Reduced Angiogenesis and Tumor Progression in Gelatinase A-Deficient Mice1

(CANCER RESEARCH.Â«. 11)48-1(151. March 1. 1998] Reduced Angiogenesis and Tumor Progression in Gelatinase A-deficient Mice1

Takeshi hÃ¶h,2Masatoshi Tanioka, Hiroshi Yoshida, Takayuki Yoshioka, Hirofumi Nishimoto, and Shigeyoshi Itohara

Shionogi Institute for Medical Science. Shionogi & Co., Lid. IT. /.. M. T., H. N.Â¡and Discovery Research Laboratories II, Shionogi & Co., Ltd. [H. Y., T. Y.J, Fukushima-ku, Osaka 553. and Institute for Virus Research. Kyoto University. Syogo-in. Sakyo-ku, Kyoto 606-01 Â¡S././. Japan

ABSTRACT normalities and are fertile, thus offering a useful system for assessing the specific role of gelatinase A in tumor progression. We show here Matrix proteolysis is thought to play a crucial role in several stages of that the rates of angiogenesis and experimental tumor growth and tumor progression, including angiogenesis, and the invasion and metas metastasis are markedly reduced in these gelatinase A-deficient mice. tasis of tumor cells. We investigated the specific role of gelatinase A This is the first direct evidence that host-derived gelatinase A plays a (matrix metalloproteinase 2) on these events using gelatinase A-deficient mice. In these mice, tumor-induced angiogenesis was suppressed accord specific role in angiogenesis and tumor progression in vivo. ing to dorsal air sac assay. When B16-BL6 melanoma cells or Lewis lung carcinoma cells were implanted intradermally, the tumor volumes at 3 MATERIALS AND METHODS weeks after implantation in the gelatinase A-deficient mice decreased by 39% for B16-BL6 melanoma and by 24% for Lewis lung carcinoma Animals. The generation of gelatinase A-deficient mice was described (P < 0.03 for each tumor). The number of lung colonies of i.v. injections previously (11). To minimize genetic variances between experimental groups, fell by 54% for B16-BL6 melanoma and 77% for Lewis lung carcinoma we backcrossed the heterozygous mutant mice five times to C57BL/6J mice before using them. The homozygous mutant and control wild-type mice were (P < 0.014 and P < 0.0015, respectively). These results indicated that host-derived gelatinase A plays an important role in angiogenesis and littermates from the crosses between hÃ©tÃ©rozygotes.Thegenotypes of the mice tumor progression, suggesting the usefulness of gelatinase A inhibitors for were determined by Southern blot analysis (11). Tumor Cells. B16-BL6 melanoma cells (12), provided kindly by Dr. I. J. anticancer chemotherapy. Fidler (M. D. Anderson Cancer Center, Houston, TX), were maintained as monolayer cultures in DMEM containing 10% fetal bovine serum. Lewis lung INTRODUCTION carcinoma (supplied from the National Cancer Institute, NIH, Bethesda, MD) was maintained by serial intradermal implantation in C57BL/6J mice. MMPs1 constitute a family of zinc-requiring matrix-degrading pro- Dorsal Air Sac Method and Image Analysis. The dorsal air sac method teinases, which include the collagenases, gelatinases, stromelysins, was carried out as described elsewhere (13). B16-BL6 melanoma cells were and MT-MMPs. These enzymes are synthesized as proenzymes and washed three times with HBSS and were suspended in HBSS at a concentra must undergo proteolytic cleavage of a NH2-terminal domain to tion of 1 X IO7 cells/ml. A Millipore chamber (diameter, 10 mm; thickness, 2 become catalytically active (1). The catalytic activities are further mm; filter pore size, 0.22 j*m; Millipore Co.) was filled with 0.2 ml of either modulated by interacting with many endogenous inhibitors (2, 3). cell suspension or HBSS and implanted s.c. into the dorsal side of the mice. At 5 days after the implantation, the mice were anesthetized and fixed in the prone Generally, a given cell expresses multiple MMPs. Although these position. A wide, rectangular incision was made in the skin on the dorsal side, enzymes are believed to have crucial roles in various physiological and the skin was carefully ablated. To locate the chamber-contacting region, a and pathological conditions, such as embryonic development and ring (Millipore) of the same shape as the chamber was placed onto the s.c. tissue repair in adults, their complexities as mentioned above have tissues adjacent to the chamber region, and the area was