Biguanides Exert Antitumoral Actions in Pituitary Tumor Cells Through AMPK-Dependent and -Independent Mechanisms

Abstract

Context

Pituitary neuroendocrine tumors (PitNETs) are a commonly underestimated pathology in terms of incidence and associated morbimortality. Currently, an appreciable subset of patients are resistant or poorly responsive to the main current medical treatments [i.e., synthetic somatostatin analogs (SSAs) and dopamine agonists]. Thus, development and optimization of novel and available medical therapies is necessary. Biguanides (metformin, buformin, and phenformin) are antidiabetic drugs that exert antitumoral actions in several tumor types, but their pharmacological effects on PitNETs are poorly known.

Objective

We aimed to explore the direct effects of biguanides on key functions (cell viability, hormone release, apoptosis, and signaling pathways) in primary cell cultures from human PitNETs and cell lines. Additionally, we evaluated the effect of combined metformin with SSAs on cell viability and hormone secretion.

Design

A total of 13 corticotropinomas, 13 somatotropinomas, 13 nonfunctioning PitNETs, 3 prolactinomas, and 2 tumoral pituitary cell lines (AtT-20 and GH3) were used to evaluate the direct effects of biguanides on cell viability, hormone release, apoptosis, and signaling pathways.

Results

Biguanides reduced cell viability in all PitNETs and cell lines (with phenformin being the most effective biguanide) and increased apoptosis in somatotropinomas. Moreover, buformin and phenformin, but not metformin, reduced hormone secretion in a cell type–specific manner. Combination metformin/SSA therapy did not increase SSA monotherapy effectiveness. Effects of biguanides on PitNETs could involve the modulation of AMP-activated protein kinase–dependent ([Ca²⁺]_i, PI3K/Akt) and independent (MAPK) mechanisms.

Conclusion

Altogether, our data unveil clear antitumoral effects of biguanides on PitNET cells, opening avenues to explore their potential as drugs to treat these pathologies.

Pituitary neuroendocrine tumors (PitNETs) are mostly benign neuroendocrine tumors that are a commonly underestimated pathology in terms of incidence and associated morbimortality. Specifically, PitNETs constitute ∼15% of all intracranial neoplasms and appear in ∼16.9% of the population (1–3). The genesis of PitNETs resides in an excessive and uncontrolled cell proliferation produced by the expansion of precursor cells. PitNETs are often accompanied by serious comorbidities related to mass effects and inappropriate secretion of pituitary hormones (1, 4–6). Transsphenoidal surgery is the first-line therapy in these patients, but ∼30% show tumor regrowth after surgery (7). The pharmacological arsenal currently available to treat PitNETs is limited mainly to synthetic somatostatin analogs (SSAs) and dopamine agonists (8), which exert their effects through the binding to their corresponding G-protein–coupled receptor families, both encoded by five genes (SSTR1–5 and DRD1–5, respectively) (9, 10). These drugs have a demonstrated efficacy in decreasing hormone hypersecretion and inducing tumor shrinkage and stabilization in functioning PitNETs (8, 11). However, some patients are (or become) unresponsive to these drugs (9, 12). For these reasons, the search for new therapies to control tumor growth or hormone secretion is crucial.

Metformin (MF), buformin (BF), and phenformin (PF) are antidiabetic drugs belonging to the biguanide family. Currently, only MF is used to treat type 2 diabetes mellitus (T2DM) (13). In addition to its well-known antihyperglycemic effect (14), it has been suggested that MF may reduce the risk of cancer and tumorigenesis in different types of neoplasms, such as brain, prostate, breast, and neuroendocrine tumors (15–20). Therefore, clinical trials of MF for nondiabetic patients with cancer have been performed. Unfortunately, some results are contradictory, especially in terms of reduction of Ki-67 expression, but other findings seem to be promising, including the reduction of the prostate-specific antigen in patients with prostate cancer (15). The precise molecular mechanisms underlying the antitumoral effects of MF are still controversial. Specifically, MF activates the AMP-activated protein kinase (AMPK), which has been associated with the modulation of cell proliferation, hormone secretion, and apoptosis in endocrine-related cancers and in PitNET cells (21, 22). Additionally, MF has been also suggested to exert some of its actions through AMPK-independent mechanisms (23).

Recent results from our group have revealed that biguanides reduce viability and secretory activity in two neuroendocrine tumor model cell lines (20). Moreover, we have also reported recently that both MF and PF exert notable direct effects in the modulation of hormonal secretion in normal pituitary cells from two primate species (24). In line with our observations, some recent reports have provided evidence that MF has a direct effect altering certain functional parameters in somatotrope GH3 and corticotrope AtT-20 pituitary cell lines (25–27). However, to date the pharmacological effects of different biguanides on human primary PitNET cell cultures are not fully elucidated. Therefore, we aimed to explore the direct effects of MF, BF, and PF on key functional parameters (cell viability, apoptosis, hormonal secretion and expression, and intracellular signaling pathways) in primary cell cultures from different human PitNETs subtypes, including adrenocorticotropin and GH secreting adenomas (ACTHomas, GHomas, respectively), nonfunctioning pituitary adenomas (NFPAs), and prolactinomas (PRLomas).

