Introduction

Pituitary adenomas (PAs) are a diverse group of benign tumors arising from the anterior pituitary gland. They are the third most common intracranial neoplasms after meningiomas and gliomas, accounting for approximately 15 % of all intracranial tumors [1]. PAs are grossly characterized by their hormonal activity as functional or nonfunctional. Depending on the produced hormone, functional adenomas are further classified as GHoma, PRLoma, ACTHoma, and TSHoma. Although nonfunctional pituitary adenomas (NFPAs) often originate form gonadotropin-secreting cells, they show no hormonal excretion. Current World Health Organization (WHO) classification of PAs is based on immunostaining and transmission electron microscopy [2, 3]. Ki-67 protein expression in nuclei of proliferating cells is also being assessed routinely in excised PAs as an indicator of tumor growth [4, 5]. Clinically, PAs result in syndromes associated with an excessive, autonomous excretion of hormones, pituitary insufficiency, or symptoms related to their mass effect and infiltration of surrounding tissues. Due to the variety of manifestations caused by pituitary adenomas, they are an interdisciplinary relevant healthcare issue. However, most of PAs are asymptomatic or display nonspecific, insidious signs being identified as incidental findings. In a recent meta-analysis, an overall prevalence of pituitary adenomas reaching 16.7 % in general population was estimated (14.4 % in autopsy and 22.5 % in radiological studies) [6]. Thus, the need for identifying factors that cause the formation of adenomas and, in the next step, mark out the symptomatic cases from asymptomatic ones seems relevant.

Tumorigenic events in PA are only partly understood. Some pituitary adenomas are components of genetic familial syndromes caused by known mutations including McCune Albright syndrome and multiple endocrine neoplasia [7–9]. However, the majority of PAs are sporadic cases with a rather elusive initiating aberration and the aforementioned mutations or mutations in well-known tumor suppressor genes or oncogenes are very rarely found [10]. Thus, studies aiming to explain the pathophysiology of pituitary adenomas have taken up many directions. One of them is the search for hyperactivation of intracellular pathways. Among them, mammalian target of rapamycin (mTOR) pathway has been recently emerging as a new target. It regulates metabolism, growth, and proliferation in normal cells and has been shown to play a critical role in proliferation, survival, and drug resistance in many human cancers. mTOR, a serine-threonine kinase, plays a central role in this pathway as it integrates stimuli from growth factors, nutrients, and energy status [11]. Although mTOR cascade has often been named PI3K/Akt/mTOR, it is clear now that besides its canonical activator, Akt, there is a line of other kinases located upstream of mTOR, including the RSK kinase, Ras/MEK/ERK pathway, AMPK (engaged in energy sensing pathway) or hypoxia response pathway (hypoxia induced REDD1 gene). Being situated in the crossroad of these pathways, mTOR acts as a sensor of parameters that define the wellbeing of the cell, determining whether the cell is able to grow and proliferate [12]. Among these activators, Ras/MEK/ERK is of special interest in pituitary pathophysiology. Upregulation of its components at each level has been shown in all types of PAs comparing to normal pituitary tissue [10, 13]. Apart from that, higher levels of Akt activity has been also demonstrated in PAs, putting forward the mTOR pathway as a potential player in pituitary tumorigenesis. The two mentioned regulatory kinases for mTOR, Akt, and Erk, displayed the highest levels of activation in NFPAs [13–15]. Their activation in this type of PA was also proposed as screening markers for identifying patients with higher risk of recurrence [16]. The phosphorylation of mTOR kinase itself in PAs was assessed directly in one study. However, it failed to detect any differences between PA and normal pituitary tissue despite strong activation of Akt and Erk [13]. In this study, we aimed to assess the activity of mTOR and Erk in pituitary adenomas at the phosphorylation level using western blotting method in different types of PAs. mTOR and Ras/MEK/ERK pathway has been extensively studied in many tumors and proved to be hyperactive [17–19]. Phosphospecific antibodies against Erk allow for detection of active form of the kinase that has the ability to phosphorylate and activate mTOR. mTOR activity was demonstrated through detection of phosphorylation of S6 ribosomal protein (S6rp). S6rp is one of the main effectors of mTOR and is often used as the read out of its activity [20–22]. Activation of S6rp exclusively by mTOR was proved in primary cultures using specific kinase inhibitors. Our findings were correlated with clinical and histological data in each adenoma in an attempt to anticipate the effect of the activation of mTOR and Erk kinase on PA features.

