Transduction effect of antisense K-ras on malignant phenotypes in gastric cancer cells

Cancer Letters 157 (2000) 1±7

www.elsevier.com/locate/canlet

Transduction effect of antisense K-ras on malignant phenotypes in gastric cancer cells Jae J. Song a, Heuiran Lee b, Eunhee Kim a, Yeon S. Kim c, Nae C. Yoo a,d, Jae K. Roh a,d, Byung S. Kim a,d, Joohang Kim a,d,* a

Institute for Cancer Research, Yonsei University College of Medicine, Yonsei Cancer Center, Seoul, 120 752, South Korea b Department of Microbiology, University of Ulsan, College of Medicine, Ulsan, South Korea c Korea Research Institute of Bioscience & Technology, Seoul, South Korea d Yonsei Cancer Center, Yonsei University College of Medicine, Seoul, South Korea Received 6 December 1999; revised 1 February 2000; accepted 24 February 2000

Abstract The antitumoral effects of antisense RNA to K-ras were investigated in gastric cancer cell lines by examining the level of Kras expression and the tumorigenicity in vitro and in vivo. Polymerase chain reaction-single strand conformation polymorphism (PCR-SSCP), DNA sequencing, and immunoblotting analysis revealed that YCC-1 gastric cancer cells overexpressed wild type K-ras, whereas YCC-2 cells had a homozygous mutation in codon 12 from GGT (glycine) to AGT (serine), while SNU-1 cells had a heterozygous mutation to GAT (asparagine) in the identical position. Both YCC-1 and YCC-2 cells were transduced by LNC-AS/K-ras containing the antisense 2.2 kb genomic K-ras DNA fragment covering exon 2 and exon 3 speci®c for K-ras. The application of antisense K-ras signi®cantly downregulated the expression of K-ras and had no in¯uence on the expression of either H-ras or N-ras. The antisense-transduced YCC-2 cells grew considerably slower than the control group transduced by LNCX, whereas the growth inhibition of antisense-transduced YCC-1 cells was less prominent than that of transduced YCC-2 cells. In addition, the tumorigenicity of YCC-2 cells transduced by LNC-AS/K-ras was totally lost. Therefore, our results imply that the speci®c inhibition of K-ras p21 protein can be accomplished by introducing the antisense covering the K-ras- speci®c region to gastric cancer cells with aberrant K-ras expression, resulting in a reduction of the growth rate and suppression of tumorigenicity. q 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Gastric cancer; K-ras; Antisense; Gene therapy

1. Introduction Human cancer development including gastric cancer involves multiple changes in oncogenes and tumor suppressor genes, which cause genetic instability and poor prognosis. Multiple genetic abnormalities including ras mutation, inactivation of p53, aberrant * Corresponding author. Tel.: 182-2-361-7622; fax: 1 82-2-3921508. E-mail address: [email protected] (J. Kim).

expression of c-met, decreased expression of p21, or c-erbB2 gene ampli®cation are often observed in gastric cancer patients [1]. The pattern in the accumulation of multiple genetic changes in gastric cancer shows a close similarity to colorectal cancer [2]. Ras encodes a 21 kDa protein resembling a guanine nucleotide-binding protein that is ubiquitously distributed in vertebrate cells [3,4]. As a potential intermediate in a signal transduction pathway, ras binds to GDP and GTP with high af®nity and possess a low level of

0304-3835/00/$ - see front matter q 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(00)00417-1

