Normal Human Pituitary Gland and Pituitary Adenomas Express Cannabinoid Receptor Type 1 and Synthesize Endogenous Cannabinoids: First Evidence for a Direct Role of Cannabinoids on Hormone Modulation at the Human Pituitary LevelThe Journal of Clinical Endocrinology & Metabolism June 1, 2001
Little is known about the expression and function of cannabinoid receptor type 1 (CB1) in the human pituitary gland. The aim of this study was to investigate CB1 expression in human normal and tumoral pituitaries by in situ hybridization and immunohistochemistry using an antibody against CB1. CB1 was found in corticotrophs, mammotrophs, somatotrophs, and folliculostellate cells in the anterior lobe of normal pituitary. After examination of 42 pituitary adenomas, CB1 was detected in acromegaly-associated pituitary adenomas, Cushing’s adenomas, and prolactinomas, whereas faint or no expression was found in nonfunctioning pituitary adenomas. Experiments with cultured pituitary adenoma cells showed that the CB1 agonist WIN 55,212–2 inhibited GH secretion in most of acromegaly-associated pituitary adenomas tested and that the CB1 antagonist SR 141716A was generally able to reverse this effect. Moreover, WIN 55,212–2 was able to suppress GHRH-stimulated GH release, and this effect was not blocked by coincubation with SR 141716A, possibly indicating a non-CB1-mediated effect. In contrast, WIN 55,212–2 was ineffective on GH-releasing peptide-stimulated GH release. In four Cushing’s adenomas tested, WIN 55,212–2 was not able to modify basal ACTH secretion. However, simultaneous application of CRF and WIN 55,212–2 resulted in a synergistic effect on ACTH secretion, and this effect could be abolished by SR 141716A, demonstrating a CB1-mediated effect. In the single case of prolactinomas tested, WIN 55,212–2 was able to inhibit basal secretion of PRL. Finally, the presence of endocannabinoids (anandamide and 2-arachidonoylglycerol) was investigated in normal and tumoral pituitaries. All tumoral samples had higher contents of anandamide and 2-arachidonoylglycerol compared with the normal hypophysis. Moreover, endocannabinoid content in the different pituitary adenomas correlated with the presence of CB1, being elevated in the tumoral samples positive for CB1 and lower in the samples in which no or low levels of CB1 were found. The results of this study point to a direct role of cannabinoids in the regulation of human pituitary hormone secretion.
Δ9-TETRAHYDROCANNABINOL (Δ9-THC), the main psychoactive component of marihuana (Cannabis sativa), is known to play a role, at hypothalamic level, in feeding behavior, thermoregulation, and pituitary hormone modulation (1, 2, 3). Studies in this field have been dramatically accelerated by the discovery of specific cannabinoid receptors and the subsequent identification of endogenous ligands of these receptors, known as endocannabinoids. To date, two different cannabinoid receptors have been characterized: type 1 (CB1), which is abundantly present in the central nervous system, and type 2 (CB2), which is mainly expressed in the immune system (4). Several endogenous compounds were shown to interact with CB1 and/or CB2, but the two major endocannabinoids described to date are anandamide [N-arachidonoylethanolamide (AEA)] and 2-arachidonoyglycerol (2-AG). These molecules are released after enzymatic cleavage of membrane phospholipid precursors. After release, both endocannabinoids are able to activate the cannabinoid receptors and are subsequently inactivated by specific mechanisms (5).
In the past both exogenous and endogenous cannabinoids have been shown to inhibit PRL and LH release (6, 7, 8) and to increase ACTH release (9), but these modulatory properties have been mainly attributed to an action at hypothalamic nuclei, rather than to a direct action on the pituitary gland (10, 11, 12). Only recently, a few reports have highlighted the possibility of a direct interaction of exogenous and endogenous cannabinoids at the level of the pituitary by describing the presence of CB1 in rat pituitary glands (13) and by providing the first evidence for the ability of the pituitary gland to synthesize endocannabinoids (14, 15). One of the aims of the present study was to analyze the expression of CB1 in human pituitary samples and to assess the ability of the human gland to synthesize endocannabinoids. Moreover, we have extended our investigation to the presence of CB1 and endocannabinoids in a large series of pituitary tumors. Pituitary adenomas are slow-growing neoplasms, but most of them are characterized by a hormonal hypersecretion, which gives rise to clinically relevant pathological syndromes (16, 17). In view of this, we have also investigated the potential role of cannabinoids to control aberrant hormone secretion in a small series of pituitary adenomas.
