The hypothalamus is the source of neuropeptides which, being secreted into the portal system, control the synthesis and the secretion of the anterior pituitary hormones. Besides the well characterized hypothalamic central control and the hormonal peripheral control, recent studies have shown, in the anterior pituitary, the expression, among many other regulatory factors, of neuropeptides that are identical to those produced by the hypothalamus and that seem involved in the local control of anterior pituitary functions through autocrine or paracrine mechanisms. The presence of the neuropeptide mRNAs, precursors and mature forms of the peptides in anterior pituitary tissues as well as the secretion of the mature peptides argue in favor of the intrinsic ability of the normal and tumoral anterior pituitary to express neuropeptides. This expression of neuropeptides occuring in tissues bearing functional receptors for these ligands, anterior pituitary control could rely, at least in part, on endogenous neuropeptides acting locally. Correlations between neuropeptide contents in the anterior pituitary and the plasma levels of anterior pituitary hormones suggest that neuropeptides of anterior pituitary origin can play a local regulatory role, complementary of the classical hypothalamic central control. In the normal anterior pituitary which remains under hypothalamic control, it is presently difficult to evaluate the relative importance of the local and central control. However, anterior pituitary hyperplasia and pituitary tumors represent two models in which the specific contribution of the local control is easier to define.

In order to gain a better understanding on the possible role of retinoic acid (RA) on human GH secretion, we have characterized the expression of its nuclear receptors in somatotropic adenoma cell extracts. By immunoblotting with rabbit polyclonal antibodies directed against RARα, β, and γ and RXRα and β, we could only detect the presence of RARα and RXRα proteins. The predominant expression of RXRα was confirmed at the mRNA level by Northern and slot‐blot analysis. When then investigated the effect of RA on GH synthesis in cell culture of adenomatous somatotrophs. In cultured cells, RA (1 μM) stimulated GH secretion, increased intracellular GH content and GH mRNA levels within 72 h, suggesting a modulation of GH synthesis by RA. J. Cell. Biochem 65:25–31. © 1997 Wiley‐Liss, Inc.

The anterior pituitary is a complex gland which controls the various peripheral endocrine glands (thyroid, adrenal cortex, ovary and testis) as well as essential functions as growth, reproduction and lactation. Its physiological role is of paramount importance and its pathology multiple, including functional dysregulations and diseases caused by lesions or tumors, in particular pituitary adenomas. Recent data show that, in addition to the classical control of anterior pituitary hormones by hypothalamic neuropeptides and by peripheral hormones, local controls of the paracrine or autocrine type exist. Although these new findings seem to increase the complexity of anterior pituirary control, they might, in the future, shed further light on the physiopathology of this gland.

TRH gene expression in the anterior pituitary has previously been reported in the human in vivo and in the rat in vitro. Until now, modulation of this synthesis with glucocorticoids and thyroid hormones has been observed in rats. The present study demonstrates for the first time that the TRH gene is also expressed, in vivo, in the rat anterior pituitary and that anterior pituitary TRH-like immunoreactivity (TRH-LI) and elongated forms of the immediate TRH progenitor sequence (TRH-elongated peptide) contents are also modulated by estrogens (E2). To investigate the presence of proTRH mRNA in the rat anterior pituitary, total RNA was reverse transcribed (RT) and the RT products were then amplified by PCR. Treatments with E2 were performed on intact and ovariectomized (OVX) rats for 2 months. TRH-LI was measured by RIA with an antibody which did not recognize the TRH-like peptide. pGlu-Glu-Pro-NH2 (< EEP-NH2) (cross-reactivity < 0.1%) and was characterized further as TRH-LI by HPLC. TRH-elongated peptides were measured by EIA and characterized by Sephadex G-50 chromatography and immunoblotting (molecular mass 25-35 kDa). The plasma prolactin levels and the pituitary sizes were increased by E2 treatment in both intact and OVX rats. Anterior pituitary TRH-LI increased in intact E2-treated rats compared with intact rats (82.7 +/- 19.0 versus 39.6 +/- 3.6 fmol/mg protein; means +/- S.E.M.; P < 0.001). This increase was greater when E2 was administered to OVX rats (599.0 +/- 98.4 after E2 treatment versus 58.6 +/- 3.6 fmol/mg protein: P < 0.001). In intact rats, anterior pituitary TRH-elongated peptide contents were not modified by E2 treatment while they were significantly decreased in OVX E2-treated rats (144.6 +/- 8.8 versus 223.7 +/- 9.5 fmol/mg protein; P < 0.001). These results demonstrate TRH gene expression in the rat anterior pituitary in vivo and suggest that E2 treatment is responsible for an increase in anterior pituitary TRH-LI, together with a decrease in TRH-elongated peptide contents.

