Lack of Imprinting of Three Human Cyclin-Dependent Kinase Inhibitor Genes'

CANCER RESEARCH 57. 926-929. March I. 19971 Lack of Imprinting of Three Human Cyclin-Dependent Kinase Inhibitor Genes'

Gregory J. Cost, Jeffrey S. Thompson, Betty-Anne Reichard, Jae-Yong Lee, and Andrew P. Feinbere

Departments of Molecular Biology & Genetics 1G. J. C'., A. P. Fl, Medicine If. S. T, B-A. R.. J-Y. L, A. P. F.], and Oncology ÃA.P. F.!, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205

ABSTRACT ternal allele in neuroblastoma, suggesting that a tumor suppressor gene on this band is imprinted (10). In this study, we examined the Genomic imprinting is an epigenetic modification In the germilne lead possibility that imprinting might be a common regulatory mechanism Ing to parental allele-specific gene expression in somatic cells. We have of CKIs, focusing on the two genes most closely related to CDKNJC, previously found that imprinted genes can be abnormally expressed or silenced in tumors and that the cyclin-dependent kinase inhibitor (CKI) namely CDKNJA and CDKNJB, as well as one of the INK4 family @ CDKNJC (p57KIP2) normally imprinted, with preferential expression of members, CDKN2C. Table 1 summarizes the CKIs studied here. the maternal allele. Here we analyze the imprinting status of three addi tional CKIs, the abnormal expression and/or chroinosomal localization of which has been implicated in human malignancy: CDKNJA, CDKNJB, MATERIALS AND METHODS and CDKN2C. Allele-specific expression was examined by reverse tran scription-PCR, using primers that span transcribed polymorphisms as Identification of Individuals Heterozygous for Restriction Site Poly well as exon/intron boundaries, to distinguish eDNA products from morphisms. Samples from 22 fetuses were obtained from the University of genomic DNA. Biallelic expression was observed for all three genes in both Washington Fetal Tissue Bank, and 16adult tissue samples were obtained from fetal and adult tissues. Thus, genomic imprinting is not a generalized snap-frozen surgical specimens. Genomic DNA was isolated as previously feature of CKIs. described (2). PCR was performed in Hot Start Micro 50 reaction tubes (Molecular Bio-Products) in a volume of 50 s.d. Reactions contained 20 mrsi Tris-HCI, pH 8.4, 50 mMKCI, 1.5 m@MgCl, 0.25 mMeach dNTP, 2.5 units INTRODUCTION of Taq polymerase (Life Technologies, Inc.), 100ng of genomic DNA, and 0.5 @.LMeach of the appropriate primers. Amplifications were performed in a Genomic imprinting is defined as an epigenetic modification of a Stratagene Robocycler and consisted of an initial denaturation at 94Â°Cfor2 specific chromosome in the gamete or zygote leading to differential mm followed by 35 cycles of denaturation at 94Â°Cfor45 s, annealing at 50 expression of maternal and paternal alleles in the somatic cells of the (CDKNJA), 58 (CDKN2C), or 62Â°C(CDKNIB)for 45 s, and extension at 72Â°C offspring. Loss of imprinting in cancer can lead to abnormal expres for 90 s. After cycling, the reactions were brought to 72Â°Cfor10 mm, then sion of growth-stimulatory genes, as well as epigenetic silencing of cooled to 6Â°C.The primers used were as follows: CDKN1A-7, 5'-CCCAGG growth-inhibitory genes (for a review, see Ref. 1). A small number of GAAGGGTGTCCFG-3' (forward); CDKNIA-8, 5'-GGGCGGCCAGGG TATGTAC-3' (reverse); CDKN1B-S3, 5'-AGTACGAGTGGCAAGAG human genomic regions have been identified that harbor multiple GTG-3' (forward); CDKN1B-A3, ATItTATGGTFGGGGAAGGGT-3' imprinted genes. One of these is I lpl5.S, which includes IGF2, (reverse); CDKN2C-2, 5'-TGCCAGGAGACTGCTA-CTI'AGAGG-3' (for expressed from the paternal allele, and H19, expressed from the ward); and CDKN2C-3, 5'-TfA1TGAAGAmGTGGCTCCCCC-3' (re maternal allele. We and our collaborators (2) recently localized the verse). Amplified products ofthe CDKNIA, CDKNIB, and CDKN2C loci were CKI3 gene CDKNIC (@57KlP2)to the same 700-kb genomic region digested with 20 units of PstI, BglI, and BstEII, respectively (New England and found that this gene is also imprinted, with preferential expression BioLabs). DNA fragments were electrophoresed in Tris-Acetate-EDTA buffer of the maternal allele. We also found that CDKNJC, like IGF2 and through 2% electrophoresis-grade agarose (Life Technologies, Inc.)