Materials and Methods

Drugs and reagents

All reagents and drugs used in this study were purchased from Sigma-Aldrich unless otherwise specified. Buformin was purchased from Santa Cruz Biotechnology (Heidelberg, Germany). Octreotide was obtained from GP-Pharm (Barcelona, Spain), and pasireotide was generously provided by Novartis (Barcelona, Spain). MF was used at 10 mM, BF and PF at 5 mM, and SSAs at 100 nM. All dosages were selected based on previous studies (20, 24, 28, 29) or based on in vitro dose-response experiments (Fig. 1).

Figure 1.

Dose-response experiment of cell viability in response to metformin (5 and 10 mM) in primary ACTHoma (n = 7), GHoma (n = 3), and NFPA (n = 7) cell cultures, measured by Alamar blue reduction. Data are expressed as percentage of vehicle-treated controls (set at 100%) within experiments. Values represent the mean ± SEM. Asterisks (*P < 0.05) indicate statistically significant differences.

Patients, samples, and primary cell cultures

This study was carried out within a project approved by our Hospital Research Ethics Committee and was conducted in accordance with ethical standards of the Helsinki Declaration of the World Medical Association. Written informed consent was obtained from each patient. Human pituitary samples were collected during transsphenoidal surgery from 42 patients (13 corticotropinomas, 13 somatotropinomas, 13 NFPAs, and 3 PRLomas). General characteristics of the patients are summarized in Table 1. In all cases, specimens were placed in sterile cold medium (S-MEM, Gibco, Madrid, Spain; supplemented with 0.1% BSA, 0.01% l-glutamine, 1% antibiotic-antimycotic solution, and 2.5% HEPES) and dispersed into single cells according to the methods previously described (29–32). The type of tumor was confirmed by two separate methods: examination by anatomopathologists and molecular screening by quantitative real-time PCR (RT-PCR), as previously described (29–32).

Table 1.

Demographic Data of Patients Included in the Study

Tumor Type	N	Sex (% of Women)	Age (Min–Max)
Corticotropinomas	13	92%	56 (18–79)
Somatotropinomas	13	54%	50 (29–64)
PRLomas	3	33%	25 (20–34)
NFPAs	13	62%	49 (24–75)

Cell lines and culturing

The two pituitary cell line models most widely used in cell biology research were used in the current study: the mouse corticotrope pituitary-derived cell line AtT-20/D16v-F2 (ATCC® CRL-1795™) and the rat somatotrope pituitary-derived cell line GH3 (ATCC® CCL-82.1™). Both were cultured and maintained in DMEM complemented with 10% fetal bovine serum, 100 U/mL penicillin/streptomycin, and 0.024 M HEPES and maintained at 37°C and 5% CO₂, under sterile conditions. Additionally, both cell lines were checked for mycoplasma contamination by PCR (33).

Analysis of cell viability

As previously reported (30, 32), Alamar blue reagent (Invitrogen, Madrid, Spain) was used to assess the effect of different biguanides alone or the combination of MF with SSAs (octreotide or pasireotide) every 24 hours until 72 hours on cell viability. We plated 10,000 cells per well (human pituitary cultures) or 6,000 cells per well (AtT-20 and GH3 cell cultures) in a 96-well plate. Treatments were daily refreshed after each measurement, and cell viability was evaluated with a FlexStation III system (Molecular Devices, Sunnyvale, CA).

Measurement of hormone release

To analyze the effect of different biguanides alone or the combination of MF with SSAs (octreotide or pasireotide) on pituitary hormone release from different primary PitNET cells or cell lines, we plated 150,000 to 200,000 cells per well onto 24-well plates in serum-containing media. Media were collected after 24 hours of incubation, and hormone secretion was measured with human [reference numbers: ACTH: EIA-3647; GH: EIA-3552; PRL: EIA-1291 (DRG, Mountainside, NJ)] and rat [reference number: GH: EZRMGH-45K (Merck Millipore, Darmstadt, Germany)] commercial ELISAs, according to the manufacturer’s instructions.

Analysis of apoptotic rate in somatotrope cells

Caspase-Glo 3/7 assay (Promega, Madrid, Spain) was used to analyze the effect of MF on apoptotic rate by measuring caspase 3/7 activity according to the manufacturer’s instructions. Thus, 25,000 cells per well were plated in 96-well white microplates and maintained for 24 hours at 37°C and 5% CO₂. Then, cells were treated with MF and vehicle and incubated for another 24 hours. After this period, 100 µL of Caspase-Glo 3/7 reagent was added to each well, and luminescence was measured at room temperature with the FlexStation III system for 3 hours.