Integrins are heterodimeric transmembrane molecules that mediate cell-cell and cell-substratum adhesion. Because alterations in the adhesive properties of tumor cells influence tumor growth and progression, the distribution of different alpha and beta integrin subunits was studied in both the parenchyma and the connective tissue in 6 normal and 25 adenomatous human anterior pituitaries. All normal parenchymal cells expressed the alpha3beta1 and alpha6beta4 integrins. By contrast, in adenomatous parenchymal cells the expression of alpha3beta1 was down-regulated and that of alpha6beta4 abrogated. Neoexpression of alphavbeta3 Occurred in the parenchyma of a subset of adenomas. All normal connective tissue cells expressed the alpha1 and beta1 subunits, a third subunit (alpha5) being present in the normal endothelium. By comparison, all adenomatous stromal cells expressed many more integrin subunits (alpha1, alpha3, alpha5, alphav, beta1 and beta3), adenomatous endothelial cells bearing additional subunits (alpha6, beta4 and beta5). Vitronectin, absent from the normal connective tissue, was constantly observed in the adenomatous stroma. To conclude, compared with cells of the normal gland, adenomatous anterior pituitary cells display a decreased expression of integrins whereas the adenomatous stroma expresses a rich repertoire of integrins. These changes are not related to the secretory type, grade or invasiveness of the adenoma. The resulting alterations in the adhesive properties of adenomatous cells could facilitate their dissemination. Enrichment of the integrin repertoire expressed by the adenomatous vasculature is indicative of its dual nature, systemic and tumoral.

In order to gain a better understanding on the possible role of vitamin A (VA) and retinoic acid (RA) on human growth hormone (GH) secretion, we used the physiological model of pituitary cells perifusion. In perifused cells from pituitary somatotropic adenomas, RA induced within minutes a peak of GH secretion. This effect was dose dependent, maximal effect being observed with 100 nM. The GH release was associated with a discharge of cAMP delayed by 4 to 8 minutes relative to the GH surge. Similar results were obtained after VA stimulation. Our observations provide the first evidence of an action of VA and RA on cAMP production. They suggest a role of RA and VA in the regulation of human GH secretion via the cAMP dependent pathway.

Arguments favor an in situ synthesis of GH-releasing hormone (GHRH) in the normal and tumoral human anterior pituitary. These tissues may express human (h) GHRH messenger RNA, contain hGHRH-(1-44)-NH2, and secrete in vitro an immunoreactive form (ir-form) of the peptide. Here, we characterize and localize the precursor of hGHRH in human anterior pituitary tissues using RIAs specific for the C-terminus or the midportion of hGHRH-(1-44)-NH2, size-exclusion chromatography, HPLC, Western blotting, and immunocytochemistry. The anterior pituitary ir-forms were compared to those found in hypothalamus, posterior pituitary, and GHRH-secreting endocrine pancreatic tumors. Three ir-forms of hGHRH with mol wt of 30-45, and 5 kilodaltons (kDa) were detected. The 30- to 45-kDa ir-form was very likely to consist of hGHRH bound to proteins. The 5-kDa ir-form represented mature forms of hGHRH. It was the major form in tissues actively synthesizing and/or secreting hGHRH. Nontumoral anterior pituitaries contained significant amounts of mature hGHRH. The 10-kDa form was identified as a hGHRH precursor ir-form. In addition to its expected presence in the hypothalamus and GHRH-secreting tumors, normal and tumoral human anterior pituitaries contained an identical ir-form of the hGHRH precursor. Cells immunoreactive for the hGHRH precursor were observed in pituitary adenomas. Evidence for precursor and mature ir-forms of hGHRH in anterior pituitary tissues provides conclusive arguments for the endogenous synthesis of the neuropeptide.

