A series of 31 pituitary adenoma samples were obtained with pathological diagnosis and informed consent. All patients (12 men, 19 women; age range 16–63, mean age 43.50±11.00) years received surgery between April 2009 and April 2011 at the Department of Neurosurgery, the First Affiliated Hospital of Henan University of Science and Technology, Luoyang, China. Of these patients, 19 had prolactinomas, four had non-functional adenomas, two had multihormonal adenomas, four had gonadotropinomas and two had GH-secreting adenomas. For each sample, one half was immediately frozen at -80°C until protein extraction, the other half was fixed with 10% formalin and embedded in paraffin for histology and immunohistochemical analysis. The study was approved by the Research Ethics Committee of Henan University of Science and Technology. Samples were made anonymous according to ethical standards.

The MMQ rat prolactin-secreting tumor cell line was used in this study, because prolactinoma is the most prevalent hormone-secreting type of pituitary adenomas and there is no mature human pituitary adenoma cell line to date (20). The MMQ rat prolactin-secreting tumor cell line was from ATCC and maintained in F12 culture medium supplemented with 5% fetal bovine serum (FBS), 10% horse serum, penicillin (100 μg/ml) and streptomycin (100 μg/ml) in a humidified incubator (37°C, 5% carbon dioxide). Human umbilical vein endothelial cells (HUVECs) were from cell culture center of Wuhan University and cultured in M200 culture medium supplemented with 20% fetal bovine serum (FBS), L-gultamine (2 mM) and heparin (50 μg/ml). All animal experiments were conducted in accordance with NIH guidelines and with the approval of the local animal use committee. Six-week-old nude mice were inoculated subcutaneously on the hind flank with MMQ cells (1×10⁶). After two weeks, xenograft tumors formed. Mice were randomized into three groups with four mice in each group. For each group, mice were injected with PBS, rhEPO (2000 U/kg, s.c., Kirin), or rhEPO plus bevacizumab (a VEGF inhibitor, 10 mg/kg, i.p., Roche) twice a week for two weeks, respectively. Tumor size was assessed by caliper measurements twice a week. Mice were sacrificed four weeks after tumor cell inoculation and tumors were excised for further analysis.

MMQ cells were seeded on 96-well plates (2000 cells/well), and then cells were treated with various concentrations (0, 1, 5 U/ml) of rhEPO for 24, 72 and 120 h in CO₂ incubator. HUVECs were seeded on 96-well plates (5000 cells/well), after attachment, cells were left untreated or treated with rhEPO (5 U/ml), rhEPO (5 U/ml) plus JAK2 inhibitor AG490 (20 μM/l) for 48 h. After treatment duration, the CCK-8 assay reagent was added to each well of the plate and incubated for another 1 h. Absorbance was read at 450 nm using a 96-well plate reader.

For immunohistochemisty, tissue sections were deparaffinized and rehydrated. Antigen retrieval was accomplished with citrate buffer (pH 6.0). Endogenous peroxidase was blocked with 3% hydrogen peroxide. Subsequently, slides were blocked with 10% normal horse serum followed by the primary antibodies incubated overnight at 4°C. Mouse anti-EPOR (M20, Santa Cruz, CA) antibody, mouse anti-PCNA (Santa Cruz) and mouse anti-CD31 antibody (Pharmingen, CA) were used. Then, slides were washed and incubated with the biotinylated secondary antibody for 30 min, followed by ABC reagent (Vector Labs) and diaminobenzidine. Slides were counterstained with hematoxylin and dehydrated by sequential ethanol and xylene. Slides were mounted and analyzed. For microvascular density (MVD) determination, two areas of most intense neovascularization were chosen at low magnification ×100. Three random visual fields (x400) in each area of high vascularisation were recorded. The final microvessel density (MVD) was the mean value of six random visual fields of the two high vascularisation areas.

