Loss of Heterozygosity in Human Primary Prostate Carcinomas: a Possible Tumor Suppressor Gene at 7Q31.11

ICANCERRESEARCH54, 6370-6373, December 15, 19941 Advances in Brief

Loss of Heterozygosity in Human Primary Prostate Carcinomas: A Possible Tumor Suppressor Gene at 7q31.11

Jean C. Zenklusen, Janet C. Thompson, Patricia Troncoso, Jacob Kagan, and Claudio J. Conti2

Department of Carcinogenesis, The University of Texas M. D. Anderson Cancer Center, Science Parkâ€”ResearchDivision, Smithville, Texas 78957 Ii. C. Z, J. C. T.. C. I. C.]. and Department of Pathology [P. T.J and Division of Laboratory Medicine if. K.], The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030

Abstract Although cytogenetic techniques are useful, they do not detect the entire spectrum of inactivating events; for example, microdeletions We studied loss of heterozygosity (LOH) on human chromosome 7q to and homologous recombination with a defective chromatid (20, 21) determine the location of a putative tumor suppressor gene (TSG) in human primaryprostatecarcinomas.Sampleswere obtainedfrom 16 are beyond the range of detection by karyotyping procedures. More primary prostate carcinomas surgically removed from patients at The sensitive molecular methods should be used to screen for genetic University of Texas M. D. Anderson Cancer Center. Paired normal and alterations in tumors and to determine the smallest chromosomal tumorDNAswereusedas templatefor PCRamplificationofa set of 14 region involved in those alterations. Therefore, LOH analysis of DNA CA microsatellite repeats on 7q21-qter. Twelve of 16 cases studied had extracted from solid tumors is the method of choice for mapping TSG LOHat one or more loci on 7q. Eighty-threepercentLOH (five of six (22). In the case of prostate carcinomas, LOH of chromosomes 8, 10, informative cases) was detected with D7S522 at 7q31.1â€”7q31.2.Percentage 16, and 18 has been found (23â€”25).Inprostate carcinomas, all of these of LOH was normally distributed around D7S522. The high incidence of chromosomes, with the exception of chromosome 16, have been LOH in primary prostate carcinomas suggests that there is a TSG rele shown by cytogenetical analysis to have alterations. vant to the developmentof prostatecancersat 7q31.1â€”31.2,confirming our previous functional evidence for a TSG at this location. Further To determine the extent and types of alterations in chromosome 7 research needs to be conducted to establish the identity and function of in prostate carcinomas, we used an extensive set of highly polymor this putative TSG. phic markers on q21-qter. By comparing the results obtained with tumor and normal DNA, we were able to narrow down the site of the Introduction TSG to a 1.2 cM region.