photographed. For made it difficult to definitively evaluate the roles of individual en histolÃ³gica! analyses, two samples were randomly taken from each skin area, zymes in vivo. However, recent progress on gene-knockout mice embedded in paraffin wax, and sliced into 3 /Â¿m-thickvertical sections, which offers a way to overcome this problem. were then stained with H&E. From each H&E-stained section, microphoto- Gelatinase A (MMP-2) and gelatinase B (MMP-9), which hydro- graphs of three randomly selected fields were taken at a final magnification of lyze type IV collagen localized in the basement membrane, have been X116. suggested to play a role in tumor progression and metastasis (4, 5). Vascular structures were recognized as luminal or slitlike structures that occasionally contained blood cells within them and were delineated by flat The idea first arose from the finding that many malignant tumor cells tened endothelial cells. Vessels beneath the museali cutÃ¡neas were traced secrete large amounts of type IV collagen-degrading enzymes (4). manually without awareness of the genotype. The traced vascular images were Furthermore, studies have suggested that either gelatinase A or gela entered into a computer, and image analyses for the area and number of vessels tinase B or both expressed by endothelial cells play a crucial role in beneath the musculi cutaneus were performed using image analysis software angiogenesis, which is considered to be an important stage in tumor (NIH Image, Version 1.61 PPC). The vascular area and number of vessels were progression (6-8). Recent studies also suggested that host-derived expressed as ;inr/mm width and number/mm width, respectively. A total of gelatinase A binds to the cell-surface proteins MT-MMP (9) and six fields per animal (two samples x three fields) were analyzed, and the mean integrin avÃŸ,(10) of tumor or endothelial cells and may modulate the value for the six analyses was used. cell entity. However, there is no clear evidence to support the hypo Intradermal Tumor Growth Assay. B16-BL6 melanoma cells were thetical roles of gelatinases in tumor progression. washed three times with HBSS and were suspended in HBSS. Lewis lung Recently, we produced gelatinase A-deficient mice by gene target carcinoma cells that had been maintained in C57BL/6J mice were treated with 1% collagenase for 10 min at room temperature, washed with HBSS three ing (11). These mice develop normally without any anatomical ab- times, and suspended in HBSS. Mice were inoculated intradermally with B16-BL6 melanoma cells or Lewis lung carcinoma cells (2 X IO5 cells/mouse Received 9/9/97; accepted 12/31/97. for either of the tumor cells). The primary tumor volume was measured by The costs of publication of this article were defrayed in part by the payment of page using a slide caliper, applying the following formula: Volume = 0.5 X charges. This article must therefore be hereby marked advertisement in accordance with (Width)2 X Length. 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was partly supported by a Grant-in-Aid for Scientific Research B from the Lung Colonization Assay. Mice were injected with 1 x IO5 cells/mouse Ministry of Education. Science. Sports and Culture of Japan (to S. 1.). (B16-BL6 melanoma cells) or 3 X 10* cells/mouse (Lewis lung carcinoma - To whom requests for reprints should be addressed, at Shionogi Research Laboratory. cells) in the lateral tail vein. Fourteen days (B16-BL6 melanoma) or 16 days Shionogi & Co., Ltd., 12-4, Sagisu 5-chome. Fukushima-ku, Osaka 553, Japan. Phone: (81) 6-458-5861; Fax: (81) 6-458-0987; E-mail: [email protected]. (Lewis lung carcinoma) later, the mice were killed, and the lungs were ' The abbreviations used are: MMP. matrix metalloproteinase; MT-MMP. membrane- removed and fixed with Bouin's fluid to count the numbers of visible tumor type MMP. colonies. 1048

HBSS Day 3 Day 5

Fig. 1. Reduced angiogenesis in gelatinase A-deficient mice. a-f. representative views from the inner side of dorsal skins in wild-type (a-c) and mutant (d-f) mice, a and d, dorsal skin of the mice 5 days after implantation of a chamber filled with HBSS for negative control. b, c, e, and/ dorsal skin of mice 3 (b and e) or 5 (c andÂ» days after implantation of a B16-BL6 melanoma cell-filled chamber. Bar. 1 mm.