2

J.J. Song et al. / Cancer Letters 157 (2000) 1±7

intrinsic GTPase activity [5]. Ras proto-oncogene is subdivided into K-, H-, and N-ras. K-ras and H-ras were ®rst identi®ed as oncogenes of acutely transforming RNA tumor viruses [4,6,7]. Aberrant expression of K-ras, such as overexpression of wild type or point mutations, notably at codons 12, 13, or 61 are frequently identi®ed in a variety of human cancers [3,8,9]. Activating mutations, such as overexpression or point mutations, are capable of transducing a strong mitogenic signal to stimulate cell proliferation, which is directly involved in the initiation and development of human neoplasia [7,10,11]. Downregulation of highly active K-ras could, therefore, result in antitumoral effects and this possibility has been extensively investigated by applying diverse strategies in a variety of human cancers, such as the treatment of antibodies to K-ras or the incorporation of a dominant-negative Kras mutant into target cells [4]. Another strategy to regulate aberrant K-ras expression is based on an antisense modality speci®c to K-ras. A number of reports have shown the growth inhibitory effect of antisense Kras on several human cancer cells [12±14]. Recently, Kita et al. reported growth inhibition of human pancreatic cancer cell lines by introducing an antisense oligonucleotide targeting a mutated K-ras region [11]. However, there have been no reports on the antitumoral effect of an antisense speci®c to K-ras in gastric cancer cells. In this study, we employed a retroviral vector to introduce K-ras antisense into human gastric cancer cell lines containing abnormal K-ras expression and examined the effect on ras expression, morphologic changes of transduced cells, growth rate, and the tumorigenicity in athymic nu/nu mice.

2. Materials and methods 2.1. Cell lines and culture conditions All the cell lines in this study were maintained in Dulbecco's modi®ed Eagle's medium supplemented with 10% fetal bovine serum, l-glutamine (2 mM), penicillin (100 IU/ml), and streptomycin (50 mg/ml). YCC-1, YCC-2, YCC-4, and YCC-6 are human gastric cancer cells established at Yonsei Cancer Center. SNU-1 and SNU-5 are human gastric cancer cells established at Seoul National University. KatoIII,

NCI-N87, and AGS are human gastric cancer cells purchased from the American Type Culture Collection (Rockville, MD). PA317 and PG13 cells are retrovirus packaging cell lines. 2.2. Polymerase chain reaction-single strand conformation polymorphism (PCR-SSCP) PCR-SSCP was performed as described previously [15]. Brie¯y, 0.1 mg of genomic DNA from gastric cancer cell lines was subjected to 30 cycles of PCR ampli®cation with oligonucleotide primers 5 0 endlabeled with 32P. Primers covering exon 1 were 5 0 primer (5 0 -ATGACTGAATATAAACTTGT-3 0 ) and 3 0 primer (5 0 -CTCTATTGTTGGATCATAT-3 0 ). Pri-mers covering exon 2 were 5 0 primer (5 0 TTCCTACAGGAAGCAAGTAGTA-3 0 ) and 3 0 primer (5 0 -AGAAA GCCTCCCCAGTCCT-3 0 ). One microliter of PCR product was resolved by 6% nondenaturing polyacrylamide gel electrophoresis. The band was visualized by exposure to X-ray ®lm overnight at 2708C and developing the ®lm using an automatic developer. 2.3. Direct nucleotide sequencing PCR products were subjected to direct nucleotide sequencing. Brie¯y, 2 mg of reampli®ed PCR product was sequenced with 5 0 end-labeled sequencing primer using the PCR product sequencing kit (Amersham Pharmacia, Piscataway, NJ). The sequencing primers (5 0 -CTGAATATAAACTTGTGG-3 0 and 5 0 -ACAGGAAGCAAGTAGTAA-3 0 ) corresponding to codon 12, 13, and 61 were utilized. 2.4. Construction of retroviral vector and transduction of gastric cancer cells A normal wild-type 2.2 kb K-ras genomic DNA segment carrying the second and third exons together with ¯anking intron sequences was ampli®ed by PCR by using 5 0 primer (5 0 -GTAATAATCCAGACTGTGTTTCTCCC-3 0 ) and 3 0 primer (5 0 -GAAGCAATGCCCTCTCAA GAGAC-3 0 ). The resultant PCR fragment was subcloned into LNCX retroviral vector containing cytomegalovirus (CMV) early promoter and SV40 polyA tail by digesting LNCX with HpaI. The antisense orientation of the cloned PCR fragment was con®rmed by the presence of a distinctive 1.8 kb