Alterations in oncogenes and TSGs3 are considered to be critical Materials and Methods in the multistep process leading to the development of tumors (1, 2), and the succession of these events is very conserved in some Tissue Samples. Prostate tumor specimens were obtained from the prostate types of cancer, such as colon carcinoma (3). Ever since the idea glands and a pelvic lymph node of 16 patients who had undergone pelvic of recessive-acting TSG was formulated (4), cytogenetic tech lymphadenectomy with or without radical prostatectomy for clinically local niques have been used to determine their locations. Although ized prostate carcinoma. The surface of each prostate was inked and the gland producing metaphase spreads of prostate carcinomas cells is tech was sectioned throughout the palpable tumor in a transverse plane perpendic nically difficult, there have been cytogenetic reports that many ular to the posterior surface. A thin slice of the tumor was made with a scalpel chromosomes (1, 3, 6, 7q, 8p, lOq, 12, 17, 18, 21, Y) are frequently and immediately frozen at â€”80Â°C.Anadjacent frozen section was examined histologically to confirm the tumor's presence and to determine the percentage altered (5, 8). Consistent deletions or inversions of part of a of the sample that was tumor. We used only samples that had a minimum of chromosome in tumor cells are indicative of inactivation of a TSG 70% tumor cells. As normal controls, we used cells from peripheral blood or in this region during neoplastic progression (9). For instance, the tip of the seminal vesicle contralateral to the tumor if it had been deletions of chromosome 7 are common in many different cancers histologically confirmed to be tumor free. including ovarian cancer, prostate carcinomas, and malignant my DNA Extraction. Specimenswere processedas describedpreviously(26). eloid disorders (10â€”16). Recently, we reported that LOH in Briefly, frozen seminal vesicle or tumor tissue (0.1â€”0.5g)was pulverized with 7q31.iâ€”31.2 is a very common event in human primary breast a mortar and pestle in liquid nitrogen. The powdered material was thawed in cancers (17). Also, our experiments with microcell fusion transfer 1 ml lysis buffer (4 Mguanidine thiocyanate, 5 mr@isodiumcitrate, pH 7.0, 0.1 of human chromosome 7 into a murine 5CC cell line indicated that M f3-mercaptoethanol, and 0.5% sarcosyl). Next, 0.48 g/ml C5CI2 was added, the inserted chromosome can delay the onset of tumors by 2-fold and the lysate was layered on top of 1.2 ml of 5.7 MCsCl2 and 0.1 MEDTA (pH 7.5) in a Beckman 5W50.i polyallomer tube. The tubes were then to 3-fold and in some cases even completely repress the tumori centrifuged at 35,000 rpm for 20 h at 25Â°C.High-molecular-weight DNA was genic potential of the 5CC cell line. In situ hybridization revealed recovered by aspirating the viscous interface. The DNA in it was precipitated that the clones that reverted to the malignant phenotype had with ethanol, and the pellet was washed with 70% ethanol, dried, and resus expelled the inserted chromosome (18). Moreover, a recent report pended in water. DNA extracted from blood leukocytes was done by standard demonstrated that insertion of an intact human chromosome 7 into procedures (27). immortalized human fibroblast cell lines that had lost the segment CA Microsateffite Repeat Amplification Analysi& FourteenCA micro 7q31â€”32restored the senescence properties of the cells (19). satellite repeats in the 7q21.3-qter region (28) were amplified in a Thermocy cler 9600 (Perkin-Elmer Cetus, Norwalk, CT). The [email protected] mixtures contained 2.5 p1 lOX standard PCR buffer, 100 ng DNA, 1 unit Taq polym Received 9/6/94; accepted 11/2/94. The costs of publication of this article were defrayed in part by the payment of page erase, 400 @Meachprimer, and 200 mMeach deoxynucleotide triphosphate. charges. This article must therefore be hereby marked advertisement in accordance with The hot-start protocol was performed with a Taq polymerase-specific antibody 18 U.S.C. Section 1734 solely to indicate this fact. that inactivates the enzyme and is released during the first denaturation cycle I Supported by National Cancer Institute Grant CA53123. (Taq Start, Clontech, CA). The DNA was amplified for 27 cycles of 20 s of 2 To whom correspondence should be addressed. @ The abbreviations used are: TSGs, tumor suppressor genes; LOH, loss of heterozy denaturation at 94Â°C, 30 s of annealing at 50Â°C, and 15 s of extension at 72Â°C. gosity; SCC, squamous cell carcinoma; SCDR, smallest common deleted region. The number of cycles used was determined to be in the linear part of the 6370

