|Home | About | Journals | Submit | Contact Us | Français|
Pituitary adenomas are classified into functioning and nonfunctioning (silent) tumors on the basis of hormone secretion. However, the mechanism of tumorigenesis and the cell of origin for pituitary adenoma subtypes remain to be elucidated. Employing a tamoxifen-inducible mouse model, we demonstrate that a novel postnatal Pax7+ progenitor cell population in the pituitary gland gives rise to silent corticotroph macro-adenomas when the retinoblastoma tumor suppressor is conditionally deleted. While Pax transcriptional factors are critical for embryonic patterning as well as postnatal stem cell renewal for many organs, we have discovered that Pax7 marks a restricted cell population in the postnatal pituitary intermediate lobe. This Pax7+ early progenitor cell population is overlapping but ontologically downstream of the Nestin+ pituitary stem cell population, yet upstream of another newly discovered Myf6+ late progenitor cell population. Interestingly, the Pax7+ progenitor cell population is evolutionarily conserved in primates and humans, and Pax7 expression is maintained not only in murine tumors but also in human functioning and silent corticotropinomas. Taken together, our results strongly suggest that human silent corticotroph adenomas may in fact arise from a Pax7 lineage of the intermediate lobe, a region of the human pituitary bearing closer scientific interest as a reservoir of pituitary progenitor cells.
The pituitary gland is central to the functions of other endocrine glands and their target tissue as a master regulator of the hypothalamus-pituitary gland-adrenal gland axis and the hypothalamus-pituitary gland-gonadal axis. Pituitary hormones such as follicle-stimulating hormone (FSH) and adrenocorticotropic hormone (ACTH), melanotroph-stimulating hormone (αMSH), and oxytocin are secreted from the anterior (AL), intermediate (IL), and posterior (PL) lobes in mammals, respectively. While well developed in most mammals, the intermediate lobe of the pituitary gland has been previously thought to be only a vestigial remnant in humans.1
Pituitary adenomas account for 10% of intracranial tumors and often cause neurological symptoms such as headache and visual impairment because of compression of the dura matter or optic chiasm.2-5 Pituitary adenomas are divided into 7 histological subtypes, and patients exhibit different clinical symptoms by subtype.6 Ten to 15% of tumors are ACTH-producing adenomas, which can cause Cushing’s disease. By contrast, silent corticotroph adenoma (SCA), which is categorized as a nonfunctioning adenoma, is defined as a tumor with positive immunoreactivity for ACTH without any signs or symptoms of Cushing’s disease, whereby plasma ACTH level is usually normal.7,8 SCA is assumed to develop from corticotroph cells in the AL or IL by the histological analysis that has been performed. However, the developmental mechanism(s) and cellular origin of SCA remain to be fully elucidated.
During embryogenesis, hormone-producing cells in the anterior pituitary are thought to arise from progenitors in Rathke’s pouch that terminally differentiate into 6 hormone-producing cells under the direction of a complex of growth factors and morphogens.9 Cell lineage commitment to pro- opiomelanocortin-expressing cells such as corticotrophs and melanotrophs is dependent on the T-box transcription factor, T-pit/Tbx19, whereas Hes1 expression is necessary for melanotroph differentiation.10-13 Recent studies have shown the presence of pituitary stem/progenitor cells in the postnatal mouse pituitary gland and their ability to terminally differentiate into hormone-producing cells in vitro and in vivo.14,15 However, whether the same transcriptional regulation found during embryogenesis is necessary for the terminal differentiation of stem/progenitor cells in postnatal pituitary gland is still unclear.
In this article, we demonstrate the presence of a novel progenitor cell population expressing Pax7 with differentiation-dependent expression of the stem cell marker Nestin in adult mouse pituitary gland. This cell population gives rise to the ACTH+ cell lineage and other cell lineages in both the AL and IL. We also have uncovered that a Pax7+ lineage in the lumen (the cystic cleft-like spaces between the IL and AL) and/or IL gives rise to SCA in mice when Rb loss is induced specifically in Pax7+ cells. The characterization of this postnatal vestigial progenitor cell population is presented in the context of its neoplastic potential.