/l% H19, can undergo abnormal imprinting and expression in cancer (3). SeaKem agarose (FMC Bioproducts). CDKNJC is the most recently identified member of a class of Assay for Allelic Expression of CKIs by RT-PCR. RNA fromheterozy gous individuals was obtained as described previously (2) or by isolation with growth-regulatory factors called CKIs. CKIs function by binding to RNAzo1B (Tel-Test Inc.) according to the manufacturer's instructions. cDNAs cyclin-dependent kinases and inactivating their proliferative activity. were synthesized in a reaction volume of 25 @lbyincubating RNA, 0.5 units CDKNJC shares conserved domains with CDKNJA (p2JWAFI/CIP!) a of RNAse inhibitor (Promega), and 20 units of DNase I (Pharmacia) in 1X target of p53-mediated growth arrest (4). CDKNJA also undergoes first-strand buffer (Life Technologies, Inc.) for 90 mm at 37Â°C. DNase I was somatic mutation in a Burkitt's lymphoma (5). CDKNJC is also subsequently thermally inactivated by heating to 94Â°Cfor 10 mm. After structurally related to CDKNIB (p27K@), which has been found only oligo(dT)15 (Boehringer Mannheim; CDKN1A and CDKN2C reactions) or rarely to be mutated in human tumors (6) but is localized to chromo CDKNIB-S3 (CDKNI B reaction) was added to a final concentration of 20 somal band l2pl3, which shows LOH in hematopoietic malignancies ng4tl, the reaction was heated to 72Â°Cfor10mm and then cooled on ice. DTF (7) and non-small cell lung cancer (8). Members of the INK4 family was added to a final concentration of 8 mM, dNTPs to a final concentration of of CKIs share homologous domains with each other but differ from 0.4 msi, and RNAse inhibitor to a final concentration of 0.04 [email protected] mixtures were then incubated at 42Â°Cfor2 mm, I p1 (200 units) of Superscript CDKNJA, CDKNIB, and CDKNJC. Four of the INK4 genes, II Reverse Transcriptase (Life Technologies, Inc.) was added, and the reactions CDKN2B (piS), CDKN2A (p16), CDKN2C (p18), and CDKN2D were allowed to proceed for 60 mm at 42Â°C.Thereaction was terminated by (p19) exhibit ankyrin repeat motifs (9). CDKN2C is localized to the addition of I unit of RNase H (Life Technologies, Inc.) followed by chromosomal band 1p32, and lp shows preferential loss of the ma incubation at 37Â°C for 20 mm. cDNAs were purified by phenol-chloroform extraction and ethanol precipitation and resuspended in IX PCR buffer (Life Received I2/I 3/96; accepted 1/6/97. Technologies, Inc.). Two-pi aliquots of the synthesized cDNAs were used for The costs of publication of this article were defrayed in part by the payment of page amplification by PCR as described above, with the addition of seven cycles (42 charges. This article must therefore be hereby marked advertisement in accordance with cycles total) to the procedure, except for CDKN 1A, for which 30 or 35 cycles 18 U.S.C. Section 1734 solely to indicate this fact. @ This work was supported by NIH Grant CA65145 (to A. P. F.) and Biochemistry and were used, to avoid heterodimer formation. Oligonucleotide primers used Cellular & Molecular Biology Training Grant GM07445 (to G. J. C.). for PCR were designed to distinguish cDNA-derived PCR products from that 2 To whom requests for reprints should be addressed, at Johns Hopkins University of contaminating genomic DNA and were as follows: CDKN1A-9, 5'- School of Medicine. Ross 1064, 720 Rutland Avenue, Baltimore, MD 21205. Phone: CAGGGGACAGCAGAGGAAGACC-3' (forward); CDKN1A-8 (reverse); (410)614-3489; Fax: (410)614-9819. 3 The abbreviations used are: CKI, cyclin-dependent kinase inhibitor; RT-PCR, reverse CDKNIB-S3 (forward), CDKNIB-3, 5'-ACATITIt'fltTGTfCTGUG transcription-PCR; LOH, loss of heterozygosity. G-3' (reverse); CDKN2C-l, 5'-TGTAAACGTCAATGCACAAAATGG-3' 926