Gelatin Zymography. Gelatin zymography was carried out as described experiments in which the chamber was filled with HBSS, vessel elsewhere (11). Briefly, the tumor tissues were homogenized in 50 mM Tris- development in mutant mice appeared to proceed normally, suggest HC1 (pH 7.5) and centrifuged at 15,000 rprn for 5 min. The protein concen ing that gelatinase A had little influence or that another enzyme tration of the supernatant was determined by a protein assay (Bio-Rad), and 20 Â¡igof supernatant proteins were applied to nonreduced SDS-PAGE using a compensated for the role of gelatinase A in normal vessel develop ment (Fig. 1, a and d). By 3 days after implantation with the tumor- 7.5% gel containing 0.1% gelatin. After electrophoresis, the gel was soaked in 50 mM Tris-HCl (pH 7.5) containing 2.5% Triton X-100 at room temperature filled chamber, the blood vessels had expanded, and newly formed with gentle shaking for 2 h and then incubated overnight in 50 mM Tris-HCl blood vessels had branched out from large blood vessels in wild-type (pH 7.5) containing 10 mM CaCU at 37Â°C.The gels were then stained with mice (Fig. \b). On the 5th day, severe hemorrhage from these vessels Coomassie Brilliant Blue. was observed (Fig. le). In mutant mice, the blood vessels were expanded, but only moderate neovascularization and hemorrhage RESULTS AND DISCUSSION were observed (Fig. 1, compare b with e and c with/). We measured Angiogenesis in Gelatinase A-deficient Mice. To examine angio angiogenesis by analyzing vertical sections of the skin under light genesis in gelatinase A-deficient mice (referred to here as "mutant microscopy. Fig. 2 shows representative views of H&E staining of mice"), we carried out a dorsal air sac assay using transplantation of skin sections at 5 days after implantation. The total vascular area and a Millipore chamber (13) filled with mouse melanoma B16-BL6 (12) number of vessels beneath the museali cutÃ¡neas per 1-mm width of cells as an angiogenesis inducer into the dorsal side of the mice. skin section were estimated using an image analyzer (Table 1). In Research has shown that B16-BL6 cells do not express gelatinase A mutant mice, the total vascular area and number of vessels per unit in vitro (14). Representative images of the area adjacent to the width were approximately 56 and 69% of those in wild-type mice, s.c.-implanted chamber are shown in Fig. 1, a-f. In the control respectively (Table 1; P < 0.013 and P < 0.03, respectively: Stu-

3$r

" -=- . â€¢¿ -Â»..--.," Ã¯â€¢:'.V > >

Fig. 2. Histological analysis of skin vertical sections. H&E stain ing of skin vertical sections 5 days after implantation of the HBSS- filled (a) and B16-BL6-filled (b) chambers in wild-type (a, b, and d) .â€¢â€¢ and mutant (c and e) mice, a-e, bottom, chamber-contacting side, d . and e, high magnifications of the same regions of b and c, respec tively, d and e, arrowheads, blood vessels beneath the museali cutaneus. Bars, 20 Â¡im(a-c) and 100 /urn (d and e).

- v Â» C :Â£\j 4*' Â£jâ€¢¿'*~ii*r" *"*" â€¢¿."Â«â€¢.

1049

Table I Mean vascular urea am! number of vessels tinase A in these tumors in in vivo studies correspond to those from in Millipore chambers were filled with 2 X IO6 BI6-BL6 cells and implanted s.c. into the vitro observations (14, 15). These results indicated that tumor-derived dorsal side of wild-type and mutant mice. After 5 days, the vascular area and number of vessels beneath the museali cutÃ¡neas were counted and expressed as fim"/mm width and gelatinase A appeared to have little or no effect in replacement of the number/mm width, respectively. Each value represents the mean Â±SD. host-derived enzyme.

