Significance of Immunohistochemical Expression of Estrogen Receptors

Human Cancer Biology

Significance of Immunohistochemical Expression of Estrogen Receptors A and B in Squamous Cell Carcinoma of the Esophagus Tadahiro Nozoe,Tsunehiro Oyama, Mitsuhiro Takenoyama,Takeshi Hanagiri, Kenji Sugio, and KoseiYasumoto

Abstract Purpose: Possible significance of sexhormone estrogen as an antitumor therapeutic arm toward esophageal squamous cell carcinoma (ESCC) cells has been suggested. The aim of the current study was to clarify the clinicopathologic significance of an immunohistochemical expression of estrogen receptors a and h (ERa and ERh)inESCC. Experimental Design: Expression of ERa and ERh were examined using an immunohistochemical methods in 73 paraffin-embedded sections collected from patients with ESCC who had been subjected to esophageal resection and digestive reconstruction without any preoperative induction therapy. Results: Forty-seven (64.4%) ESCCs had a positive cytoplasmic expression of ERa and 21 (28.8%) ESCCs had a positive nuclear expression of ERh. Univariate analysis showed that both positive ERa expression (P = 0.0001) and negative ERh expression (P = 0.026) were unfavorable prognostic indicators in ESCC. Moreover, multivariate analysis showed that ERa-positive/ ERh-negative expression (P = 0.003) and progression of tumor stage (P =0.014)werefound to be independent unfavorable prognostic indicators in ESCCs. Conclusions: Immunohistochemical expression of ERa and ERh were found to be observed in ESCC. Positive expression of ERa in addition to negative expression of ERh proved to be an unfavorable independent prognostic indicator in ESCC.

Previous reports have indicated that there is a possible Subtypes of ER, ERa and ERh, were initially cloned in 1986 correlation of estrogen with the biological potential of and 1996, respectively. ERa and ERh are found to be located on esophageal squamous carcinoma cells (1, 2). An epidemiologic human chromosome 6q25 (6) and chromosome 14q22-24 (7), investigation showed that the male to female rate of the respectively. These ERs function as the factor for ligand-activated incidence of esophageal squamous cell carcinoma (ESCC) transcription. Both ERa and ERh contain some functional reached to 3:1 or 4:1 (3). domains, including a central DNA-binding domain conversed The previous investigation showed that the prognosis of between ERa and ERh. Therefore, it is suggested that these sex esophageal carcinoma was found to be significantly better in hormone receptor might contain different biological function. female than in male (4). Therefore, estrogen and estrogen Since the discovery of these ERs, the biological significance of receptors (ER) are considered to be possibly involved in these sex hormone receptors in human tumors, including breast tumorigenesis and/or progression of squamous cell carcinoma carcinoma, has been investigated, and the relationship between of the esophagus. Previous report showed that inhibitory effect the expression of ER subtypes and the malignant potential of the against the proliferation of esophageal squamous carcinoma tumors have been reported (8–10). ERh is known to bind to ERa cells might be regulated by the function of ER (5). However, an and to have a suppressive function toward ERa (11, 12). immunohistochemical expression of ER on esophageal squa- Therefore, these sex hormones have been considered to mutually mous carcinoma cells has not been studied fully. contain inverse biological function. The purpose of this study is to elucidate the clinicopathologic and/or prognostic significances of an immunohistochemical Authors’ Affiliation: Second Department of Surgery, University of Occupational expression of ERa and ERh in ESCC. and Environmental Health, Kitakyushu, Japan Received 2/21/07; revised 4/16/07; accepted 5/7/07. The costs of publication of this article were defrayed in part by the payment of page Materials and Methods charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Patients and samples. Seventy-three specimens of ESCCs collected Requests for reprints: Tadahiro Nozoe, Second Department of Surgery, University of Occupational and Environmental Health, 1-1, Iseigaoka, Yahata- from the patients, who had been subjected to esophageal resection and NishiWard, Kitakyushu 807-0855, Japan. Phone: 81-093-691-7442; Fax: 81- reconstruction of the digestive tract in our department between 1993 093-692-4004; E-mail: [email protected]. and 2004, were enrolled. Neoadjuvant therapy was not done to any F 2007 American Association for Cancer Research. patients. These 66 men and 7 women had a median age of 64 years doi:10.1158/1078-0432.CCR-07-0449 (range, 42-83 years).