Samples (30 μg of proteins) were electrophoresed on a 10–12% SDS-PAGE gel under reducing conditions and electroblotted onto nitrocellulose membrane. The membrane was blocked in 5% non-fat milk and incubated with primary antibodies overnight at 4°C. Primary antibodies used were anti-JAK2, anti-phospho-JAK2, anti-STAT3, anti-phospho-STAT3 (Millipore, MA), anti-ERK2, anti-phospho-ERK2, anti-AKT1, anti-phospho-AKT1 (Santa Cruz, CA), anti-VEGF (Pharmingen, CA). After washing, blots were incubated in secondary antibody (1:5000) for 1 h. ECL substrate was used for signal detection and GAPDH for internal control. The intensity of bands was quantified by densitometric analysis. Results were expressed as the ratio of intensity to that of internal control.

Total RNA was isolated from cultured cells using TRIzol reagent (Invitrogen) according to the manufacturer’s protocol, complementary DNA was synthesized using reverse transcription reagent kit (Applied Biosystems). RT-PCR analysis was performed using the following primers: human JAK2 forward 5′-TTATGGACAACAGTCAAACAACAATTC-3′ and reverse 5′-CTTACTCTCGTCTCCACAAAA-3′. Human STAT3 forward 5′-CAAAACCCTCAAGAGCCAAGG-3′ and reverse 5′-TCACTCACAATGCTTCTCCGC-3′. Human VEGF forward 5′-CGAAGTGGTGAAGTTCATGGATG-3′ and reverse 5′-TTCTGTATCAGTCTTTCCTGGTGAG-3′. Human GAPDH forward 5′-GCTTTTAACTCTGGTAAAGTGG-3′ and reverse 5′-TCACGCCACAGTTTCCCGGAGG-3′. The PCR reactions were initiated with denaturation at 95°C for 5 min; followed by 30 amplification cycles at 95°C (30 sec), 58°C (30 sec) and 72°C (30 sec), then finally 72°C (10 min) and 4°C. The relative amount of gene was normalized against GAPDH mRNA.

Chicken eggs were kept in a 37°C, 60% humidity incubator for 9 days. The CAM was dropped and the window sealed with stretchy tape. Fibrous membranes diluted with rhEPO (5 U/ml), rhEPO plus AG490 (20 μM/l) or PBS were placed on the avascular area of the Chicken chorioallantoic membrane. After 3 days, the chicken chorioallantoic membrane was fixed in 4% paraformaldehyde in PBS, dissected and photographed using stereomicroscope equipped with a Digital camera for further angiogenesis analysis.

Statistical analysis was conducted with the aid of SPSS 16.0 software. Data are expressed as mean ± SEM. Data obtained from two groups were analyzed by Student’s t-test. P-values <0.05 were considered significant.

Since the EPO-EPOR signaling has never been investigated in human pituitary adenomas, we first assessed the EPOR expression in 31 human pituitary adenoma samples using immunohistochemistry and Western blotting. HepG2 human hepatoma cells and glioma tissue, known EPOR-positive cells and tissue, were used as positive control. In this study, we found no EPOR protein expression in pituitary adenoma samples (Fig. 1). Moreover, similar results were also found in MMQ pituitary adenoma cells (Fig. 1). These data demonstrate that, at least in our cases, pituitary adenomas are EPOR-negative tumors and MMQ cells are EPOR-negative cells.

To further analyze the effect of rhEPO on pituitary adenomas, the potential proliferative effect of rhEPO was evaluated on MMQ cells in vitro. MMQ cells were starved in F12 medium containing 1% FBS for 48 h, and then co-cultured with rhEPO (0, 1, 5 U/ml) for 24, 72 and 120 h. As expected, no proliferative response was observed in MMQ cells when rhEPO were used as stimulator (Fig. 2A), indicating that rhEPO has no proliferative effect on MMQ cells in vitro.

JAK2/STAT3, MAPK/ERK, and PI3K/AKT pathways are known to be the classic EPO-activated signaling pathways. Moreover, some of these pathways have been reported to be activated in some non-proliferative responding cancer cell lines. Therefore, we determined whether 48 h rhEPO (5 U/ml) stimulation can activate these classic pathways using Western blotting and found no increased tyrosine phosphorylation of JAK2, ERK2 and AKT in MMQ cells treated with rhEPO (Fig. 2B). These results indicate that rhEPO has no influence on these classic EPO-induced signaling pathways in MMQ cells in vitro.