Materials and Methods
This study was performed after approval of the ethics committee of the Max Planck Institute, and informed consent was received from each patient or their relatives. Our study comprised 7 normal human pituitaries (3 males and 4 females; taken 8–12 h after death from autopsy cases after sudden death without any evidence of endocrine disease) and 42 pituitary adenomas (21 males and 21 females). Tissue fragments of normal pituitaries and pituitary adenomas were shock-frozen on dry ice for in situ hybridization (ISH) and immunohistochemistry (IHC). Tumors were diagnosed by clinical, biochemical, radiological, and surgical findings and by IHC and were classified into acromegaly-associated pituitary tumors (ACRO; 13 cases), prolactinomas (PROL; 5 cases), corticotropinomas (CUSH; 4 cases), and clinically nonfunctioning adenomas (NFPA; 20 cases); the later was divided after immunopathological examination into gonadotropinomas (9 cases), null cell adenomas (10 cases), and silent corticotroph adenoma (1 case). All tumors were benign and graded according to a modified Hardy’s classification (18) (Table 1⇓). Although a general classification of the tumors into major groups was possible, each tumor was clinically and immunohistochemically different from the others. Due to this heterogeneity, each tumor was analyzed separately for its response to cannabinoids and the presence of endocannabinoids.
ISH was performed as previously described with minor modifications (19). In brief, 8-μm sections were thaw-mounted onto sterile poly-L-lysine-coated slides, fixed in 4% phosphate-buffered paraformaldehyde, and stored in 96% ethanol at 4 C until use. Three oligodeoxynucleotides (ODN; MWG Biotech, Ebersberg, Germany) complementary to coding parts of CB1 messenger ribonucleic acid (mRNA) were 3′-end labeled with[α -33P]deoxy-ATP (NEN Life Science Products, Boston, MA) by terminal transferase (Roche, Mannheim, Germany). The sequences are shown in Table 2⇓. After hybridization and washing, sections were exposed to a phosphorimager analyzer, dipped in Ilford K5 photoemulsion (Ilford, Dreieich, Germany), and developed after 28 days. For negative control, a 100-fold excess of nonlabeled ODN was added to the radioactive probe and applied to the adjacent section on the same slide.
Hybridization signals were quantified by measuring the optical density (OD) of the phosphorimager pictures by digital analysis (Tina 4.0, Raytest, Munich, Germany) and subtracting the OD of the negative control (in the presence of a 100-fold excess of nonlabeled ODNs) in the same slide. Values were arbitrarily divided into five categories:− , 0–5 OD/mm2; +, 5–10 OD/mm2; ++, 10–20 OD/mm2;+++ , 20–50 OD/mm2; and ++++, 50–100 OD/mm2.