RNA isolation, reverse transcription, and quantitative RT-PCR of human and primate transcripts

Details of RNA extraction, quantification, reverse transcription, and quantitative RT-PCR with a specific set of primers included in this study have been previously reported elsewhere by our group (31, 32). New primer sequences were used in the current study to amplify rHprt1 (sense, AGCTTGCTGGTGAAAAGGAC and antisense, TCCACTTTCGCTGATGACAC; accession number, NC_005120.4; product size, 153 pb), rPpia (sense, CGTCTGCTTCGAGCTGTTT and antisense, GGAACCCTTATAGCCAAATCCT; accession number, NC_005113.4; product size, 97 pb), rSst₁ (sense, TGCCCTTTCTGGTCACTTCC and antisense, AGCGGTCCACACTAAGCACA; accession number, NC_005105.4; product size, 135 pb), rSst₂ (sense, CCCATCCTGTACGCCTTCTT and antisense, GTCTCATTCAGCCGGGATTT; accession number, NC_005109.4; product size, 134 pb), rSst₅ (sense, TCATTGTGGTCAAGGTGAAGG and antisense, AAGAAATAGAGGCCGGCAGA; accession number, NC_005109.4; product size, 199 pb), rAmpk (sense, CTGTAAACACGGGAGGGTTG and antisense, ACGTTCTCTGGCTTCAGGTC; accession number, NC_000070.6; product size, 120 pb), mAmpk (sense, TCGGCTGGTTGTAGTGAATG and antisense, TCTCCTTCTGTTTGGCACCT; accession number, NC_000071.6; product size, 106 pb), and hAMPK (sense, AGATTGTATGCAGGCCCAGA and antisense, TGGTCATCATCAAATGGAAGG; accession number, NC_000005.10; product size, 92 pb). To control for variations in the amount of RNA used in the reverse transcription reaction and the efficiency of the reverse transcription reaction, the expression level (copy number) of each transcript was adjusted with a normalization factor calculated from β-actin, hypoxanthine-guanine phosphoribosyltransferase 1, and glyceraldehyde 3-phosphate dehydrogenase expression levels in PitNETs and calculated from Actb, Hprt, and Ppia (peptidylprolyl isomerase A) expression levels in cell lines.

Measurement of free cytosolic calcium ([Ca²⁺]_i) kinetics

Kinetics of free cytosolic calcium ([Ca²⁺]_i) were measured in response to different biguanides alone in human pituitary primary cell cultures, as previously described (30–32). Thus, 50,000 cells per coverslip were plated, and changes in [Ca²⁺]_i in single cells were measured with fura-2 acetoxymethyl ester (Molecular Probes, Eugene, OR).

Measurement of signaling pathways by Western blotting

In short, 500,000 cells per well were cultured in 12-well plates and incubated for 8 minutes with different biguanides alone and vehicle-treated controls. Proteins were extracted, separated by SDS-PAGE, and transferred to nitrocellulose membranes (Millipore), as previously reported (31). Then, blocked membranes were incubated with the primary antibodies [antiphosphoAMPKα (SC-33524) and anti-ERK1/2 (SC-154; Santa Cruz, CA) and anti-AMPKα (2532S), antiphosphoERK1/2 (4370S), anti-Akt (9272S), and antiphosphoAkt (4060S) from Cell Signaling (Danvers, MA)] and appropriate secondary antibody (anti-rabbit antibody from Cell Signaling). Proteins were developed with an enhanced chemiluminescence detection system (GE Healthcare, Madrid, Spain) with dyed molecular weight markers. A densitometric analysis of the bands was carried out with ImageJ software. Relative phosphorylation was estimated from normalization of p-AMPK, p-ERK1/2, or p-Akt against the total AMPK, ERK1/2, or Akt, respectively.

Statistical analysis

Statistical differences were evaluated by paired parametric t test or two-way ANOVA followed by Tukey test for multiple comparison (according to normality evaluated by Kolmogorov-Smirnov test). All data are expressed as mean ± SEM. As previously reported (30), to normalize values within each treatment and minimize intragroup variations in the different in vitro experiments (i.e., different age of the tissue donor or metabolic environment), the values obtained were compared with vehicle-treated controls (set at 100%). All experiments were performed in a minimum of three independent primary pituitary cultures from different patients (three or four replicates per treatment per experiment), unless otherwise indicated. P values ≤0.05 were considered statistically significant. A trend for significance was indicated when P values ranged between >0.05 and <0.1. All statistical analyses were performed in GraphPad Prism 6 (GraphPad Software, La Jolla, CA).