Materials and Methods

Patients

In total, 53 patients who underwent elective transsphenoidal surgery (26 women and 27 men) were enrolled in this study. Patients’ age ranged between 28 and 83 (mean age, 54.6 ± 15.7). Participants gave written informed consent for the protocol, which was approved by the local ethical committee. Tissue obtained during surgery was immediately flash-frozen for further analysis. Collected samples included: 14 GHomas, 33 NFPAs, and five ACTHomas (clinical and morphopathological details of each case are given in Electronic Supplementary Material (ESM) 1). In several cases, concurrently with the frozen sampling, tissue was immediately processed in attempt to establish primary cultures, as described below. Control pituitary tissue was obtained from five cases of micro-ACTHomas where the tumor was not visualized clearly on magnetic resonance imaging (MRI). The suspected tumor was excised with the surrounding tissues. After the exploratory surgery, resected tissue was examined in the laboratory, where tissue of the non-adenomatous appearance was separated. This was confirmed in histopathological examination; therefore, in further studies, these tissue samples were used as the pituitary control.

Immunostaining and Histopathology

Histopathological diagnosis was based on conventional light microscopy and immunohistochemical staining for pituitary hormones in accordance with the WHO classification of tumors of the endocrine system [3]. Briefly, material was fixed in 10 % formalin, embedded in paraffin and routinely stained with hematoxylin and eosin. Immunohistochemical staining was performed on paraffin-embedded specimens according to the labeled EnVision Flex Visualization System (Dako, Denmark) with DAB (3,3′-diaminobenzidine) as chromogen, using antibodies against: prolactin (PRL, dilution 1:200), growth hormone (GH, dilution 1:500), ACTH (dilution 1:500), β-TSH (dilution 1:200), β-FSH (dilution 1:500), β-LH (dilution 1:500) (all from LAB VISION, Thermo Fisher Scientific, USA), the glycoprotein α-subunit (dilution 1:100, Novocastra, UK), Ki-67 (MIB1, dilution 1:100; Dako, Denmark) and phosphorylated S6 ribosomal protein (Ser235/236; dilution 1:400, Cell Signaling Technology, USA). MIB-1 labeling index was routinely established in all pituitary adenoma specimens. In GHomas, antibodies against somatostatin receptor proteins: 2A type (sstr2A; SS-800) and 5 type (sstr5; SS-890) were also used (Biotrend, Germany).

Electron Microscopy

Small pieces of each GHoma were fixed in glutaraldehyde, postfixed in osmium tetroxide, and embedded in Epon. Ultrathin sections counterstained with uranyl acetate and lead citrate were examined with a Philips CM120 BioTWIN electron microscope.

Protein Extraction and Western Blotting

Protein lysates were prepared from approximately 5 mm3 of tissue. Samples were pesteled on ice in 200 μl of chilled RIPA buffer supplemented with 0.2 mM phenylmethylsulfonyl fluoride, 0.2 mM sodium orthovanadate, phospatase inhibitor cocktail (PhosSTOP, Roche, Basel, Switzerland), and protease inhibitors (Complete, Roche, Basel, Switzerland). The grinded tissue was rested for 15 min on ice. Lysates were cleared by centrifugation at 12,000×g at 4 °C. The supernatants were assayed for protein concentration by Lowry method and stored at −80 °C until further analysis. For western blotting, 50 μg of the protein was separated on 15 % SDS-PAGE and transferred to PVDF membrane (Immobilon-P; Millipore, Watford, Herts, UK). The membrane was blocked then with 5 % non-fat dry milk in Tris-buffered saline/Tween 20 and incubated with primary antibodies overnight at 4 °C. A cocktail of primary antibodies against phospho-Akt (Ser473), phospho-Erk1/2 (Thr202/Tyr204), and phosphorylated S6 ribosomal protein (Ser235/236) was used (Cell Signaling Technology, USA). Primary antibody against eIF4E included in the cocktail was used for protein loading control according to manufacturer (Cell Signaling Technology, USA). Its stability, sufficient for this purpose, has been proven previously [23, 24]. After incubations with primary antibody cocktail, membranes were washed and incubated with secondary, HRP-labeled antibody for 45 min and, after washing, the phosphorylated proteins were detected using the western blotting luminol reagent (Santa Cruz, CA, USA) and visualized on X-ray films. The films were analyzed using Image J software (NIH, USA) [25]. The activation of each kinase was normalized to the expression of a housekeeping gene-elF4E in the same sample and was expressed as the ratio of signal from a given phospho-protein to the signal form elF4E.