Several regulatory neurotactors, classically associated with the hypothalamus, may be synthesized in the anterior pituitary (AP). Dopamine (DA) is the main prolactin‐inhibiting factor. Its de novo synthesis in the normal AP has not been proved, although the TH transcript has been previously demonstrated by RT/PCR in the AP. We investigated tyrosine hydroxylase (TH) gene expression at both the protein and mRNA levels in the AP of normal random cycling female rats and in a catecholaminergic tissue, the adrenal gland (AG). The Western blot analysis of AP homogenates revealed two immunoreactive forms of TH in the AP, both differing from the TH present in the AG. RT/PCR products from AP and AG mRNA were subcloned and sequenced. In addition to the full‐length form, we identified two TH transcripts generated by alternative splicing either involving the use of a new alternate splice‐donor site within exon 2 or skipping exon 11. The form lacking exon 11 was not isolated from the AG. In the AP, all three forms were present. Although the AP contained the full‐length TH mRNA, the expected size protein was not detected. Thus, there is alternative splicing of the TH primary transcript, and putative additional poisttranslational regulation may yield TH proteins with no enzymatic activity, at least in non‐catecholaminergic tissues. ©1995 Wiley‐Liss, Inc.

We have previously shown that the somatostatin (SRIH) precursor (pro-SRIH) and SRIH are present in the normal human pituitary, whereas mature SRIH is never detected in GH-secreting adenomas associated with high plasma GH levels, and pro-SRIH is absent in 50% of them. Considering the fact that GHRH is present and released in vitro in higher amounts from GH adenomas than from normal human pituitaries, the possibility of a negative control exerted by GHRH on pituitary SRIH was investigated. When GH-secreting adenomas were incubated for 4 h in the presence of GHRH (10(-8) mol/L), the amount of pro-SRIH, quantified after Sephadex G-50 filtration of the acetic acid extracts, was significantly decreased. The percent inhibition was negatively correlated to the amount of endogenously released GHRH measured in the control incubation medium, suggesting a local negative feedback control exerted by pituitary GHRH on pituitary SRIH. This was further demonstrated when adenomas were incubated with a GHRH antibody. Indeed, the presence of the GHRH antibody resulted in a significant increase in the content of pro-SRIH in the adenoma. Similar results were obtained for in vitro SRIH release; exogenous GHRH induced an inhibition of SRIH release from GH-secreting adenomas, and the GHRH antibody had the opposite effect. Using a perifusion system, we showed the concomitance between the increase in GH release and the decrease in SRIH release after GHRH stimulation. Together, these results show in vitro a negative control exerted by GHRH (both exogenous and locally released) on adenomatous pituitary SRIH. This further amplifies the importance of autocrine or paracrine controls in these tumors.

The purpose of this study was to determine whether gonadotropin-releasing hormone (GnRH) may exert a direct action on human prolactinomas. On a series of 17 adenomas, we studied the effect of GnRH on the in vitro prolactin (PRL) secretion of dispersed and perifused cells of 10 cases and the [125I]GnRH agonist binding on frozen sections of three out of the adenomas studied in perifusion and on the membrane preparations of seven other cases. Two 20-min pulses of GnRH (10(-7) mol/l) stimulated the in vitro PRL secretion of three adenomas out of 10 (increase of 200, 444 and 205%, respectively, above basal levels). The GnRH receptors of three adenomas bound GnRH agonist (Des-Gly10-(D-Ala6)-GnRH ethylamide). The binding was specific, with a high affinity (Kd = 0.60, 0.48 and 0.40 nmol/l) similar to that of two human anterior pituitaries obtained post-mortem (Kd = 0.70 and 0.40 nmol/l). Indirect immunoperoxidase revealed that the majority of the cells (60-90%) in all the adenomas studied contained immunoreactive PRL. Four of them also contained cells immunoreactive to the alpha-subunit of the glycoprotein hormones. In none of the prolactinomas were cells immunoreactive to antiserum of anti-beta-luteinizing hormone, anti-beta-follicle-stimulating hormone or anti-beta-thyrotropin. All the prolactinomas that were responsive to GnRH in perifusion experiments and/or bound specifically to [125I]GnRH agonist were also immunoreactive for alpha-subunit. These results show that GnRH, via GnRH specific receptors, exerts a stimulation on in vitro PRL secretion in a subset of prolactinomas characterized by the presence of alpha-subunit.