Results

Biguanides reduce cell viability in PitNET cells

In general, administration of biguanides produced a reduction of cell viability in all types of PitNET cell cultures and a decrease in cell proliferation in the pituitary cell lines tested (Fig. 2). In particular, MF (10 mM) significantly reduced cell viability after 72 hours of incubation in ACTHomas cells, and BF and PF (5 mM) also decreased cell viability in ACTHomas after 24, 48, or 72 hours of incubation, with PF being the most effective compound (Fig. 2A). Moreover, MF (the only type of biguanide currently used in the clinical practice) was also evaluated in combination with SSAs (octreotide or pasireotide) in primary ACTHoma cultures (Fig. 2B). Thus, MF alone significantly reduced cell viability by 49.7% after 72 hours of incubation. In contrast, although treatment with octreotide and pasireotide alone apparently reduced cell viability at 72 hours (35.1% and 18.4% reduction, respectively), this reduction did not reach statistical significance in any case (Fig. 2B). Coincubation of MF with both SSAs did not significantly alter the inhibitory actions of MF (Fig. 2B). In the same line, MF significantly reduced cell proliferation after 72 hours of incubation in AtT-20 cells (Fig. 2A). BF and PF (5 mM) also decreased cell proliferation in AtT-20 cells after 24, 48, and 72 hours of incubation, with PF being the most effective compound (Fig. 2A). In this point, it should be clarified that Fig. 2B shows only the ACTHoma samples that were treated with metformin, SSAs, and the combination of them (n = 3 experiments), whereas Fig. 2A shows the results of all the ACTHoma samples treated with metformin (n = 10 cell cultures), including the tumors that appear in Fig. 2B. Therefore, although the final results and conclusion are similar in both graphics (i.e., metformin treatment significantly decreased cell viability in primary cell cultures from ACTH-producing tumors), direct comparisons between Figs. 2A and 2B should be done with caution, because the number of tumors included in each figure differs.

Figure 2.

Measurement of cell viability (24 to 72 hours) and apoptosis in response to different biguanides alone or in response to the combination of metformin with somatostatin analogs (octreotide or pasireotide) in primary PitNET cell cultures and pituitary cell lines. (A) Effect of MF (10 mM), BF (5 mM), and PF (5 mM) on cell viability in primary ACTHoma cell cultures (MF: n = 10; BF: n = 6; PF: n = 5) and in the corticotropinoma AtT-20 cell line (MF: n = 9; BF: n = 6; PF: n = 7), measured by Alamar blue reduction. (B) Effect of MF alone or in combination with octreotide or pasireotide on cell viability in ACTHomas (n = 3). (C) Effect of MF, BF, and PF in primary GHoma cell cultures (MF: n = 9; BF and PF: n = 4) and in the somatotropinoma GH3 cell line (n = 4). (D) Effect of MF alone or in combination with octreotide or pasireotide on cell viability in GHomas (n = 5). (E) Effect of biguanides in primary NFPA cell cultures (MF: n = 11; BF: n = 7; PF: n = 8). (F) Effect of MF alone or in combination with octreotide or pasireotide on cell viability in NFPAs (n = 2). (G) Effect of biguanides in primary PRLoma cell cultures (n = 3). (H) Effect of MF on apoptosis (24-hour treatment) in primary GHoma cell cultures measured by Caspase-Glo 3/7 assay. Data are expressed as percentage of vehicle-treated controls (set at 100%) within experiments. Values represent the mean ± SEM. Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. In cases where fewer than three experiments were performed, no significance tests were performed.

In primary GH-secreting PitNET cell cultures, administration of MF did not alter cell viability at any of the incubation times tested; in contrast, BF, and especially PF, clearly decreased cell viability after 24 to 72 hours of incubation (Fig. 2C). In the GH3 cell line, all biguanides significantly decreased cell proliferation (Fig. 2C). Coadministration of MF with SSAs was also evaluated in primary GH-secreting PitNET cells (Fig. 2C). As previously observed, MF did not alter cell viability in GHoma cell cultures; however, octreotide and pasireotide alone decreased cell viability by 35.8% and 33.2% at 72 hours of incubation, respectively. The combination therapy of MF with octreotide or pasireotide did not alter the inhibitory effect of both SSAs (Fig. 2D).

In primary NFPA cell cultures, treatment with MF and BF significantly decreased cell viability after 48 to 72 hours of incubation, and PF decreased this parameter after 24 to 72 hours of incubation in a time-dependent manner (Fig. 2E). Interestingly, although cell viability was not reduced in response to octreotide and pasireotide in NFPA cell cultures, combination therapy of MF with both SSAs seemed to produce a higher decrease in cell viability as compared with the different treatments alone (Fig. 2F); however, this reduction did not reach statistical significance, probably because we could test this combination therapy only in two primary NFPA cell cultures. Finally, PF, but not MF and BF, significantly decreased cell viability in primary PRLoma cell cultures (Fig. 2G).

Metformin increases apoptotic rate in GHoma cells

Because of the limited number of cells obtained after dispersions, the effect of MF on apoptosis was evaluated only in primary GH-secreting adenomas. Specifically, MF (10 mM) significantly increased caspase 3/7 activity, a robust indicator of apoptosis, as compared with vehicle-treated controls after 24 hours of incubation (Fig. 2H).

Effect of biguanides on hormone secretion in PitNET cells

MF, BF, and PF did not alter ACTH secretion after 24 hours of incubation in ACTHoma cell cultures (Fig. 3A). In GHoma cell cultures, BF and PF, but not MF, reduced GH secretion (Fig. 3B). In line with the cell viability results previously observed, the three biguanides significantly reduced GH secretion in the GH3 cell line after 24 hours of incubation (Fig. 3E). Additionally, we had the opportunity to measure the effect of MF alone or in combination with SSAs (octreotide and pasireotide) in cell cultures from GHomas (Fig. 3C). The results showed that treatment with SSAs alone tended to decrease GH release and that the combination therapy of SSAs with MF did not modify the inhibitory action of both SSAs when tested alone (Fig. 3C). Finally, BF and PF, but not MF, treatment tended to decrease PRL secretion in cell cultures from PRLomas as compared with vehicle-treated controls after 24 hours of incubation (Fig. 3D).