Primary antibodies (Ab) and dilutions were: mouse monoclonal antihuman βFSH, 1:500; antihuman βLH, 1:500; antihuman βTSH, 1:800; antihuman PRL, 1:400; antihuman α-subunit, 1:500 (all from Immunotech, Karlsruhe, Germany); antihuman ACTH, 1:100 (DAKO Corp., Hamburg, Germany); antihuman GH, 1:800 (gift from Dr. C. J. Strasburger, Ludwig Maximilian University, Munich, Germany); and anti-S-100, 1:20 (Biogenesis, Poole, UK). Rabbit polyclonal anti-CB1 (1:800) was a gift from Dr. K. Mackie, University of Washington (Seattle, WA). Single IHC was performed as previously described (20). The specificity of the anti-CB1 Ab was examined by preabsorption with the respective antigen peptide (gift from Dr. K. Mackie). As negative controls, the primary Ab was omitted, and in the case of CB1, sections were also incubated with preimmune serum. Double IHC was performed by incubating the slides for 30 min in goat serum (diluted 1:10) in Tris-based buffer. The Vector Avidin/Biotin Blocking Kit (Vector Laboratories, Inc., Burlingame, CA) was used to block endogenous biotin or biotin-binding proteins present in the section, according to the manufacturer’s instructions. The anti-CB1 Ab was incubated together with the monoclonal Ab against each of the pituitary hormones overnight at 4 C. After washing in Tris-based buffer, sections were incubated for 30 min in a mixture of goat antirabbit biotinylated antibody (1:300; Vector Laboratories, Inc.) and antimouse IgG (1:100; Sigma, Deisenhofen, Germany) and in biotin-peroxidase complex (Vector Laboratories, Inc.) together with mouse alkaline phosphatase-anti-alkaline phosphatase (1:50; Sigma) two times for 30 min each time. Immunoreactivity was visualized using 1 mg/mL diaminobenzidine (Sigma) with 0.01% hydrogen peroxide for CB1 and the Vector Red reaction kit (Vector Laboratories, Inc.) for the pituitary hormones. Levamisole (10 mmol/L) was used to block the endogenous alkaline phosphatase activity. After washing in water, the sections were counterstained with toluidine blue, fixed in xylol, and coverslipped using Entellan (Merck, Darmstadt, Germany). Controls were performed by omitting one of the two, or both, primary Abs.
Analysis of endocannabinoid levels by gas chromatography-mass spectrometry
Frozen tissue was weighed and extracted with chloroform/methanol (2:1, vol/vol) three times. The extracting solvent contained 1 nmol each of d8-AEA and d8-2-AG (Cayman Chemicals, Ann Arbor, MI) as internal standards. The lyophilized lipid extracts were purified by a sequence of open bed silica chromatography and normal phase high performance liquid chromatography as described previously (14, 21). Normal phase high performance liquid chromatography fractions with the same elution time as 2-AG and AEA standards were derivatized and subjected to gas chromatography-electron impact mass spectrometry as described previously in detail (14, 21). Endogenous AEA and 2-AG were quantified by the isotope dilution procedure previously described (14, 21).
Unless stated otherwise, materials and reagents were obtained from Sigma, Life Technologies, Inc. (Eggenstein, Germany), NUNC (Wiesbaden, Germany), and Falcon (Heidelberg, Germany). To establish primary cell cultures, 10 pituitary adenoma tissues (5 ACRO, 4 CUSH, and 1 PROL) were dispersed as previously described (22, 23). After mechanical and enzymatic dispersions, the cells were washed by repetitive centrifugations and finally resuspended in DMEM (pH 7.3) supplemented with 10% FCS, 26.2 mmol/L NaHCO3, 10 mmol/L HEPES, 2 mmol/L glutamine, 10 mL/L nonessential amino acids, 10 mL/L MEM vitamins, 5 mg/L insulin, 20 mg/L sodium selenite, 5 mg/L transferrin, 30 pmol/L T3 (Henning, Germany), 2.5 mg/L amphotericin, and 105 U/L penicillin/streptomycin (Biochrom, Berlin, Germany). Cell viability was consistently more than 90%, as assessed by acridine orange/ethidium bromide staining. Cells were plated in 48-well plates (100,000 cells/well in 0.5 mL culture medium) and incubated in a 5% CO2 atmosphere at 37 C. After the cells had attached to the plates (48 h), the culture medium was replaced by stimulation medium (DMEM, pH 7.3, containing 26.2 mmol/L NaHCO3, 10 mmol/L HEPES, 2 mmol/L glutamine, and 1 g/L BSA). After a washout period of 24 h, fresh stimulation medium was added to the cells together with the drug treatment. Immunocytochemical investigation of each cultured adenoma was performed to compare the pattern of hormone expression to that observed in the adenoma tissue.
