CANCER STEM CELLS Erythropoietin Receptor Expression in Non-Small Cell Lung Carcinoma: A Question of Antibody Specificity W. MARK BROWN,a PERRY MAXWELL,a,b ALASTAIR N.J. GRAHAM,c ANITA YAKKUNDI,a ELAINE A. DUNLOP,a ZHANZHONG SHI,a PATRICK G. JOHNSTON,a TERENCE R.J. LAPPINa a
Centre for Cancer Research and Cell Biology, Queen’s University, Belfast, United Kingdom; bDepartment of Pathology and cNorthern Ireland Regional Department of Thoracic Surgery, Royal Group of Hospitals Trust, Belfast, United Kingdom Key Words. Erythropoietin receptor • Non-small cell lung carcinoma • C20 Antibody • Heat shock proteins
ABSTRACT Immunohistochemical studies on formalin-fixed, paraffin-embedded (FFPE) tissue utilizing polyclonal antibodies form the cornerstone of many reports claiming to demonstrate erythropoietin receptor (EPOR) expression in malignant tissue. Recently, Elliott et al. (Blood 2006;107:1892–1895) reported that the antibodies commonly used to detect EPOR expression also detect non-EPOR proteins, and that their binding to EPOR was severely abrogated by two synthetic peptides based on the sequence of heat shock protein (HSP) 70, HSP70-2, and HSP70-5. We have investigated the specificity of the C20 antibody for detecting EPOR expression in non-small cell lung carcinoma (NSCLC) utilizing tissue microarrays. A total of 34 cases were available for study. Antibody absorbed with peptide resulted in marked suppression of cytoplasmic staining com-
INTRODUCTION Erythropoietin (EPO), a 30.4-kDa glycoprotein, induces erythropoiesis and is produced mainly by the kidney and the liver in response to hypoxia. EPO binds to the erythropoietin receptor (EPOR), a type 1 transmembrane protein belonging to the cytokine receptor superfamily, thereby promoting the survival, proliferation, and differentiation of the erythroid cells. Subsequent downregulation of the signal is effected by internalization and degradation of the receptor-ligand complex via both lysosomal and proteasomal pathways [1]. The EPOR gene is localized to chromosome 19p13.2–p13.3 and encodes a protein of 508 amino acid residues. The first report of EPO and EPOR expression in neoplasia was in clear and chromophilic cell renal carcinoma [2]. Their combined occurrence was also reported to be present in ductal carcinoma of the breast but absent in normal ductal epithelium and in benign pathologies, such as hyperplasia of usual type, ductal papilloma, fibrocystic change [3], and sclerosing adenosis [4]. Expression of EPOR has also been claimed in melanoma [5] and cervical squamous [6], papillary thyroid [7], endometrial [8], and head and neck squamous [9] carcinomas and numerous tumor cell lines [10, 11]. Many cancer patients suffer from anemia, and recombinant human EPO (rHuEPO) is now widely used therapeutically, as it improves hematocrit, lowers transfusion requirements, and im-
pared with nonabsorbed antibody. Four tumors that initially showed a membranous pattern of staining retained this pattern with absorbed antibody. Positive membranous immunoreactivity was also observed in 6 of 30 tumors that originally showed a predominantly cytoplasmic pattern of staining. Using the C20 antibody for Western blots, we detected three main bands, at 100, 66, and 59 kDa. Preincubation with either peptide caused abolition of the 66-kDa band, which contains non-EPOR sequences including heat shock peptides. These results call into question the significance of previous immunohistochemical studies of EPOR expression in malignancy and emphasize the need for more specific anti-EPOR antibodies to define the true extent of EPOR expression in neoplastic tissue. STEM CELLS 2007;25:718 –722 proves quality of life. Erythropoiesis-stimulating agents ameliorate chemotherapy-associated anemia and are recommended in guidelines issued by the National Comprehensive Cancer Network as treatment options for patients with cancer, cited in LaMontagne et al. [12]. Normal erythropoiesis is dependent on basal levels of secreted EPO, which are equivalent to 0.8 – 4.0 pmoles/l of EPO (5–25 U/l) in plasma. The widespread practice of treating cancer patients with rHuEPO has raised concerns about the possible modulation of tumor growth from