Mean area (finr/ Mean number/mm The data presented here have important implications for the role of gelatinase A in the progression of malignant tumors. The suppressed Wild-type (n = 6) 17.750 Â±6170 58 Â±17 Mutant (n = 7) 10.020 Â±2990" 40 Â±9* tumor-induced angiogenesis, tumor growth, and lung colonization in " Value is significantly different from that of wild-type mice (P < 0.0131, as deter mutant mice indicated that gelatinase A plays an important role in mined by Student's / test. these events. It should be emphasized that host-derived but not tumor- h Value is significantly different from that of wild-type mice (P < 0.03), as determined by Student's ( test. derived gelatinase A contributed significantly to growth and lung colonization of experimental tumors. Because recent studies indicated that tumor progression could be restricted solely by antiangiogenic dent's / test). These results indicated that gelatinase A plays an therapy (16-19), the suppression of tumor progression in the mutant important role in tumor-induced angiogenesis. mice may partly be attributed to the weak angiogenesis. It is also In the skin sections of wild-type mice in which the HBSS-filled possible that host-derived gelatinase A binds membrane-bound mol chamber was implanted, slight proliferation of fibroblasts and infil ecules, such as MT-MMP (9) or integrin avÃŸ,(10), at the surface of tration of inflammatory cells, principally macrophages, were observed tumor cells, and thus modulates the tumor cell entity to increase its in the subcutis immediately beneath the museali cutÃ¡neas (Fig. 2a). invasive nature. Perhaps the weak angiogenesis in the host and the Neither granulation tissue formation nor hemorrhage beneath the museali cutÃ¡neas was seen. In the skin sections of wild-type and mutant mice in which the B16-BL6-filled chamber was implanted, histological hallmarks of inflammation consisted of inflammatory cell infiltration and granulation tissue formation containing newly formed small blood vessels and also proliferation of fibroblasts in the subcutis immediately beneath the museali cutÃ¡neas (Fig. 2, b-e). Principal infiltrating cell types for wild-type and mutant mice were mainly macrophages and polymorphonuclear leukocytes with a small number of lymphocytes and were qualitatively identical. Dilation of preexist ing small vessels in the museali cutÃ¡neas and hemorrhage beneath the museali cutÃ¡neas were also noted. In general, the extent of the Time after implantations (weeks) granulation tissue formation and hemorrhage was greater in wild-type mice than in mutant mice. d Tumor Progression in Gelatinase A-deficient Mice. To examine cT*Â£0E>o3 tumor growth in mutant mice. B16-BL6 cells or Lewis lung carcinoma nEE.93O cells were implanted intradermally. Fig. 3, a and b, shows the growth â€ž¿â€ž': rates of B16-BL6 melanoma and Lewis lung carcinoma, respectively. 8o_a The growth rates of both tumors in mutant mice are significantly slower than those in wild-type mice (P < 0.017 for B16-BL6 mela >og3r-5000-4000-3000-20001000-n D.Oo9 o_y noma and P < 0.01 for Lewis lung carcinoma by two-way repeated- e0 measures ANOVA). At 3 weeks postinjection, the average volumes of 1-6000-5000-4000-3000-2000-Ã‹B- the tumors in mutant mice were reduced by 39% for B16-BL6 -DD_D melanoma and by 24% for Lewis lung carcinoma (P < 0.03 for either tumor by Student's t test; Fig. 3, c and d). We also injected these au tumor cells i.v. to examine the efficacy of lung colonization in mutant mice. After about 2 weeks, the macroscopic tumor colonies formed in 40-30-20-10-n-DB the lung were counted. The number of colonies was significantly D0D|B o fewer in mutant mice (Fig. 3, e and/). The average numbers of tumor o colonies in mutant mice were 54% fewer for B16-BL6 melanoma and u 77% fewer for Lewis lung carcinoma (P < 0.014 and P < 0.0015, respectively, by Mann-Whitney U test). wD 9o Gelatinase Secretion of Implanted Tumor Cells. To investigate whether the two tumor cell lines themselves secrete gelatinase A in vivo, we carried out gelatin zymography using soluble fractions from Fig. 3. Reduced progression of transplanted tumor cells in gelatinase A-deficient mice. tumors grown from implanted seeds in wild-type and mutant mice. In a-f, growth of B16-BL6 melanoma (a, c. and f) and Lewis lung carcinoma (b. (I, and/) B16-BL6 melanoma extracts, gelatinase A activity was not detected in in wild-type (D) and mutant (O) mice, a and b, average growth rate of BI6-BL6 the tumor from mutant mice (Fig. 4, Lane 2). Thus, the gelatinase A melanoma (wild-type, n = 9; mutant, n = 9; Â«)and Lewis lung carcinoma (wild-type. n = 9; mutant, /; = 8; h) after intradermal implantations. *, P < 0.017; **, P < 0.01 activity seen in the extract from the tumor implanted in wild-type (two-way repeated-measures ANOVA). c and tl. tumor volumes after 3 weeks of intra mice was likely to have been derived from the host cells that had dermal implantations. The mean B16-BL6 melanoma volumes were 3220 mm1 (wild type) and 1950 mm1 (mutant). The mean Lewis lung carcinoma volumes were 5080 mm3 (wild infiltrated into the tumor tissues (Fig. 4, Lane I). In the case of Lewis type) and 3860 mm' (mutant). *, P < 0.03; *Â»,P < 0.03 (Student's l test), e and/ number lung carcinoma extracts, gelatinase A activity was detected in extracts of tumor colonies at the lung surface after 14 days (BI6-BL6 melanoma; c) or 16 days both from wild-type and mutant mice (Fig. 4, Lanes 3 and 4), (Lewis lung carcinoma;/) of i.v. injections. The mean numbers of B16-BL6 melanoma colonies were 21.2 (wild type) and 9.7 (mutant). The mean numbers of Lewis lung suggesting that Lewis lung carcinoma cells themselves secrete detect carcinoma colonies were 53 (wild type) and 12 (mutant). *. P < 0.014; **, P < 0.0015 able levels of gelatinase A in vivo. The expression patterns of gela (Mann-Whitney U test), c-f, bars, means. 1050