WIN 55,212–2, a CB1 agonist, and SR 141716A, a CB1 antagonist (RBI, Sigma), were dissolved in dimethylsulfoxide (Sigma) as a 10 mmol/L stock solution. WIN 55,212–2 and SR 141716A were used at a 1 μmol/L concentration. SR 141716A was added 45 min before the start of the experiment. GHRH, GH-releasing peptide (GHRP), CRF, and TRH (Bachem, Heidelberg, Germany) were used at a concentration of 10 nmol/L. The same amount of dimethylsulfoxide was also added to the controls. The final volume of stimulation medium was 500 μL/well in each case. The incubation varied from 1–48 h, and after this period, the supernatant was removed, and the hormone content was determined. Due to the limited amount of tumoral cells in culture, not all in vitro experiments could be performed with each tumor. Cell numbers were determined at the end of the stimulation experiments using the Cell Proliferation Reagent Kit WST-1 (Roche) according to the manufacturer’s instructions. Values obtained for hormone secretion were normalized to cell number.
Human ACTH was determined by RIA as previously described (24). Human PRL and GH levels were determined with RIA kits from DPC Biermann (Bad Nauheim, Germany) according to the manufacturer’s instructions.
Experiments were performed in triplicate wells. Results are expressed as the mean Â± SEM. Statistics were performed by ANOVA in combination with ScheffÃ©’s test.
CB1 expression in normal human pituitary
In all normal pituitaries examined, hybridization signals for CB1 mRNA were strong in the adenohypophysis (Fig. 1A⇓), whereas only trace amounts of transcripts were found in the neurohypophysis (data not shown). Detailed observation at higher magnification revealed that not all endocrine cells contained CB1 transcripts and that the hybridization signals were distributed in only a fraction of the endocrine cells. The three different ODNs revealed no variation in the hybridization signal in all pituitaries tested.
Figure 1. (see link below) Expression of CB1 mRNA and protein in normal anterior human pituitary. A, Histoautoradiograph of CB1 mRNA (ISH) in normal human pituitary cells. Hybridization signals (silver grains) for CB1 mRNA are present in the anterior lobe. The small inset shows the negative controls (incubation of an adjacent section with a 100-fold excess of unlabeled probe). B, Cytoplasmatic immunoreactivity for CB1 protein (brown) is detectable in many cells of normal human anterior pituitary. C, Signal specificity is demonstrated by using preabsorbed serum (Pre). D–F, Immunohistochemical costaining of CB1 protein and PRL, ACTH, or LH in normal anterior pituitary. CB1 protein (brown) is seen in ACTH-immunopositive (D), PRL-immunopositive (E), and GH-immunopositive cells (F; red). Open arrows, Cells expressing only the hormone; arrows, cells expressing only CB1; double arrows, cells expressing both CB1 and the hormone. Counterstaining was performed with toluidine blue.
These results were confirmed at the protein level using an Ab against CB1 (Fig. 1B⇑), after testing the specificity of the antibody by preabsorption with the respective antigen peptide (Fig. 1C⇑). Immunoreactivity was found in about 20–30% of cells in the anterior lobe. To investigate which anterior pituitary cell population shows CB1 immunoreactivity, double IHC using Abs against CB1 and anterior pituitary hormones was performed. CB1 was present in ACTH-producing (Fig. 1D⇑), PRL-producing (Fig. 1E⇑), and GH-producing (Fig. 1F⇑) cells, whereas no immunoreactivity was found in LH-, FSH-, and TSH-positive cells (data not shown). Interestingly, only a small percentage of PRL-positive cells expressed CB1 in the normal pituitaries screened (Fig. 1E⇑). CB1 staining was detected in the majority of GH-producing (Fig. 1F⇑) and ACTH-producing (Fig. 1D⇑) cells. Almost all of the corticotrophs located in the proximity of the posterior lobe displayed CB1 immunoreactivity. A fraction of folliculostellate cells, as recognized by positive S-100 staining, was also positive for CB1 (data not shown). The neural lobe was nearly devoid of CB1 immunoreactivity. The same distribution pattern of CB1 was observed in tissues derived from both genders.