Figure 3.

Hormone secretion in response to different biguanides alone or to metformin in combination with octreotide or pasireotide in human PitNET cell cultures and cell lines. (A) Effect of MF (10 mM), BF (5 mM), and PF (5 mM) on ACTH secretion in primary ACTHoma cell cultures (24 hours treatment; MF: n = 6; BF and PF: n = 3), determined by commercial ELISA kit. (B) Effect of MF, BF, and PF on GH secretion in primary GHoma cell cultures (MF: n = 8; BF: n = 4; PF: n= 3). (C) Effect of MF alone or in combination with octreotide or pasireotide on GH secretion in primary GHoma cell cultures (n = 4). (D) Effect of biguanides on PRL secretion in primary PRLoma cell cultures (n = 2). (E) Effect of MF on GH secretion in GH3 cell line (n = 5), determined by commercial ELISA kit. Data are expressed as percentage of vehicle-treated controls (set at 100%) within experiments. Values represent the mean ± SEM. Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. In cases where fewer than three experiments were performed, no significance tests were performed.

Effect of biguanides on mRNA expression of relevant genes in PitNETs

We evaluated the direct effect of treatment with different biguanides alone on mRNA levels of pathologically relevant genes in cell cultures of corticotropinomas (primary ACTHoma cells and AtT-20 cells) and somatotropinomas (primary GHoma cells and GH3 cells). MF and BF did not modify the mRNA expression levels of POMC (ACTH precursor) in ACTHoma or AtT-20 cell cultures (Fig. 4A); in contrast, PF significantly reduced Pomc expression levels in AtT-20, but not ACTHomas, cell cultures (Fig. 4A). Interestingly, PF but not MF or BF treatment tended to increase the expression levels of somatostatin receptors (SST₁, SST₂, and SST₅) in ACTHoma cell cultures (Fig. 4B, left panel). Similar results were observed in AtT-20 cell cultures wherein PF significantly increased Sst₂ expression levels (both isoforms identified in rodents) but not Sst₅ [Fig. 4B, right panel; Sst₁ was not expressed in AtT-20 cell line (data not shown)].

Figure 4.

Measurement of mRNA expression levels of key genes in response to different biguanides in human PitNET cell cultures and cell lines. (A, B, E) Effect of MF (10 mM), BF (5 mM), and PF (5 mM) on the expression levels of different genes in primary ACTHoma cell cultures (MF: n = 3; BF and PF: n = 2) and in the corticotropinoma AtT-20 cell line (n = 4). (C, D, F) Effect of MF, BF, and PF on expression levels in primary GHoma cell cultures (n = 4) and in the somatotropinoma GH3 cell line (n = 4). Expression levels were measured by quantitative RT-PCR and adjusted by normalization factor (NF). Data are expressed as percentage of vehicle-treated controls (set at 100%) within experiments. Values represent the mean ± SEM. Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) indicate statistically significant differences. In cases where fewer than three experiments were performed, no significance tests were performed.

Administration of PF, but not MF or BF, increased GH mRNA levels after 24 hours of incubation in GHoma and GH3 cell cultures (Fig. 4C). Interestingly, treatment with PF, but not MF and BF, increased SST₂ and SST₅ but not SST₁ expression levels in GHoma cell cultures (Fig. 4D, left panel). In GH3, Sst₂ expression seemed to be upregulated in response to the treatment with all biguanides, although this increase was statistically significant only in response to MF. Moreover, PF increased Sst₅ and reduced Sst₁ expression levels (Fig. 4D, right panel).

To further explore the mechanisms involved in biguanide actions, we measured the effects of MF, BF, and PF on the expression of AMPK, because this enzyme has been typically considered the central mediator of MF effects (34). Results revealed that administration of MF and BF did not modify AMPK expression in ACTHoma or GHoma cell cultures (i.e., primary ACTHoma and GHoma cell cultures and AtT-20 cells, but with the exception of GH3, wherein MF increases Ampk expression levels). By contrast, PF treatment significantly increased AMPK expression levels in all cellular models of ACTHoma or GHoma tested (Fig. 4E and 4F, respectively).