pharmacological doses of EPO that result in plasma concentrations several logs higher than the normal physiological level. Indeed, a study of the effects of EPO in anemic head and neck cancer patients undergoing radiotherapy indicated poorer locoregional progressionfree survival in those treated with EPO [13]. The Breast Cancer Erythropoietin Survival Trial (BEST), which used overall survival as the study endpoint, was halted early because of unacceptably high mortality in the EPO treatment group [14]. This sets in context the need for specific tools for the detection of EPOR in malignancy. Reports of EPOR expression in cancer rely heavily on the use of anti-EPOR antibodies in immunostaining and immunoblotting. In particular, anti-EPOR antibodies from Santa Cruz Biotechnology (C20; catalog number SC-695; Santa Cruz, CA, http://www.scbt.com/) that recognize the last 20 amino acids of the receptor have been used in most of the reported studies (Table 1). Recently, Elliott et al. [15] reported that C20 fails to
Correspondence: Perry Maxwell, Ph.D., Department of Pathology, Institute of Pathology, Royal Group of Hospitals Trust, Grosvenor Road, Belfast BT12 6BA, United Kingdom. Telephone: ⫹44 2890635074; Fax: ⫹44 2890632763; e-mail:
[email protected] Received October 26, 2006; accepted for publication November 8, 2006; first published online in STEM CELLS EXPRESS November 16, 2006. ©AlphaMed Press 1066-5099/2007/$30.00/0 doi: 10.1634/stemcells.2006-0687
STEM CELLS 2007;25:718 –722 www.StemCells.com
Brown, Maxwell, Graham et al.
719
Table 1. Reported immunohistochemical studies in malignancy using the C20 anti-erythropoietin receptor antibody Site/malignancy
References
Breast (DCIS and invasive ductal/lobular adenoCa) Cervix (squamous Ca) Thyroid (papillary adenoCa) Endometrium (endometrioid, serous, and clear cell adenoCa) Head and neck (squamous Ca) Lung (non-small cell Ca)
[3, 4, 11] [6] [7] [8] [9] [16]
Abbreviations: Ca, carcinoma; DCIS, ductal carcinoma in situ.
detect peptides at the calculated molecular weight of mature EPOR (59 kDa), but does detect peptides at 35, 66, and 100 kDa. They found that the 66-kDa band represents heat shock proteins (HSPs) and that binding of the C20 antibody was abolished by two synthetic peptides based on the sequence of HSP70. Clearly, confirmation of this report would call into question the immunohistochemical findings often cited as evidence for the occurrence of EPOR in neoplasia. Utilizing the C20 antibody, Dagnon et al. [16], examined samples from 29 patients with non-small cell lung carcinoma (NSCLC) and detected EPOR in 28 (96%) by immunohistochemistry. Similarly, in a cohort of 66 patients with histopathologically confirmed NSCLC, we detected positive immunoreactivity to EPOR in all cases with the same antibody (unpublished data). In the light of the report by Elliott et al, we decided to investigate the specificity of the C20 antibody for detection of EPOR in NSCLC. Consequently, we used tissue microarrays (TMAs) to determine whether the two peptides, HSP70-2 and HSP70-5, can affect C20 binding to formalin-fixed, paraffin-embedded (FFPE) tissue both with and without heat-mediated antigen retrieval (HMAR). In addition, to investigate the assertion that C20 cannot detect EPOR at 59 kDa [15], Western blotting analysis of protein lysates from three different cell lines using C20 as primary antibody was performed.
MATERIALS
AND
METHODS
Tissue Microarrays Formalin-fixed, paraffin-embedded TMAs were obtained from the National Cancer Institute Tissue Microarray Research Program (TARP). Arrays were fifth-generation (TARP5), T-CL-1-containing cores from NSCLC, complete with associated clinicopathological data.
Heat Shock Peptides Both heat shock peptide sequences identified by Elliott et al. [15] were synthesized in the Protein and Peptide Chemistry Laboratory at Cancer Research U.K. (London, United Kingdom). They were purified by high-performance liquid chromatography and supplied as lyophilized powder, which was dissolved in 50 mM Tris-HCl (pH 7.2) and 150 mM NaCl (TBS) to a final concentration of 5 mg/ml. The sequence of HSP70-2 was QQGRVEILANDQGNRTTPSYVAFTDTER and of HSP70-5 was EIIANDQGNRITPSYVAFTPEGERLIGDAA.