CB1 expression in pituitary adenomas
A total of 42 pituitary tumors were analyzed for CB1 mRNA by ISH. The intensity of CB1 expression in each tumor, as determined after quantification of the signal derived from ISH, is shown in Table 1⇑. Similar findings were obtained in two other independent experiments. Hybridization signals for CB1 mRNA were most intense in ACRO and CUSH (Table 1⇑ and Fig. 2⇓, A and C), whereas PROL showed lower values (Fig. 2B⇓). In all NFPA the signals for CB1 mRNA were weak or absent (see one example in Fig. 2D⇓). The three different ODNs revealed no variation in hybridization signal in each tumor tested. In agreement with our findings at the mRNA level, CB1 protein staining was highly variable among the 42 tumors tested, being highest in ACRO (Fig. 2E⇓) and CUSH (Fig. 2G⇓), moderate in PROL (Fig. 2F⇓), and absent in NFPA (data not shown). Interestingly, the only case of NFPA showing a high amount of CB1 protein was immunohistologically identified as a silent corticotroph adenoma (Fig. 2H⇓), which is a neoplasm histologically indistinguishable from Cushing’s adenoma (25), but has lost the ability to secrete ACTH.
Figure 2. (see link below) Expression of CB1 mRNA and protein in different types of pituitary adenomas. A–D, Histoautoradiographs of CB1 mRNA in representative examples of ACRO (no. 3), PROL (no. 15), CUSH (no. 22), and NFPA (no. 27). CB1 mRNA expression is intense in ACRO (A), PROL (B), and CUSH (C), whereas it is nearly absent in NFPA (D). The small insets show negative controls after incubation of an adjacent section with a 100-fold excess of unlabeled probe. E–H, CB1 immunostaining in representative examples of ACRO (no. 4; E), PROL (no. 16; F), CUSH (no. 19; G), and a silent corticotroph adenoma (Silent; no. 23; H). Cytoplasmatic immunoreactivity for CB1 (brown) is detected in all of these types of tumors. The small insets show negative controls in the absence of primary antibody. Nuclei were counterstained with toluidine blue. The numbers of tumors refer to those in Table 1.
Effects of exogenous administration of cannabinoids on hormonal secretion in primary pituitary adenoma cell cultures
Ten pituitary adenomas (five ACRO, four CUSH, and one PROL) were able to provide enough material to establish primary cell cultures to analyze the putative involvement of CB1 in hormonal secretion.
As shown in Fig. 3⇓, all five ACRO were responsive to GHRH (10 nmol/L) and/or GHRP (10 nmol/L) by increasing GH secretion. In the same figure, it is demonstrated that CB1 agonist WIN 55,212–2, at a concentration of 1 μmol/L, was able to inhibit the basal release of GH in three of five ACRO after 4 h of incubation, whereas the secretion was unaffected at 1 and 48 h (data not shown). The specific CB1 antagonist SR 141716A was tested at a dose of 1 μmol/L in combination with WIN 55,212–2 (in two cultures) and alone (in four cultures). As shown in Fig. 3B⇓, SR 141716A was able to counteract the inhibitory effect of WIN 55,212–2 on GH secretion. Interestingly, SR 141716A alone was able to stimulate GH release in one primary culture where WIN 55,212–2 had no effect (Fig. 3A⇓). Moreover, in this tumor WIN 55,212–2 blocked this stimulatory effect of SR 141716A (Fig. 3A⇓). In two ACRO primary cultures, enough cells were available to test the effect of WIN 55,212–2 on GHRH- and/or GHRP-stimulated GH release. Interestingly, in one case WIN 55,212–2 was able to significantly abrogate GHRH-stimulated GH release (Fig. 3A⇓); in the other case a reduction of GH release was observed, although it was not statistically significant (Fig. 3E⇓). The inhibitory action of WIN 55,212–2 on GHRH-stimulated GH release was not blocked by coincubation with SR 141716A (Fig. 3A⇓), possibly indicating a CB1-independent effect. WIN 55,212–2 was unable to modify GHRP-stimulated GH release in both tumors tested (Fig. 3⇓, A and D).
