Pituitary Phenotypes of Mice Lacking the Notch Signalling Ligand Delta-Like 1 Homologue

The Notch signalling pathway ligand delta-like 1 homologue (Dlk1, also named Pref1) is expressed throughout the developing pituitary and becomes restricted to mostly growth hormone (GH) cells within the adult gland. We have investigated the role of Dlk1 in pituitary development and function from late embryogenesis to adulthood using a mouse model completely lacking the expression of Dlk1. We confirm that Dlk1-null mice are shorter and weigh less than wild-type littermates from late gestation, at parturition and in adulthood. A loss of Dlk1 leads to significant reduction in GH content throughout life, whereas other pituitary hormones are reduced to varying degrees depending on sex and age. Both the size of the pituitary and the proportion of hormone-producing cell populations are unchanged, suggesting that there is a reduction in hormone content per cell. In vivo challenge of mutant and wild-type littermates with growth hormone-releasing hormone and growth hormone-releasing hexapeptide shows that reduced GH secretion is unlikely to account for the reduced growth of Dlk1 knockout animals. These data suggest that loss of Dlk1 gives rise to minor pituitary defects manifesting as an age- and sex-dependent reduction in pituitary hormone contents. However, Dlk1 expression in other tissue is most likely responsible for the weight and length differences observed in mutant animals.

The anterior pituitary gland has a central role in physiology, regulating various processes such as growth, lactation and pregnancy, reproduction, metabolism, and response to stress. The five different hormone-producing cell types within the gland are specified during embryogenesis by a cascade of transcription factors and signalling pathways (1), although the number of cells of each population increases after birth (2) and can vary throughout life in response to physiological demand (3). These changes in cell number and function are under the control of hypothalamic and peripheral signals (3,4), although paracrine and autocrine factors are also considered to be involved (5). Paracrine factors are likely to regulate differentiation of the recently described multipotent progenitor, or stem cells, of the pituitary (6), especially for cell types such as somatotrophs (producing growth hormone; GH), where the receptor for the primary proliferative signal is restricted to the differentiated cell type (7).
The Notch pathway is a conserved signal transduction mechanism first identified in Drosophila (8) that is activated by the binding of a transmembrane ligand to the transmembrane Notch receptor on a neighbouring cell. This interaction results in cleavage of the receptor to form the Notch intracellular domain, which translocates to the nucleus, binds RBPjj and initiates transcription of target genes such as Hes1 and Hey1 (9). Various components of the Notch pathway are expressed during pituitary development, including the Notch2 and Notch3 receptors, the ligand Jagged1, and the downstream effector Hes1 (10,11). Overexpression of Hes1 in differentiating pituitary cells in vivo has been shown to inhibit gonadotroph and thyrotroph differentiation in mice (12). Conversely, Hes1-deficient mice display increased cell cycle exit and increased expression of cyclin-dependent kinase inhibitors such as p27 in the pituitary (13), whereas Hes1 and Prop1 double-mutants show premature differentiation of corticotrophs (14). Persistent expression of the receptor Notch2 during embryogenesis causes a reduction in the number of thyrotrophs and delays gonadotroph differentiation, although the gonadotroph population is rescued as the mice develop to maturity (15). Conditional deletion of the Notch effector RBPjj in the developing mouse embryo leads to premature differentiation of corticotrophs and, conversely, overexpression of the active Notch receptor inhibits terminal differentiation (11). Taken together, this evidence points towards Notch signalling as a regulator of differentiation timing within pituitary hormone cell types.
The nonclassical ligand delta-like 1 homologue (Dlk1), a paternally-imprinted gene on mouse chromosome 12 (16), is expressed throughout the developing Rathke's pouch from embryonic day (E) 10.5 (17) and in the adult anterior pituitary, as well as in bone, b-cells in pancreatic islets, placenta and adrenal glands (17)(18)(19)(20)(21). The protein is expressed in the majority of GH cells in the pituitary gland, and a low proportion of all other hormone cell types (22,23), as well as the Sox2-expressing putative stem/progenitor cells, which can form pituispheres in culture (24,25). In vitro studies have previously shown that, in somatolactotroph GH3 cells overexpressing Dlk1, GH expression and secretion are down-regulated (26). Expression of Dlk1 is also increased in human hormone-secreting pituitary tumours (27), whereas silencing of the Dlk1/MEG3 imprinted locus is detected in nonfunctioning pituitary adenomas (28,29). This pattern of expression suggests a role for Dlk1 in normal pituitary development and function.
One of the observed phenotypes of mice lacking Dlk1, generated by deletion of exons 2 and 3, is growth retardation (30), which was later confirmed in a similar but independently-generated Dlk1-null mutant deleting the promoter and exons 1-3 (31). A recent study using a mouse model with altered expression of several imprinted genes, including overexpression of Dlk1, reported a reduced weight of transgenics at weaning associated with a failure to thrive (32). Therefore, either increased or decreased expression of the Dlk1 gene may have an effect on the growth of the mice. A recent study using the Dlk1-null mutant generated by Raghunandan et al. (31) found fewer GH-, prolactin (PRL)-and follicle-stimulating hormone (FSH)-immunoreactive cells in the adult pituitary, and an increased serum leptin concentration (23). Interestingly, the organisation of somatotrophs, an important factor in robust GH secretion (33), was found to be altered, suggesting that the growth phenotype in null animals may be caused in part by an altered GH axis. In the present study, we have expanded on these previous studies by assessing the effect of loss of Dlk1 on growth and pituitary hormones, in particular investigating whether the growth retardation could be accounted for by a deficient secretion of GH.