Clin Cancer Res 2007;13(14) July 15, 2007 4046 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2007 American Association for Cancer Research. ER a and b in Esophageal Carcinoma

The follow-up for patients was continued until their death and only Results patients who died of ESCC were included in the tumor-related deaths. Follow-up periods after surgical treatment ranged from 3 months to Forty-seven (64.4%) ESCCs had a positive cytoplasmic 11 years and 3 months with a mean of 2 years and 8 months. expression of ERa and the other 26 (35.6%) did not (Fig. 1A Pathologic features were shown based on the guidelines for clinical and pathologic studies on carcinoma of the esophagus established by and B). In the normal mucosa concomitant with carcinoma a the Japanese Society for Esophageal Diseases (13, 14). tissue, cytoplasmic expression of ER was observed in only Immunohistochemical expression of ERa and ERb. Four-micrometer- some cells (Fig. 1C). Relationship between ERa expression and thick sections sliced from paraffin-embedded specimen were prepared on clinicopathologic features was shown in Table 1. Proportion of the slide glasses precoated with silane. After removing the paraffin with male among patients with ESCCs expressing ERa was signifi- xylene and washing in a graded series of ethanol, the sections were placed cantly higher than that among patients without ERa expression in TBS for 10 min. Endogenous peroxidase activity was blocked for (P = 0.004). Lymph node metastasis and venous invasion of 10 min in methanol containing 0.3% H2O2, and then the slides were ESCCs with ERa expression were significantly more frequent placed in TBS. The sections were incubated with TBS including 1% than those of tumors without ERa expression (P = 0.023 and concentration of bovine serum albumin for 10 min to block nonspecific 0.017, respectively). Stage of tumors was significantly more binding of the immunoreagents. After washing in TBS, the sections were a P incubated with 1:50 diluted rabbit anti-human polyclonal ERa antibody advanced in ESCCs with ER expression ( = 0.027). Five-year a (HC-20; Santa Cruz Biotechnology) or 1:20 diluted rabbit anti-human survival rate of patients with ESCCs with ER expression polyclonal ERh antibody (H-150; Santa Cruz Biotechnology). All (23.2%) was significantly worse than that of patients with incubations proceeded overnight at 4jC. After washing in TBS, an ESCCs without ERa expression (68.6%; P = 0.0001; Fig. 2). immunoperoxidase staining was done by an EnVision antibody complex Twenty-one (28.8%) ESCCs had a positive nuclear expression method (15) using Envision kit (DAKO Ltd.). Finally, the localization of of ERh and the other 52 (71.2%) did not (Fig. 1D and E). In the ERs was visualized with diaminobenzidine tetrahydrochloride. normal mucosa concomitant with carcinoma tissue, nuclear In accordance with the criteria in the report of Wu et al. (16) about expression of ERh was observed in only some cells (Fig. 1F). ER expression in lung carcinoma, ESCCs with at least 50% of the tumor There was no significant difference about clinicopathologic a cells having cytoplasmic expression of ER and nuclear expression of features between ERh-positive and ERh-negative patients ERh were regarded as ERa and ERh positive, respectively (Fig. 1). Statistical analysis. m2 t (Table 2). Five-year survival rate of 57.5% in patients with The test and Student’s test were used to h compare the clinicopathologic data. The cumulative survival rates were ESCCs with ER expression was significantly better than that of h calculated by the Kaplan-Meier method and the survival curves were 32.6% in patients with ESCCs without ER expression tested by the Mantel-Cox method. The multivariate survival analysis was (P = 0.026; Fig. 3). calculated according to Cox’s proportional hazards model in a forward Proportion of positive and negative expression of ERa among stepwise manner. A P value of <0.05 was considered significant. ESCCs with positive ERh expression was 42.9% (9 of 21) and

Fig. 1. Photo of immunohistochemical expression of ERa and ERh. A, representative photo of cytoplasmic expression of ERa. B, negative expression of ERa. C, cytoplasmic expression of ERa in some cells in the normal epithelium. D, representative photo of nuclear expression of ERh. E, negative expression of ERh. F, nuclear expression of ERh in some cells in the normal epithelium.