AnalysisNameAliasLocationSuggested Table I CKI

associationReferenceCDKNIA tumor target; Burkitt lymphoma 5 6â€”8CDKN2Cp181CDKNIB@2jWAFI/Ci@I @27KlP16p21 l2pl3p53 Hematopoietic malignancies; NSCLC4, p32NeuroblastomaI 0

(forward); and CDKN2C-3 (reverse). Optimal annealing temperatures were as Analysis of CDKN2C. A similar approach was used for CDKN2C follows: CDKN1A, 58Â°C;CDKN1B, 54Â°C;andCDKN2C, 58Â°C.RT-PCR as for CDKNJA and CDKNJB. In this case, primers were used that products were gel purified by electrophoresis, the bands were excised, and the spanned intron 1 and a polymorphic BstEII site (12). Three of 22 DNA was extracted and concentrated with the QiaQuick Gel Extraction Kit fetuses were heterozygous for this polymorphism, corresponding to a (Qiagen) according to the manufacturer's recommendations, with an additional 7% allele frequency. All tissues examined (including the liver, lung, centrifugation to remove all traces of silica resin. Isolated RT-PCR products and placenta) from this individual showed biallelic expression, and were digested with 20 units of the appropriate enzyme for 12 h, followed by quantitation again showed equal representation of both alleles (Fig. 2). incubation with an additional 20 units for 4 h. Electrophoresis was carried out as described above, and restricted fragments were transferred to Nytran nylon We had shown previously that the human CDKNJC gene is nor membrane (Amersham Corp.). Membranes were probed with 32P-labeledoh mally imprinted with preferential expression of the maternal allele (2), gonucleotides homologous to sequences internal to the primers used for suggesting that genomic imprinting might be a regulatory feature of amplification. Quantification of the expression of each allele was determined functionally related genes. The major result ofthis study is that at least by digital integration with a Molecular Dynamics Phosphorlmager. three other CKIs, two of which (CDKNJA and CDKNIB) show strong homology to CDKNIC, show no evidence of imprinting. All three RESULTS AND DISCUSSION genes studied map to chromosomal regions that show LOH in tumors: CDKNJBis on l2pl3, whichshowsLOHin hematopoieticmalig Analysis of CDKNJA. To assay allele-specific expression of nancies (7) and non-small cell lung cancer (8); CDKNJA is on 6p2l, CDKNJA, we designed primers that spanned a known PstI polymor which shows LOH in ovarian carcinomas (13); and CDKN2C is on phism in exon 3 (11). To identify patients heterozygous for the chromosome lp32, which shows LOH in breast cancer (14). lp32 also polymorphism, primers were designed within intron 2 and exon 3 (see undergoes preferential LOH of the maternal allele in advanced neu â€œseeMaterials and Methodsâ€•).DNA was extracted from 22 samples from separate fetal specimens, PCR was performed, and the products were digested with PstI. From this analysis, two fetuses were identi CDKN1A fled as being heterozygous for the PstI site. The overall frequency of the noncutting allele was observed to be 6%, very close to the 9% FP Pst frequency observed by Law and Deka (11). RNA was prepared from the lungs, intestines, liver, and kidneys of each of these fetuses, F Exon2f. I Exon3 reversed transcribed, and PCR amplified using primers spanning