Figure 3. (see link below) A–E, Effect of 4-h incubation of 1 μmol/L of the CB1 agonist WIN 55,212–2 (WIN) and 1 μmol/L of the CB1 antagonist SR 141716A (SR) on basal and GHRH- and/or GHRP-stimulated (10 nmol/L each) GH secretion by five individual ACRO cell cultures (no. 7, 10, 11, 5, and 13). Basal GH secretion was set at 100%, and results are expressed as a percentage of the control value. The mean basal control values Â± SEM for GH secretion for each adenoma were 61 ng/mLÂ·100,000 cells (no. 7), 180 ng/mLÂ·100,000 cells (no. 10), 31 ng/mLÂ·100,000 cells (no. 11), 146 ng/mLÂ·100,000 cells (no. 5), and 16 ng/mLÂ·100,000 cells (no. 13). *, P < 0.05; **, P < 0.01 (vs. control). a, P < 0.05; b, P < 0.01. ND, Not determined due to insufficient number of cells. The numbers of tumors refer to those in Table 1.
Four CUSH provided a sufficient number of cells to perform release experiments. All of the tumors were stimulated with CRF (10 nmol/L). Only two of the four responded by increasing ACTH secretion (Fig. 4⇓, A and B), whereas the other two did not show any stimulatory effect (data not shown) and were consequently excluded from further investigation. In these two tumors in cultures, WIN 55,212–2 was not able to modify basal ACTH secretion after 1, 4, and 48 h (Fig. 4⇓, A and B, and data not shown). Similarly, SR 141716A alone was ineffective in influencing basal ACTH secretion (Fig. 4⇓, A and B). In the same tumors, simultaneous application of CRF and WIN 55,212–2 resulted in a synergistic effect (Fig. 4⇓, A and B). This effect was abolished by adding SR 141716A, indicating that the effect of the exogenous cannabinoid was presumably CB1 mediated (Fig. 4⇓, A and B).
Figure 4. (see link below...) A and B, Effect of 4-h incubation of 1μ mol/L WIN 55,212–2 (WIN) and 1 μmol/L SR 141716A (SR) on basal and CRF-stimulated (10 nmol/L) ACTH secretion by two individual CUSH cell cultures (no. 19 and 21). The mean basal control values for ACTH secretion for each adenoma were 42 ng/mLÂ·100,000 cells (no. 19) and 122 ng/mLÂ·100,000 cells (no. 21) and was set at 100%. C, Effect of 4-h incubation of 1 μmol/L WIN 55,212–2 (WIN) on basal and TRH-stimulated (10 nmol/L) PRL secretion in one PROL cell culture (no. 15). The mean basal control value for PRL secretion was 23 ng/mLÂ·100,000 cells. *, P < 0.05; **; P < 0.01 (vs. control). a, P < 0.05. The numbers of tumors refer to those in Table 1.
Due to the low number of cells, the single case of PROL was tested only in two conditions: the cells were normally responsive to TRH-induced PRL secretion, and WIN 55,212–2 was able to significantly decrease basal PRL release at 4 h (Fig. 4C⇑). At 1 and 48 h, WIN 55,212–2 had no effect on PRL secretion (data not shown).