Primary Cell Culture

Excised pituitary adenomas were processed within 3 h from the surgery. A small, 5 mm3 fragment of adenoma was saved for protein analysis by western blotting as described above. The remaining tissue was cut into small pieces, washed in PBS and centrifuged for 5 min at 1,500×g three times. Cells were dispersed enzymatically in 0.5 % collagenase and 0.05 %/0.02 % trypsin/EDTA in M-199 medium for 1 h at 37 °C being vigorous shaken every 10 min. After digestion FSC was added to the final concentration of 10 % and the cell suspension was centrifuged at 800×g for 20 min in room temperature. The cell pellet was dispersed then in M-199 medium supplemented with 10 % fetal calf serum (FCS) and penicillin/streptomycin. Matrigel (BD, Bedford, MA, USA) was mixed with serum free cell culture medium in 1:4 ratio, poured on the plates and incubated for polymerization for 30 min in 37 °C. Primary PA cells were plated on matrigel-coated 6-well plates, and cultured in 5 % CO2 atmosphere at 37 °C for 48 h before tests.

Drug Tests

To analyze the effect of specific kinases on mTOR activation, 20 nM rapamycin (mTOR inhibitor), 100 nM wortmanin (Akt inhibitor), and 10 μM U0126 (Erk inhibitor) were added to the cell cultures as described previously [26] (indicated are respective final concentrations of each compound). Pure solvent (DMSO; final v/v concentration 0.1 %) was added to the “untreated” control culture. To enhance test sensitivity, cells were incubated in serum-free medium for 2 h before inhibitors were used. Forty-five minutes after inhibitors were administrated, culture medium was supplemented with FCS (final concentration, 10 %). 10 min later cells were harvested and lyzed. The extraction of the protein and the WB was carried out as described above.

Tumor Invasiveness and Size Estimation

Tumor invasiveness was based on MRI imaging and intraoperative findings and expressed according to the Knosp’s grading [27]. Tumor diameter and its approximate volume were calculated using MRI scans. To simplify calculations, adenoma shape was considered ellipsoid. Using ellipsoid volume formula V = 4/3πabc, where a, b, and c are the three semi-axes, tumor volume was estimated.

Statistical Analysis

Statistical analysis was performed with Statistica software (Statsoft, Tulsa, USA) with the use of the Fisher exact probability test, Spearman’s rank correlation coefficient, U Mann–Whitney test, and Kruskal–Wallis test. Statistical significance was defined at p < 0.05.

Results

mTOR is Active in Pituitary Adenomas

The activation of mTOR was calculated as pS6rp/elF4E ratio in all tissue samples (see Table in ESM 1, which contains the detailed measurements). An adenoma was considered to have active mTOR when pS6rp/elF4E >0 (Table 1, Figs. 1 and 2).

Table 1 The frequency of mTOR active adenomas presented as percentage of samples with active mTOR in the given type of adenoma and, in parenthesis, as number of active samples/number of analyzed samples. The data are non-normally distributed
Full size table
Fig. 1
figure 1

mTOR and pErk1/2 in a sample of pituitary adenomas. The phosphorylation of S6 ribosomal protein (pS6rp) is used as the read out of mTOR activity. No pS6rp was detected in any of the NFPAs whilst GHoma-1 and GHoma-3 have a strong signal for it. pErk1/2 was detected in both types of adenomas and is not coupled with mTOR activity. elF4E served as loading control

Full size image
Fig. 2
figure 2

mTOR activity in pituitary adenomas. mTOR kinase activity was calculated as pS6rp/elF4E ratio. The control group consisted of five specimens, with a zero value for pS6rp/elF4E ratio in four of them. GHomas had the highest level of mTOR activity, that reached statistical significance in comaprison to NFPA (p = 0.04)

Full size image

GHoma samples had the highest frequency of mTOR activity comparing to other types of adenomas (GHoma vs NFPA p = 0.018, GHoma vs NFPA + ACTHoma p = 0.02 in Fisher exact probability test). Moreover, GHomas showed higher level of mTOR activity compared to NFPAs (p = 0.04 in U Mann–Whitney test). In control pituitary tissue, one sample out of five had active mTOR (pS6rp/elF4E = 0.085).

pErk Activity is High in all Adenoma and Control Pituitary Tissue Samples

Erk was activated in most pituitary samples, including control samples (see Table in the ESM 1, which contains the detailed measurements). The level of Erk activity, calculated as pErk/elF4E ratio, was the highest in control pituitary samples reaching statistical significance (p = 0.003 in Kruskal–Wallis test comparing all groups of adenomas and normal pituitary tissue; Table 2, Figs. 1 and 3).

