Occurrence of the Tumor-Specific, Calcium-Binding Protein, Oncomodulin, in Virally Transformed Normal Rat Kidney Cells1

(CANCER RESEARCH 43. 5390-5394, November 1983]

Occurrence of the Tumor-specific, Calcium-binding Protein, Oncomodulin, in Virally Transformed Normal Rat Kidney Cells1

J. P. Durkin,2 L. M. Brewer, and J. P. MacManus

Cell Physiology Group, Division ot Biological Sciences, National Research Council of Canada, Ottawa, Ontario K1A OR6, Canada

ABSTRACT of cells isolated from the kidney of a normal Osborne Mendel rat. The transformed ASV-NRK line was derived from infection of NRK cells with Oncomodulin, an apparently tumor-specific calcium-binding the B77 strain of ASV. Both cell lines were kindly supplied by Dr. P. K. protein, has been detected in many chemically induced rat hep- Vogt (University of Southern California, School of Medicine, Los Angeles, atomas. It is now possible to detect, by radioimmunoassay and Calif.). Spontaneously transformed lines (Spon 1-NRK and Spon 3-NRK) immunofluorescence, the presence of Oncomodulin in normal rat were isolated from NRK cultures which had been passaged repeatedly at high density. Stock cultures of all cells were maintained at 36Â° in kidney cells virally transformed by avian sarcoma virus. By medium containing 85% (v/v) Dulbecco's modified Eagle's medium contrast, it was not detected in uninfected, nonneoplastic normal (Grand Island Biological Co., Grand Island, N. Y.), 15% (v/v) bovine calf rat kidney cells. The protein was isolated and purified by a novel serum (Colorado Serum Co., Denver, Colo.), and the antibiotic gentami- high-performance liquid chromatography procedure and shown cin. ASV-NRK cells were highly tumorigenic: 107 of them produced large to be identical to that isolated previously from rat hepatoma. The (2 to 3 cm) fibrosarcomas within 5 weeks after s.c. injection into athymic cellular levels of oncomodulin approached the levels of calmo- BALB/c nude mice. Spon 1-NRK and Spon 3-NRK cells also formed dulin in avian sarcoma virus-transformed normal rat kidney cells, tumors in athymic mice, but the parent NRK line had failed to do so by suggesting that the total calcium-binding activity of the cell may 6 months after injection. play a role in expression of the transformed phenotype. Oncomodulin. Rat oncomodulin was purified to homogeneity from Morris hepatoma 5123tc by conventional means (10). The purity of 1- INTRODUCTION mg aliquots was shown to be greater than 99% by reverse-phase HPLC with monitoring at 220 nm. After elution of the major peak, only trace A unique calcium-binding protein found in the extracts of many amounts of 220-nm-absorbing material could be washed from the column chemically induced rat hepatomas, but not normal liver, has been support. Anti-Rat Hepatoma Oncomodulin IgG. The crude immunoglobuhn purified to homogeneity and shown to be an acidic protein with a molecular weight of 11,500, possessing 2 calcium-binding sites fraction obtained from an immunized goat has been described (11 ). Forty per molecule (9-11,13,15). The cellular function of this calcium- ml of IgG (20 mg/ml) in 20 mw KH2P04, pH 8.0, were further purified by passage through a column (1.5 x 30 cm) of DEAE-Affigel Blue equili binding protein remains unknown, but it is capable of modulating brated in 20 HIM KH2P04 buffer. The material passing through the column the activity of calmodulin-dependent phosphodiesterase in vitro was collected and dialyzed against 125 ITIMborate:75 