LOH IN PROSTATE CARCINOMAS

TablecarcinomasNo. 1Chromosome 7 LOH in16 primary prostate 1). Fig.1 shows representative photographs of CA microsatellite am casesMarkers(% of informative plifications of the prostate carcinoma samples. As can be seen, the LOHD7S5278 of total)No. of cases with LOH% cases in which LOH occurred were obviously different from the (50)112.5D7S51811(68.7)318.2D7S4968(50)337.5D7S5237 others, as most of the time they had only one of the bands present in the normal DNA lane. (43.8)228.5D7S4866 Our results indicated that the loss of part or all of 7q was a common (37.5)233.3D7S6337 event in prostate cancer. LOH occurred in at least one locus on 7q in 12 (43.8)342.8D7S6778 (50)562.5D7S5226 (75%)of 16tumors.As canbe seenin Fig.2, themostfrequentmarker (37.5)583.3D7S65514(87.5)642.9D7S48011(68.7)327.7D7S4906 lost was D7S522 (7q31.1â€”31.2)(83.3%). Furthermore, the segment flanked by the CA repeats D7S633 and D7S655 (7q22-q31.2), which (37.5)116.2D7S4878 includes D7S522, was lost in 57.8% of the prostate carcinomas studied. (50)112.5D7S4984(25)00D7S5508 Analysis of a histogram of the LOH data from our tumor samples by the Kolmogorov-Smirnov test (30) indicated that they were nor (50)112.5 mally distributed, as would be expected from a stochastic process such as the inactivation of a tumor suppressor gene. amplification process, i.e., before product saturation, permitting us to assume Fig. 3 is a schematic representation of the deleted regions of that equal amounts ofboth alleles would be amplified if no LOH had occurred. The PCR products were separated in a 3.5% Metaphor agarose human chromosome 7 in nine cases in which large deletions could (FMC Bioproducts, Rockland, ME) gel at 5.5 V/cm2 for 3 h in Tris-borate be predicted by the molecular techniques used. The probability of EDTA buffer (89 mMTris-borate, 89 mMboric acid, and 2 mMEDTA, pH7.5) three or more allelic losses in the same fragment being caused by with 0.5 mg/mi ethidium bromide in Tris-borate EDTA buffer and a standard independent events is very small, so such a series of LOH in loading buffer (27). The gel was photographed with a Fotodyne 3â€”4400UV contiguous markers is more likely due to deletion of the entire transilluminator (Fotodyne Inc., New Berlin, WI) and Polaroid Positive-Neg segment. In most of our samples, these deletions were interstitial. ative 4 X 5 Instant Film (Polaroid Corp., Cambridge, MA). Comparison of the deletions revealed a 1.2 cM SCDR that is Determination of LOH. We considered a sample to have LOH if an entire flanked by the CA repeats D7S677 and D7S522. Only one case band was absent or if the band was less than 30% as intense as the normal band with LOH (42636) had not lost this fragment. In that case, how (29). The signal intensity of fragments was determined by densitometry, by ever, a deletion was detected proximal to the SCDR that seems to visual examination by two reviewers, or both. Although PCR amplification cannot be considered to be quantitative, we optimized the PCR conditions so be unrelated to the inactivation of the TSG. that equal amounts of template should have produced equal amounts of amplified product. We used 27 amplification cycles, which we demonstrated to Discussion be in the linear part of the amplification process, i.e., before product saturation Although prostate cancer is the most prevalent neoplasia among (data not shown). We also conducted a series of titrations with different men (31), there is very little information on its etiology and mecha proportions of homozygous and heterozygous templates to assess the influence of stromal tissue contamination of our amplification reactions. We determined nisms of progression. LOH of chromosomes 8, 10, 16, and 18(23â€”25) that we could detect as little as 30% contamination by heterozygous template has been consistently reported in prostate cancers, suggesting the in homozygous DNA (data not shown). Thus, our limit of 30% intensity for presence of TSGs on those chromosomes, but none of these genes LOH is very conservative. have been cloned to date. Many reports have indicated that the p53 Statistical Analysis. The normality of the percentage of LOH distribution gene is somehow involved in the progression of prostate tumor toward was tested with the Kolmogorov-Smirnov continuous cumulative distribution a metastatic, more malignant phenotype (32â€”34).A metastasis sup test (30). pressor gene has been localized in the short arm of chromosome 11 by Results microcell fusion transfer of human chromosome 11 into a highly metastatic rat prostate cell line (35). Fourteen CA microsatellite repeats (28) were amplified by PCR to Since deletions in the long arm of human chromosome 7 are also a screen 16 prostate carcinomas samples for LOH on 7q21-qter (Table frequent trait of prostate carcinomas, we studied 16 primary prostate

52 W 40 40 42 850 10732266 11 743 213 636 870 3303 L NTNTNTNTNTNTNTNTNT @!@@-,..@*.. @@@@ 506/516 bp â€” .-.- ._ v..@ A 220 bp â€” 200 bp Fig. 1. Representative PCR amplifications of DNA of prostate carcinoma. The case numbers are shown at the top of the lanes; T and N, matched DNA samples isolated from tumor tissue and pe ripheral leukocytes, respectively. A, D7S527; B, E @@@ 220 bp @â€˜@-:. . â€”.â€” .. , @I a .._ B D7S522; C, D7S487. 200 bp â€” 154 bp â€”