Table 2 The frequency of pErk active adenomas presented as percentage of samples with active mTOR in the given type of adenoma and, in parenthesis, as number of active samples/number of analyzed samples. The data are non-normally distributed
Full size table
Fig. 3
figure 3

pErk1/2 activity in pituitary adenomas. Erk kinase activity was calculated as pErk1/2/elF4E ratio. Control pituitary tissue had the highest level of pErk1/2 (p = 0.003)

Full size image

Correlation Analysis of Kinases Activity and Morphopathological Features of Adenoma Samples

Correlation analysis was performed (Table 3). Additionally in NFPAs and GHomas, mTOR active and non-active adenomas were analyzed separately. The level of mTOR activity did not show any significant correlation with any of the following parameters: tumor volume, tumor largest dimension, Knosp’s grading, Ki-67 %, and pErk activity in any set of samples.

Table 3 Correlation analysis of kinase activity and tumor features (Spearman R). Correlations that reached statistical significance (p < 0.05) are italicized. Due to the small number of samples, ACTHoma group was included only in the general analysis
Full size table

In NFPAs, the activity of pErk showed a medium level of inverse correlation with Knosp’s grading that reached statistical significance (R Spearman = −0.31, p = 0.018). In mTOR non-active NFPAs, this correlation became even stronger(R Spearman = −0.48, p = 0.022) and was not present in mTOR-active samples.

In GHomas, pErk showed a strong level of correlation with somatostatin receptor subtype 2 A (SSTR2A) expression (R Spearman = 0.57, p = 0.04). Among GHomas, 6/14 were densely granulated, 3/14 were sparesly granulated in electron microscopy, and 3/14 were GH+/PRL+. There were no significant differences in between GHoma types.

mTOR Activation in Primary Cultures of Pituitary Adenomas can be Mediated by Mechanisms Other Than Akt and Erk Kinases

Primary cultures form mTOR active GHoma, NFPA, and ACTHoma were setup on matrigel. All derived primary cell cultures displayed identical pattern of activation of mTOR, Erk, and Akt. When selective inhibitors were used, all cell lines showed mTOR inhibition in response to rapamycin while inhibition of Erk did not affect mTOR pathway despite the complete inhibition of targeted kinase. Also, inhibition of PI3K/Akt in ACTHoma line, did not modulate mTOR activation. Two adenomas with primarily non-active mTOR were also cultured. Despite the lack of mTOR activity in the tumor, cell cultures exhibited high level of S6rp phosphorylation with a pattern of kinase inhibition similar to active adenomas (Fig. 4).

Fig. 4
figure 4

Western blot from the primary cultures (48 h) on matrigel derived from adenomas with mTOR active in vivo (left) and mTOR non-active in vivo (right). All cultures had active mTOR and Erk. Akt activity was detected in three of them. All cell lines showed mTOR inhibition in response to rapamycin while inhibition of Erk and Akt did not affect mTOR activity

Full size image

pS6rp Immunostaining

Of each of GHoma and NFPA, seven samples were subjected to immunostaining for pS6rp and assessed by an experienced pathologist. The level of pS6rp staining was clearly higher in GHoma group (Fig. 5 and ESM 2).

Fig. 5
figure 5

Immunostaining for pS6rp as a readout of mTOR activity was performed in seven GHomas and 7 NFPAs. Representative pictures are shown. The staining is more intense in GHomas (left) comparing to NFPAs (right). Microscopic photographs for all 14 samples are provided in ESM 2. Scale bar = 50μm

Full size image

Discussion

Since mTOR is a key regulator of cell growth, mTOR inhibitors are potential therapeutics for treatment of cancer. mTOR pathway plays a major role in numerous hematological malignancies but recently it is getting more attention in regard to the management of solid tumors. This interest is supported by clinical application of mTOR inhibitors—temsirolimus and everolimus—that have been approved for the treatment of advanced renal clear cell carcinoma and pancreatic neuroendocrine tumors [28, 29].

Previous studies concerning mTOR kinase activation in pituitary adenomas focused mainly on cell line models. A work by Zatelli and colleagues on NFPA primary cultures showed that everolimus (a rapamycin derivative) decreased the activity of mTOR and led to the reduction of cell viability affecting apoptosis and proliferation [30]. Another study on NFPAs showed relative insensitivity to rapamycin treatment that was overcome with the somatostatin analog—octreotide [31]. Similar studies were conducted on GHoma-derived cell lines. Rapamycin and everolimus had antiproliferative effect on two GHoma-related rat cell lines (GH3 and MtT/S) and primary pituitary cells isolated form dispersed GHomas, where eight out of nine (88 %) cultures showed reduction in cell viability upon drug administration [32]. However, these in vitro studies have been carried out in culturing conditions that may not be a reliable model for tumor in situ environment. In this study we showed that primary cell lines derived from adenomas that did not exhibit mTOR activation in fresh frozen samples, have high level of mTOR activity in vitro, in standard culturing conditions. This may suggest that data from primary cell cultures and cell lines may not be representative for the in situ adenoma environment.