J.J. Song et al. / Cancer Letters 157 (2000) 1±7

band after digestion with XbaI (data not shown). To generate recombinant retrovirus, the amphotropic packaging cell line PA317 was transfected with LNC-AS/K-ras or LNCX (used as a negative control) by a calcium phosphate coprecipitation method. After 48 h, the transfected cells were placed in a medium containing 400 mg/ml of G418. About 2 weeks later, G418-resistant colonies were con®rmed for the production of recombinant retrovirus and expanded into large cultures. Another packaging cell line, PG13, was then infected with the virus soup from PA317 producer cells, selected, and con®rmed for virus production. Recombinant retrovirus of a higher viral titer was obtained by this procedure, routinely providing about 5 £ 10 6 colony forming units (cfu)/ ml. Then, YCC-1 or YCC-2 cells were infected with LNCX or LNC-AS/K-ras recombinant retrovirus and selected in 1 mg/ml of G418. 2.5. PCR for neomycin selection marker Cells (1 £ 10 7) were lysed in lysis buffer and genomic DNA was recovered as described in the manufacture's manual (Qiagen, Valencia, CA). Five microliters of DNA was used for the ampli®cation of the target gene by PCR 5 0 primer (5 0 -GGCACAACAG ACAATCGGCTA-3 0 ) and 3 0 primer (5 0 -CTACGGACGAACGGCTTATA-3 0 ) covering a part of the neomycin phosphotransferase gene generated 516 bp PCR product in the presence of the neomycin gene. 2.6. Immunoblotting analysis of ras protein YCC-1 (2 £ 10 6 or 1 £ 10 7) and YCC-2 (1 £ 10 7) cells were lysed in 1 ml of TBS (10 mM Tris, (pH 7.5) containing 100 mM NaCl, 1 mM phenylmethylsulfonyl ¯uoride (PMSF), 1% Nonidet P-40, 1% deoxycholate). Twenty-®ve microliters of precleared lysates were resolved by 12% reduced sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) and transferred to polyvinylidene di¯uoride (PVDF) membranes (Amersham Pharmacia, Piscataway, NJ). Immunodetection was performed with 1± 5 mg of mouse monoclonal antibody speci®cally recognizing K-, N-, or H-ras (K-ras; Calbiochem, Cambridge, MA, N-ras and H-ras; Santa Cruz, Santa Cruz, CA) with peroxidase-labeled anti-mouse secondary antibody using the Pierce enhanced-chemi-

3

luminescence system (Amersham Pharmacia, Piscataway, NJ). 2.7. Determination of the growth rate in cultured cells and the tumorigenicity in nu/nu mice Cells (1 £ 10 3) were seeded in 48-well culture plates, harvested, and counted at 48 h intervals to determine the growth rate. Male athymic nu/nu mice were obtained at 5±6 weeks of age from the Korea Research Institute of Chemical Technology, housed, and handled in accordance with the Animal Research Committee Guidelines at Yonsei University. After being quarantined for 1±3 weeks, 1 £ 10 7 in 100 ml of cells was implanted subcutaneously into the abdominal wall. Tumors were measured twice weekly with linear calipers in two orthogonal directions for 2 months. Tumor volume was measured by a following equation: (minor axis) 2 £ major axis £ 0.523. 3. Results and discussion 3.1. Abnormality of K-ras in human gastric cancer cell lines To explore the status of K-ras in nine gastric cancer cells, we performed PCR-SSCP and direct sequencing analysis and immunoblotting analysis (Figs. 1 and 2). Fig. 1A indicated that the PCR fragment for exon 1, including codons 12 and 13 of K-ras, in YCC-2 showed altered mobility, while the PCR fragment for exon 2 covering codon 61 did not show any distinct change in mobility compared with the wild type of K-ras (data not shown). Direct sequencing analysis showed that homozygous mutation from GGT (glycine) to AGT (serine) in YCC-2 and heterozygous change to GAT (asparagine) in SNU-1 at codon 12 had occurred (Fig. 1B). There have been several reports, indicating that detection accuracy of mutated genotype by PCR-SSCP mainly depends on the nature of target sequence, reaction condition for PCR, or electrophoresis condition [16,17]. Previously Orita et al. showed that they could detect ten of 12 mutated sequences for the repair enzyme of Bacillus subtilis (83% accuracy) by their signi®cantly altered mobility [16]. Therefore, it is possible that the heterozygously mutated genotype of K-ras in SNU-1 was not detected under the experimental condition in the