www.aacrjournals.org 4047 Clin Cancer Res 2007;13(14) July 15, 2007 Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2007 American Association for Cancer Research. Human Cancer Biology

Table 1. Relationship between ERa expression and clinicopathologic features

ERA positive, ERA negative, P n =47(%) n =26(%) Gender Male 46 (69.7) 20 (30.3) 0.004 Female 1 (14.3) 6 (85.7) Age (y) 64.1 F 7.8 63.8 F 7.40.907 Location of tumors Upper 9 (64.3) 5 (35.7) 0.310 Middle 29 (70.7) 12 (29.3) Lower 9 (50.0) 9 (50.0) Histology Well 7 (70.0) 3 (30.0) 0.892 Moderately 32 (62.7) 19 (37.3) Poorly 8 (66.7) 4(33.3) Depth of the tumors

Tis,T1 17 (68.0) 8 (32.0) 0.139 T2 4(40.0) 6 (60.0) T3 16 (64.0) 9 (36.0) T 10 (76.9) 3 (23.1) Fig. 2. Survival curves of patients with ESCCs with and without ERa expression. 4 a bold line Lymph node metastasis Survival rates of patients with ESCCs with ER expression ( )was significantly worse than that of patients with ESCCs without ERa expression Positive 31 (75.6) 10 (24.4) 0.023 (thin line ; P =0.0001). Negative 16 (50.0) 16 (50.0) Lymphatic permeation Positive 39 (68.4) 18 (31.6) 0.180 Previous experimental investigations showed that sex hor- Negative 8 (50.0) 8 (50.0) Venous invasion mone estrogen might have a biological function to suppress the Positive 28 (77.8) 8 (22.2) 0.017 progression and proliferation of ESCC (1, 2). Utsumi et al. (5) Negative 19 (51.4) 18 (48.6) reported that growth of transplanted tumors derived from Stage of tumors ESCC cells expressing ER in female nude mice were significantly 0, I 12 (63.2) 7 (36.8) 0.027 enhanced when they were oophorectomized, which suggested II 12 (50.0) 12 (50.0) III 23 (76.7) 7 (23.3) Table 2. Relationship between ERh expression and clinicopathologic features

57.1% (12 of 21), respectively. On the other hand, proportion ERB positive, ERB negative, P of positive and negative expression of ERa among ESCCs n =21(%) n =52(%) with negative ERh expression was 73.1% (38 of 52) and 26.9% Gender (14 of 52), respectively. There was a significantly inverse Male 17 (25.8) 49 (74.2) 0.081 correlation between ERa and ERh expression (P = 0.15; Table 3). Female 4(57.1) 3 (42.9) Comparison of clinicopathologic features between ERa- Age (y) 64.0 F 6.9 64.0 F 7.9 0.981 positive/ERh-negative ESCCs and tumors showing other expres- Location of tumors a Upper 3 (21.4) 11 (78.6) 0.233 sion patterns (both positive, both negative, and ER negative/ Middle 10 (24.4) 31 (75.6) ERh positive) was shown in Table 4. Proportion of male among Lower 8 (44.4) 10 (55.6) patients with ERa-positive/ERh-negative ESCCs was significantly Histology higher than that among patients with ESCCs of other expressing Well 3 (30.0) 7 (70.0) 0.951 patterns (P = 0.028). Lymph node metastasis and venous Moderately 15 (29.4) 36 (70.6) a h Poorly 3 (25.0) 9 (75.0) invasion of ER -positive/ER -negative ESCCs were significantly Depth of the tumors more frequent than those with other expressing patterns Tis,T1 6 (24.0) 19 (76.0) 0.645 (P = 0.028 and 0.045, respectively). Five-year survival rate of T2 2 (20.0) 8 (80.0) 22.1% in patients with ERa-positive/ERh-negative ESCCs was T3 10 (40.0) 15 (60.0) T4 3 (23.1) 10 (76.9) significantly worse than that of patients with ESCCs with other Lymph node metastasis expressing patterns (58.1%; P < 0.0001; Fig. 4). Positive 10 (24.4) 31 (75.6) 0.350 Multivariate analysis showed that stage of the tumor (P = Negative 11 (34.4) 21 (65.6) 0.014) and ERa-positive/ERh-negative expression (P = 0.003) Lymphatic permeation proved to be independent prognostic indicators of patients Positive 17 (29.8) 40 (70.2) 0.706 Negative 4(25.0) 12 (75.0) with ESCCs (Table 5). Venous invasion Positive 11 (30.6) 25 (69.4) 0.739 Discussion Negative 10 (27.0) 27 (73.0) Stage of tumors In spite of the recent remarkable progress in diagnosis and 0, I 5 (26.3) 14(73.7) 0.615 II 7 (29.2) 17 (70.8) therapeutic options, including surgical treatment in ESCC, the III 9 (30.0) 21 (70.0) prognosis of the disease still remains dismal (17).