Our study that bases on analysis of the adenoma samples obtained freshly from the patients demonstrates the increased activity of mTOR kinase in all adenomas; however, this activity was observed with the highest frequency in GHomas. Also, the strength of this activity was the highest in GHomas (p = 0, 04 GHoma vs. NFPA in U Mann–Whitney test). This result was confirmed in immunochemistry. To the best of our knowledge, the same approach was undertaken in one study conducted by Dworakowska and colleagues. However, it revealed no differences in the phosphorylation of mTOR kinase between different types of adenomas and in comparison to healthy pituitaries obtained from autopsy [13]. This discrepancy may be due to the different methods and samples used as controls in the studies. In our study the activity was measured as the level of S6rp phosphorylation—a protein that is robustly phosphorylated by mTOR kinase and is often used as the readout of its activation. Validity of this approach is further proved by complete abolition of S6rp phosphorylation using specific inhibitor of mTOR—rapamycin—in primary cells derived from mTOR active adenomas. However, despite finding that mTOR is hyperactive in these tumors, we did not observe any correlation between the activity of mTOR and morphopathological features in PAs, such as tumor volume, largest dimension, Ki-67 index, Knosp’s grading or somatostatin receptor (SSTR) expression. In the primary cell cultures, we assessed three mTOR active adenomas (GHoma, NFPA, ACTHoma) and two mTOR inactive adenomas (two NFPAs). As mentioned above, all primary cultures from adenomas showed strong S6rp phosphorylation that was abolished in response to very low doses of highly selective mTOR inhibitor. In the experiments utilizing selective inhibitors of two major upstream kinases, mTOR activity showed to be independent of Erk or Akt. The observation that these two chief activators of mTOR do not contribute to its activity in PA suggests involvement of other regulators of mTOR activity. Lack of Erk involvement in mTOR activity is further supported by the absence of correlation between Erk and S6rp phosphorylation in fresh frozen samples.

The activity of pErk in fresh frozen samples showed a surprising pattern. It was active in all types of adenomas, but very high activity was detected in the control pituitary tissue. Our correlation analysis also supports this finding where Erk showed an inverse correlation with Knosp’s grading (a marker for adenoma invasiveness). Our data are in many aspects novel and not in line with some of the previously published reports. The cause of discrepancies may be the tissue used as the pituitary control. So far most studies analyzing intracellular pathways in PAs used autopsy obtained specimen for this purpose [13–15]. However, cadaveric samples may underestimate the level of kinase activation due to the short half-life of phosphorylation [33]. In the preliminary experiments we found, that phosphoproteins of mTOR pathway are hard to detect in the pituitary autopsy samples (ESM 3). Therefore, since biopsy is not an option, it seems that fresh control tissue can be obtained only during pituitary surgery as was done in our study. Although the nature of such control tissue may be considered a limitation, since pituitary hormones in these individuals have abnormal serum levels, to our knowledge, there is no evidence for ACTH or glucocorticoids (GCs) clear effect on Erk in pituitary tissue. Furthermore, in studies performed on nonpituitary models GCs were proved to decrease Erk activity [34, 35].

Conclusion

The role of mTOR activation in PA pathogenesis still remains inconclusive. Despite some limitations of this study such as limited number of control samples, we present here, for the first time, the data indicating the high rate of mTOR activity in GHomas compared to NFPAs. Although, further studies and statistical analysis revealed lack of correlation between this activity and adenoma morphopathological features. However, as the subjects of our study were benign cases of pituitary adenomas, mTOR may still pose an interesting target for recurrent, invasive GHomas or malignant tumors. In this course, in a recent study, performed on GH3 adenoma cell line, mTOR inhibition-sensitized cell cultures to radiation suggesting possible use of such agents to overcome radioresistance [36]. Another research in this regard assessed the effect of Nelfinavir in locally invasive pituitary adenomas in vitro and in vivo [37]. This agent caused a statistically significant radiation sensitization alongside with the inhibition of S6rp phosphorylation. Hence, the role of mTOR pathway seems to be promising in radiation combined therapy of locally invasive adenomas. mTOR was also proposed as a new target for the treatment of pituitary carcinomas. However, in at least one study, a combined everolimus–octreotide therapy failed to improve the outcome of an aggressive temozolomide-resistant ACTH-secreting carcinoma [38]. Therefore, further researches elucidating the role of mTOR in pituitary neoplasms seem indispensible.