intron 2 and the PstI polymorphism (Fig. 1). PCR products were DNA 1235bp digested with PstI and analyzed by gel electrophoresis. In all tissues examined, biallelic expression was observed (Fig. 2). To confirm the cDNA @-â€”@â€”-â€”-â€”.â€”@ 335 bp apparent similarity of expression of the two alleles, DNA was ampli fled in an analogous manner, transferred by Southern blotting, and hybridized with an internal probe for CDKN1A. The filter was then CDKN1B subjected to Phosphorlmager analysis, which revealed identical am plification of both DNA and RNA. Thus, CDKNJA showed no cvi F BgII P dence of imprinting in two independent fetuses. â€”xl - @ To rule out the possibility that genomic imprinting might be man I Exoni I Exon2 ifest as allele-specific expression postnatally, 16 additional adult colons were analyzed for heterozygosity of the polymorphism, of DNA 826 bp

which 3 heterozygotes were identified. RNA was extracted from these cDNA @@@â€¢â€”Â°@- 304 bp three samples and analyzed as described above. As observed in the fetal tissue samples, biallelic expression was seen in each case (Fig. 2). Thus, there was no evidence for genomic imprinting of CDKNJA. CDKN2C Analysis of CDKNJB. CDKNJB, like CDKNJA, shares homology with the CM domain of CDKNJC and was therefore analyzed for P BstEII evidence of genomic imprinting. To assay allele-specific expression, primers in exon 1 and intron 1 were designed that spanned a known I Exoni @- â€”I-Exon2 BglI polymorphic site (Ref. 6; Fig. 1). Analysis of DNA from 12 fetal

specimens and 16 adult specimens identified 1 fetal specimen and 5 DNA >900bp adult specimens that were informative for this polymorphic marker. @ cDNA 424 bp RNA was extracted from liver, heart, placenta, and colon, and sub jected to RT-PCR and BglI digestion. The results of these experiments performed in duplicate showed biallelic expression in all tissues Fig. 1. Schematic representation of the partial genomic loci of CDKN1A, CDKNJB, and CDKN2C. Boxes, exons; lines, introns. The positions of restriction site polymor examined. Again, Southern hybridization performed with a probe phisms used to distinguish between alleles are annotated, as are the primers used for internal to CDKNJB followed by Phosphorlmager analysis revealed amplification of cDNAs (arrows); F, forward; R, reverse. The locations of oligonucleo tides used as probes (P) for quantitation of allele-specific expression are shown as bars. equal expression ofboth alleles of the gene (Fig. 2). Thus, no evidence Beneath each locus is shown the expected amphicon size for both DNA (approximate) and for genomic imprinting was found. reverse-transcribed cDNA. 927

DNA RNA DNA ANA DNA RNA DNA RNA DNA lu nt liv kd p1 col cot @ a- â€¢ _ @-. . b.1@O@ I@

Fig. 2. RT-PCR analysis of CDKNJA, CDKNJB, and CDKN2C CDKN1B expression. Amplified cDNA or genomic DNA products derived from the indicated fetal tissues or adult colon were restricted with the appropriate enzyme as detailed in â€œMaterialsandMethods.' F-i A-i Â£4,kidney; lit. heart; ml. intestine; lu, lung; lit', liver; p1, placenta; col, colon; F, fetus: A, adult. The two DNA bands of the CDKN2C b allele are nearly identical in size (165 and 187 bp) and could not RNA DNA be resolved by electrophoresis. The genomic DNA products are liv @l ht larger than the cDNA products. but are aligned for clarity.