344 bp 298 bp â€” Et@ C @@@@@@ â€”,, -. @. ,-@- : ,u@, @-- 154 bp â€” 142 bp 6371

LOH IN PROSTATE

@ CA â€” l.J .,@ .1- !Ji ..J @( inprostateresults not only confirmed and expanded previous reports 1_i Repeats cancer, they also supported our other LOH studies of other a21.3 . .CARCINOMASOur types of tumors (17). Because deletions of chromosome 7q are observed cytogenetic event in many different types of TSG22 â€”D7S527commonly tumors (10â€”16)and because we had functional evidence of a 7q (18), we studied 7q LOH in breast cancer (17). That tumor type had a very similar pattern of allele loss, with a normal @ 1D7s518 IITI distribution around a peak at CA microsatellite repeat D7S522. The extent of allele loss was 81.8% in primary breast carcinomas. The differences in the percentages of LOH found may be due to different patterns of inactivation of the TSG in different tumor @ I__I D7S496 types, or to differences in the amount of stromal contamination in @ 31 1 ::@:::D7S523D7S486 the tumor tissues used. . D7S633 The high frequency of LOH occurrence at 7q31.1â€”31.2in prostate @ 31 2 r:i D7S677 carcinomas shown in this report, a similar pattern of LOH in other @ D7S522 .@ .I types of neoplasias (17), and our repression of tumorigenicity by . D7S655 microcell-mediated transfer of chromosome 7 to a murine SCC cell D7S480 line (18) indicate that human chromosome 7 harbors a TSG distal to D7S490 c-met in band 7q31.1â€”31.2.This TSG seems to be relevant to several types of human neoplasias, as can be inferred by our previous reports D7S487.on cytogenetic-,evidenceof LOH in human primary breast cancer (17) and by (10â€”16).Extensiveof chromosome 7 deletions in other neoplasias ispresently work on the identity and function of this putative TSG being conducted in our laboratory.W4

32 5 2 4483002731726001413 2 0 2 8 7 6 3â€”D7S4983 5 3 6 336:Iâ€”D7S550Fig. 0 3 213 34

22 35

31.1

36 31.2

313 microsatelliterepeats2. Representation of 7q21.3-qter and approximate position of the theinformative(29). Histogram shows the percentage LOH for each of these microsatellites in studied.carcinomasprostate samples 32 indicatethat to determine the extent of LOH on 7q. Our data 7qloci.12 (75%) of 16 tumors studied presented LOH at one or more 33 This incidence is higher than that found at other frequently â€” D7S498 deleted regions in these tumor types (36, 37). Five (83.3%) of six informative cases had LOH in CA microsatel 34 34 1 I lite repeat D7S522 (7q31.1â€”31.2). Previous allelotypic studies of prostate cancer (23â€”25)did investigate the long arm of chromosome 35 [@ 7 for LOH and revealed some LOH, but the percentages ranged from 8 to 10%. However,theseresultsdonotconflictwith ourfindings 36 because the markers used (one in each study) are several cM away from our SCDR. Moreover, the percentages of LOH we obtained with 36 U@D7S55O the markers closest to the ones used by Kunimi et a!. (23) and Carter Fig. 3. Deletions suggested by the allele losses. Shaded columns, region retained; et al. (24) are very similar (12.5% for D7S527 and 18.2% for D7S518) open columns, loss of the region. Length of any column is arbitrary since we possess no tooverimportance the percentages obtained by those two groups. In fact, theâ€˜@â€˜@exactinformationofthe preciselimitsof@ja determinate deletion. Closediscircles of the 7q LOH is reinforced by the finding by Carter et a!. columns, retention of both alleles; open circles open Non informative cases for each marker are symbolized the absence of circles at that (24), that in one of the cases studied the only alleic aberration found position. SCDR obtained by this analysis is indicated at the right.columns,CaseLOH.numbers are was the loss of the 7q marker.shaded displayed at the top of each column.byover 6372