We studied the boundary between adenoma and peritumoral anterior pituitary tissues in order to understand their mutual interactions during tumour progression. We selected 18 adenomas of different secretory type, grade and invasiveness in which fragments of peritumoral anterior pituitary were still attached to the adenoma. Immunohistochemistry was performed on serial sections with markers of the basement membranes (type IV collagen), the hormone-producing cells of the normal and neoplastic anterior pituitary, and the folliculo-stellate cells (S-100 protein). In passing from tumour to gland, localized areas of passive compression of the normal gland were seen in only 3 cases. In all the tumours, the boundary consisted partly or solely of a transitional zone characterized by the presence of enlarged cell-cords. Openings in the basement membrane of these enlarged cell-cords were seen in contact with the tumour tissue. Normal and neoplastic cells intermingled in the transitional zone. Normal residual cells could be seen in the central area of the tumour but no adenomatous cells were observed in the gland around the tumour. Folliculo-stellate cells were concentrated in the vicinity of the transition zone. These findings favour the existence of an active process of adenoma expansion within the normal parenchyma, without noticeable infiltration of tumour cells into surrounding gland.

BACKGROUND: Laminin (LM) is an integral component of basement membranes (BM), with important roles in various aspects of cell biology. Several different isoforms of LM have been described, and each comprises a molecule consisting of three subunit polypeptides, the A, B1, B2, Mor S chain.EXPERIMENTAL DESIGN: The distribution of different LM subunits was studied in human nontumoral anterior pituitaries obtained postmortem and in pituitary adenomas by immunocytochemical methods using specific monoclonal antibodies.RESULTS: In normal tissue, the A, B1, and B2 chains had a ubiquitous localization in both parenchymatous and vascular BMs and in the pericapillary connective tissue space also. The S chain seemed to have principally a vascular localization, in contrast to the M chain that was mostly localized in the parenchymatous BM. The same antigens were investigated in 23 human pituitary adenomas of different secretory type, grade and invasiveness. In all cases, the five monoclonal antibodies gave a positive staining. The presence or the localization of LM chains did not display a specific pattern in the adenomas according to their secretory type, grade or extent of local invasion. In the adenomas, the anti-A, -B1, and -B2 monoclonal antibodies stained the stroma and in particular the vascular BMs as well as the sparse elements of parenchymatous BMs when they were present. The staining with anti-S antibody was localized in the stroma and around all the blood vessels. In contrast, the immunoreactive material to anti-M antibody was associated to the sparse fragments of parenchymatous BMs and it delineated the boundary of adenoma cells and stroma. Within the stroma, the anti-M antibody immunoreactivity was regularly associated with the arterial walls but rarely with the venule walls.CONCLUSIONS: The human normal and tumoral anterior pituitary express all five LM subunits and thus contain several LM isoforms with different patterns of localization. The most striking difference between the human normal and tumoral anterior pituitary concerns the peculiar expression of LM isoforms by adenomatous neovessels.

In order to see whether, in thyrotropic adenomas with thyrotoxicosis, plasma thyroid hormones regulate the thyrotropin-releasing hormone (TRH) binding sites and the thyrotropin (TSH) response to TRH, we investigated: the presence of TRH binding sites in two cases of thyrotropic adenomas associated with hyperthyroidism and in one case of thyrotropic adenoma secondary to thyroid failure: and the in vitro effect, in a perifusion system, of triiodothyronine (T3) on the response of TSH to TRH in three cases of TSH-secreting adenomas associated with hyperthyroidism. The TRH binding sites were absent in the adenomas associated with high levels of circulating thyroid hormones, whereas they were present in the adenoma secondary to primary thyroid failure (Kd = 47 nmol/l, Bmax = 40 nmol/kg membrane proteins). In vitro, the three adenomas spontaneously released TSH in the perifusion medium (1.49 +/- 0.06 (mean +/- SEM), 7.25 +/- 0.12 and 16.73 +/- 0.36 mIU.l-1 x 10(6) cells-1 x 2 min-1) and exhibited an ample TSH response to 10(-7) mol/l TRH pulses. In two cases, tumoral secretion of fragments was compared with those of fragments maintained since the time of surgical removal in the presence of 10(-8) mol/lT3. The TSH responses to TRH were abolished in the presence of T3 in these two cases. We conclude that thyrotropic adenomas associated with hyperthyroidism are still controlled in vivo by T3. In particular, T3 regulates the TSH response to TRH, probably via a down-regulation of the TRH binding sites.

