Anti-Epo receptor antibodies do not predict Epo receptor expression

Erythropoietin (Epo) binds to the Epo receptor (EpoR) on the surface of erythroid precursor cells, stimulating their proliferation, survival, and differentiation into mature red cells.¹  Recently, EpoR was reportedly detected in human tumors,^2-12  suggesting that Epo activates EpoR on tumor cells, thereby promoting growth. The anti-EpoR antibodies used to measure EpoR protein in these studies detected 66- to 78-kDa proteins; however, the mature form of human EpoR has a calculated molecular weight of between 56 and 57 kDa. This discrepancy in size suggested that these antibodies cross-reacted with non-EpoR proteins. We therefore examined the specificity of anti-EpoR antibodies used in those studies.

The MCF-7, 769-P, SH-SY5Y, HeLa, and Caki-2 cell lines from breast, kidney, brain, cervix, and kidney tumors, respectively, were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The megakaryoblastic leukemia cell line, UT-7/Epo,¹³  was a gift from Dr Norio Komatsu (Jichi Medical School, Minamikawachi, Japan). The murine leukemia cell line, 32D/Epo, required Epo for growth.^14,15 

We tested 4 commercially available rabbit polyclonal antipeptide anti-EpoR antibodies, including C-20 (sc-695; anti–human EpoR), M-20 (sc-697; anti–mouse EpoR), and H-194 (sc-5624; anti–human EpoR) from Santa Cruz Biotechnology (Santa Cruz, CA), and 07-311 (anti–mouse EpoR) from Upstate Biotech (Waltham, MA). M-20 and 07-311 may also recognize human EpoR. Anti-Flag (M-2) and anti–cyclophilin B were from Sigma-Aldrich (St Louis, MO) and AbCam (Cambridge, MA), respectively.

Tagged versions of EpoR contained an 8-amino acid N-terminal FLAG sequence attached to full-length EpoR (FLAG-EpoR) or to EpoR with the C-terminal 40 amino acids deleted (FLAG-EpoR Δ40). COS-7 cells were transiently transfected with pcDNA3.1 containing these constructs or with pcDNA3.1 only, using Lipofectamine 2000 (Invitrogen, Carlsbad, CA).

Cells were sonicated in 50 mM Tris-HCl (pH 7.2), 150 mM NaCl, 1% Triton X-100, 0.1% sodium deoxycholate, 1 × complete protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN), 100 μg/mL Pefabloc (Roche Applied Science), and 10 μg/mL pepstatin. Reduced and denatured proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to PVDF membranes, blocked with 5% milk, and incubated with primary antibody. The secondary antibody was either antimouse or antirabbit conjugated to horseradish peroxidase (Amersham Biosciences, Piscataway, NJ). ECL-Plus (Amersham Biosciences) was used for detection. Imaging was done using Hyperfilm (Amersham Biosciences) and an x-ray developer. For blocking experiments, C-20 was preincubated with peptides and the immunoblots were processed as described. Peptides were synthesized based on the sequence analysis of proteins detected by C-20 and included HSP70-2 (QQGRVEILANDQGNRTTPSYVAFTDTER) and HSP70-5 (EIIANDQGNRITPSYVAFTPEGERLIGDAA).

Sections of fixed cell pellets and whole mouse embryos were blocked with CAS BLOCK (Zymed Laboratories, San Francisco, CA) and incubated with primary antibody. Sections were washed, incubated with biotinylated goat antirabbit antibody (Vector Laboratories, Burlingame, CA), quenched with 3% hydrogen peroxide, and incubated with avidin-biotin complex conjugated to HRP (Vector Laboratories). Reaction sites were visualized with DAB (Dako, Carpinteria, CA) and counterstained with hematoxylin. Images were acquired using a Nikon Eclipse E600 microscope with a Nikon DXM1200F digital camera attached through a Nikon Plan Apo 2 ×/0.1 numeric aperture lens, using ACT-1 software for DXM1200 and DXM1200F (all from Nikon, Melville, NY). Composite images were assembled using Adobe Photoshop version 5.5 (Adobe Systems, San Jose, CA).

Proteins immunoprecipitated by C-20 from UT-7/Epo cells were separated by SDS-PAGE, excised from the gel, reduced and carboxyamidomethylated, and subjected to overnight tryptic in-gel digestion as described.¹⁶  Peptides were separated by high-performance liquid chromatography (HPLC) and analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) on a Thermo Finnigan LCQ DECA XP (San Jose, CA). Peptide sequences determined from MS/MS fragment ion spectra were searched against a nonredundant database to identify full-length proteins. Protein sequences that had at least a 5-amino acid identity to the C-20–immunizing peptide (C-20p) were analyzed further.