Endocannabinoids in normal and tumoral pituitary samples
Before starting the examination of endocannabinoids, all tissues were checked for preservation of nuclear and cytosolic structures by performing IHC to detect the integrity of nuclear transcription factors and membrane surface receptors known to be present in pituitary tissues (data not shown). The content of endocannabinoids was analyzed in two normal pituitaries (both derived from male subjects). AEA levels were 14.6 and 28.1 pmol/g tissue, whereas the levels of 2-AG were 0.38 and 1.32 nmol/g tissue. The limitations set by the small size of tumoral samples prevented us from studying a high number of pituitary adenomas; thus, our investigation was limited to a representative sample for each subclass of tumors (one ACRO, one PROL, one CUSH, and one NFPA). The specimens were all derived from male patients and were devoid of any contamination by normal pituitary, as demonstrated by IHC performed before analysis of the endogenous cannabinoid content. Interestingly, all tumor samples had higher contents of AEA compared with normal pituitaries, and the values differed among the neoplastic samples. PROL showed the highest level of AEA (223.5 pmol/g tissue), followed by ACRO and CUSH, which contained 76.4 and 67.4 pmol/g tissue AEA, respectively. NFPA showed the lowest level of AEA (39.2 pmol/g tissue; Table 3⇓). The 2-AG content was also increased in the tumors compared with the normal glands. The tumors showing the highest amount of 2-AG were ACRO (39.1 nmol/g), CUSH (9.33 nmol/g), and PROL (6.15 nmol/g), whereas the NFPA showed a value closer to that of normal pituitary (1.66 nmol/g; Table 3⇓). Together, these results indicate that the content of endocannabinoids in pituitary tumors is higher than that in normal human pituitary.
Natural or synthetic cannabinoids have been shown to affect hormonal pituitary release in several in vivo and in vitro rodent models (6, 26, 27). Only recently has a direct action of cannabinoids on the pituitary gland been proposed, because CB1 and endocannabinoids were identified in rat pituitary gland (13, 14). The first aim of this investigation was to document the presence and cellular localization of CB1 in the normal human pituitary gland. CB1 was predominantly expressed in the adenohypophysis and was localized in GH-, PRL-, and ACTH-producing cells and in folliculostellate cells. These data are partially in contrast with previous findings in the rat pituitary, in which CB1 was detected in LH- and PRL-secreting cells (13). Differences in the experimental design and use of different Abs may account for the discrepancy in the results between the study in rat hypophysis and our present investigation in the human pituitary gland. However, we cannot exclude that these discrepancies might be due to species differences in the expression pattern of CB1.
The cellular localization of CB1 in pituitary adenomas was similar to that detected in normal human pituitary glands. CB1 was found in somatotroph or mammosomatotroph, corticotroph, and mammotroph cells of ACRO, CUSH, and PROL, respectively. No or very low CB1 expression was found in NFPA.
To attribute a functional significance to CB1, primary tumor cell cultures were stimulated with cannabinoids in the presence and absence of physiological stimulants. The cannabinoid agonist WIN 55,212–2 inhibited GH secretion in most of the ACRO tested, and this effect was generally reversed by the specific CB1 antagonist SR 141716A, suggesting that cannabinoids are able to directly influence basal GH secretion through CB1 activation in three of five ACRO tested. Intracerebroventricular injection of Δ9-THC in rats was able to inhibit GH via hypothalamic activation, butΔ 9-THC had no effect on GH secretion in rat primary pituitary cell culture, excluding a direct pituitary action of cannabinoids (11). By contrast, our findings show that at least in a subgroup of ACRO, cannabinoids are able to directly affect GH secretion. The reason for the heterogeneity in the GH response after cannabinoid stimulation among tumors sharing the same clinical characteristics is still unclear and needs further investigation. Interestingly, WIN 55,212–2 was able to suppress the stimulatory effect on GH release produced by GHRH, but not that caused by GHRP. The GH secretagogue properties of these two hypothalamic factors have been attributed to the activation of different intracellular pathways, i.e. cAMP for GHRH and inositol triphosphate for GHRP (28). Therefore, our findings seem to be in agreement with the known concept of cross-talk between cannabinoids and cAMP-activated pathways (4). However, the cannabinoid inhibitory effect of adenylate cyclase dependent-GH stimulation was not blocked by coincubation with SR 141716A. This may be explained by the recent hypothesis that pituitary cells may contain another type of cannabinoid receptor (29). Indeed, evidence for non-CB1/non-CB2 brain WIN 55,212–2 receptors coupled to G proteins and adenylate cyclase inhibition has been recently reported in rodents (30, 31), although their presence in the pituitary has not been investigated.