4

J.J. Song et al. / Cancer Letters 157 (2000) 1±7

variety of cancer cells [7]. These changes could be involved in the tumorigenicity of these gastric cancer cells, but not the other gastric cells. 3.2. Transduction of YCC-1 and YCC-2 cells with retrovirus containing antisense K-ras

Fig. 1. (A) Single strand conformation polymorphism analysis of cK-ras exon 1. Genomic DNAs were isolated from nine different gastric cancer cell lines and PCR was carried out with 5 0 endlabeled primers speci®c to exon 1 or 2 of K-ras. PCR products were then separated on 6% non-denaturing polyacrylamide gel to detect the abnormal band mobilities. The PCR fragment containing exon 1 of YCC-2 only showed the abnormal band mobility as indicated by an arrow. (B) DNA sequencing analysis of YCC-2 and SNU-1 in c-K-ras exon 1 PCR products. The PCR products were sequenced with 5 0 end-labeled sequencing primer using the PCR product sequencing kit.

present study. In addition, the expression levels of Kras in gastric cells were investigated by immunoblotting analysis speci®c to K-ras, indicating the overexpression of wild-type K-ras in YCC-1, as well as the detectable expression in YCC-2 or AGS, but not the other gastric cancer cells, such as Kato III, NCI-N87, or SNU-1 (Fig. 2B). The expression level of K-ras in YCC-1 was over 20 ^ 10 times higher than YCC-2 or AGS (data not shown). Therefore, these results indicated that three out of nine gastric cancer cells we tested had aberrant expression of K-ras, such as the overexpression of wild-type K-ras (YCC-1) or the point mutation at codon 12 (YCC-2, SNU-1), which is one of the amino acids most frequently mutated in a

To investigate the effect of antisense K-ras on gastric cancer cells with aberrant K-ras expression, we transduced YCC-1 and YCC-2 cells by infecting them with the retrovirus (LNC-AS/K-ras) containing an antisense K-ras of 2.2 kb wild type genomic DNA segment covering the second and third exons of the genome. The YCC-1 or YCC-2 cells, containing a selection marker, the neomycin phophotransferase gene, were selected by placing the infected cells in a medium containing 1 mg/ml of G418. Fig. 2A showed that a neomycin phosphotransferase gene was readily detected as a 516 bp fragment by PCR in the cells transduced by both LNC-AS/K-ras (lane 4 and 7) and LNCX (lane 3 and 6) as a negative control group but not in non-transduced parental cells (lane 2 and 5). This result implied that both YCC-1 and YCC-2 cells were indeed transduced by the retrovirus containing antisense K-ras. Next, we examined the expression level of K-ras by immunoblotting analysis using K-ras speci®c monoclonal antibody (Fig. 2B). The data indicated that the expression of K-ras in YCC-1 cells was extremely high compared with that of YCC-2 cells. However, the expression of K-ras was consistently downregulated in both YCC-1 (80 ^ 10% reduction rate by densitometer) and YCC-2 cells (90 ^ 5% reduction rate) transduced with LNC-AS/K-ras. The level of K-ras p21 protein expression in YCC-1 cells was yet higher than that in YCC-2 cells as indicated by the intensity of the band in Fig. 2B. The carcinogenic nature of YCC-1 cells despite having a wild type K-ras, could be linked to the overexpression of K-ras protein. In fact, there are several reports implying that preferentially increased expression levels of the wild type K-ras gene were involved in the carcinogenesis of multiple cancer types [14]. In addition, we examined whether the introduction of antisense K-ras in¯uenced the expression level of c-H-ras or c-N-ras by immunoblotting analysis (Fig. 2B). The similar level of c-H-ras and c-N-ras were detected in YCC-1 cells, but not in YCC-2 cells, regardless of the presence of antisense