CDKNC2 F-i

DNA RNA liv lu p1 a-

roblastoma (10). Although there are likely to be imprinted genes on at with mutational analysis. As new candidate genes for cancer are least lp32, these CKIs do not appear to be imprinted. Thus, genomic identified and the tumors are scanned for mutation by, for example, imprinting is unlikely to contribute to the role of these genes in tumor single-strand conformational polymorphism-PCR, polymorphisms progression. that are identified incidentally can easily be exploited to determine These results do not exclude the possibility that other CKIs are whether the gene is imprinted or not by RT-PCR analysis of the paired imprinted. Alternatively, epigenetic silencing of these genes may not normal tissues. involve imprinting. For example, CDKN2A (p16) may undergo epi genetic silencing in tumors because some show absence of expression ACKNOWLEDGMENTS and hypermethylation of this gene (15, 16). However, it is important to note that alterations of methylation could be secondary to gene We thank the University of Washington Fetal Tissue Bank, Stanley regulation and influenced by trans-acting factors. Furthermore, mice Hamilton for tissues, and J. Patey for preparing the manuscript. harboring a CDKN2A heterozygous knockout showed no decreased survival or abnormal phenotype, whereas homozygous knockout mice REFERENCES showed an increased incidence of malignancy (17). If CDKN2A were imprinted in mice, one would expect the same phenotype in half of the I. Feinberg, A. Genomic imprinting and gene activation in cancer. Nat. Genet.. 4: 110â€”113.1993. independently derived heterozygotes as in the homozygous knockout 2. Matsuoka, S.. Thompson. J. S.. Edwards. M. C., Barletta. J. M.. Grundy. P., Kalikin, animals. Similarly, lack of imprinting of CDKNJB in mice is sug L. M.. Harper, J. W.. Elledge. S. J.. and Feinberg. A. P. Imprinting of the gene encoding a human cyclin-dependent kinase inhibitor, @57K11@2,onchromosome gested by gene disruption experiments, which do not result in the I Ipl5. Proc. Nail. Acad. Sci. USA. 93: 3026â€”3030. 1996. generation of affected heterozygotes (18, 19). 3. Thompson. J. S., Reese, K. J., DeBaun, M. R., Perlman, E. J., and Feinberg. A. P. The data reported here exclude imprinting of the specific genes Reduced expression of the cyclin-dependent kinase inhibitor gene @57K@inWilms' tumor. Cancer Res., 56: 5723â€”5727, 1996. studied, which are of considerable interest to human cancer genetics. 4. El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B.. Parsons, R., Trent, J. M., They are also consistent with the observation that imprinted genes Lin, D., Mercer, W. E., Kinzloer, K. W., and Vogelstein, B. WAFJ. a potential tend to cluster within defined chromosomal domains or may involve mediator of p53 tumor suppression. Cell, 75: 817â€”825, 1993. 5. Bhatia, K., Fan, S., Spangler, G., Weintraub, M., O'Connor, P. M., Judde, J-G., and genes within the same signaling pathway, such as 1GF2 and JGF2R, Magrath. I. A mutant p2! cyclin-dependent kinase inhibitor isolated from a Burkitt's rather than involving separate genes of the same functional class. lymphoma. Cancer Res., 55: 143 1â€”1435,1995. 6. Kawamata. N., Morosetti, R., Miller, C. W.. Park, D., Spirin. K. S.. Nakamaki, T.. Finally, the experiments reported here may serve as a useful para Takeuchi, S., Hatta, Y., Simpson, J., Wilczynski. S., Lee, Y. Y., Bartram, C. R., and digm for rapid investigation of imprinting performed concurrently Koeffler, H. P. Molecular analysis of the cyclin-dependent kinase inhibitor gene 928