Clin Cancer Res 2007;13(14) July 15, 2007 4048 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2007 American Association for Cancer Research. ER a and b in Esophageal Carcinoma

Fig. 3. Survival curves of patients with ESCCs with and without ERh expression. a h h bold line Fig. 4. Survival curves of patients with ER -positive/ER -negative ESCCs and Survival rates of patients with ESCCs with ER expression ( )was other patients. Survival rates of the former (bold line) was significantly worse than significantly better than that of patients with ESCCs without ERh expression that of the latter (thin line ; P < 0.0001). (thin line ; P = 0.026).

investigation for an immunohistochemical expression of these the inhibitory effect of estrogen to the proliferation of ESCC. ER subtypes in ESCC has not been reported and this is the first These results also suggested the clinical application of sex report about this point. hormone estrogen as a possible antitumor agent in ESCC. Both ERa and ERh are ligands to estradiol, but they differ in Experimental investigations to date have shown the antitumor effect on transcription at activator protein-1 sites (22). effect of tamoxifen for squamous cell carcinoma, mainly of the head and neck. However, the effect could not be ascribed to the estrogen antagonism but to some other biological genetic Table 4. Comparison of clinicopathologic features alteration (18, 19). between patients with ERa-positive/ERh-negative Whereas estrogen has been known to accelerate the malignant tumors and the others potential of breast cancer, the ER has been chiefly focused as the target of the hormonal therapy in breast cancer. Moreover, the ERA positive/ERB Other expressing P prognosis of patients with breast cancer–expressing ER has been negative, types, reported to be more favorable than that of patients with tumors n =38(%) n =35(%) without ER expression, and the expression of ER has been Gender acknowledged as a marker of candidates for hormonal therapy in Male 37 (56.1) 29 (43.9) 0.028 patients with breast carcinoma (20). According to the panel of Female 1 (14.3) 6 (85.7) Age (y) 64.3 F 8.0 63.7 F 7.2 0.724 the investigations, showing a difference between functions of sex Location of tumors hormone to squamous cell carcinoma of some organs and breast Upper 9 (64.3) 5 (35.7) 0.158 carcinoma, the biological significance of ERs in ESCC might Middle 23 (56.1) 18 (43.9) differ from that in breast carcinoma. Lower 6 (33.3) 12 (66.7) a h Histology Recently, expression of ER subtypes, ER and ER , in human Well 5 (50.0) 5 (50.0) 0.891 malignant tumors has come to be focused. Clinical and Moderately 26 (51.0) 25 (49.0) experimental investigations to date have generally shown that Poorly 7 (58.3) 5 (41.7) expression of ERa and/or loss of ERh expression could be Depth of the tumors correlated with malignant potential of human tumors, includ- Tis,T1 13 (52.0) 12 (48.0) 0.080 T 4(40.0) 6 (60.0) ing non–small lung cancer, colorectal cancer, and prostate 2 T3 11 (44.0) 14 (56.0) cancer (8–10, 16, 21). However, to the best of our knowledge, T4 10 (76.9) 3 (23.1) Lymph node metastasis Positive 26 (63.4) 15 (36.6) 0.028 Table 3. Relationship between ERa expression Negative 12 (37.5) 20 (62.5) h Lymphatic permeation and ER expression Positive 32 (56.1) 25 (43.9) 0.259 Negative 6 (37.5) 10 (62.5) ERA expression Venous invasion Positive, Negative, Positive 23 (63.9) 13 (36.1) 0.045 n =47(%) n =26(%) Negative 15 (40.5) 22 (59.5) Stage of tumors ERh expression 0, I 9 (47.4) 10 (52.6) 0.116 Positive (n = 21) 9 (42.9) 12 (57.1) II 10 (41.7) 14 (58.3) Negative (n = 52) 38 (73.1) 14(26.9) III 19 (63.3) 11 (36.7)

www.aacrjournals.org 4049 Clin Cancer Res 2007;13(14) July 15, 2007 Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2007 American Association for Cancer Research. Human Cancer Biology