mM NaCI, pH 8.4. (12) and of mimicking the effects of calmodulin by releasing T51B This IgG was further purified by antigen-affinity chromatography on an rat liver cells from a d-S transition block induced by low-calcium oncomodulin:Sepharose column (0.9 x 10 cm), prepared according to medium (1). The development of antisera to this protein purified the method of Klee and Krinks (7) and containing 0.8 mg of oncomodulin from rat hepatoma has enabled immunological methods to es per ml of gel. Antioncomodulin was desorbed by reverse flow with 100 tablish the presence of immunoreactive material in tumors arising mÂ« glycine, pH 2.7. The eluted IgG was neutralized, desalted on a from a variety of chemical carcinogen-treated tissues and animal Sephadex G-25 column, and dialyzed against 100 mw NaCI:10 mw imidazole, pH 7.4. For immunocytology, rabbit anti-rat hepatoma onco species (13, 15). Such material has not been detected in normal modulin was produced and purified to the DEAE-Affigel Blue step as tissue or in nonneoplastic rodent and human cell lines by specific described above. radioimmunoassay (13, 15). Because of its similarity to calmo Radioimmunoassay. The oncomodulin radioimmunoassay using an dulin and its apparent tumor specificity, this calmodulin-binding tigen-purified goat anti-rat hepatoma oncomodulin IgG has been de protein has been named oncomodulin (1, 15). scribed (13, 15). The radioimmunoassay using either goat or rabbit IgG We now report the detection, by immunofluorescence and was not affected by the presence of a 2000-fold excess of other calcium- radioimmunoassay, of high levels of oncomodulin in NRK3 cells binding proteins such as calmodulin, troponin C, or parvalbumin (13,15). transformed by ASV. This protein has been isolated, purified by The preparation of extracts of cells and tumors for oncomodulin assay has been described (13-15). a novel and rapid method using HPLC, and shown to be identical Immunocytofluorescence. NRK and ASV-NRK cells were plated at a to that isolated previously from the chemically induced rat hep density of 3 x 103/sq cm on tissue culture chamber slides (Lab-Tek; atoma, 5123tc. Miles Laboratories, Inc., Naperville, III.) in stock medium and incubated at 36Â°for 48 hr. Logarithmically growing cells were rinsed in PBS, fixed MATERIALS AND METHODS in 3.7% (v/v) formaldehyde: 10% acetic acid in PBS for 15 min at room temperature, rinsed 3 times in distilled water, immersed in acetone at Cell Line. Nonneoplastic NRK fibroblast-like cells were descendants -20Â° for 10 min, and air dried. Apparent oncomodulin-specific immuno fluorescence was detected by the indirect antibody approach. Slides 1National Research Council of Canada No. 21315. supported in part by Grant were preincubated in 0.5% Triton X-100 in PBS for 20 min, flooded with X-ROI-CA-31898 of the National Cancer Institute, NIH. Affigel Blue-purified rabbit anti-rat hepatoma oncomodulin IgG (100 ^g 2 To whom requests for reprints should be addressed. 3 The abbreviations used are: NRK, normal rat kidney; ASV, avian sarcoma IgG per ml), and incubated for 90 min at room temperature in a humidified virus; HPLC, high-performance liquid chromatography; PBS, phosphate-buffered atmosphere. This concentration of antibody was determined by serial saline. dilution to yield maximum specific staining while maintaining an accept Received March 1, 1983; accepted July 29, 1983. able nonspecific background fluorescence. The slides were washed in

