homozygous at three loci: DNF15S2, RAFZ, and D3S18. This was demonstrated directly by comparison of normal and neoplastic tissue in 14 patients. Among tumors without paired normal tissue, it was shown by analysis of frequency of heterozygosity. Two of the loci have been placed on the physical map of 3p: DNF15S2 is located at 3p21 (Carritt et al., 1986); RAFZ is located a t 3p25 (Bonner et al., 1984). Six SCLC cell lines were heterozygous at one or more 3p loci. Loci with retained heterozygosity were D3S4, D3S30, D3S3, D3S2, D3S32, THRB, and D3S17. Some of these loci have been placed on the physical map of 3p: D3S2 is located at 3p21 (Naylor et al., 1984); THRB is located at 3p21-25 (Gareau et al., 1988) or 3~22-24.1(Drabkin et al., 1988). Regions harboring these loci may not be required for development of SCLC. Previous cytogenetic and molecular studies have emphasized 3p14-23 as the region lost in all SCLC tumors (Whang-Peng et al., 1982; Kok et al., 1987; Ibsen et al., 1987). However, our data suggest loss in a more distal 3p region represented by the loci RAFZ and D3S18. Although Drabkin et al. (1988) found only variable deletion of the THRR gene, Leduc et al. (1989) showed frequent loss of heterozygosity at the THRB locus in 16/17 SCLC samples.' Sithanandam et al. (1989) found that RAFZ was deleted in 5 of 5 SCLC samples. These results, along with our work, suggest that loci distal to 3p21 should be critical in the development of SCLC. T h e interpretation of our results depends on the alignment of 3p loci. There is no certainty about orientation and sequential order of some loci. Drabkin et al. (1989) suggested that the D3S2 locus should be located between the loci D3S3 and DNF15S2. Studies on a derivative chromosome 3 from a patient with Greig polysyndactyly showed presence of the D3S3 and D3S2 loci but absence of 'Among the SCLC samples examined by Leduc et al. (1989), seven were also in our SCLC series.

245

the DNF15S2 locus. This order is in agreement with our results on three cytogenetically defined somatic cell hybrids, 45H, A9 Sul-2, and Rag Su31-2-3 (Brauch et al., 1989), but it contradicts the order established earlier by linkage analysis (Leppert et al., 1987). Assuming that D3S2 is proximal to DNF15S2, our data, with one exception, show that the SCLC locus should be distal to 3p21.1 in a region including the loci DNF15S2, THRB, RAFZ,and D3S17. This region contains the VHL gene (Seizinger et al., 1988; Tory et al., 1989). It is possible that the VHL gene is involved in lung carcinoma. Previously we reported that at least 25% of nonSCLC tumors had 3p deletion. T o avoid normal cell contamination in primary non-SCLC tumor samples, which could decrease the observed frequency of loss of heterozygosity, we determined frequency of heterozygosity at 3p loci in 24 nonSCLC cell lines. Reduction in frequency of heterozygosity was significant at most 3p loci tested, with the exception of D3S4, R A F l , and D3S18. We found frequently retained heterozygosity at the loci D3S4, D3S30, DNF15S2, RAFZ, D3S18, and D3S17. This is in contrast to the findings of Kok et al. (1987), who reported allele loss at the DNF15S2 locus in 100% of SCLC and non-SCLC samples. Comparison of the SCLC and non-SCLC panels at 3p loci showed that there were differences between these two groups. Whereas the loci DNF15S2, RAFZ, and D3S18 were always affected in SCLC, they were less frequently affected in nonSCLC. T h e loci most often affected in non-SCLC were D3S3, D3S2, D3S32, and THRB. Retained heterozygosity at THRB (Leduc et al., 1989) and RAFl (Sithanandam et al., 1989) in non-SCLC was previously reported. We conclude that non-SCLCs follow a different pattern of loss of heterozygosity at 3p loci than SCLCs. Non-SCLC may be characterized by a preferential loss of alleles at the loci D3S2 and D3S32. Preferential loss of alleles at the D3S2 locus was recently also reported for carcinoma of the uterine cervix (Yokota et al., 1989). Thus the site of chromosomal rearrangement responsible for the loss of heterozygosity may be different in SCLC and in non-SCLC, and the genes critical for their development appear to be located at different loci. Useful future lines of investigation will be to determine whether there is a common regulatory element located on the short arm of chromosome 3, loss of which could lead to tumors in different tissues. Knowing the critical loci involved in 3p deletion in SCLC and non-SCLC tumors will increase our understanding of these neoplasms and

246

BRAUCH ET AL.