Acknowledgments 18. Zenklusen, J. C., Barrett, J. C., Oshimura, M., and Conti, C. J. Inhibition of tumor igenicity of a murine squamous cell carcinoma (5CC) cell line by a putative tumor The authors wish to thank Dr. Maureen Goode for her skillful revision of suppressor gene on human chromosome 7. Oncogene, 9: 2817â€”2825,1994. 19. Ogata, T., Ayusawa. D., Namba, M., Takahashi, E., Oshimura, M., and Oishi, M. this manuscript. Chromosome 7 suppresses indefinite division of nontumorigenic immortalized human fibroblast cell lines KMST-6 and SUSM-1. Mol. Cell. Biol., 13: 6036â€”6043,1993. References 20. James, C. D., Caribom, E., Nordenskjold, M., Collins, V. P., and Cavenee, W. K. Mitotic recombination of chromosome 17 in astrocytomas. Proc. NatI. Acad. Sci. 1. Knudson, A. G. Antioncogenes and human cancer. Proc. Nail. Acad. Sci. USA, 90: USA, 86: 2858â€”2862,1989. 10914â€”10921, 1993. 21. Mon, N., Yokota, J., Oshimura, M., et al. Concordant deletions of chromosome 3p 2. Weinberg, R. A. Tumor suppressor genes. Science (Washington DC), 254: 1138â€” and loss of heterozygosity for chromosomes 13 and 17 in small cell lung carcinoma. 1146, 1991. Cancer Res., 49: 5130â€”5135,1989. 3. Fearon, E. R., and Vogelstein, B. A genetic model for colorectal tumorigenesis. Cell, 22. Vogelstein, B., Fearon, E. R., Kem, S. E., et al. Allelotype of colorectal carcinomas. 61: 759â€”767,1990. Science (Washington DC), 244: 207â€”211,1989. 4. Knudson, A. G. Mutation and cancer: statistical study of retinoblastoma. Proc. Natl. 23. Kunimi, K., Bergerheim, U. S., Larsson, 1. L, Ekman, P., and Collins, V. P. Acad. Sci. USA, 68: 820â€”823,1971. Allelotyping of human prostatic adenocarcinoma. Genomics, 11: 530â€”536,1991. 5. Sandberg, A. A. Chromosomal abnormalities and related events in prostate cancer. 24. Carter, B. S., Ewing, C. M., Ward, W. S., et a!. Allelic loss ofchromosomes 16q and Hum. Pathol., 23: 368â€”380,1992. lOq in human prostate cancer. Proc. Nail. Acad. Sci. USA, 87: 8751â€”8755, 6. Limon, J., Lundgren, R., Elfving, P., et a!. Double minutes in two primary 1990. adenocarcinomasoftheprostate.CancerGenet.Cytogenet.,39:191â€”194,1989. 25. Isaacs,W. B.,andCarter,B. S. Geneticchangesassociatedwithprostatecancerin 7. Brothman, A. R., Peehi, D. M., Patel, A. M. and McNeal, J. E. Frequency and pauern humans. Cancer Sun'., 11: 15â€”24,1991. of karyotypic abnormalities in human prostate cancer. Cancer Res., 50: 3795â€”3803, 26. Kagan, J., Pisters, L. L., Troncoso, P., Joe, Y., Parat, J. M., Babaian, R. J., and von 1990. Eschenbach, A. C. Genetic instability in microsatellite sequences in prostate cancer. 8. Micale, M. A., Mohamed, A., Sakr, W., Powell, I. J. and Wolman, S. R. Cytogenetics Int. J. Oncol., in press, 1994. of primary prostatic adenocarcinoma. Clonality and chromosome instability. Cancer 27. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., Genet. Cytogenet., 61: 165â€”173,1992. and Struhl, K. Current Protocols in Molecular Biology, Vol. 1, Chap. 2. New York: 9. Knudson, A. G. The Ninth Gordon Hamilton-Fairley Memorial Lecture. Hereditary John Wiley and Sons, 1992. cancers: clues to mechanisms of carcinogenesis. Br. J. Cancer, 59: 661â€”666, 28. Weissenbach, J., Gyapay, G., Dib, C., et aL A second-generation linkage map of the 1989. human genome. Nature (Lond.), 359: 794â€”801,1992. 