Genetically-modified mice strains
We have characterised pituitary phenotypes in previously-generated Dlk1null mutant mice (31) on a C57BL/6J background. Genomic DNA from ear biopsies was used for polymerase chain reaction genotyping using previously described primers. Because Dlk1 is a paternally-imprinted gene (16,17,34), we compared wild-type with Dlk1 +/ÀPat mice, with the null allele paternally inherited (referred to as Dlk1-null mice), except when comparing heterozygotes with homozygous null mice. Mutants show no noticeable impairment in fertility and litter size.

In vivo experiments
Mice were given access to water and chow ad lib., and experiments were performed in accordance with Institutional and Home Office legislation and guidelines. Weights were recorded weekly between age-matched littermates after weaning at 3 weeks of age. Body lengths were measured after mice were sacrificed. Pituitary response to acute challenge by GH secretagogues was performed as described previously (35).

Radioimmunoassays
Total pituitary hormone contents were assayed using a previously described method (36) using mouse-specific reagents kindly provided by A. L. Parlow [National Hormone and Pituitary Program (NHPP), Torrance, CA, USA].

Cell dispersion
Pituitary glands were dispersed as previously described (37) and all cells plated onto 13-mm diameter coverslips coated with polylysine (Sigma, St Louis, MO, USA). Cell counts of dispersed cells were performed manually after immunofluorescence imaging.

Total pituitary protein measurement
Total protein from pituitary samples was measured using the Micro BCA e Protein Assay Reagent Kit (Thermo Scientific Pierce, Rockford, IL, USA) in accordance with the manufacturer's instructions.
with ANOVA to test the overall effect of genotype on growth followed by Tukey's post-hoc tests.

Loss of Dlk1 expression leads to growth retardation
Immunohistochemistry for Dlk1 confirmed expression in normal pituitary, a loss of expression in Dlk1 null animals, as well as expression in somatotrophs in wild-type animals, with some nonsomatotroph cells also expressing the protein, although the majority of Sox2-positive cells do not express Dlk1 (Fig. 1). As reported previously (30), a loss of Dlk1 led to a visible reduction in body size. Weight measurements beginning at 3 weeks of age demonstrated that males have a significantly reduced weight at the majority of timepoints, and females had a lower weight from weaning to 14 weeks of age ( Fig. 2A, i). However, there is no difference in the weight gained between wild-types and mutants between 4 and 14 weeks of age (weight gained between 4 and 14 weeks of age: males +/+14.73 AE 0.67 g versus +/À13.64 AE 0.75 g; females +/+9.91 AE 0.58 g versus +/À9.34 AE 0.71 g; n = 7-10; wild-type versus heterozygote not significant), showing that the rate of growth of the mutants post-weaning is normal. The nose-tail length of Dlk1-null mice at 14 weeks of age was reduced compared to wild-type littermates in both male and female mice ( Fig. 2A, ii), consistent with them having a reduction in body size from birth without subsequent catch-up growth. Embryos at E18.5 were lighter than wild-type littermates (Fig. 2B, i) and also had a reduced pituitary GH content (Fig. 2B, ii).