5390 CANCER RESEARCH VOL. 43

8 12 16 C.S's.-\ 1 0.3[ Il D^>-,.-*"Ã€ lid/1/1y11\

Cz 0.2iVIjii KX)m'a1 â€ž*540 0 ' 50 250SÂ¡^Â¿. sh\*o ;3030 Â¿r20m \4Jt\ O5 ^Â§20 xIOW 'B77 O-1i50SoC"fi_-K ~s1 f\ikâ„¢!â„¢â„¢P v*i\8 PTIO 5 '/V_"^"-â€¢'â€¢""' |4 320 j l"^^^") 1 1 â€”"M in iJ) ÃŒ0_12 19 |4 '0iU â€¢â€” 4 8s. 1620MINUTES1VI2Â¿-I-(V.t2)500iooÂ§ )MIINU1624MINUTES0.280.1Â¡iiJio(~ 18 20 22 4 8 12 16 202MINUTES,.',-'â€¢'.--'"'y|28 32^â€”_^_136450

Chart 1. Purification of immunoreactive material from ASV-NRK solid tumors by HPLC methods. A, ion-exchange HPLC on AX300 at 1.0 ml/min. The hatched area in A indicates the immunoreactive fractions pooled and subjected to B, size exclusion HPLC on a TSK-125 column at 0.5 ml/min. The hatched area in B indicates immunoreactive fractions pooled and subjected to reverse-phase HPLC. C, calibration of TSK-125 column. Standards used were: 1, bovine serum albumin (M, 68,000); 2, ovalbumin (M, 45,000); 3, chymotrypsinogen A (M, 25,000); 4, myoglobin (M, 17,500); 5, cytochrome c (M, 12,500); 6, aprotinin (M, 6,800); 7, insulin chain B (M, 3,500); and 8, insulin chain A (M, 2,500). x, retention time of rat hepatoma oncomodulin (M, 11,500) running anomalously at 18 min; j, retention time of immunoreactive material from ASV-NRK tumors. D, reverse-phase HPLC on RP-P at 0.8 ml/min. Material from the single peak eluting at 35.5 min was collected, lyophilized, and subjected to amino acid analysis (see Table 2).

PBS for 20 min and incubated 1 hr with fluorescein isothiocyanate- purified by HPLC methods using 2 Beckman 10OA pumps controlled by conjugated sheep anti-rabbit IgG (Cappel Laboratories, West Chester, a Model 420 programmer and monitored at 280 nm with a Beckman Pa.) diluted 1:100 in PBS. Cells were washed in PBS, mounted in antifade Model 165 UV detector. Ion exchange HPLC was performed first on an mounting medium (6), and examined on a Leitz Orthoplan microscope AX300 column (4.1 x 300 mm; Synchrom, Inc., Linden, Ind.) equilibrated equipped with a Pleompak 2 fluorescence incidence illuminator and a in 20 mw Tris-acetate, pH 8.0, and eluted with a gradient (0 to 500 mw) standard fluorescein filter system. Both preimmune rabbit serum (100 Â¿Â¿gof sodium acetate. The relevant fractions (Chart 1/1) were pooled and IgG per ml) and oncomodulin-adsorbed serum served as immunological lyophilized. The dried material was dissolved in 1 ml of 20 mivi controls. Adsorbed serum was produced by the addition of purified rat NaH2PO4:100 mM Na2SO4 (pH 6.8), and 200-^1 aliquots were subjected hepatoma oncomodulin to Affigel Blue-purified rabbit anti-oncomodulin sequentially to size exclusion HPLC on a BioSil TSK-125 column (7.5 x IgG (340 n\, 7.5 mg/ml) to a final concentration of 2 x 10~6 M and 300 nm; Bio-Rad, Mississauga, Ontario, Canada) equilibrated in the incubation for 1 hr at room temperature and then overnight at 4Â°. above phosphate buffer. Finally, the immunoreactive fractions (Chart 1S) Following dilution to 100 ng IgG per ml, the immunoprecipitate was from the separate TSK-125 column runs were pooled (2.5 ml), lyophilized, pelleted by centrifugation at 35,000 x g for 1 hr, and the supernatant redissolved in 500 n\ of 10 mM KH2PO4, pH 7.0, and injected onto a was used in the indirect immunofluorescent technique described above. reverse-phase column (RP-P, 4.1 x 250 mm, 300 A pore; Synchrom, Purification of ASV-NRK Oncomodulin. Twenty-five g, wet weight, Inc.) equilibrated with 10 mM KH2PO4, and eluted with a gradient of n- of ASV-NRK solid tumors carried as xenografts in athymic mice were propyl alcohol (Chart 1D). The protein eluting at 35.5 min in 2.0 ml was extracted as described (9, 11). The 154 ml of supernatant were treated recovered by lyophilization. with ammonium sulfate to 65% saturation, and the proteins remaining in Amino acid analyses were performed after hydrolysis for 24, 48, and solution were precipitated by adjusting the pH to 4.0 with sulfuric acid 72 hr in 6 N HCI (Aristar; BDH, Mississauga, Ontario, Canada) at 100Â°in (10). This precipitate was dissolved in 20 mM Tris-acetate, pH 8.0, and the laboratory of Dr. Yaguchi (Division of Biological Sciences, National dialyzed against this same buffer. The immunoreactive material, as Research Council, Ottawa, Canada) using a one-column procedure on a sessed by radioimmunoassay, in the dialyzed sample (12.5 ml) was Durram D-500 analyzer.