Immunohistochemistry C20 antibody was used for immunohistochemistry. C20 (2 g/ml) was incubated at 4°C overnight with peptide 1 or 2 (5 mg/ml), a mix of both peptides (5 mg/ml), or TBS as control. Eight T-CL-1 slides were divided into two groups of four. One group was subjected to HMAR, whereas the other was not. All other details of the protocol were identical. Slides were deparaffinized using xylene and,
www.StemCells.com
Figure 1. A squamous carcinoma showing cytoplasmic staining with C20 (A) which is abolished after preabsorption with HSP70-5 (B). Abbreviation: HSP, heat shock protein.
depending on group, were microwaved for 22 minutes in 0.01 M citrate buffer (pH 6.0) for antigen retrieval. The slides were overlaid with the antibody/peptide mix and localized using an Envision peroxidase system (DAKO, Glostrup, Denmark, http://www.dako. com). All sections were counterstained in hematoxylin.
Cell Culture The OCIM-1 cell line was grown in Iscove’s Dulbecco’s minimal essential medium (MEM). The UT-7 cells were cultivated in the alpha modification of MEM supplemented with 2 ng/ml granulocyte macrophage colony-stimulating factor. The MCF-7 cell line was grown in Dulbecco’s MEM containing 4.5 g/l glucose. All media contained 10% fetal calf serum and penicillin/streptomycin (100 g/ml).
Western Blots Cells (2 ⫻ 106) from each line were lysed using radioimmunoprecipitation assay buffer with protease inhibitors (Roche Diagnostics GmbH, Mannheim, Germany, http://www.roche-diagnostics.com/) followed by sonication and centrifugation. Total protein was estimated using the Pierce BCA Protein assay kit (Pierce, Rockford, IL, http://www.piercenet.com/). Protein (25 g) from each line was added to 3⫻ Laemmli buffer, boiled before loading onto three individual Bis-Tris gels (4%–12%; Invitrogen, Paisley, United Kingdom, http://www.invitrogen.com), and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by blotting onto polyvinylidene fluoride membranes (Invitrogen). C20 antibody (1:1,000, 5 l) was preincubated overnight at 4°C with either HSP70-2, HSP70-5 (5 mg/ml, 5 ml), or TBS (5 ml) as control.
NSCLC and EPOR Antibody Specificity
720
Figure 2. A bronchioloalveolar carcinoma showing membranous staining with C20 (A), which is maintained after preabsorption with HSP70-5 (B). Abbreviation: HSP, heat shock protein.
The blots were incubated at room temperature for 1 hour with one of the three mixes, washed with TBS/Tween, and detected using horseradish peroxidase-conjugated anti-rabbit antibody at 1:2,000 (Dakocytomation, Glostrup, Denmark, http://www.dako.com/) for 1 hour at room temperature. The membranes were washed, and proteins were visualized using the ECL plus Western Blotting Detection System (Amersham, Little Chalfont, United Kingdom, http:// www.amersham.com/). To assess uniform protein loading, the membranes were stripped using methanol and washed and probed for pan-actin expression (Cell Signaling Technology. Inc., Danvers, MA, http://www.cellsignal.com/). The membranes were then washed and proteins visualized using SuperSignal West Dura (Pierce).
RESULTS The cornerstone of immunohistochemical evidence for EPOR in tumors is the C20 antibody, the use of which on formalin-fixed paraffin-embedded NSCLC tissue produces a typical, variably granular, positive cytoplasmic immunoreactivity. However, the observation by Elliott et al. [15] that such staining could be related to HSP challenges the interpretation that this positivity represents EPOR. To investigate the impact of preabsorbing C20 with HSP on immunohistochemical staining patterns in NSCLC, we examined 34 cases. Positive cytoplasmic and membranous immunoreactivity by the unabsorbed C20 polyclonal antibody was seen in FFPE
Figure 3. A squamous carcinoma showing cytoplasmic staining with C20 (A), which undergoes relative clearing with release of membranous staining after preabsorption with HSP70-5 (B). Abbreviation: HSP, heat shock protein.