In all CUSH tested, WIN 55,212–2 alone was not able to influence basal ACTH secretion, but together with CRF it had an additive effect on ACTH release that was specifically blocked by SR 141716A, therefore indicating a CB1-mediated effect. CRF is known to exert its effects through the cAMP pathway; thus, these data appear to be in contrast to the well established cAMP inhibitory role attributed to CB1 (32). On the other hand, a series of recent reports suggested that under certain conditions activation of CB1 cannabinoid receptors can also involve stimulatory Gs proteins, thus resulting in cAMP accumulation (33, 34, 35). At any rate, natural and synthetic cannabinoids were shown to increase ACTH and corticosterone secretion in rodents in vivo (10, 36, 37). Activation of the hypothalamus-pituitary-adrenal axis has been proposed to be a consequence of cannabinoid modulation at the hypothalamic level via an activation of CRF-producing neurons (10). Our data showing a positive modulation of cannabinoids on CRF-induced ACTH secretion are of interest because they point to a direct involvement of the pituitary gland in the stress-associated responses to cannabinoid administration (3).
Cannabinoids can modulate PRL secretion (8, 11, 27), but it is still controversial whether this is a direct pituitary action or an indirect activation of central neurotransmitters. Studies in PROL cell cultures were limited by the small size of the samples. Nevertheless, in the single case we studied, WIN 55,212–2 was able to inhibit basal PRL secretion. A recent report has highlighted the ability of cannabinoids to inhibit TRH and vasoactive intestinal polypeptide-stimulated PRL release in the rat mammosomatotroph tumoral cell line GH4C1 (38). Taken together, these findings clearly favor the hypothesis that cannabinoids are able to directly influence PRL secretion at the pituitary level.
A further finding was the presence of CB1 in folliculostellate cells. The mouse folliculostellate-like cell line TtT/GF was also found to express CB1 (our unpublished observations). Folliculostellate cells are known to be a large reservoir of cytokines and growth factors in the anterior pituitary gland (39, 40). Therefore, it is tempting to speculate that cannabinoids acting at CB1 might also influence the paracrine milieu of the adenohypophysis by regulating the secretory activity of these cells.
After showing that cannabinoids are able to regulate hormonal secretion from tumoral pituitary cells, we asked whether normal and tumor pituitary glands are able to synthesize the two endocannabinoids, AEA and 2-AG. In agreement with previous findings in the rat pituitary gland (14), in human normal pituitaries both endocannabinoids were present, thus indicating the ability of the gland to provide in situ synthesis of these compounds. Interestingly, all tumor samples had higher contents of AEA and 2-AG than the normal hypophysis. Moreover, the endocannabinoid content in pituitary adenomas correlates with the presence of CB1 by being elevated in the CUSH, ACRO, and PROL, which were the tumors positive for CB1, and lower in NFPA, which are characterized by low or absent CB1 expression. Together, these findings allow us to postulate the existence of an auto/paracrine cannabinoid loop in pituitary adenomas that may have an important role in modulating hormone overproduction.
Recent observations have demonstrated an antitumor effect of cannabinoids on breast and prostate cancer cells in vitro (41, 42) and on gliomas in vivo (43). Proliferation studies are difficult to perform on primary pituitary adenoma cells because of the low proliferation rate. However, preliminary results from our laboratory showed cannabinoid-mediated antiproliferative effects on the mouse tumor corticotroph cell line AtT-20 (unpublished observations). Therefore, growth inhibitory effects of cannabinoids on pituitary adenomas might represent an interesting prospect for future research.
In conclusion, we report for the first time the expression of CB1 in the human pituitary gland and human pituitary adenomas. Furthermore, increased levels of endocannabinoids in pituitary tumors and cannabinoid-induced modulation of hormonal release were shown. Together, these results indicate a potentially important role of the endocannabinoid system in pituitary pathophysiology, thus opening new perspectives in the management of pituitary adenomas.
Source with Charts, Graphs and Links: Normal Human Pituitary Gland and Pituitary Adenomas Express Cannabinoid Receptor Type 1 and Synthesize Endogenous Cannabinoids: First Evidence for a Direct Role of Cannabinoids on Hormone Modulation at the Human Pituitary Level