Dlk1-null mutants have age-and sex-specific reductions in all pituitary hormones
Pituitaries from E18.5 embryos and 2-, 6-and 14-week-old mice were dissected for total pituitary hormone content measurements by radioimmunoassay. A loss of Dlk1 led to a 30% reduction in pituitary GH content in E18.5 embryos, which remained reduced in both male and female mutants (Fig. 2B, ii), from juveniles (2 weeks) to fully-adult mice (14 weeks) ( Table 1).
Other pituitary hormones are also reduced to varying extents in the mutants in an age-and sex-dependent manner. Mutant hormone contents expressed as a percentage of the wild-type content  Statistical analysis by mixedeffects model of weight versus age using genotype as fixed factors and subjects (mice) as random factors, with ANOVA to test the overall effect of genotype on growth followed by Tukey's post-hoc tests (A) and unpaired t-test (B). *P < 0.05; **P < 0.01; ***P < 0.001 Dlk1-null versus wild-type littermates. Pituitaries from 2-, 6-and 14-week old Dlk-1 null mutant and wild-type were assayed for GH, prolactin (PRL), thyroid-stimulating hormone (TSH), adrenocorticotrophic hormone (ACTH), luteinising hormone (LH) and follicle-stimulating hormone (FSH). All hormones show a mild reduction in content at some time during the mice's life, although not all remain deficient. Data are shown as the mean AE SEM of seven to ten mice. * P < 0.05; * * P < 0.01; * * * P < 0.001 Dlk1-null versus wild-type littermates. NS, not significant.
show the progression of hormone reduction phenotypes over time (Fig. 3). GH and FSH contents are reduced in both sexes at all three ages; PRL content is reduced in 2-and 6-week-old females; TSH content is significantly reduced at all time-points, except at 2 weeks of age in females; ACTH content is reduced at 6 weeks of age in males and at 2 weeks of age in females; and LH contents are reduced at all time-points except 6 weeks of age in females. PRL and ACTH contents in Dlk1-null mutant females return to normal by 14 weeks of age. There was no difference in the pituitary hormone contents of homozygous null animals compared to paternally-inherited Dlk1 heterozygote littermates at 6 weeks of age (total GH content per pituitary: males +/À67.08 AE 8.18 lg versus À/À68.10 AE 7.87 lg; females +/À34.79 AE 3.75 lg versus À/ À37.90 AE 3.40 lg; n = 5-6; heterozygote versus homozygote not significant).
To address the possibility that the Dlk1-null mutants have reduced pituitary hormone content simply as a result of a smaller pituitary gland, we measured total pituitary protein content in wild-type and Dlk1-null male and female mice at 6 weeks of age. There was no significant reduction in the total protein content of the Dlk1-null mice compared to wild-type littermates (males +/+0.90 AE 0.053 lg versus +/À0.84 AE 0.033 lg; females +/+0.81 AE 0.041 lg versus +/À0.73 AE 0.037 lg, n = 4-9, wildtype versus heterozygote not significant).

Proportions of cell populations unchanged in Dlk1-null female mutants
A reduction in pituitary hormone content could result from a reduced number of hormone-producing cells. To determine whether this was the case after loss of Dlk1, pituitaries from 6-week-old female wild-type and null mutants were dispersed and plated for immunohistochemistry for individual hormones, allowing for imaging and manual quantification of hormone populations. No  Table 1 expressed as a percentage of the corresponding wild-type content at 2, 6 and 14 weeks of age. Data are shown as the mean AE SEM of seven to ten mice. *P < 0.05; **P < 0.01; ***P < 0.001 Dlk1-null versus wild-type littermates. ACTH, adrenocorticotrophic hormone; FSH, follicle-stimulating hormone; GH, growth hormone; LH, lutesinsing hormone; PRL, prolactin; TSH, thyroid-stimulating hormone.
significant differences that could account for the reduction in hormone content was observed in any of the hormone-producing cell populations in the mutant mice (Fig. 4). There was no evidence for adenoma formation in pituitaries, either at this age or in the pituitaries of 14-week-old animals.