may be a step in the identification of tumor suppressor genes. ACKNOWLEDGMENTS

We are indebted to Dr. J. Eggleston for providing tissue samples from case 44984-96; to Dr. P.G. Rausch for providing blood samples; to C. Riggs for statistical support; to Drs. P. OConnell, Y. Nakamura, and R. White for probes to the loci D3S4, D3S30, D3S3, D3S32, and D3S2; to Dr. B. Carritt for the probe for DNF15S2; to Dr. W.E.C. Bradley for the probe for THRB; and to Dr. U. Rapp for the probe for RAFl. REFERENCES Armitage P (1971) References from proportions. In: Statistical Methods in Medical Research. Oxford: Blackwell Scientific Publications, pp 7 f 7 7 , 111-115. Barker D, Schafer M, White R (1986) Restriction sites containing CpG show higher frequency of polymorphism in human DNA. Cell 36:131-138. Banner T, OBrian J, Nash WG, Rapp UR, Morton CC, Leder P (1984) T h e human homologues of the raf (mil) oncogene are located on human chromosome 3 and 4. Science 223:71-74. Brauch H, Johnson B, Hovis J, Yano T, Gazdar A, Pettengill OS, Graziano S, Sorenson GD, Poiesz BJ, Minna J, Linehan M, Zbar B (1987) Molecular analysis of the short arm of chromosome 3 in small-cell and non-small-cell carcinoma of the lung. N Engl J Med 317: 1109-1 113. Brauch H, Lerman MI, Glenn G, Hampsch K, Grzeschik KH, McBride WO, Zbar B (1989) A somatic cell hybrid panel for physical assignment of 3p probes. Paper presented at the Tenth International Workshop on Human Gene Mapping, New Haven, Connecticut. Carrit B, Welch HM, Parry-Jones NJ (1986) Sequences homologous to the human D l S l locus present on human chromosome 3. Am J Hum Genet 38:428436. de Leij L, Postmus PE, Buys CHCM (1985) Characterization of three new variant type cell lines derived from small cell carcinoma of the lung. Cancer Res 45:6024-6033. Dobrovic A, Houle B, Belouchi A, Bradley WEC (1988) erbArelated sequence coding for DNA-binding hormone receptor localized to chromosome 3p21-3p25 and deleted in small cell lung carcinoma. Cancer Res 45:46824685. Donis-Keller H, Green P, Helms C, Cartinhour S, Weiffenbach B, Stephens K, Keith TP, Bowden DW, Smith DR, Lander ES, Botstein D, Akots G, Rediker KS, Gravius T , Brown VA, Rising MB, Parker C, Powers JA, Watt DE, Kauffman ER, Bricker A, Phipps P, Muller-Kahle H, Fulton T R , Ng S, Schumm JW, Braman JC, Knowlton RG, Barker DF, Crooks SM, Lincoln SE, Daly MJ, Abrahamson J (1987) A genetic linkage map of the human genome. Cell 51:319-337. Drabkin H, Kao FT, Hartz J, Hart I, Gazdar A, Weinberger C, Evans R, Gerber M (1988) Localization of human ERBAL to the 3p22-3p24.1 region of chromosome 3 and variable deletion in small cell lung cancer. Proc Natl Acad Sci USA 85:925%9262. Drabkin H, Sage M, Helms C, Green P, Gemmill R, Smith D, Erickson P, Hart I, Ferguson-Smith A, Ruddle F, Tommerup N (1989) Regional and physical mapping studies characterizing rhe Greig polysyndactyly 3;7 chromosome translocation, r(3;7) (pZl.l;p13). Genomics 4518-529. Fujimoto E, Nakamura Y, Gill J , O’Connel P, Leppert M, Lathrop GM, Lalouel JM, White R (1988) Isolation and mapping of a polymorphic DNA sequence (pEFD145) on chromosome 3 (D3S32). Nucleic Acids Res 16:9357. Gareau JLP, Houle B, Leduc F, Bradley WEC, Dobrovic A (1988) A frequent Hind111 RFLP on chromosome 3~21-25detected by a genomic ErbA beta sequence. Nucleic Acids Res 16:1223. Gazdar AF, Carney DN, Russell EK, Sims HL, Baylin SB, Bunn PA Jr, Guccion JG, Minna J D (1980) Establishment of continuous, clonable cultures of small-cell carcinoma of the lung which have amine precursor uptake and decarboxylation cell properties. Cancer Res 40:3502-3507.