J.J. Song et al. / Cancer Letters 157 (2000) 1±7

5

Fig. 2. (A) PCR analysis of YCC-1 and 2 cells transduced by retroviral vector. The genomic DNAs of parental YCC-1, 2 were extracted and PCR was carried out with primers speci®c to neomycin phosphotransferase gene. Lane 1, LNCX as a positive control; lane 2, parental YCC-1; lane 3, YCC-1 with LNCX; lane 4, YCC-1 with LNC-AS/K-ras; lane 5, parental YCC-2; lane 6, YCC-2 with LNCX; lane 7, YCC-2 with LNCAS/K-ras. B. Immunoblotting analysis of ras p21 proteins in transduced YCC-1, 2 cells. Cells were lysed in lysis buffer and the cell lysates corresponding to 2.5 £ 10 5 cells, except YCC-1 for K-ras (5 £ 10 4) were loaded on 12% SDS-polyacrylamide as indicated by the expression level of a-actin. Proteins were transferred onto polyvinylidene di¯uoride membranes (PVDF) and probed with a-actin and ras-speci®c monoclonal antibodies. Lane 1, YCC-1 with LNCX; lane 2, YCC-1 with LNC-AS/K-ras; lane 3, YCC-2 with LNCX; lane 4, YCC-2 with LNC-AS/K-ras.

K-ras, implying that antisense K-ras did not alter other ras expression. However, the level of H-ras and N-ras expression was considerably lower than that of K-ras in YCC-1 cells. Taken together, our data indicated that the antisense RNA can speci®cally inhibit K-ras expression by not altering the expression of H- or Nras, which is consistent with other studies [18±20]. In addition, this type of antisense could not distinguish single amino acid substitution shown in YCC-2 cells. Therefore, LNC-AS/K-ras was capable of downregulating K-ras expression in both YCC-1 cells with overexpressed wild type p21 and YCC-2 cells with mutated p21. Finally, the data demonstrated that K-

ras was more abundantly expressed in gastric cancer cell lines compared with H- or N-ras, which supported the fact that K-ras was most prevalently expressed in the gut, lung, and thymus [21]. 3.3. Morphological changes and growth inhibition of gastric cells by antisense K-ras When the cells were established by infecting and selecting then with retrovirus LNC-AS/K-ras or LNCX, no morphological change was observed (data not shown). In addition, the transduced YCC-1 or YCC-2 with LNCX showed no detectable change

6

J.J. Song et al. / Cancer Letters 157 (2000) 1±7

in morphology and growth rate compared with original parent cell lines, YCC-1 or -2 (data not shown).However, we noticed that the YCC-1 or YCC-2 transduced cells with LNC-AS/K-ras grew much more slowly. Fig. 3 showed that the growth rate of YCC-2 cells with LNC-AS/K-ras was signi®cantly reduced (8.3 ^ 1.1% of control group). The growth rate of YCC-1 cells with LNC-AS/K-ras was reduced, but the degree of the reduction rate (48 ^ 6% of control group) was less severe compared with that of YCC-2 cells. These results showed that the antiproliferative effect of antisense K-ras RNA was clearly observed in cancer cells with aberrant expression of K-ras. Yet this downregulating effect was more profound in the cells with a mutated K-ras gene, such as YCC-2 cells, than in the cells with the overexpression of K-ras, such as YCC-1 cells. Previous studies have indicated that the integration of antisense K-ras into the cells without K-ras mutation didn't affect the growth rate of the cells in culture [18,22] or affected them less signi®cantly [23]. Taken together with our results, these facts may suggest that the cells with the wild type K-ras gene better tolerated the reduction of K-ras p21 protein.