As rhEPO demonstrated no direct effect on MMQ cells in vitro, we further examined whether rhEPO regulate tumor growth of pituitary adenomas in vivo using a nude mouse xenograft model of MMQ prolactin-secreting pituitary adenoma cells. Contrary to our in vitro results, we found that a two-week rhEPO administration (2000 U/kg, twice per week) significantly accelerated tumor growth (Fig. 3). Because rhEPO has been shown to promote proliferation of human endothelial cells and cerebral arteries express EPOR, in seeking the underlying mechanism of the in vivo rhEPO-induced tumor growth acceleration, we evaluated the microvessel density and tumor cell proliferation using immunohistochemisty staining of anti-CD31 and anti-PCNA antibodies and found significantly higher microvessel density (P<0.05, Fig. 4A–D) and increased tumor cell proliferation (P<0.05, Fig. 4E and F) in rhEPO treated xenograft tumors than control ones. Moreover, immunoreactivity for PCNA in EPO treated xenografts showed that tumor cells in peri-vessel areas displayed higher proliferation, which was indicated as increased PCNA staining (Fig. 4F). These results suggest a key role of angiogenesis in the rhEPO induced tumor growth acceleration in MMQ cell xenografts.

Since rhEPO promotes angiogenesis in MMQ cell xenografts in vivo without direct effect on MMQ cells in vitro, and VEGF has been reported to play a key role in angiogenesis, in order to characterize signaling pathways involved in rhEPO induced in vivo tumor growth acceleration, Western blot analysis was carried out to measure VEGF expression of rhEPO treated xenografts. Being classic EPO proangiogenic signaling pathway, phosphorylation of JAK2 was also evaluated (21). We found that significant enhanced VEGF expression and phosphorylation of JAK2 in rhEPO treated xenograft tumors (P<0.05, Fig. 4G and H). Based on these observation, we hypothesized that rhEPO induced VEGF signaling may play a key role in its angiogenic effect. As STAT3 has been reported to be required for EPO induced VEGF upregulation (22), we further explored this issue and found significant increased phosphorylation of STAT3 in rhEPO treated xenograft tumors (P<0.05, Fig. 4G and H). The data demonstrate that the EPO-JAK2-STAT3-VEGF signaling axis may be involved in rhEPO induced angiogenesis.

To further verify the vital role of angiogenesis in rhEPO induced tumor growth of pituitary adenoma cell xenografts, we administered rhEPO (2000 U/kg) plus bevacizumab (10 mg/kg), a VEGF inhibitor, to nude mice bearing xenograft tumors. After a two-week administration, the bevacizumab administration significantly attenuated EPO-dependent tumor growth and angiogenesis (P<0.05, Figs. 3 and 4). These observations further confirmed the key proangiogenic role of rhEPO in its in vivo tumor growth acceleration.

To further verify the involvement of EPO-JAK2-STAT3-VEGF signaling axis in rhEPO-induced angiogenesis, we cultured HUVECs with rhEPO (5 U/ml), rhEPO (5 U/ml) plus AG490 (20 μM/l) or low-serum medium for 48 h, and then CCK8 assay, Western blot analysis and RT-PCR were performed. We found that rhEPO significantly promoted the proliferation (P<0.05, Fig. 5A) and activated JAK2-STAT3-VEGF signaling in HUVECs (P<0.05, Fig. 5B–D). Moreover, AG490 significantly inhibited these EPO-induced effects in vitro (P<0.05, Fig. 5). These results further confirmed the vital role of EPO-JAK2-STAT3-VEGF signaling axis in rhEPO-induced angiogenesis.

To further demonstrate the proangiogenic role of rhEPO more intuitively, we investigated its effect on neovascularisation in the developing chorioallantoic membrane (CAM) of the chicken embryo. As a trigger for angiogenesis, fibrous membrane diluted with rhEPO (5 U/ml) or PBS were placed on the CAM of chicken embryos. After 72 h, vessel formation was significantly increased in EPO treated CAM compared to the control (P<0.05, Fig. 6). As shown in Fig. 6, in contrast, when rhEPO (5 U/ml) plus AG490 (20 μM/l) were further incubated on the CAM for 72 h, AG490 significantly attenuated the EPO induced vessel formation (P<0.05).