Minor defect in sustained GH secretion in Dlk1-null male mutants
Six-week-old male wild-type and Dlk1-null mutants were assessed for their ability to secrete GH in response to the GH secretagogues growth hormone-releasing hormone (GHRH) and growth hormonereleasing hexapeptide (GHRP-6) (Fig. 5). Basal plasma GH concentrations before the administration of secretagogues were low in both wild-type and mutant animals, as expected. Both GHRP-6 and GHRH caused GH secretion into the circulation in both wild-type and mutant mice. The GH response in mutants is normal at all time-points except one, with the only significant reduction observed being at 15 min after administration of the second secretagogue (GHRH).

Dlk1 in the placental labyrinth and its role in intrauterine growth restriction (IUGR)
Because Dlk1 mutant neonates have a reduced weight at birth and Dlk1 has previously been shown to be expressed in the embryonic vessels of the placenta, we examined the placentas of Dlk1 mutant mice for any obvious aetiologies for intrauterine growth restriction. Immunohistochemistry of placentas of wild-type mice at E18.5 confirmed the expression of Dlk1 in foetal endothelia, shown by its colocalisation with platelet endothelial cell adhesion molecule-1 (PECAM-1) (Fig. 6A-C). Dlk1 is not expressed in the placental labyrinth of Dlk1 mutant embryos, as expected, although there is no obvious alteration in PECAM immunostaining, indicating normal foetal vessel formation (Fig. 6D-F). Higher magnification of the labyrinth of mutant animals shows there is no overt defect in the branching of the trophoblast layer and the formation of the foetal endothelial cells (Fig. 6G-L). Placental weights were also measured at E18.5 to study placental insufficiency as a cause of IUGR. Although it appears that placentas of Dlk-1 null animals may have a reduced weight compared to wild-type littermates (P = 0.09) (Fig. 6M, i), expressing this as a percentage of the embryo weight shows a proportional placental weight reduction (Fig. 6M, ii).