2. Docherty. A. J. P.. Lyons. A.. Smith. B. J.. Wright, E. M.. Stephens. P. E.. and Harris. 1234 T. J. R. Sequence of human tissue inhibitor of metalloproleinases and its identity to erythroid-potenliating activity. Nature (Land.), 318: 66-69, 1985. 3. Stetler-Stevenson. W. G., Krutzsch, H. C.. and Liotta. L. A. Tissue inhibitor of metalloproteinase (TIMP-2). J. Biol. Chem.. 264: 17374-17378. 1989. 4. Liotta. L. A.. Tryggvason. K.. Garbisa. S.. Hart. L, Foltz, C. M., and Shafie. S. Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature (Lond.). 2X4: 67-68. 1980. Gel B 5. Liotta. L. A. Tumor invasion and mÃ©tastases:role of the extracellular matrix: Rhoads Memorial Award Lecture. Cancer Res., 46: 1-7, 1986. 6. Moses. M. A.. Sudhalter. J.. and Langer. R. Identification of an inhibitor of neovascularization from cartilage. Science (Washington DC). 248: 1408-1410. 1990. Gel A 7. Johnson. M. D.. Kim, H-R. C.. Chesler, L.. Tsao-Wu. G., Bouck. N.. and Polverini. P. J. Inhibition of angiogenesis by tissue inhibitor of metalloproteinase. J. Cell. Physiol. 160: 194-202, 1994. 8. Unemori, E. N.. Bouhana. K. S., and Werb, Z. Vectorial secretion of extracellular matrix proteins, matrix-degrading proteinases. and tissue inhibitor of meialloprotein- Fig. 4. Gelatinase secretion of implanted tumor cells. Gelatin /ymography of B16-BL6 ases by endothelial cells. J. Biol. Chem.. 265: 445-451, 1990. melanoma (Lanes I and 2) and Lewis lung carcinoma (Lanes 3 and 4) extracts. Lanes I 9. Sato, H.. Takino. T.. Okada. Y.. Cao. J.. Shinagawa. A.. Yamamoto. E.. and Seiki. M. and 3. extracts of the tumor grown in a wild-type mouse. Lanes 2 and 4, extracts of the A matrix metalloproteinase expressed on the surface of invasive tumour cells. Nature tumor grown in a mutant mouse. Arrows, gelatinolytic bands of gelatinase A (Gel A) and (Lond.). 370: 61-65. 1994. B (Gel B). 10. Brooks. P. C., Stromblad. S.. Sanders. L. C., von Schalscha. T. L., Aimes. R. T.. Stetler-Stevenson. W. G.. Quigley, J. P.. and Cheresh. D. A. Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrili poor invasive nature of the tumor cells synergistically prevented the av/33. Cell. Â«5.683-693. 1996. 11. Itoh. T.. Ikeda. T.. Gomi. H.. Nakao. S. Suzuki. T.. and Itohara. S. Unaltered secretion growth and lung colonization of implanted tumor cells in the mutant of /3-amyloid precursor protein in gelatinase A (matrix metulloproteinase 2)-deficient mice. mice. J. Biol. Chem.. 272: 22389-22392. 1997. The suppression of angiogenesis and the growth and lung coloni 12. Hart, I. R. The selection and characterization of an invasive variant of the B16 zation of implanted tumor cells were partial in the mutant mice. melanoma. Am. J. Pathol., 97: 587-600. 1979. 