p27/Kipi in human malignancies. Cancer Res., 55: 2266â€”2269,1995. loss of heterozygosity on chromosome 6 in human ovarian carcinoma. Br. J. Cancer, 7. Sato, Y., Suto, Y., Pietenpol, J., Golub, T. R., Gilliland, D. G., Davis, E. M., Le Beau, 67: 551â€”559,1993. M. M., Roberts, J. M., Vogelstein, B., Rowley, J. D., and Bohlander, S. K. TEL and 14. Bieche, I., Champeme, M., Matifas, F., Cropp, C. S., Callahan, R., and Lidereau. R. KIP1definethesmallestregionofdeletionsonl2pl3 inhematopoieticmalignancies. Two distinct regions involved in lp deletion in human primary breast cancer. Cancer Blood,86:1525â€”1533.1995. Res., 53: 1990â€”1994, 1993. 8. Takeuchi, S., Mori, N., Koike, M., Slater, J., Park, S., Miller, C. W., Miyoshi, I., and 15. Herman, J. G., Merlo, A., Mao, L., Lapidus, R. G., Issa, J-P. J., Davidson, N. E., Koeffler, H. P. Frequent loss of heterozygosity in region of the KIPI locus in Sidransky, D., and Baylin, S. B. Inactivation of the CDKN2/p16/MTSI gene is non-small cell lung cancer: evidence for a new tumor suppressor gene on the short frequently associated with aberrant DNA methylation in all common human cancers. arm of chromosome 12. Cancer Res., 56: 738â€”740,1996. Cancer Res., 55: 4525â€”4530. 1995. 9. Guan, K-L., Jenkins, C. W., Li, Y., Nichols, M. A., Wu, X., O'Keefe, C. L., 16. Gonzalez-Zulueta, M., Bender, C. M., Yang. A. S., Nguyen, T., Beart, R. W., Matera, A. G., and Xiong, Y. Growth suppression by p18, a @J6@4/MTSIand Van Tornout. J. M., and Jones, P. A. Methylation of the 5' CPG island of the pI4INK4FS/MTS2relatedCDK6 inhibitor, correlates with wild-type pRb function. Genes p16/CDKN2 tumor suppressor gene in normal and transformed human tissues corre Dev., 8: 2939â€”2952,1994. lates with gene silencing. Cancer Res.. 55: 4531â€”4535, 1995. 10. Fong. C., Dracopoli, N. C., White, P. S., Merrill, P. T.. Griffith, R. C., Housman, 17. Serrano, M., Lee, H-W., Chin, L., Cordon-Cardo, C., Beach, D., and DePinho, R. A. D. E., and Brodeur, G. M. Loss of heterozygosity for the short arm of chromosome 1 in human neuroblastomas: correlation with N-myc amplification. Proc. NatI. Acad. Role of the INK4Alocus in tumor suppression and cell mortality. Cell, 85: 27â€”37, Sci. USA. 86: 3753-3757, 1989. I996. @ 11. Law, J. C., and Deka, A. Identification of a PstI polymorphism in the 18. Kiyokawa, H., Kineman, R. D., Manova-Todorova, K. 0., Soares, V. C., Hoffman, cyclin-dependent kinase inhibitor gene. Hum. Genet., 95: 459â€”460, 1995. E. S., Ono, M., Khanam, D., Hayday, A. C., Frohman, L. A.. and Koff, A. Enhanced 12. Okamoto, A., Hussain, S. P., Hagiwara, K., Spillare, E. A., Rusin, M. R., Demetrick, growth of mice lacking the cyclin-dependent kinase inhibitor function of @27Ki@/â€¢ D. J., Serrano, M., Hannon, G. J., Shiseki, M.. Zariwala, M., Xiong, Y.. Beach, D. H., Cell, 85: 721â€”732,1996. Yokota, J., and Harris, C. C. Mutations in the p161@4/MTS1/CDKN2, @j5INK4lJ/ 19. Fero, M. L., Rivkin, M., Tasch, M., Porter, P., Carow, C. E., Firpo, E., Polyak. K., MTS2, and p!8 genes in primary and metastatic lung cancer. Cancer Res., 55: Tsai, L-H., Broudy, V., Perlmutter, R. M., Kaushansky, K., and Roberts, J. M. A 1448â€”1451,1995. syndrome of multiorgan hyperplasia with features of gigantism. tumorigenesis, and 13. Foulkes, W. D., Ragoussis, J., Stamp, 0. W., Allan, 0. J., and Trowsdale, J. Frequent female sterility in p27K@@/@deficientmice.Cell, 85: 733â€”744,1996.

929

Gregory J. Cost, Jeffrey S. Thompson, Betty-Anne Reichard, et al.

Cancer Res 1997;57:926-929.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/57/5/926

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/57/5/926. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.