Discussion
The present study investigated physiological and pituitary phenotypes in a previously-generated Dlk1-null mutant mouse strain (31).
Although initially generated for a study on B lymphocytes, Raghunandan et al. (31) also noted the growth phenotype of the null mutants. A recent study using this mouse model investigated the cellular and subcellular localisation of Dlk1 within hormone-producing cells, and the effect of its loss within those cells (23), finding an effect on the somatotroph, lactotroph and gonadotroph axes, and elevated circulating leptin concentrations. In the present study, we have shown that a loss of Dlk1 leads to effects on all pituitary hormones, although the reduction in GH secretion is unlikely to cause the size reduction associated with a loss of Dlk1. We have confirmed the paternal expression of Dlk1 in the pituitary because the inheritance of a normal maternal allele and a null paternal allele led to a loss of pituitary expression, with the antigen of the antibody in an unaltered part of the Dlk1 protein. We also show that the majority of Sox2-positive cells do not express Dlk1, although they have been reported to be co-expressed in the same cell population (24). It has previously been reported that certain  genetic and epigenetic factors may cause reactivation of the silent maternal allele of Dlk1 in the brain (38) and other imprinted genes have been shown to undergo post-natal loss of imprinting (39); however, at 6 weeks of age, we could not detect any reactivation in the pituitary. To confirm that reactivation in other tissues or at other ages was not affecting the phenotypes analysed, we also compared homozygous Dlk1-null animals with heterozygous animals with a paternally inherited null allele. There was no difference  in weight, growth rate, lengths or any hormone contents between the heterozygous and the homozygous null mice at 6 weeks of age, allowing us to use heterozygotes with a paternally-inherited null allele as functionally null mutants in all our subsequent studies. It has been previously reported that the Dlk1 null-mutant mice generated by Moon et al. (30) are visibly smaller compared to agematched littermates and weighed less from 20 to 120 days of age. Monitoring of the weight of the Dlk1-null mutants used [from Raghunandan et al. (31)] from weaning at 3 weeks of age until 14 weeks of age shows a reduced weight throughout life but a normal rate of growth. The mutant mice continue to have reduced weight compared to wild-type littermates as they become older. This differs to the growth phenotype described for the Dlk1-null mutant strain generated by Moon et al. (30) where mutants eventually return to weights similar to those of wild-type animals. However, the Dlk1-null strain generated by Moon et al. (30) deletes a larger portion of the Dlk1 gene and the background strain was BALB/cJ (27), which may lead to subtle differences in phenotype. Background effects on Dlk1-null phenotypes were previously identified by Moon et al. (30) with an eyelid defect on a BALB/cJ background that is absent on a C57BL/6 background. This may also be the cause of phenotypic differences between the present study and those observed by both Raghunandan et al. (31) and Puertas-Avendano et al. (23), who used the Dlk1-null strain on a 129/SvJ background.
The length of Dlk1-null mutants was reduced compared to wildtype littermates, similar to those found in a mouse strain with a low GH content as a result of targeted ablation of GH cells (40). GH deficiency would be consistent with the post-weaning weight and length reductions of Dlk1-null mutants but does not account for the reduced size of E18.5 embryos because other mice with more severe GH deficiency do not show any size difference at this age (41). However, we are unable to exclude a mechanism whereby a combination of a loss of Dlk1 with GH deficiency could cause growth retardation in the embryo. It is also unlikely that any effects on body size at older ages are simply a result of reduced pituitary GH content because other models with a similar reduction have normal body length and weight from weaning to 14 weeks. Because total pituitary GH content is far in excess of the amount secreted per pulse, a mild reduction in total GH protein may be insufficient to cause a growth defect.
The reduction in weight of Dlk1-null embryos at E18.5 compared to their wild-type littermates suggests a role for Dlk1 in intrauterine growth retardation (IUGR). Dlk1 is expressed in the endothelial cells of the placental labyrinth (21), the interface between the maternal and foetal circulations that allows nutrient exchange with the foetal vasculature (42), prompting us to determine whether the altered function of these endothelial cells in Dlk1-null animals may be resonsible for the IUGR. In the Dlk1 mutant embryos at E18.5, the expression of Dlk1 found in the placenta of wild-type embryos is lost but endothelial cells are able to form normally and there is no observed defect in vascular expression of PECAM-1. This suggests that Dlk1 is not required for branching morphogenesis of the trophoblast layer that forms the placental labyrinth. The ratio of the placental weight to the embryo weight is not different in the Dlk1-null embryos, meaning that the Dlk1-null embryos and placentas are both reduced in weight to the same extent. Previous studies that induced IUGR by diet restriction in pregnant rats showed 20% reductions in both placental and embryo weights, which persists into later life (43). A recent study showed no IUGR in mice with a conditional deletion of Dlk1 in placental endothelial cells (44), suggesting that the reduced placental size in Dlk1-null animals is most likely a result of loss of Dlk1 in other embryonic cell types. It is also possible that the reduced placental size is a consequence of other defects in DLK1-dependent embryonic development, which also leads to IUGR.