NOVEMBER 1983 5391

RESULTS AND DISCUSSION virally transformed ASV-NRK cells contain the same oncomodulin molecule as that found in chemically transformed rat hepatoma We have isolated and purified oncomodulin previously from 5123tc. Morris hepatoma 5123tc (10) and have detected its presence in Specific radioimmunoassay indicated that nonneoplastic NRK 10 different rat hepatomas induced by a variety of chemical cells in culture do not contain oncomodulin at levels above the carcinogens (9, 11, 13, 15). The presence of oncomodulin in rat limits of detection (2 ng/mg protein) of the assay (Table 3). hepatoma has been confirmed recently by other investigators However, viral transformation of these cells by ASV caused (4).4 These observations prompted us to investigate whether cellular oncomodulin levels to increase by at least 600-fold to oncomodulin was also present in virally transformed cells and 1200 ng/mg protein. This high level persisted when ASV-NRK the tumors they produce. Material binding to anti-rat oncomo cells were carried as solid tumors in athymic mice (Table 3), dulin IgG was localized in ASV-transformed NRK cells by the indirect immunofluorescence technique (Fig. 1). Specific immu- Table 1 nofluorescence was detected in the nucleus and, to a lesser Purification ot immunoreactive material Â¡romASV-NRK solid tumors, 25 g, wet extent, the cytoplasm of all cells in a logarithmically growing weight, total ASV-NRK population (Fig. 16). No specific immunofluorescence Oncomodulin was observed (i.e., was similar to that found with preimmune rabbit serum; Fig. 1C) when antiserum preadsorbed with rat cen-t ra hepatoma oncomodulin was used, indicating that the observed tion8(ng/nO251902108001000Total(mg)3.72.42.12.02.0xg/mgprotein7.615.4161.5869.5 fluorescence was probably due to oncomodulin. Thus, onco stepHeated,Purification modulin appears to be localized mainly in the nucleus of cultured 100,000 x g superna ASV-NRK cells. It occurred in all cells, and thus, the expression tantpH of oncomodulin was not a random event in a subpopulation of dialyzedAX3004 precipitate, exchange)TSK-125(ion cells. By contrast, no specific immunofluorescence was detect exclusion)RP-P(size able in uninfected, nonneoplastic NRK cells (Fig. 1, Â£and F) (reverse phase)Volume(ml)15412.5102.52Totalprotein(mg)477156132.3Con- grown under identical culture conditions. Measured by radioimmunoassay.

While these observations point to the presence of oncomodulin Table 2 in NRK cells transformed by ASV, the question remained as to Amino acid composition of immunoreactive material from NRK solid tumors and whether anti-rat hepatoma oncomodulin IgG was recognizing the comparison with rat hepatoma oncomodulin same molecule as was isolated previously from rat hepatoma Molar ratio3 Oncomod 5123tc (10). This question could be answered only upon isolating ulin hepa- Proposed resi- toma and characterizing the immunoreactive material present in ASV- Amino acid 24 hr 48 hr 72 hr dues/molecule 5123tci)