Table 5. Multivariate analysis to determine factors independently associated with unfavorable prognosis of patients

Variables Regression SE Odds ratio P coefficient (95% confidence interval) Tumor stage (III and IV versus I and II) 1.104 0.449 3.012 (1.250-7.299) 0.014 ER expression pattern 1.169 0.391 3.215 (1.495-6.944) 0.003 (ERa positive/ERh negative versus others)

Although ERa activates the transcription at activator protein-1 In lung cancers, the proportion of ERa expression in female site, ERh inversely inhibits the transcription. ERh is known to patients has been reported to be higher than in men, showing bind to ERa and to have a suppressive function toward ERa that the ERa expression could be gender dependent (24). (11, 12). Although the number of female patients enrolled in the current In the current study, expression of ERa was found to be also study is small, the proportion of male among patients with inversely correlated with that of ERh in ESCC. Similar results ESCCs expressing ERa-positive and ERh-negative expression was have been shown about lung cancer (10). Expression of ERa significantly higher than in female. has been reported to be crucially correlated with progression of In conclusion, positive expression of ERa in addition to colorectal carcinoma (8), and the significant correlation negative expression of ERh was found to be an unfavorable between loss of ERh expression and advanced Dukes’ stage independent prognostic indicator in ESCC. Sex hormone was shown, which might suggest the significance of ERh (estrogen) therapy can be considered as one of candidates to expression as an indicator of favorable prognosis and a improve prognosis of ESCC. Although results in the current protective role of ERh against colon cancer progression (9). study might provide the hint for possibility of sex hormone Expression of ERa and ERh were observed in the cytoplasm and therapy for selected patients with ESCC, relationship between the nucleus of the ESCC cells, respectively, which was consistent expression of ERa and ERh and possible antitumor effect derived with the previous reports (16, 23). However, the reason for an by sex hormone in ESCC should be established using experi- intracellular localization of ERa and ERh remains unclear. mental and clinical investigations in the future proposition.

