Recent evidence confirms the presence of erythropoietin receptors on a variety of cancer cells. This has raised concerns about the use of erythropoiesis-stimulating agents in the treatment of cancer-related anemia. Having previously identified expression of functional erythropoietin receptors in a non-small cell lung carcinoma cell line, H838, which activated key signaling pathways in response to erythropoietin stimulation, we now demonstrate impaired downregulation of the erythropoietin receptor in these tumor cells. The erythropoietin receptor is not ubiquitinated following erythropoietin stimulation in this cancer cell line, and there is no turnover of the receptor in either unstimulated or stimulated cells. Compounding this blunted response is impaired SOCS3 induction downstream of erythropoietin stimulation and an extremely delayed SOCS1 response. If this finding in non-small cell lung carcinoma is a widespread phenomenon, then impaired erythropoietin receptor downregulation and degradation in tumor cells has clinical implications for those patients receiving erythropoiesis-stimulating agents for cancer-related anemia.
The erythropoietin receptor (EpoR) is a type I cytokine receptor belonging to a family of proteins including the prolactin, growth hormone, and interleukin 3 receptors . It is a 508-amino acid transmembrane protein encoded by a gene on chromosome 19p13.2. Erythropoietin (Epo) binding to the extracellular EpoR domain causes a conformational change in the receptor, resulting in Jak2-induced phosphorylation of key tyrosine residues in the cytoplasmic domain. This phosphorylation activates three main signaling pathways, namely Janus kinase 2/signal transducer and activator of transcription 5 (Jak2/STAT5), phosphatidylinositol 3-kinase (PI3K)/Akt, and Ras/mitogen-activated protein kinase kinase/extracellular signal-regulated kinase. EpoR is essential for the maturation of erythroid precursor cells and was previously thought to be expressed exclusively in the hematopoietic compartment. EpoR expression has since been detected in numerous nonerythroid cells, such as neurons , myoblasts , and endothelial cells . In the nervous system, Epo and EpoR are neuroprotective following stroke .
In addition, EpoR has also been detected on numerous tumors and cancer cell lines [6, –8]. Some groups report phosphorylation and activation of signaling pathways following Epo stimulation of these cell lines [7, 9, 10], and in some circumstances, this corresponded with increased cellular proliferation [6, 7, 10]. However, others have reported no downstream effects of Epo on cancer cell lines . We have previously reported activation of the three key signaling pathways (STAT5, Akt, and ERK) in the non-small cell lung carcinoma (NSCLC) cell line H838 , but we did not detect a proliferative advantage when the cells were treated with Epo.
In the erythroid system, downregulation mechanisms are activated rapidly following EpoR stimulation. These mechanisms include dephosphorylation of the receptor ; the generation and inhibitory actions of Cis, SOCS1, and SOCS3 ; and the ubiquitination and subsequent degradation of the receptor itself by both proteasomes and lysosomes . The phosphotyrosines of the receptor can be dephosphorylated by phosphatases such as SHP1  and PTP1B , thereby removing the docking sites of SH2-containing signaling proteins. SOCS family members are upregulated by Epo signaling and act to rapidly turn off activated receptors by inhibiting Jak2 catalytic activity or by blocking STAT binding . The EpoR is also ubiquitinated following stimulation, which targets the receptor to the proteasome for degradation, comprehensively downregulating the receptor . Although these mechanisms are well understood in erythroid cells, EpoR downregulation in cancer cells has not been reported. As erythropoiesis-stimulating agents (ESAs) are commonly used in the treatment of cancer-related anemia, it is important to investigate the potential role of EpoRs in cancer biology.
We examined the downregulation and turnover of the EpoR in the NSCLC cell line H838 and compared the data with that from UT-7 erythroid line, whose mechanism of degradation was previously reported . We provide evidence that EpoR downregulation in NSCLC is compromised with impaired SOCS3 and SOCS1 function, lack of EpoR ubiquitination, and no detectable EpoR degradation following Epo stimulation.
Materials and Methods
The NSCLC cell line H838 was maintained in RPMI 1640 + l-glutamine medium supplemented with 10% fetal calf serum (FCS), 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Invitrogen, Paisley, U.K., http://www.invitrogen.com). The UT-7 cell line was grown in minimal essential medium-α modification with l-glutamine and nucleosides supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 5 ng/ml granulocyte macrophage-colony stimulating factor.