Somatuline, in common with other SRIH analogues, exerts antiproliferative and antisecretory activities on various tumors. Our purpose was to test the effectiveness of a slow-release formulation of somatuline on lactotroph hyperplasia and PRL hypersecretion induced by estrogens (17 beta E2) in rats. Female rats were primed with 17 beta E2 for 6 weeks before receiving somatuline (2 mg/kg) intramuscular injections every 10 days for one month. The mean anterior pituitary weight was 11.22 +/- 0.32 mg (mean +/- SEM) in non-estrogenized rats, 29.62 +/- 1.63 mg in 17 beta E2-primed rats and 23.58 +/- 1.26 mg in 17 beta E2-primed somatuline-treated rats. Mean plasma PRL level was 5.63 +/- 0.97 ng/ml, 182.37 +/- 27.55 ng/ml and 113.89 +/- 15.07 ng/ml in the same groups respectively. Thus, the 17 beta E2-induced pituitary enlargement and hyperprolactinemia were 20% and 38% lower respectively when animals were treated with somatuline during the last month of estrogenization. The 17 beta E2-induced increase in PRL cell density was also reduced by somatuline treatment. We conclude that the slow-release formulation of somatuline impedes 17 beta E2-induced hyperprolactinemia and pituitary enlargement concomittantly, at least in part by acting on lactotroph proliferation.

Thyroliberin (TRH) has been reported to be synthesized in anterior pituitary. Modulations of anterior pituitary concentrations of TRH and its precursor (proTRH) were investigated in rats rendered hypothyroid for 21 and 48 days. In hypothyroid rats, as expected, plasma TSH levels were increased and plasma free T4 concentrations were reduced, whereas anterior pituitary concentrations of TRH were markedly increased (mean +/- SEM fmol/mg protein, 106.9 +/- 21.1 at 21 days and 181.8 +/- 59.0 at 48 days vs 42.2 +/- 13.4 in controls). Significant increases of proTRH concentrations were found in hypothyroid rats (mean +/- SEM fmol/mg protein, 266.5 +/- 16.7 at 21 days and 320.4 +/- 51.1 at 48 days vs 214.1 +/- 18.3 in controls). Daily administration of L-T4 for the last week of PTU treatment (days 41 to 48) reversed hypothyroidism by lowering plasma TSH levels and by increasing plasma free T4 concentrations. Concomitantly, T4 administration significantly decreased anterior pituitary TRH contents (114.4 +/- 35.9) as well as proTRH contents (250.4 +/- 11.7). The fact that hypothyroidism regulates anterior pituitary TRH and proTRH contents strengthens the view that TRH might play a role in local regulatory mechanisms within the anterior pituitary on the thyrotropic function.

Protein kinase-C (PKC) is a ubiquitous eukaryotic kinase that plays a key role in transmembrane signaling and influences important cellular processes, such as proliferation. Increases in its activity and expression have been demonstrated in adenomatous human pituitaries, with protein expression being the highest in invasive tumors (1). Moreover, in these same invasive tumors, the mean increase in expression (8.9-fold) does not correlate with the mean increase in activity (2.6-fold), suggesting a dysfunction in PKC in these tumors. Here, we show that the PKC alpha-isoform (alpha PKC) is overexpressed in human pituitary tumors. The complete sequencing of the PKC cDNA from four invasive tumors has revealed a point mutation that is absent in the noninvasive tumors analyzed. The point mutation is located at position 294 of the protein, in the V3 region, leading to a substitution of a negatively charged aspartic acid by an apolar glycine. Thus, not only is alpha PKC overexpressed in human pituitary tumors, but it is also structurally altered in the invasive subpopulation of these tumors.