10. Osella, P., Carlson, A., Wyandt, H. and Milunsky, A. Cytogenetic studies of eight 29. Cropp, C. S., Lidereau, R., Campbell, G., Champene, M. H., and Callahan, R. Loss squamous cell carcinomas of the head and neck. Deletion of 7q, a possible primary of heterozygosity on chromosomes 17 and 18 in breast carcinoma: two additional chromosomal event. Cancer Genet. Cytogenet., 59: 73â€”78,1992. regions identified. Proc. NatI. Acad. Sci. USA, 87: 7737â€”7741, 1990. 11. Atkin, N. B., and Baker, M. C. Numerical chromosome changes in 165 malignant 30. Zar, J. H. Biostatistical Analysis, Chap. 6. Englewood Cliffs, NJ: Prentice Hall, 1991. tumors. Evidence for a nonrandom distribution of normal chromosomes. Cancer 31. Parkin, D. M., Pisani, P., and Ferlay, J. Estimates of the worldwide incidence of Genet. Cytogenet., 52: 113â€”121,1991. eighteen major cancers in 1985. mt. J. Cancer, 54: 594â€”606,1993. 12. Kibbelaar, R. E., Mulder, J. W., Dreef, E. J., et aL Detection of monosomy 7 and 32. Navone, N. M., Troncoso, P., Pisters, L. L, et aL p53 protein accumulation and gene trisomy 8 in myeloid neoplasia: a comparison of banding and fluorescence in situ mutation in the progression of human prostate carcinoma. J. NatI. Cancer Inst., 85: hybridization. Blood, 82: 184â€”193,1993. 1657â€”1669,1993. 13. Atkin, N. B., and Baker, M. C. Chromosome 7q deletions: observations on 13 33. Bookstein, R., MacGrogan, D., Hilsenbeck, S. G., Sharkey, F., and Allred, D. C. p53 malignant tumors. Cancer Genet. Cytogenet., 67: 123â€”125,1993. is mutated in a subset of advanced-stageprostatecancers. Cancer Res., 53: 14. Kere, J., Ruutu, T., Davies, K. A., et a!. Chromosome 7 long arm deletion in myeloid 3369â€”3373,1993. disorders: a narrow breakpoint region in 7q22 defmed by molecular mapping. Blood, 34. Bookstein, R., and Allred, D. C. Recessive oncogenes. Cancer (Phila.), 71: 73: 230â€”234,1989. 1179â€”1186,1993. 15. Lundgren, R., Kristoffersson, U., Heim, S., Mandahl, N., and Mitelman, F. Multiple 35. Ichikawa, T., Ichikawa, Y., Dong, J., et a!. Localization of metastasis suppressor structural chromosome rearrangements, including del(7q) and del(lOq), in an adeno gene(s) for prostatic cancer to the short arm of human chromosome 11. Cancer Res. carcinoma of the prostate. Cancer Genet. Cytogenet., 35: 103â€”108,1988. 52: 3486â€”3490,1992. 16. Lukeis, R., Ball, D., Irving, L., Garson, 0. M., and Hasthorpe, S. Chromosome 36. Bova, G. S., Carter, B. S., Bussemakers, M. J., et al. Homozygous deletion and abnormalities in non-small cell lung cancer pleural effusions: cytogenetic mdi frequent allelic loss of chromosome 8p22 loci in human prostate cancer. Cancer Res., cators of disease subgroups. Genes Chromosomomes & Cancer, 8: 262â€”269, 53: 3869â€”3873, 1993. 1993. 37. Cleton-Jansen, A. M., Moerland, E. W., Kuipers-Dijkshoorn, N. J., Callen, D. F., 17. Zenklusen, J. C., BiÃ¨che,I.,Lidereau, R., and Conti, C. J. CA microsatellite repeat Sutherland, G. R., Hansen, B., Devilee, P., and Cornelisse, C. J. At least two different D7S522 is the most commonly deleted region in human primary breast cancer. Proc. regions are involved in ailetic imbalance on chromosome arm 16q in breast cancer. NatI. Acad. Sci. USA, 91: 12155â€”12158,1994. Genes Chromosomomes & Cancer, 9: 101â€”107,1994.

6373

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1994 American Association for Cancer Research. Loss of Heterozygosity in Human Primary Prostate Carcinomas: A Possible Tumor Suppressor Gene at 7q31.1

Jean C. Zenklusen, Janet C. Thompson, Patricia Troncoso, et al.

Cancer Res 1994;54:6370-6373.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/54/24/6370

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/54/24/6370. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.