Effect of biguanides on intracellular signaling pathways in PitNET cells

To test the ability of biguanides to modulate intracellular signaling pathways in PitNETs, we first evaluated the dynamics of [Ca²⁺]_i concentration in single cells derived from ACTHomas, GHomas, NFPAs, and PRLomas in response to the treatment with MF, BF, and PF (Table 2). Specifically, all biguanides elicited a calcium response in >50% of ACTHomas analyzed, evoking a similar reduction in [Ca²⁺]_i that ranged between 23.34% and 27.11%. However, differences were found regarding the percentage of responsive cells. Thus, BF was the most effective compound because 46.80% of ACTH-secreting cells responded, compared with 30.25% or 9.8% of responsive cells to PF and MF, respectively. In contrast, treatment with biguanides did not elicit any [Ca²⁺]_i response in the GHomas analyzed. In NFPAs, MF reduced [Ca²⁺]_i by 30.62% in 50% of the samples, BF decreased it by 20.6% in 66.7% of NFPAs, and PF was the most effective compound, reducing [Ca²⁺]_i by 44.7% in 32.4% of responsive cells of the 66.7% NFPAs analyzed. Finally, in PRLomas, BF and PF did not alter [Ca²⁺]_i dynamics, and MF reduced [Ca²⁺]_i by 18.4% in one of the two tested PRLomas.

Table 2.

Results From Free Cytosolic Calcium Dynamic Assays in PitNET Cells in Response to Biguanides

Treatment	No. Samples	Cells Analyzed	% PRC	% PRM ± SEM	Time (s) ± SEM
ACTHomas
MF	3/5	236	9.80%	23.34 ± 1.77	49.38 ± 2.67
BF	3/4	154	46.80%	25.5 ± 1.64	61.55 ± 3.10
PF	2/3	119	30.25%	27.11 ± 2.14	65.98 ± 2.78
GHomas
MF	0/3	150	0	—	—
BF	0/3	79	0	—	—
PF	0/3	81	0	—	—
NFPAs
MF	2/4	144	16.00%	30.62 ± 2.62	54.7 ± 5.67
BF	2/3	106	18.90%	20.6 ± 3.2	54.7 ± 3.46
PF	2/3	105	32.40%	44.7 ± 4.92	62.2 ± 2.2
PRLomas
MF	1/2	60	6.70%	18.4 ± 0.92	61.25 ± 2.07
BF	0/2	49	0	—	—
PF	0/2	58	0	—	—

No. samples is the number of responsive samples of the total of samples analyzed. Cells analyzed indicates the total number of individual cells analyzed. % PRC is the percentage of responsive cells in responsive samples. Time: indicates the time of response to biguanide administration.

Furthermore, we could analyze the combination therapy with octreotide in two ACTHoma and two GHoma cell cultures available (Table 3). In ACTHoma cells, combined incubation with MF and octreotide did not significantly affect the [Ca²⁺]_i reduction elicited by MF or octreotide alone (28.9% vs 20% or 23.9%, respectively). In GH-secreting cells, MF did not elicit changes in calcium dynamics, as mentioned earlier, and octreotide slightly reduced calcium levels in 10% of the cells analyzed. Interestingly, the coadministration of MF and octreotide produced a higher reduction of [Ca²⁺]_i (37.6% vs 0% or 21.6%) and an increase of responsive cells compared with MF or octreotide administered alone (27.9% vs 0% or 10%).

Table 3.

Results From Free Cytosolic Calcium Dynamic Assays in PitNET Cells in Response to MF Alone or in Combination With Octreotide

Treatment	No. Samples	Cells Analyzed	% PRC	% PRM ± SEM	Time (s) ± SEM
ACTHomas
MF	1/2	58	12.1%	20 ± 1.77	51.43 ± 1.95
Octreotide	1/2	67	4.5%	23.9 ± 3.62	110 ± 0
MF + Octreotide	1/2	66	4.5%	28.9 ± 5.44	45 ± 4.08
GHomas
MF	0/2	70	0	—	—
Octreotide	2/2	70	10%	21.6 ± 0.8	89.4 ± 3.71
MF + Octreotide	1/2	68	27.9%	37.6 ± 3.89	85 ± 0

No. samples is the number of responsive samples of the total of samples analyzed. Cells analyzed indicates the total number of individual cells analyzed. % PRC is the percentage of responsive cells in responsive samples. Time is the time of response to biguanide administration.

To better understand the effects observed in response to biguanides, we explored several signaling pathways in GHomas (Fig. 5). Our results show that a short-term incubation with BF and PF numerically increased phosphorylation levels of AMPKα as compared with vehicle-treated controls, although this difference did not reach statistical significance. In contrast, MF did not alter the phosphorylation levels of AMPKα, which is in accordance with the results obtained at mRNA levels in GHomas. We also measured other signaling pathways intimately related to proliferation and survival in tumor pathologies, including PitNETs (35, 36). Thus, we observed an increase in phosphorylation levels of Akt and ERK1/2 in response to BF and PF, with this increase being significant only in ERK1/2 in response to BF. On the other hand, MF did not alter phosphorylation levels of Akt or ERK1/2 (Fig. 5).

Figure 5.

Measurement of phosphorylation levels of AMPK, Akt, and ERK1/2 in response to different biguanides in primary GHoma cell cultures. Representative Western blots and quantification of levels of p-AMPK/total AMPK (n = 3), p-Akt/total Akt (n = 2), and p-ERK1/2/total ERK1/2 (n = 3) in response to MF (10 mM), BF (5 mM), and PF (5 mM) in GHomas. Data are expressed as percentage of vehicle-treated controls (set at 100%) within experiments. Values represent the mean ± SEM. Asterisks (*P < 0.05) indicate statistically significant differences. In cases where fewer than three experiments were performed, no significance tests were performed.