NRK cells. To do this, the immunoreactive material in extracts AsxThreonineSerineGlxProlineGlycineAlanineCysteineValineMethionineIsoleucineLeucineTyrosinePhenylalanineHistidineLysineArginineTryptophanTotal of ASV-NRK solid tumors carried as xenografts in athymic mice was purified to homogeneity by a novel procedure using ion- exchange, size-exclusion, and reverse-phase HPLC (Chart 1). The purification is summarized in Table 1. Following ion-ex change HPLC, the immunoreactive material occurred only as a single peak coeluting with authentic rat hepatoma oncomodulin between 43 and 50 min (Chart 1A). These fractions were pooled, lyophilized, and fractionated further by size-exclusion HPLC (Chart 10). Again, the immunoreactive material occurred only as a single peak coeluting with authentic rat hepatoma oncomodulin at 18 min (Chart 1, ÃŸand C). From previous work (9-11), and from the amino acid sequence,5 it is known that the molecular weight of oncomodulin is 11,700. The elution time on size- exclusion HPLC of 18 min (equivalent to apparent M, 14,000, residues17.04.58.718.02.06.29.01.12.02.65.78.51.67.41.07.02.3017.14.28.018.01.46.29.01.32.12.45.88.41.47.61.17.42.3016.23.46.518.01.46.19.31.32.11.95.98.61.17.61.06.72.10175C10-11C18269123C682C81720107-108175111826912368281720108 Chart 1C) was anomalous for oncomodulin. The reason for this a For the molar ratio, Glx was assumed to be 18 residues/molecule. early elution is unknown, but it should be emphasized that the 6 From amino acid sequence. c After extrapolation to zero time of hydrolysis. immunoreactive material from the B77-NRK tumors was exactly the same apparent molecular weight as authentic rat hepatoma oncomodulin. This material from size-exclusion HPLC was Table 3 Oncomodulin content of normal and transformed NRK cells in culture and in solid pooled and subjected to reverse-phase HPLC (Chart 10). It was tumors essentially homogenous, with the major peak eluting at 35.5 min, tumors8 exactly the retention time of rat hepatoma oncomodulin. This (ng/mgprotein)NO" wt)194.5(iig/g wet (3)c material eluting at 35.5 min was shown to be immunoreactive NRK 1200Â±80d ASV-NRK Â±24.3 (12) (not shown). Finally, amino acid analysis of this material from Spon1-NRK reverse-phase HPLC revealed the composition to be identical to 300 Â±10 (3) 22.6 Â± 2.4(13) Spon 3-NRKCells 200 Â±10 (3)Solid 16.8 Â± 1.9(16) that of oncomodulin from rat hepatoma 5123tc (Table 2). Thus, a Carried as xenografts in athymic mice. 6 ND, not detectable (i.e., <2 ng/mg protein). c Numbers in parentheses, number of separate estimates made or separate 4 P. Burtin and J. G. DÃ©maille,personal communication. tumors analyzed. 5J. P. MacManus. D. C. Watson, and M. Yaguchi, manuscript in preparation. a Mean Â±S.D.