We performed a literature search on EpoR expression in tumors and identified 4 commercially available anti-EpoR antibodies, H-194, C-20, M-20, and 07-311. Each antibody detected recombinant full-length FLAG-EpoR (Figure 1), but the observed size of this protein was 59 kDa, significantly less than the size of EpoR suggested by investigators and the manufacturers of the antibodies (64-72 kDa).^6-8,10 

Each antibody detected additional proteins in tumor cells, ranging from 3 for C-20 to over 20 for H-194 (Figure 1). C-20 detected 3 dominant proteins of approximately 35, 66, and 100 kDa in all tumor cell lines and detected a protein similar in size to FLAG-EpoR in UT-7/Epo cells (Figure 1F). This is consistent with a report that C-20 detects a 66-kDa protein in tumor cell lines.^6,10  However, using a variety of methods, we showed that the 35-, 66-, and 100-kDa proteins were not EpoR nor derived from EpoR; knockdown of EpoR expression in SH-SY5Y cells did not prevent C-20 from detecting these proteins (Figure S1, available on the Blood website; see the Supplemental Figures link at the top of the online article), and C-20 immunoprecipitated the 59-kDa protein from UT-7/Epo cells, but not the 66-kDa protein. The immunocomplexed 59-kDa protein was detected by 07-311 and M-20 (Figure S2), suggesting it was EpoR.

Sequence analysis confirmed EpoR peptide sequences in the 59-kDa protein band but not in the 66-kDa protein band. One third of the peptides identified in the 66-kDa band were from HSPs including HSP70-2 and HSP70-5. These proteins contained regions of similarity to the immunizing peptide (C-20p), suggesting that C-20 cross-reacted to these regions. This was confirmed in peptide-blocking experiments, where the HSP70-2 (Figure 1G) and HSP70-5 (data not shown) peptides specifically inhibited C-20 binding to the 66-kDa protein. In contrast, C-20p blocked binding of C-20 to all proteins on the blot (Figure 1H). HSP70 was elevated in many human carcinomas^17,18  and increased with hypoxia.^19,20  Thus, reports that EpoR was overexpressed in tumor cells may be attributed to detection of HSP70 by C-20.

In Western analysis, M-20 detected a 59-kDa protein in all human tumor cell lines except 769-P (Figure 1D). M-20 did not detect the 59-kDa protein in fetal liver cells from EPOR^-/- knockout mice (Figure S3), and expression was reduced in SH-SY5Y cells that expressed EPOR shRNA (Figure S2). These results confirmed that M-20 can be used for detection of EpoR in Western blots and that the 59-kDa protein is EpoR.

However, using M-20, there was no difference in staining of tissue from EPOR-knockout (EPOR^-/-) and wild-type (EPOR^+/+) mice (Figure 2A,D). Thus, although M-20 was suitable for detection of EpoR in immunoblots, it was not suitable for immunohistochemistry. When M-20 was preincubated with a 5-fold mass excess of M-20–immunizing peptide (M-20p), staining was absent in both wild-type and knockout tissues (Figure 2B,E), suggesting the M-20 staining was due to cross-reactivity to non-EpoR proteins. As expected, no staining was observed with IgG control antibodies (Figure 2C,F).

C-20 has also been used extensively to detect EpoR in tumor cells using immunohistochemistry methods, and investigators have reported that EpoR is overexpressed in tumor cells.^2-6,10-12  Using C-20, we observed similar intensities of staining in UT-7/Epo, MCF-7, and 769-P cells that contained high, moderate, and undetectable levels of EpoR, respectively, indicating that C-20 has insufficient specificity to determine EpoR expression. Cell staining was eliminated by preincubating C-20 with C-20p (Figure 2H,K,N), and no staining was observed with IgG control antibody (Figure 2I,L,O). Therefore, these controls do not validate the use of C-20 for immunohistochemistry.

In conclusion, this study shows that only M-20 is suitable for detection of the 59-kDa EpoR by immunoblotting, and none of the antibodies are suitable for detection of EpoR by immunohistochemistry. Thus, studies alleging EpoR expression using these antibodies should be interpreted with caution.

We thank Kathy Christensen, James McCabe, Ivonne Archibeque, John Hui, and Shinichi Hara for technical assistance; Kathryn Boorer for assistance in writing this report; and Dr Harvey Lodish for useful discussions.