tissue only when pretreated by HMAR (Fig. 1A). Nontumor components also showed staining. Positive immunoreactivity within plasma cell cytoplasm, sparing the perinuclear hof region, was observed, whereas bronchial respiratory epithelium showed cilial staining, in stark contrast to the lack of staining in adjacent areas of immature squamous metaplasia. After preincubation of C20 with HSP70-2, essentially all tumor cell cytoplasm was rendered negative with diminution of cilial staining. In 21 of the 34 cases examined, HSP70-5 abolished all staining, including cilial positivity (Fig. 1B). In four carcinomas, positive membranous immunoreactivity was retained after preincubation with HSP70-5 (Fig. 2). Positive cytoplasmic immunoreactivity seen in 30 cases (Fig. 3A) converted to a membranous pattern in 6 (Fig. 3B), whereas 3 cases initially showing cytoplasmic positivity preserved this pattern. These data are summarized in Table 2. Therefore, after preabsorption with heat shock peptides, positive cytoplasmic immunoreactivity disappears in the majority of tumors. However, a small number retain staining at the membrane or reveal such staining due to diminution of cytoplasmic positivity, creating a dichotomy of tumor type. Such staining is unlikely to be HSP, given the membranous location, but, on the basis of this work, cannot be assigned to genuine EPOR expression. In addition to sequences at 59 and 66 kDa, C20 also detects a higher molecular weight product that could not be distinguished from EPOR using immunohistochemistry alone.
Brown, Maxwell, Graham et al.
721
Table 2. Summary of the immunohistochemical staining patterns seen in 34 cases of non-small cell lung carcinoma stained with C20 preabsorbed with heat shock protein 70-5 Staining pattern
Cases
Cytoplasmic to negative
Cytoplasmic to membranous
Membranous (no change)
Cytoplasmic (no change)
Total
21
6
4
3
34
Given the differential staining pattern between native and absorbed C20 in immunohistochemistry, it was considered important to further validate the C20 antibody by Western blotting. All three cell lines (OCIM-1, UT-7, and MCF-7) produced bands at 59, 66, and 100 kDa. Whereas the 66- and 100-kDa bands showed similar intensity in all three of the cell lines, the band at 59 kDa was much stronger in OCIM-1 than in either UT-7 or MCF-7 cells. Preincubation with either peptide caused abolition of the 66-kDa band, but C20 still detected proteins at 59 and 100 kDa, albeit with a slight reduction in intensity in the case of HSP70-2 and a more obvious reduction with HSP70-5 (Fig. 4A, 4B). Utilizing the FLAG-EPOR vector, Elliott et al. [15] detected a single strong EPOR band at 59 kDa. In 769-P, MCF-7, and UT-7 cells, which are recognized to express low, medium, and high levels of EPOR, respectively, they noted bands of similar intensity at 59 kDa. Using MCF-7, UT-7, and OCIM-1 (which shows high expression of EPOR), we found a clear gradation of intensity, suggesting that C20 does indeed detect EPOR with the ability to discriminate between medium and high levels of expression. This conclusion is also in agreement with sequencing data presented by Elliott et al. [15], which indicate the presence of EPOR in the 59-kDa band. Both studies concur that a further band is present at 100 kDa, but we have been unable to detect the band found by Elliott et al. at 35 kDa [15].
CONCLUSION Preabsorption of C20 with synthetic HSP sequences causes a marked change in immunohistochemical staining pattern in formalin-fixed NSCLC tissue. The majority of tumors lose all staining, but a small number reveal membranous positivity, and this is mirrored in those tumors initially showing this pattern, retaining it. Western blots show three major components, of which the 66-kDa HSP fraction is lost after preabsorption. Crucially, a 59-kDa component is retained, together with a higher molecular weight fraction. Many questions remain regarding the significance of the EPOR in the context of in vivo tumors of different types and their response to rHuEPO. The use of rHuEPO to treat anemia in cancer patients is now common practice, and the European Organisation for Research and Treatment of Cancer has recently issued guidelines for the use of erythropoietic proteins in such patients [17]. Given the potential implications for this therapeutic approach, more specific tools for detection of EPOR in FFPE tissue are urgently required.
REFERENCES
Figure 4. Western blot showing presence of three bands at 59, 66 and 100 kDa (A) with abolition of the 66 kDa band on preabsorption with HSP70 –2 (B). Blot was exposed at 18 hours. Loading control is panactin. Abbreviations: HSP, heat shock protein; M, MCF-7; O, OCIM-1; U, UT-7.