References 1. MatsuokaH,SugimachiK,UeoH,KuwanoH, 9. Jassam N, Bell SM, Speirs V, Quirke P. Loss of small cell lung cancers. J Thorac Cardiovasc Surg Nakano S, Nakayama M. Sexhormone response of a expression of oestrogen receptor h in colon cancer 2005;130:979^86. newly established squamous cell line derived from and its association with Dukes’ staging. Oncol Rep 17. TewWP, Kelsen DP, Ilson DH. Targeted therapies for clinical esophageal carcinoma. Cancer Res 1987;47: 2005;14:17^21. esophageal cancer. Oncologist 2005;10:590 ^ 601. 4134 ^ 40. 10. Kawai H, Ishii A,Washiya K, et al. Estrogen receptor 18. Ferrandina G, Ranelletti FO, Larocca LM, et al. Ta- 2. Ueo H, Matsuoka H, Sugimachi K, Kuwano H, Mori a and h are prognostic factors in non-small cell lung moxifen modulates the expression of Ki67, apoptosis, M, Akiyoshi T. Inhibitory effects of estrogen on the cancer. Clin Cancer Res 2005;11:5084 ^ 9. and microvessel density in cervical cancer. Clin Can- growth of a human esophageal carcinoma cell line. 11. Liu MM, Albanese C, Anderson CM, et al. Oppos- cer Res 2001;7:2656 ^ 61. Cancer Res 1990;50:7212^ 5. ing action of estrogen receptors a and h on cyclin 19. Hoffmann TK, Bojar H, Eckel J, et al. Effects of 3. Yang PC, Davis S. Incidence of cancer of the esophagus D1 gene expression. J Biol Chem 2002;277: tamoxifen on human squamous cell carcinoma lines in the US by histologic type. Cancer1988;61:612^ 7. 24353 ^ 60. of the head and neck. Anticancer Drugs 2002;13: 4. SugimachiK,MatsuokaH,MatsufujiH,MaekawaS, 12. Hayashi SI, Eguchi H,Tanimoto K, et al. The expres- 521 ^31. KaiH,OkudairaY.Survivalratesofwomenwithcarci- sion and function of estrogen receptor a and h in hu- 20. Greene GL, Nolan C, Engler JP. Monoclonal anti- noma of the esophagus exceed those of men. Surg man breast cancer and its clinical application. Endocr bodies to human estrogen receptor. Proc Natl Acad Gynecol Obstet 1987;164:541 ^ 4. Relat Cancer 2003;10:193^ 202. Sci U S A 19 8 0;77:5115 ^ 9. 5. Utsumi Y, Nakamura T, Nagasue N, Kubota H, 13. Japanese Society for Esophageal Diseases. Guide- 21. Cheng J, Lee EJ, Madison LD, Lazennec G. Expres- Morikawa S. Role of estrogen receptors in the growth lines for the clinical and pathologic studies on carci- sion of estrogen receptor h in prostate carcinoma cells of human esophageal carcinoma. Cancer 1989;64: noma of the esophagus. 9th ed. Preface, general inhibits invasion and proliferation and triggers apopto- 88^93. principles, part I. Esophagus 2004;61-88. sis. FEBS Lett 2004;566:169 ^72. 6. Woo IS, Park MJ, Choi SW, et al. Loss of estrogen 14. Japanese Society for Esophageal Diseases. Guide- 22. Paech K, Webb P, Kuiper GG, et al. Differential li- receptor-a expression is associated with hypermethy- lines for the clinical and pathologic studies on carci- gand activation of estrogen receptors ERa and ERh lation near its ATG start codon in gastric cancer cell noma of the esophagus. 9th ed. part II. Esophagus at AP1sites. Science 1997;277:1508 ^ 10. lines. Oncol Rep 2004;11:617^ 22. 2004;107-25. 23. Stabile LP, Davis AL, Gubish CT, et al. Human non- 7. Peng B, Lu B, Leygue E, Murphy LC. Putative func- 15. Sabattini E, Bisgaard K, Ascani S, et al. The En- small cell lung tumors and cells derived from normal tional characteristics of human estrogen receptor-h Vision++ system: a new immunohistochemical lung express both estrogen receptor a and h and isoforms. J Mol Endocrinol 2003;30:13^ 29. method for diagnostics and research. Critical com- show biological responses to estrogen. Cancer Res 8. Cavallini A, Notarnicola M, Giannini R, et al. Oestro- parison with the APAAP, ChemMate, CSA, LABC, 2002;62:2141 ^ 50. gen receptor-related receptor a (ERRa)andoestro- and SABC techniques. J Clin Pathol 1998;51: 24. Fasco MJ, Hurteau GJ, Spivack SD. Gender-depen- gen receptors (ERa and ERh) exhibit different gene 506^11. dent expression of a and h estrogen receptors in hu- expression in human colorectal tumour progression. 16. Wu CT, Chang YL, Shih JY, Lee YC. The significance man nontumor and tumor lung tissue. Mol Cell Eur J Cancer 2005;41:1487 ^ 94. of estrogen receptor h in 301 surgically treated non- Endocrinol 2002;188:125^40.

Clin Cancer Res 2007;13(14) July 15, 2007 4050 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2007 American Association for Cancer Research. Significance of Immunohistochemical Expression of Estrogen Receptors α and β in Squamous Cell Carcinoma of the Esophagus

Tadahiro Nozoe, Tsunehiro Oyama, Mitsuhiro Takenoyama, et al.

Clin Cancer Res 2007;13:4046-4050.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/13/14/4046

Cited articles This article cites 22 articles, 12 of which you can access for free at: http://clincancerres.aacrjournals.org/content/13/14/4046.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/13/14/4046.full#related-urls

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://clincancerres.aacrjournals.org/content/13/14/4046. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.