5392 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1983 American Association for Cancer Research. OncomodulÃ¬nin Virally Transformed Cells making these cells as rich a source of oncomodulin as rat a critical assessment of the evidence. Adv. Cyclic NÃºcleo.Res.,75: 193-294, 1983. hepatoma (11, 13, 15). In addition, 2 neoplastic transformants which arose "spontaneously" upon repeated passage of NRK 3. Chafouleas, J. G., Bolton, W. E., Hidaka, H., Boyd, A. E., and Means, A. R. Calmodulinand the cell cycle: involvement in regulation of cell cycle progres cultures were cloned and found to contain lower, but measurable, sion. Cell, 28: 41-50,1982. 4. Criss, W. E., and Kakiuchi, S. C. Calcium: calmodulin and cancer. Fed. Proc., levels of oncomodulin whether carried as solid tumors or in tissue 47:2289-2291,1982. culture (Table 3). 5. Hidaka, H., Sasaki, Y., Tanaka, T., Endo, T., Ohno, S., Fujii, Y., and Magala, T. fV-<6-Aminohexyl)-5-chloro-1-naphthalenesulfonamide,acalmodulin antag We conclude from these observations that oncomodulin can onist, inhibits cell proliferation. Proc. Nati. Acad. Sei. U. S. A., 78:4354-4357, be detected in both virally and spontaneously transformed NRK 1981. cells, as well as in chemically induced rat liver tumors (9-11,13, 6. Johnson, G. D., and Arauja, G. M. N. A simple method of reducing the fading of immunofluorescence during microscopy. J. Immunol. Methods, 43: 349- 15). Thus, it would appear that the means by which cells are 350,1981. transformed does not determine oncomodulin production. Al 7. Klee, C. B., and Krinks, M. H. Purification of cyclic 3',5'-nucleotide phosphc- though the cellular function of oncomodulin remains unknown, it diesterase inhibitory protein by affinity chromatography on activator protein coupled to Sepharose. Biochemistry, 17:120-126,1978. is clearly produced as a direct consequence of neoplastic trans 8. LaPorte, D. C., Gidwitz, S., Weber, M. J., and Storm, D. R. Relationship formation. The levels of oncomodulin in transformed cells ap between changes in the calcium-dependentregulatory protein and adenylate proach that of calmodulin (4, 13, 15), the ubiquitous calcium- cyclase during viral transformation. Biochem. Biophys. Res. Commun., 86: 1169-1177. 1979. binding protein implicated in the regulation of normal cell prolif 9. MacManus,J. P. Occurrenceof a low-molecular-weightcalcium-bindingprotein eration (2-5). The fact that cellular calmodulin levels also increase in neoplastic liver. Cancer Res., 39: 3000-3005,1979. 10. MacManus, J. P. The purification of a unique calcium-binding protein from dramatically upon neoplastic transformation (4, 8, 14-17) sug Morris hepatoma 5123tc. Biochim. Biophys. Acta, 627: 296-304,1980. gests that changes in the total calcium-binding activity of the cell 11. MacManus, J. P. Developmentand use of a quantitative immunoassayfor the may be important to the expression of the transformed pheno- calcium-bindingprotein (molecularweight, 11,500) of Morris hepatoma 5123. Cancer Res., 47: 974-979, 1981. type. 12. MacManus,J. P. The stimulation of cyclic nucleotide phosphodiesteraseby a M, 11500 calcium binding protein from hepatoma. FEBS Lett., 726: 245-249, 1981. ACKNOWLEDGMENTS 13. MacManus, J. P. Calcium-binding proteins and cell proliferation. In: A. L. Boynton, W. H. McKeehan, and J. F. Whitfield (eds.), Ions, Cell Proliferation, Our thanks are due to H. Trinh for the radioimmunoassay,B. M. Bracelandfor and Cancer, pp. 489-498. New York: Academic Press, 1982. technical assistance, D. Watson for amino acid analysis, D. J. Gillan for the 14. MacManus,J. P., Braceland, B. M., Rixon, R. H., Whitfield, J. F., and Morris, illustrative material, and D. PichÃ©fortyping the manuscript. We also thank Dr. J. H. P. An increase in calmodulin during growth of normal and cancerous liver F. Whitfield for his constructive review of this manuscript and his constant encour in vivo. FEBS Lett., 733: 99-102, 1981. agement and support. 15. MacManus,J. P., Whitfield, J. F., Boynton, A. L., Durkin, J. P., and Swierenga. S. H. H. Oncomodulinâ€”awidely distributed, tumor-specific, calcium-binding protein. Oncodev. Biol. Med., 3: 79-90, 1982. 16. Watterson, D. M., Van Eldik, L. J., Smith, R. E., and Vanaman,T. C. Calcium- REFERENCES dependent regulator protein of cyclic nucleotide metabolism in normal and transformed chicken embryo fibroblasts. Proc. Nati. Acad. Sei. U. S. A., 73: 1. Boynton, A. L., MacManus, J. P., and Whitfield. J. F. Stimulation of liver cell 2711-2715,1981. DMA synthesis by oncomodulin, an MW 11500 calcium-binding protein from 17. Wei, J. W., and Hickie, R. A. Increased content of calmodulin in Morris hepatoma. Exp. Cell Res., 738: 454-458,1982. hepatoma 5123t.c.(h). Biochem. Biophys. Res. Commun., 700: 1562-1568, 2. Boynton, A. L., and Whitfield, J. F. The rote of cyclic AMP in cell proliferation: 1981.

NOVEMBER 1983 5393

'-â€¢v^:< <Ã-^ ,"

â€¢

Fig. 1. Immunological localization of immunoreactive material in NRK cells transformed by ASV. Nonneoplastic NRK cells and ASV-NRK cells were treated with antioncomodulin IgG or preimmune IgG and then with fluorescein isothiocyanate-conjugated anti-rabbit IgG as described in "Materials and Methods." ASV-NRK (A) and NRK (D) cells under phase contrast; ASV-NRK (B) and NRK (5) cells treated with rabbit anti-rat hepatoma oncomodulin IgG and fluorescein Â¡sothiocyanate-conjugated sheep anti-rabbit IgG; and ASV-NRK (C) and NRK (F) cells treated with preimmune rabbit IgG and fluorescein isothiocyanate-conjugated sheep anti-rabbit IgG.

5394 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1983 American Association for Cancer Research. Occurrence of the Tumor-specific, Calcium-binding Protein, Oncomodulin, in Virally Transformed Normal Rat Kidney Cells

J. P. Durkin, L. M. Brewer and J. P. MacManus

Cancer Res 1983;43:5390-5394.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/43/11/5390

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/43/11/5390. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.