ACKNOWLEDGMENTS This work was funded by the Research and Development Office of the Health and Personal Social Services in Northern Ireland, the Northern Ireland Leukemia Research Fund, and the Elimination of Leukemia Fund. W.M.B. was supported by Cancer Research U.K., E.A.D. by the European Social Fund, and A.Y. by Johnson and Johnson, Raritan, NJ. We thank the staff at the Protein and Peptide Research Laboratory at Cancer Research U.K. for synthesis of HSP70-2 and HSP70-5; the U.S. National Cancer Institute Tissue Array Research Program (TARP) for supplying the TMAs; and Dr. Lynn McCallum for useful advice on Western blotting protocols.
DISCLOSURES The authors indicate no potential conflicts of interest.
2 3
1
Walrafen P, Verdier F, Kadri Z et al. Both proteasomes and lysosomes degrade the activated erythropoietin receptor. Blood 2005;105: 600 – 608.
www.StemCells.com
4
Westenfelder C, Baranowski RL. Erythropoietin stimulates proliferation of human renal carcinoma cells. Kidney Int 2000;58:647– 657. Acs G, Acs P, Beckwith SM et al. Erythropoietin and erythropoietin receptor expression in human cancer. Cancer Res 2001;61:3561–3565. Acs G, Zhang PJ, Rebbeck TR et al. Immunohistochemical expression of erythropoietin and erythropoietin receptor in breast carcinoma. Cancer 2002;95:969 –981.
722
5
Selzer E, Wacheck V, Kodym R et al. Erythropoietin receptor expression in human melanoma cells. Melanoma Res 2000;10:421– 426. 6 Acs G, Zhang PJ, McGrath CM et al. Hypoxia-inducible erythropoietin signaling in squamous dysplasia and squamous cell carcinoma of the uterine cervix and its potential role in cervical carcinogenesis and tumor progression. Am J Pathol 2003;162:1789 –1806. 7 Eccles TG, Patel A, Verma A et al. Erythropoietin and the erythropoietin receptor are expressed by papillary thyroid carcinoma from children and adolescents. Expression of erythropoietin receptor might be a favourable prognostic indicator. Ann Clin Lab Sci 2003;33:411– 422. 8 Acs G, Xu X, Chu C et al. Prognostic significance of erythropoietin receptor expression in human endometrial carcinoma. Cancer 2004;100: 2376 –2386. 9 Arcasoy MO, Amin K, Chou SC et al. Erythropoietin and erythropoietin receptor expression in head and neck cancer: Relationship to tumor hypoxia. Clin Cancer Res 2005;11:20 –27. 10 Yasuda Y, Fujita Y, Matsuo T et al. Erythropoietin regulates tumour growth of human malignancies. Carcinogenesis 2003;24:1021–1029. 11 Arcasoy MO, Amin K, Karayal AF et al. Functional significance of erythropoietin receptor expression in breast cancer. Lab Invest 2002;82:911–918.
NSCLC and EPOR Antibody Specificity
12 LaMontagne KR, Butler J, Marshall DJ et al. Recombinant epoetins do not stimulate tumor growth in erythropoietin receptor-positive breast carcinoma models. Mol Cancer Ther 2006;5:347–355. 13 Henke M, Laszig R, Rube C et al. Erythropoietin to treat head and neck cancer patients with anaemia undergoing radiotherapy: Randomised, double-blind, placebo controlled trial. Lancet 2003;362: 1255–1260. 14 Leyland-Jones B, Semiglazov V, Pawlicki M et al. Maintaining normal hemoglobin levels with epoetin alfa in mainly nonanemic patients with metastatic breast cancer receiving first-line chemotherapy: A survival study. J Clin Oncol 2005;23:5960 –5972. 15 Elliott S, Busse L, Bass MB et al. Anti-EPO receptor antibodies do not predict EPO receptor expression. Blood 2006;107:1892–1895. 16 Dagnon K, Pacary E, Commo F et al. Expression of erythropoietin and erythropoietin receptor in non-small cell lung carcinomas. Clin Cancer Res 2005;11:993–999. 17 Bokemeyer C, Aapro MS, Courdi A et al. EORTC guidelines for the use of erythropoietic proteins in anaemic patients with cancer. Eur J Cancer 2004;40:2201–2216.