Discussion

PitNETs are commonly considered benign tumors because of their frequent slow growth and moderate proliferative capacity (37). However, the incidence of PitNETs is increasing, also because of the enhanced diagnostic capacity of novel imaging techniques and improved diagnostic technologies, which have increased the sensitivity to detect pituitary neoplasms (38). Likewise, the severe morbimortality associated with PitNETs has been clearly established, reinforcing the need to develop novel therapeutic options, especially for invasive, recurrent, or functioning tumors. In this context, there is an emerging interest in biguanides, particularly in MF, because of their potential as antitumoral compounds, related to their beneficial effects in increasing insulin sensitivity and decreasing oxidative phosphorylation (39). Thus, we recently described that different biguanides exert direct actions in the pituitary cells of two nonhuman primate species, suggesting that the well-known metabolic effects of biguanides might be, at least in part, influenced by their actions at the pituitary level (24). However, the antitumoral actions of these compounds on different human PitNETs are poorly understood. Thus, we evaluated the direct antitumoral effects of different biguanides on primary cultures of functioning and nonfunctioning PitNETs and explored their possible underlying mechanisms.

The antiproliferative capacity of MF has been described previously in different endocrine cancer settings, including neuroendocrine tumor cells (20, 40–42), and in various animal models (42), and is currently being tested as adjuvant therapy in several clinical trials (43). However, MF does not alter epithelial proliferation in Barrett esophagus (44), suggesting that this drug might exert cell- and tissue-type-dependent antiproliferative effects. In our study, ACTHoma cells were more sensitive to MF than GHoma and NFPA cells. Moreover, PF was the most effective compound in reducing cell viability in all cases, which is in accordance with a recent report from our group in neuroendocrine tumors (20). Additionally, results reporting a marked reduction of cell viability in response to MF in AtT-20 (27) and GH3 cells (26) are consistent with our results in these cell lines. However, our present findings in GHomas are not totally in line with previous reports, where MF reduced cell viability in seven of eight somatotropinomas (26). These differences could be caused by disparities in the experimental design or in the basal characteristics of the GHomas analyzed. Nevertheless, although we did not observe significant changes in cell viability, there was a significant increase in apoptotic rate in response to MF in GHomas after 24 hours of incubation, a finding that is consistent with results reported in GH3 cells (26). These results might be seen contradictory because most reports seem to associate cell proliferation with apoptosis through the modulation of common or distinct proteins (45, 46). However, several studies have shown an imbalance between cell survival and apoptosis in pathological conditions such as breast or lung cancer (47–49). In line with the results observed in GHomas, a similar pattern of response to biguanides in terms of cell viability was found in PRLomas, which could be caused by the common developmental lineage of these cells (4). However, these results differ from the growth reduction observed in response to MF in the lactotrophic MMQ cell line (50), supporting again the idea that the effect of biguanides is highly cell type dependent.

As previously mentioned, SSAs represent important tools in the medical treatment of some PitNET types, especially in GH-secreting tumors, in relapsed or persistent disease (51); however, these treatments are in many cases ineffective (52, 53), and in vitro data suggest a mild effect on cell proliferation, especially when first-generation SSAs have been evaluated (54). For these reasons, the search for pharmacological alternatives to control tumor growth or hormone secretion has been crucial in recent years. In this context, treatment with SSAs combined with other pharmacological therapies, such as dopamine agonists or pegvisomant, is often used to control hormone-related or other symptoms in PitNETs (55, 56). Here, we tested the combined therapy of MF with SSAs but found that this combination did not alter the inhibitory effect of MF or SSAs alone in functioning PitNETs (GHomas and ACTHomas). Similar results have been reported using AMP mimetic compound 5-aminoimidazole-4-carboxamide ribonucleoside (AMPK activator) and somatostatin-14, where only one in eight GHomas showed an additive effect of both compounds (22). Of note, however, our study indicates that the combination therapy of MF and SSAs seems to have a stronger effect in reducing cell viability in NFPA cell cultures, a tumor type where SSAs have shown poor efficacy (57). Although more experiments are necessary to confirm and elucidate this interesting observation, we might speculate that this additive effect of MF and SSAs could be a potential therapeutic combination for patients with NFPAs because these tumors are the most resistant PitNETs to the therapeutic options currently available (29, 58).

With regard to hormone secretion, previous reports have described a time- and dose-dependent effect of MF on ACTH and GH secretion in AtT-20 and GH3 cell lines (26, 27). In our study, BF and PF but not MF treatment clearly reduced GH and PRL secretion without altering ACTH release. These results are in part comparable to a previous recent report from our group performed in neuroendocrine tumor cells, in which MF had no effect on hormone secretion in BON-1 or QGP-1 cell cultures, whereas PF decreased hormone secretion in BON-1 but not in QGP-1 cell cultures (20). However, we found that all biguanides significantly reduced GH secretion in the GH3 cell line, supporting again the notion of a cell type–specific effect of different biguanides on pituitary cells. In line with the results observed in cell viability, the combination therapy with SSAs did not increase the effectiveness of the SSA in monotherapy at the level of hormone secretion. Indeed, the fact that MF did not affect SSAs at the cell viability or hormone secretion level might suggest shared mechanisms of action between MF and SSAs in pituitary tumor cells.