Graziano SL, Cowan BY, Carney DN, Bryke CR, Mitter NS, Johnson BE, Mark GE, Planas AT, Catino JJ, Comis RL, Poiesz BJ (1987) Small cell lung cancer cell line derived from primary tumor with a characteristic deletion of 3p. Cancer Res 47:214& 2155. Ibsen JM, Waters JJ, Twentyman PR, Bleehen NM, Rabbitts PH (1987) Oncogene amplification and chromosomal abnormalities in small cell lung cancer. J Cell Biochem 33267-288. Kok K, Osinga J, Carritt B, Davis MB, van der Hour AH, van der Veen AY, Landsvater RM, de Leij LFMH, Berendsen HH, Postmus PE, Poppema S, Buys CHCM (1987) Deletion of a DNA sequence at the chromosomal region 3p21 in all major types of lung cancer. Nature 330:576581. Leduc F, Brauch H, Haij C, Dobrovic A, Kaye F, Gazdar A, Harbour JW, Pettengill OS, Sorenson GD, van den Berg A, Kok K, Campling B, Paquin F, Bradley WEC, Zbar B, Minna J, Buys C, Ajoub J (1989) Loss of heterozygosity in a gene coding for a thyroid hormone receptor in lung cancers. Am J Hum Genet 44~282-287. Leppert M, OConnel P, Nakamura Y, Camright P, Lathrop M, Lalouel JM, White R (1987) Two linkage groups on chromosome 3. Cytogenet Cell Genet 46:648. Mode F, Kloepfer C, Moisan JP, Mandel JL, Weil D, Grzeschik KH (1984) Assignment of two unique DNA sequences which define restriction fragment length polymorphisms to chromosome 3 and 18. Cytogenet Cell Genet 37:544. Nakamura Y, Culver M, Gillilan S, OConnel P, Leppert M, Lathrop GM, Lalouel JM, White R (1987) Isolation and mapping of a polymorphic DNA sequence pYNZ86.1 on chromasome 3 (D3.530). Nucleic Acids Res 15:10079. Naylor SL, Sakaguchi AY, Barker D, White R, Shows T B (1984) DNA polymorphic loci mapped to human chromosomes 3, 5, 9, 11, 17, 18, and 22. Proc Natl Acad Sci USA 81:2447-2451. Naylor SL, Johnson BE, Minna JD, Sakaguchi AY (1987) Loss of heterozygosicy of chromosome 3p markers in small-cell lung cancer. Nature 329:451454. Pettengill OS, Sorenson GD, Wurster-Hill DH, Curphey TJ, Noll WW, Care CC, Maurer L H (1980) Isolation and growth characteristics of continuous cell lines from small-cell carcinoma of the lung. Cancer 45:906-918. Rider SH, Gnrman PA, Shipley JM, Moore G, Vennstrom B, Solomon E, Sheer E (1987) Localization of the oncogene cerbA2 to human chromosome 3. Ann Hum Genet 51:153-160. Seizinger BR, Rouleau GA, Ozelius LJ, Lane AH, Farmer GE, Lamiell JM, Haines J, Yuen JWM, Collins D, Majoor-Krakauer D, Banner T , Mathew C, Rubenstein A, Halperin J, McConkieRosell A, Green JS, Trofatter JA, Ponder BA, Eierman L, Bowmer MI, Schimke R, Oostra B, Aronin N, Smith DI, Drabkin H, Waziri MH, Hobbs WJ, Martuza RL, Conneally PM, Hsia YE, Gusella JF (1988) Van Hippel-Lindau disease maps to the region of chromosome 3 associated with renal cell carcinoma. Nature 332268-269. Sithanandam G, Dean M, Brennscheidt U, Beck T, Gazdar A, Minna JD, Brauch H, Zbar B, Rapp UR (1989) Loss of heterozygosity at the c-raf locus in small cell carcinoma. Oncogene 4:451455. Shih C , Weinberger RA (1982) Isolation of a transforming sequence from a human bladder carcinoma cell line. Cell 29:161-169. Steel RGD, Torrie J H (1980) Fisher’s exact test. In: Principles and Procedures of Statistics: A Biometrical Approach. New York: McGraw-Hill Book Company, pp 504-507. Tory K, Brauch H, Linehan M, Barba D, Oldfield E, Filling-Katz M, Nakamura Y, White R, Seizinger B, Lerman MI, Zbar B (1989) Specific genetic change in tumors associated with van Hippel-Lindau disease. J Natl Cancer Inst 81:1097-1101. Whang-Peng J, Kao-Shan CS, Lee EC, Bunn PA, Carney DN, Gazdar AF, Minna J D (1982) Specific chromosomal defecr associated with human small-cell lung cancer: deletion 3p( 14-23). Science 215:181-182. Yokota J , Wada M, Shimosato Y, Terada M, Sugimura T (1987) Loss of heterozygosity on chromosome 3, 13, and 17 in smallcell carcinoma and on chromosome 3 in adenocarcinoma of the lung. Proc Natl Acad Sci USA 84:9252-9256. Ynkota J, Tsukada Y, Nakajima T , Gotoh M, Shimosato Y, Mori N, Tsunokawa Y, Sugimura T, Terada M (1989) Loss of heterozygosity on the short arm of chromosome 3 in carcinoma of the uterine cervix. Cancer Res 49:3598-3601. Zech L, Bergh J, Nilsson K (1985) Karyotypic characterization of established cell lines and short-term cultures of human lung cancers. Cancer Genet Cytogenet 15:335-347.