An additional possibility to reconcile a smaller body size with only a mild GH content reduction is a defect in the ability of GH cells to secrete GH in response to GH secretagogues. Ultrastructural localisation of Dlk1 has been shown in the paranuclear rough endoplasmic reticulum and secretory vesicles (23), suggesting a possible involvement in hormonal secretion. However, in 6week-old male Dlk1-null mice, we found no defect in the immediate response to acute challenge by both GHRP-6 and GHRH. There is an approximately 50% reduction in plasma GH concentration 15 min after administration of both GHRP-6 and GHRH, although this only reached significance after GHRH challenge (P = 0.035). The reduction is again not as severe as that seen in GH-M2 mice with normal growth (40). The organisation of the GH cell network has previously been found to be important for sustained calcium activity in response to GHRH stimulation (33). The reduction in plasma GH concentration at the 15-min time point would be consistent with the alteration in the organisation of GH cells reported by Puertas-Avendano et al. (23) affecting the coordinated secretion of GH. In addition, the larger reduction of GH release in response to GHRH stimulation may suggest differential effects of loss of Dlk1 on signalling pathways utilised by the GHRH-receptor (adenylate cyclase activation) (7) and the GHRP-6 stimulated ghrelin receptor (mitogen-activated protein kinase pathways) (45).
Dlk1-null mice were found to have mild reductions in GH contents during all stages in both sexes throughout life, from late embryonic development to adulthood. All other pituitary hormones were found to display mild reductions in an age-and sex-specific manner. It is possible that there is an interaction between the pituitary hormone deficiencies leading to reduced growth, especially from reduced TSH content. However, in mice with a similarly low TSH content (40), plasma thyroxine concentrations were not reduced, suggesting that the Dlk1 mutant may similarly not have reduced plasma thyroid hormone. The relative proportion of each hormone cell type was unchanged at 6 weeks and, because the total pituitary protein suggests no change in the total number of pituitary cells in Dlk1-null mutants, it is likely that the total number of each of the hormone cell types is unchanged. This suggests that it is the hormone content per cell that is reduced in Dlk1-null mutants. Since the Notch signalling pathway has a role in regulating hormone cell differentiation (11,12,15,46), a difference in the proportion of endocrine cells might have been expected; however, it is possible that other Notch ligands compensate for the loss of Dlk1 to allow correct specification of pituitary endocrine cells. It is unclear whether the reductions in pituitary hormone contents reflects a common role for Notch signalling in the regulation of gene expression in the different cell types of the pituitary, or whether this is a secondary effect of loss in nonpituatary tissues. However, a recent study with somatotroph specific deletion of Dlk1 with no effect on GH gene expression suggests that secondary effects from nonpituitary tissue is at least part of the cause of reduced hormone content (44).
Reduced pituitary FSH contents at all ages concurs with previous reports of reduced FSH immunoreactivity and mRNA content in the Dlk1-null mutant pituitary (23). Although we observe reductions in pituitary gonadotrophins, there is no noticeable defect in fertility or litter size, and the onset of puberty appears to be normal with mutant mice breeding normally at 6 weeks of age. Pituitary PRL content was unaffected in male Dlk1-null mice, which is consistent with the normal gene expression reported by Puertas-Avendano et al. (23), although that study reported a reduced PRL immunoreactivity. Our finding that pituitary GH content in adult animals is reduced with normal cell number is also different to the findings of Puertas-Avendano et al. (23) who found normal mRNA expression but a reduction in somatotroph cell number. The reason for these discrepancies between our measurements of hormone content and the GH expression and PRL immunoreactivity of Dlk1-null pituitaries found by Puertas-Avendano et al. (23) is unclear but may be a result of differences in the background strain used or age of mice examined.
Dlk1 has been previously shown to colocalise in mostly GH cells and only a small proportion of other hormone-producing cell types (22,23), which we have confirmed by immunohistochemistry. This would suggest that the effects of loss of Dlk1 on non-GH pituitary cells is indirect and may result from paracrine effects, either through Dlk1 itself or a secondary factor secreted by GH cells, on the other cell types. Dlk1 expression has also been described in the putative pituitary progenitor population expressing Sox2 and Sox9 (24,25), although, in our studies, the majority of Sox2-positive cells are not Dlk1-positive, suggesting that only a subset of Sox2-positive cells express Dlk1. Because Notch signalling has been implicated in the regulation of endocrine differentiation during development (11,12,15,46), an alteration in the proliferation and differentiation of these progenitor cells could also lead to the pituitary phenotype of Dlk1-null mice. Although there is a complete loss of Dlk1 expression beginning in the embryo, we see no effect on the specification of cells into hormone-producing cell types in the developed pituitary gland, possibly suggesting a functionally redundant role of Dlk1 during the development of the developing Rathke's pouch with other Notch ligands. Because the model used in the present study leads to a global loss of Dlk1, however, it is also possible that the changes in all pituitary hormones (including GH) may be secondary to the loss of Dlk1 in other tissues. For example, the altered serum leptin found by Puertas-Avendano et al. (23) may have an impact on all hormones in the pituitary that could vary with age and sex. It is also possible that the alterations in embryonic development, demonstrated by altered embryo size and pituitary GH content, could result in altered programming of the pituitary hormone axes.
The data reported in the present study have shown an effect of loss of Dlk1 on all pituitary hormone axes, although without overt changes to normal physiological function. These minor pituitary phenotypes are likely to be secondary to loss of Dlk1 elsewhere in the body. These Dlk1-null mice may reflect human pathologies lacking expression of Dlk1 and suggest a minor contribution of the pituitary to pathological conditions, having implications in clinical considerations and treatments of human Dlk1-null conditions. This is supported by the description of a model with a conditional loss of Dlk1 in somatotrophs, which had no growth phenotype or change in GH gene expression (44). A conditional loss of Dlk1 in placental endothelial (44) also does not recapitulate the growth defects observed in the present study. Further work using this conditional allele of Dlk1 could prove crucial for unravelling the mechanisms that lead to the primary causes of the pituitary deficiencies.