Fig. 3. Growth curve of transduced gastric cancer cells. Transduced gastric cancer cells (1 £ 10 3) were seeded in 48-well plates, harvested and counted by a hemocytometer at 24 h intervals. (A) Transduced YCC-1 cells. (B) Transduced YCC-2 cells.

3.4. Effects of antisense K-ras on tumorigenicity in a nude mouse model To investigate the tumorigenicity of gastric cancer cells transduced by LNC-AS/K-ras, 1 £ 10 7 cells of YCC-2/LNC-AS/K-ras as the experimental group or YCC-2/LNCX as the control group were implanted into nu/nu mice subcutaneously and examined for tumor development (Fig. 4). The initial mass continuously decreased for 20 days postimplantation in both groups. However, tumor mass then sharply increased

Fig. 4. Tumorigenicity of YCC-2 cells transduced with LNC-AS/Kras in nu/nu mice. The nude mice (six animals per each group) were injected subcutaneously in the abdominal wall with 1 £ 10 7 transduced YCC-2 cells and tumors were measured with a caliper at the indicated times. (A) Changes in tumor volume. (B) Photographs 2 months postimplantation of nu/nu mice implanted with YCC-2/ LNCX as a control group (upper) and with LNC-AS/K-ras cells (lower).

J.J. Song et al. / Cancer Letters 157 (2000) 1±7

in the control group, while initial mass completely disappeared in experimental group of YCC-2/LNCAS/K-ras. In all six experimental mice did not develop any tumor up to 4 months and remained healthy, whereas mice in ®ve of six control mice developed tumors and eventually died. Therefore, the data indicated that the tumorigenicity of YCC-2 cells in nu/nu mice was completely inhibited by the subsequent reduction of K-ras expression. In summary, this study showed that the antisense covering exon 2 and exon 3 of the K-ras genome is capable of downregulating the expression of K-ras, but not H- or N-ras in gastric cancer cells with the aberrant expression of K-ras. The speci®c downregulation of K-ras gene expression subsequently resulted in both the reduction of the cell proliferation rate and the loss of tumorigenicity of YCC-2 cells with a mutated K-ras. Therefore, this study has interesting implications for the prevention and therapy of gastric cancers, especially having a K-ras mutation. Acknowledgements Drs J.J. Song and H. Lee contributed equally to this work. This paper was supported by the NON Directed Research Fund of the Korea Research Foundation (1995±1996). References [1] E. Tahara, Molecular biology of gastric cancer, World J. Surg. 19 (1995) 484±488. [2] E. Tahara, H. Kuniyasu, W. Yasui, H. Yokozaki, Gene alterations in intestinal metaplasia and gastric cancer, Eur. J. Gastroenterol. Hepatol. 6 (Suppl. 1) (1994) S97±102. [3] M. Barbacid, Ras genes, Ann. Rev. Biochem. 56 (1987) 779± 827. [4] D.R. Lowy, Function and regulation of ras, Annu. Rev. Biochem. 62 (1993) 851±891. [5] F. McCormick, GTPase activating protein: signal transmitter and signal terminator, Cell 56 (1989) 5±8. [6] R.W. Ellis, D. DeFeo, T.Y. Shih, M.A. Gonda, H.A. Young, N. Tsuchida, D.R. Lowy, E. Scolnick, The P21 src genes of Harvey and Kirsten sarcoma viruses originates from divergent members of a family of normal vertebrate genes, Nature 292 (1981) 506±511. [7] G.M. Cooper, Guanine nucleotide binding proteins, in: G.M. Cooper (Ed.), Oncogenes, 2nd ed., Jones and Bartlett, Boston, MA, London, 1995, pp. 222±242.