Inhibition of Protein Synthesis Experiments
Cells were plated out at 25,000 cells per cm2 and allowed to adhere overnight. Cells were washed three times in sterile phosphate-buffered saline (PBS) and serum-starved for 24 hours. Cell treatment was with 10 μg/ml cycloheximide (CHX) (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) for the required length of time or with 10 μg/ml CHX for 15 minutes with the addition of 10 U/ml recombinant human Epo (rHuEpo) (Eprex; Ortho Biotech, Bridgewater, NJ, http://www.orthobiotech.com) for 0–120 minutes. The lysosomal inhibitor methylamine was used at a concentration of 10 mM.
Proteasome Inhibition Experiments
Cells were serum-starved for 24 hours and then treated with 1 μM proteasome inhibitor MG132 (Sigma-Aldrich) for 105 minutes, with 10 U/ml rHuEpo added 30, 60, or 90 minutes prior to the end of the incubation .
Real-Time Polymerase Chain Reaction of SOCS Genes
Cells were washed in PBS and serum-starved for 24 hours prior to stimulation with 10 U/ml rHuEpo for 0, 30, 60, or 120 minutes. Control cells were treated with PBS/1% bovine serum albumin for 60 minutes. Cells were harvested into TRIzol (Invitrogen), and RNA was extracted according to the manufacturer's protocol. RNA was reverse-transcribed into cDNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen), and 0.2 μl of cDNA was analyzed by real-time polymerase chain reaction (PCR). The primers and probes were obtained from Applied Biosystems (Foster City, CA, http://www.appliedbiosystems.com) (Hs00269575_s1, Hs00705164_s1, and Hs99999901_s1), and reactions were performed on the Applied Biosystems 7700 system using Taqman-based chemistry.
Whole cell lysates were obtained by harvesting cells into Laemmli buffer, followed by sonication and boiling for 5 minutes. Samples were separated on 10% gels, transferred to polyvinylidene difluoride membranes, and probed with anti-EpoR antibody M-20 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com). Detection was by horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibody (DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com) and ECL plus Western Blotting Detection System (Amersham Biosciences, Little Chalfont, U.K., http://www.amersham.com).
For immunoprecipitation, cells were harvested into Buffer A (20 mM Tris pH 8, 137 mM NaCl, 2.7 mM KCl, 1% Nonidet P40, 10% glycerol, and protease inhibitors), adjusted to a protein concentration of 1 mg/ml, and immunoprecipitated using anti-EpoR antibody (R&D Systems Inc., Abingdon, U.K., http://www.rndsystems.com) and Protein A beads (Upstate, Lake Placid, NY, http://www.upstate.com). Following washing, beads were resuspended in 2× SDS sample buffer (62.5 mM Tris pH 6.8, 40 mM dithiothreitol [DTT], 2% SDS, 10% glycerol, 0.1% bromphenol blue) and boiled. Western blotting was performed on the samples which were probed using monoclonal anti-ubiquitin antibody (Biomol, Exeter, U.K., http://www.biomol.com) and anti-mouse HRP-conjugated secondary (DakoCytomation).
Cells were passed 24 hours prior to the experiment to ensure that they were in log phase when harvested. Approximately 1 × 107 cells were lysed in 1 ml of proteasome buffer (50 mM MgCl2, 100 mM Tris-HCl [pH 7.8]) with 500 μM DTT and 5 mM ATP added fresh on each occasion the buffer was made. The lysates were incubated on ice for 10 minutes and sonicated at 6 μm for 15 seconds, and particulate matter was removed by centrifugation at 4°C for 10 minutes. The activity of each part of the proteasome was assayed using fluorogenic substrates specific for each activity. The chymotrypsin-like substrate was Succ-LLVY-AMC (Sigma-Aldrich), trypsin-like was Z-ARR-AMC (Calbiochem), and peptidylglutamyl peptide hydrolyzing (PGPH) was Z-LLE-AMC (Calbiochem). The assay was carried out in a black 96-well plate, with wells containing 5 mM EDTA (pH 8), 100 μM substrate, 50 μg of lysate, and proteasome buffer to a final volume of 200 μl. Fluorescence was detected every minute for 35 minutes, using excitation at 395 nm and emission at 460 nm, with the plate shaken during the procedure.