13. lv..ih.in i M.. Nakayama. Y.. Tanaka. N. G.. Goryo. M.. and Okada. K. Quantification Activated endothelial cells secrete many MMPs. including gelatinase of tumour-induced angiogenesis by image analysis. Int. J. Exp. Palhol., 77: 109-114, B (MMP-9; Refs. 8 and 20). Some of these enzymes might partly 1996. compensate for the role of gelatinase A. A series with knockout mice 14. Kalo. Y.. Nakayama. Y., Umeda. M., and Miyazaki. K. Induction of 103-kDa gelatinase/type IV collagenase by acidic culture conditions in mouse metasialic of these genes will help to elucidate the issue. melanoma cell lines. J. Biol. Chem., 267: 11424-11430. 1992. In conclusion, this study suggested that tumor-induced angiogene 15. Brodt. P.. Reich. R.. Moroz. L. A., and Chambers. A. F. Differences in the sis and tumor progression can be controlled by inhibiting gelatinase A repertoires of basement membrane degrading enzymes in two carcinoma sublines with distinct patterns of site-selective metastasis. Biochim. Biophys. Acta. /139: activity, thus encouraging the use of gelatinase A inhibitors for 77-83, 1992. chemotherapy of cancer and other diseases. The data suggested that 16. Kim, K. J., Li, B., Winer, J.. Aramanini. M.. Gillett. N.. Phillips. H. S.. and Ferrara. systematic use of such inhibitors might be effective, because host- N. Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses derived gelatinase A activities, which could be supplied by either local tumor growth in viro. Nature (Lond.). 362: 841-844. 1993. 17. Millauer, B.. Shawver. L. K.. Plate. K. H.. Risau. W., and Ullrich. A. Glioblastoma tissues or circulation, apparently affect the progression of experimen growth inhibited in virÂ«by a dominant-negative FLK-I mutant. Nature (Lond.). 367: tal tumors. In this context, it is noteworthy that the gelatinase A-de- 576-579, 1994. 18. Parangi, S.. O'Reilly. M.. Christofori. G.. Holmgren. L.. Grosfeld. J. Folkman. J.. and ficient mice developed normally without any anatomical abnormali Hanahan, D. Antiangiogenic therapy of transgenic mice impairs de nim> tumor ties (11), suggesting few side effects of the specific inhibitors. growth. Proc. Nati. Acad. Sci. USA, 93: 2002-2007, 1996. 19. Tanaka. T.. ManÃ³me. Y.. Wen. P.. Kufe. D. W., and Fine, H. A. Viral vector-mediated transduction of a modified platelet factor 4 cDNA inhibits angiogenesis and tumor REFERENCES growth. Nat. Med.. 3: 437-442. 1997. 1. Stetler-Stevenson, W. G., Krutzsch. H. C.. Wacher. M. P.. Margulies. I. M. K.. and 20. Fisher, C.. Gilbertson-Beadline. S.. Powers. E. A.. Petzold, G. Poorman. R.. and Liotta. L. A. The activation of human type IV collagenase proenzyme. J. Biol. Chem., Mitchell. M. A. Interstitial collagenase is required for angiogenesis in vitro. Dev. 264: 1353-1356. 1989. Biol.. /62: 499-510. 1994.

1051

Takeshi Itoh, Masatoshi Tanioka, Hiroshi Yoshida, et al.

Cancer Res 1998;58:1048-1051.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/58/5/1048

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/58/5/1048. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.