7

[8] J.L. Bos, Ras oncogenes in humans cancer: a review, Cancer Res. 49 (1989) 4682±4689. [9] J. Heighway, P.S. Hasleton, c-Ki-ras ampli®cation in human lung cancer, Br. J. Cancer 53 (1986) 285±287. [10] R.W. Ruddon, Oncogenes, in: R.W. Ruddon (Ed.), Cancer Biology, 3rd ed., University Press, New York, 1995, pp. 277±317. [11] K. Kita, S. Saito, C.Y. Morioka, A. Watanabe, Growth inhibition of human pancreatic cancer cell lines by anti-sense oligonucleotides speci®c to mutated K-ras genes, Int. J. Cancer 80 (1999) 553±558. [12] E.H. Chang, P.S. Miller, C. Cushman, K. Devadas, K.F. Pirollo, P.O.P. Ts'o, Z.P. Yu, Antisense inhibition of ras p21 expression that is sensitive to a point mutation, Biochemistry 30 (1991) 8283±8286. [13] B.P. Monia, J.F. Johnston, D.J. Ecker, M.A. Zounes, W.F. Lima, S.M. Freier, Selective inhibition of mutant Ha-ras mRNA expression by antisense oligonucleotides, J. Biol. Chem. 267 (1992) 19954±19962. [14] G. Chen, S. Oh, B.P. Monia, D.W. Stacey, Antisense oligonucleotides demonstrate a dominant role of c-Ki-ras proteins in regulating the proliferation of diploid human ®broblasts, J. Biol. Chem. 271 (1996) 28259±28265. [15] J.Y. Cho, J.H. Kim, Y.H. Lee, K.Y. Chung, S.K. Kim, S.J. Gong, N.C. You, H.C. Chung, J.K. Roh, B.S. Kim, Correlation between K-ras gene mutation and prognosis of patients with nonsmall cell lung carcinoma, Cancer 79 (1997) 462±467. [16] M. Orita, T. Sekiya, K. Hayashi, DNA sequence polymorphisms in Alu repeats, Genomics 8 (1990) 271±278. [17] Y. Suzuki, T. Sekiya, K. Hayashi, Allele-speci®c polymerase chain reaction: a method for ampli®cation and sequence determination of a single component among a mixture of sequence variants, Ann. Biochem. 192 (1991) 82±84. [18] T. Mukhopadhyay, M. Tainsky, A.C. Cavender, J.A. Roth, Speci®c inhibition of K-ras expression and tumorigenicity of lung cancer cells by antisense RNA, Cancer Res. 51 (1991) 1744±1748. [19] Y. Zhang, T. Mukhopadhyay, L.A. Donehower, R.N. Georges, J.A. Roth, Retroviral vector-mediated transduction of K-ras antisense RNA into human lung cancer cells inhibits expression of the malignant phenotype, Hum. Gene Ther. 4 (1993) 451±460. [20] R. Alemany, S. Ruan, M. Kataoka, P.E. Koch, T. Mukhopadhyay, R.J. Cristiano, J.A. Roth, W. Zhang, Growth inhibitory effect of anti-K-ras adenovirus on lung cancer cells, Cancer Gene Ther. 3 (1986) 296±301. [21] J. Leon, I. Guerrero, A. Pellicer, Differential expression of the ras gene family in mice, Mol. Cell. Biol. 7 (1987) 1535±1540. [22] R.N. Georges, T. Mukhopadhyay, Y. Zhang, N. Yen, J.A. Roth, Prevention of orthotopic human lung cancer growth by intratracheal instillation of a retroviral antisense K-ras construct, Cancer Res. 53 (1993) 1743±1746. [23] K. Aoki, T. Yoshida, N. Matsumoto, H. Ide, T. Sugimura, M. Terada, Suppression of Ki-ras p21 levels leading to growth inhibition of pancreatic cancer cell lines with Ki-ras mutation but not those without Ki-ras mutation, Mol. Carcinog. 20 (1997) 251±258.