EpoR Turnover Is Impaired in H838 Cells
To assess the turnover and degradation of the EpoR in the NSCLC cell line, H838 cells were treated with CHX to inhibit protein synthesis and were probed for the presence of EpoR at various time points in both unstimulated and stimulated cells (Fig. 1). Reduced EpoR levels were detected after 180 minutes in unstimulated UT-7 cells, with a further decrease at 240 minutes. EpoR levels remained static in H838 cells over the 4 hours of the experiment. It would be expected that Epo stimulation of the cells would cause more rapid degradation of the receptor, and this was indeed the case for UT-7 cells (Fig. 1B), with EpoR levels falling to less than 50% of the initial level after 90 minutes. A similar decrease in UT-7 EpoR levels following CHX and Epo treatment has been reported previously . In contrast, no change was detected in H838 EpoR levels despite inhibition of protein synthesis and activation of the receptor by Epo.
Lysosomal Degradation of EpoR
It is known that lysosomes are involved in the degradation of EpoR. Walrafen et al. have demonstrated that addition of the lysosomal inhibitor methylamine results in a delay in EpoR degradation in UT-7 cells . When H838 cells were treated with CHX, followed by an incubation with 10 U/ml Epo and 10 mM methylamine, no degradation of the EpoR could be detected, even over a protracted time course (Fig. 2). Only following 24 hours of incubation with CHX and Epo could a reduction in EpoR levels be seen. The addition of methylamine did not significantly change the EpoR degradation pattern in H838 cells.
Ubiquitination of EpoR
If the EpoR downregulation is proteasome-dependent, cells would need to ubiquitinate EpoR to promote its degradation. The EpoR has been shown to be ubiquitinated in UT-7 cells prior to targeting to the proteasome , so it was relevant to see whether this process also occurred in H838 cells or whether some abnormal ubiquitination could account for the lack of degradation observed in H838 cells. The cells were treated with proteasome inhibitor and Epo over a time course, and whole cell lysates were probed for the presence of EpoR using the M-20 antibody (Fig. 3A). This antibody was the only EpoR antibody found to be specific by Elliott et al. , and we have confirmed that it detects an EpoR construct at the expected size by Western blot (data not shown). High molecular weight streaks that increased in intensity with time could clearly be seen in UT-7 cells in addition to the EpoR bands at approximately 62 kDa. This is consistent with previously reported data . This “streaking” pattern is characteristic of ubiquitination of a protein, as the different numbers of ubiquitin molecules attached to each protein result in structures with a continuum of molecular weights, and therefore no discrete bands are seen. In contrast to the UT-7 response, no streaks are seen in H838 cell lysates; the only bands present are those of the nonubiquitinated receptor and some nonspecific bands.
To confirm that the streaks were indeed caused by ubiquitination, EpoR was immunoprecipitated from the lysate using a monoclonal EpoR antibody, and then the samples were probed with an anti-ubiquitin antibody (Fig. 3B). A definite increase in ubiquitin staining can be seen with the UT-7 cells over time, indicating continuous ubiquitination following Epo stimulation. However, in H838 cells, although some staining can be seen in this blot because of the increase in EpoR concentration from the IP procedure, there is no alteration in staining intensity with prolonged incubation with Epo.
SOCS3 and SOCS1 Induction in Response to Epo
Ubiquitination and subsequent receptor degradation are not the only mechanisms for downregulation of the EpoR signal. Soon after activation, expression of members of the SOCS family of proteins are induced and function to abrogate the activated EpoR by inhibiting Jak2 catalytic activity and blocking STAT binding . SOCS3 is important in the downregulation of the EpoR following activation by Epo , and its levels increase as soon as 30 minutes after cytokine stimulation . To examine this in NSCLC, real-time quantitative PCR was used to determine levels of SOCS3 mRNA following Epo stimulation of the cells. The expected induction of SOCS3 transcription is seen in UT-7 cells (Fig. 4A), as levels of SOCS3 mRNA are increased 30, 60, and 120 minutes post-Epo stimulation (represented by a decrease in CT values). This increase in SOCS3 is statistically significant (p < .05) using the Mann-Whitney U test at all three time points compared with t0. In contrast, SOCS3 levels in H838 cells were unchanged following Epo stimulation of the cells (Fig. 4B), thus attenuating the normal SOCS3 downregulation of EpoR signaling.