In contrast with the results published recently by our group in normal pituitary glands from nonhuman primate species (24), MF and BF did not modify GH, POMC, or SSTs at mRNA expression levels. On the contrary, PF significantly increased GH expression levels, which could occur as a feedback loop mechanism to compensate for the striking reduction of GH secretion observed in GHomas and GH3 cell lines in response to this compound. PF also increased SST₂ and SST₅ expression levels in ACTH- and GH-secreting cells, which is in line with the similar increase observed in normal pituitary gland in Papio anubis (24). These data indicate that the actions of different biguanide types in PitNET cells are not only confined to the regulation of cell viability or hormonal secretions but also included the regulation of synthesis of key genes involved in regulating pituitary pathophysiology (i.e., pituitary hormones and SSTs). Although additional experimental work is necessary to confirm, elucidate, and achieve a definitive interpretation of the biological meaning of these results, our data adds compelling evidence of the direct effects biguanides exert on the expression profile of relevant pituitary genes, which might pave the way toward identification and validation of additional biomarkers or therapeutic targets in these pathologies.

Our report also provides information about the signaling pathways underlying the effects of biguanides in PitNET cells. Calcium is a relevant second messenger for pituitary cell physiology, which has been classically associated with pituitary hormone secretion (59). In this sense, our results showed an inhibitory action on [Ca²⁺]_i levels predominantly in ACTHomas and NFPAs and not in GHomas or PRLomas. Intriguingly, the calcium dynamics observed in our study do not seem to be associated to hormone secretion, because biguanides did not modify ACTH secretion from corticotropinoma cells. Therefore, our results suggest that calcium kinetics in response to biguanides could be more related to another functional parameter, such as cell viability (60, 61). Additionally, combination therapy of MF with SSAs did not alter the effect of MF as monotherapy in ACTHomas but seemed to increase the effect of octreotide in GHomas, although this increase was not enough to be associated with any functional endpoint. Furthermore, as expected, we found a regulation on the phosphorylation levels of AMPK [considered the central mediator of MF effects in different tissues and organs (34)] in response to BF and PF (which is in line with our results at mRNA expression levels). However, this increase in AMPK phosphorylation levels was not observed in response to MF, which is in contrast with the increase reported in GH3 and AtT-20 cell lines (25–27). In the same line, in this study we found an overall increase in phosphorylation levels of Akt and ERK1/2 in response to BF and PF but not MF. It should be mentioned that an increase in the phosphorylation levels of these two pathways has been related in several reports with a reduction of cell proliferation in response to short-period of incubations with SSAs or dopamine analogs in PitNETs (62, 63). Therefore, our findings demonstrate that some biguanides act through signaling pathways that have not been linked to AMPK, suggesting that some actions of these compounds could be exerted through AMPK-independent mechanisms in PitNET cells.

In conclusion, our study provides primary evidence that biguanides exert important antiproliferative and antisecretory effects in some PitNET cell types through the modulation of AMPK-dependent ([Ca²⁺]_i dynamics, PI3K-Akt pathway) and independent (ERK pathway) mechanisms. Moreover, combination of MF and SSA treatment did not exert additional antitumoral effects in functioning PitNETs, which suggests shared mechanisms of action. Of note, combined therapy with MF and SSAs might represent a potential therapeutic approach for patients with NFPAs. Taken together, our results unveil a clear overall antitumoral effect of different biguanides on PitNET cells and pave the way to consider these compounds as a potential option in the treatment of these severe pathologies.

Acknowledgments

Financial Support: This work has been funded by the following grants: Junta de Andalucía (CTS-1406 to R.M.L., BIO-0139 to J.P.C.), Ministerio de Ciencia, Innovación y Universidades (BFU2016-80360-R to J.P.C., FJCI-2016-30825 to A.I.C.), and Instituto de Salud Carlos III, co-funded by European Union (ERDF/ESF, “Investing in Your Future”: PI16/00264 to R.M.L., CP15/00156 to M.D.G., and CIBERobn). CIBER is an initiative of Instituto de Salud Carlos III.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations:

Abbreviations:
 
• AMPK

  AMP-activated protein kinase

 
• BF

  buformin

 
• [Ca²⁺]_i

  free cytosolic calcium

 
• MF

  metformin

 
• NFPA

  nonfunctioning pituitary adenoma

 
• PF

  phenformin

 
• PitNET

  pituitary neuroendocrine tumor

 
• PRL

  prolactin

 
• RT-PCR

  real-time PCR

 
• SSA

  synthetic somatostatin analog

 
• T2DM

  type 2 diabetes mellitus