Normal and tumoral human pituitaries release in vitro SRIH and contain messenger RNAs encoding preproSRIH. In the present study, we document the presence and characterization of the SRIH precursor in both human normal pituitaries and in GH-secreting adenomas. Molecular sieve filtration of normal pituitary and adenoma acid extracts revealed the presence of three immunoreactive SRIH peaks, distinct from SRIH 1-28 and SRIH 1-14. Western blot analysis performed under both nonreducing and reducing conditions confirmed the presence of the three different immunoreactive forms with respective molecular masses estimated as 21, 17, and 12 kilodaltons. When analyzed in reverse phase high performance liquid chromatography, the three immunoreactive SRIH material contained in the three peaks coeluted with the proSRIH extracted from rat hypothalamus and from a neuroblastoma cell line (N2A) transfected with the human preproSRIH complementary DNA. None of these three forms could bind concanavalin A. Digestion of the three forms with a specific endopeptidase resulted in the generation of SRIH 1-14, indicating that they can be considered as SRIH precursor or related forms. Estimation of proSRIH amount after molecular sieve filtration indicated that normal human pituitaries (n = 4) contained proSRIH in variable amounts (from 26.5-303 pg/mg proteins), the 21-, 17-, and 12-kilodalton forms being always present, in addition to SRIH 1-28. In contrast, among the 21 adenomas tested, only 11 contained proSRIH material (from 9-4480 pg/mg proteins). Moreover, neither SRIH 1-28 nor SRIH 1-14 could be detected in these adenomas. These data demonstrate the presence of high molecular mass SRIH in normal pituitaries which represent unprocessed proSRIH material. This provides an additional argument for a local synthesis of SRIH. Additionally, the fact that only 50% of GH-secreting pituitary adenomas here studied contain exclusively proSRIH may reflect an alteration in messenger RNA translation or in proSRIH processing, or both, resulting in the absence of mature SRIH in these adenomas.

To determine the cellular mechanism(s) of the variability in GH responsiveness to octreotide in acromegaly, somatostatin (SRIH) receptor status was studied in 37 GH-secreting adenomas. SRIH receptor binding was always present in all GH-secreting adenomas either in membrane preparations (Exp A; n = 17) or by quantitative autoradiography (Exp B; n = 20). In membranes, maximal binding capacities ranged from 83-2331 fmol/mg protein; affinity was in the nanomolar range (Kd, 1.3 +/- 0.2 nmol/L). By quantitative autoradiography, SRIH-14 and octreotide were equally active in displacing [125I]SRIH binding in tumors (Spearman correlation rho = 0.92). IC50 values ranged from 3-22 nmol/L (mean +/- SE, 8.0 +/- 1.3 nmol/L). In Exp A, basal adenylate cyclase (AC) activity was high in 7 tumors (841 +/- 306 pmol/L cAMP x 30 min/mg protein) compared to that in the other 10 (252 +/- 92 pmol/L cAMP x 30 min/mg protein). In these 7 tumors, GH-releasing hormone (0.1 mumol/L) stimulation of AC was lower (53 +/- 11% vs. 297 +/- 48%), whereas SRIH (1 mumol/L) inhibition was higher (52 +/- 5% vs. 34 +/- 5%). Similar results were obtained with Exp B tumors. In both experiments, no correlation was apparent between SRIH-binding capacity and inhibition of AC. In Exp B, a variable decrease in mean plasma GH levels was observed (> or = 80% in 5 patients, between 50-80% in 8 patients, and < or = 50% in 5 patients) after a single sc injection of octreotide (100 micrograms). A modest correlation was found between the GH response to octreotide and SRIH-binding capacity (rho = 0.48) or SRIH inhibition of AC (rho = 0.61). The IC50 values to displace SRIH binding were lower in poorly responsive patients than in highly responsive ones (IC50, 4.6 +/- 1.9 and 13.9 +/- 2.7 nmol/L, respectively). These data indicate that an absence of SRIH receptors cannot account for the weak response to SRIH therapy in 20-30% of acromegalic patients. Alternatively, the weak correlation between either SRIH binding or SRIH inhibition of AC with octreotide inhibition of plasma GH levels might reflect the heterogeneity of SRIH receptor subtypes in GH-secreting adenomas.