To determine whether another SOCS family member was able to function as expected to terminate EpoR activation, SOCS1 levels were examined by real-time PCR, as there are reports of SOCS1 activation following EpoR stimulation . As with SOCS3, SOCS1 mRNA levels increased in UT-7 cells following 30, 60, or 120 minutes stimulation with Epo (Fig. 4C). These changes were statistically significant (p = .031 at 30 minutes, p = .003 at 120 minutes). H838 SOCS1 levels remained relatively constant, with induction only seen after 120 minutes of stimulation (p = .014 compared with t0) (Fig. 4D). This delayed response would only lead to diminution of the activated EpoR after 2 hours rather than the expected 15–30 minutes.
As the proteasome is involved in EpoR degradation, a lack of proteasomal activity could account for the impaired turnover of the EpoR in H838 cells. An assay was performed to detect the three enzymatic activities of the proteasomes in both H838 and UT-7 cells, as shown in Figure 5. All three activities were functional, with trypsin-like activity being highest in both cell types. There was a general trend of higher proteasomal activity in H838 cells compared with UT-7 cells.
When a receptor is stimulated, intracellular signaling pathways are activated that lead to transcription of downstream genes and a phenotypic response. Simultaneously, downregulation mechanisms are also switched on to return signaling proteins to their basal levels. Such downregulation is crucial to prevent cellular hyperstimulation. With the demonstration of functional EpoRs in cancer cell lines that show the potential for increased proliferation following Epo treatment, it is important to know whether these receptors are downregulated efficiently.
Cells harboring EpoR mutations are often hypersensitive to Epo, and this results in conditions such as congenital erythrocytosis [1, 21]. The EpoR in H838 cells has previously been shown to be wild-type , but hypersensitivity could result from impaired downregulation of the Epo-induced signal.
Proteasomal and lysosomal degradation of the EpoR in erythroid cells has recently been elucidated , and it would be expected that the same downregulation mechanisms would apply in EpoR-positive cancer cells. In addition, proteasomal degradation of the prolactin receptor has been demonstrated in breast cancer following prolactin stimulation , so it would be expected that EpoR would be downregulated and degraded in a similar fashion in cancer cells, as both receptors are members of the same superfamily. Using the UT-7 erythroleukemic cell line as an erythroid model, it can be demonstrated that the Epo receptor can be downregulated by ubiquitination and subsequent degradation of the receptor. Signaling pathways are also abrogated by the induction of SOCS3 and SOCS1.
In contrast, our detailed analysis of the downregulation of EpoR in the NSCLC cell line H838 showed no EpoR turnover in either resting or Epo-stimulated cells, suggesting impaired EpoR degradation. This was not attributable to an inactive proteasome, as all three enzymatic activities of the proteasome were capable of degrading substrates and accomplished this much faster than the UT-7 cell line, which did turn over the EpoR. The lack of turnover could be caused by aberrant ubiquitination of the EpoR in H838 cells, as no increase in ubiquitination of the receptor could be detected following Epo stimulation. However, other components of the proteasomal pathway in H838 cells may not be functioning and could be contributing to defective EpoR turnover. A decrease in EpoR levels could only be detected following 24 hours of incubation with CHX and Epo, suggesting that the EpoR in these cells is extremely resistant to degradation. This implies that the proteasomal and lysosomal degradation pathways have a limited capacity to turn over the EpoR in H838 cells. The additional downregulation provided by SOCS3 and SOCS1 was also impaired, with SOCS3 not induced at all and SOCS1 having an extremely delayed response to Epo stimulation.
A previous report indicated that SOCS3 was underexpressed in H838 cells. This was due to hypermethylation of the 5′ promoter region of the gene . SOCS3 expression was detected at a moderate level by real-time PCR in this study, but hypermethylation of the SOCS3 promoter could account for the lack of induction of this gene in H838 cells. Similarly, SOCS1 hypermethylation is found in different solid tumors and hematological malignancies, resulting in silencing of the gene , so it may be that hypermethylation is also causing altered induction of SOCS1 in the NSCLC cell line.
If the aberrant downregulation of EpoR activation seen in the NSCLC cells in vitro is a general feature of tumor cell behavior in vivo, then this has potentially serious clinical implications. The widespread use of ESAs for cancer-related anemia may, in some instances, augment tumor growth or survival. The improved control of anemia symptoms with ESAs may be offset by worse overall patient outcomes if tumor cells expressing dysfunctional EpoRs have an ESA growth or survival advantage.
The authors indicate no potential conflicts of interest.
E.A.D. is funded by the European Social Fund. We thank Dr. Lisa Crawford for advice on proteasome assays.