Pax transcriptional factors are critical for embryonic patterning and postnatal stem cell renewal of many organs, including eye and muscle.16-18 In this study, we are the first to demonstrate that Pax7 is expressed in adult pituitary gland at a level comparable to the known expression in adult skeletal muscle (Fig. 1A). In the latter, Pax7 is known to be specifically expressed in quiescent and newly activated satellite cells and plays a critical role in maintaining this tissue-specific stem cell population.18,19 We therefore speculated that Pax7 might play a similar role in the maintenance of a pituitary-specific stem cell population. To address this possibility, we sought to identify the cell population expressing Pax7 in young adult pituitary gland by immunohistochemistry in postnatal day 30 (P30) mice. Pax7-expressing (Pax7+) cells were localized throughout IL, yet Pax7 expression was entirely absent in PL and AL (Fig. 1B). Strikingly, the majority of cells in the IL expressed Pax7 (78%; Fig. 1C). In addition, some Pax7+ cells were present in the lumen (cleft) margin (Fig. 1B, arrowhead). The frequency of Pax7+ cells in the pituitary and their site restriction in the IL led us to investigate whether Pax7+ cells are endocrine cells. Given that cells of the IL have been reported to be melanotrophs, which can be ACTH-immunoreactive, we carried out immunohistochemistry for Pax7 and ACTH on young adult pituitary gland (P30). ACTH-positive (ACTH+) cells were localized in both the IL and AL, and about 60% of ACTH+ cells in the IL also expressed Pax7. Pax7+/ACTH+ cells were not detected in the AL (Fig. 1D,,EE).
Developmentally, pituitary organogenesis begins at embryonic day 9 (E9), and all hormone-producing cells are thought to be derived from Rathke’s pouch.20 However, the development of the pituitary does not stop at birth; while the pituitary gland of newborn animals has a full set of terminally differentiated hormone-producing cells, the size of the gland dramatically increases after birth via proliferation of hormone-producing cells.21 In our studies, we observed that proliferating Ki67+ cells are present in young adult pituitary gland (age 4 weeks) and that some ACTH+ cells in the IL are Ki67+ (Fig. 1F, arrowhead in upper panel). Furthermore, about 13% of Pax7+ cells in the IL were Ki67+ (Fig. 1F,,G),G), suggesting that Pax7+ cells in the IL contribute to postnatal pituitary growth.
Adult pituitary is composed of both endocrine and nonendocrine cell populations (examples of the latter include folliculo-stellate cells and side population [SP] cells).22,23 In addition, stem/progenitor cell populations exist in adult pituitary gland, which express Nestin and/or Sox2. These cells are located at the margin of the lumen (cleft) and contribute to repopulation of all hormone-producing cells after birth.14,15 Because some Pax7+ cells were noticed nearby the lumen, as shown in Figure 1B (arrowhead), we examined whether Pax7+ cells in the young adult pituitary gland express Nestin by immunohistochemistry. Although the majority of pituitary stem/progenitor cells (Nestin+) localized to the margin of the lumen (cleft) were Pax7− (Fig. 2A, arrowhead), Pax7+/Nestin+ cells were present only on the IL side of the lumen (17% of Nestin+) (Fig. 2A, arrow).
These results led us to hypothesize that Pax7+/Nestin+ cells located nearby the lumen are at a more advanced stage of differentiation than stem cells. Recently, Gleiberman et al.14 reported that isolated pituitary stem cells can self-renew and differentiate into hormone-producing cell lineages in vitro. To test our hypothesis, pituitary stem cells were isolated from 4-week-old mice and examined by immunocytochemistry for Nestin and Pax7 expression at day 0 and day 3 after plating. Pax7 expression was never detected in Nestin+ cells at day 0 (data not shown) or day 3 (Fig. 2B). This result suggests that Pax7+ cells are a different cell population than stem cells in the postnatal pituitary gland. Because isolated pituitary stem cells can be induced to undergo terminal differentiation to endocrine cells in differentiation media,14 we examined whether Pax7 expression is induced during differentiation. At day 7 after cell plating, culture media were changed to the differentiation media, and cells were cultured for an additional 7 days. ACTH-producing cells were observed at day 14 after plating (Fig. 2D), although these cells were never seen at day 0 and day 3 (data not shown), indicating that isolated pituitary stem cells were able to differentiate into endocrine cells, consistent with the previous study.14 Interestingly, Pax7+ cells were also detected in the cultures exposed to differentiation media for 7 days (Fig. 2C, arrowhead). The majority of these Pax7+ cells were Nestin+ (Pax7+/Nestin+), and a minority of cells were Nestin− (Pax7+/Nestin−) (Fig. 2C). Furthermore, we detected Nestin−/Pax7+/ACTH+ cells at day 7 under differentiation conditions (Fig. 2D, arrowhead). Taken together, these results suggest that Pax7+ progenitor cells in the IL of the postnatal pituitary gland are derived from pituitary stem cells and that Pax7 expression is sustained, at least transiently, in terminally differentiated ACTH+ cells.
Tamoxifen-inducible CreER systems allow temporal and spatial control of genetic modifications and gene expression in vivo.24 Recently, we reported a Pax7CreER (Pax7CreERp/WT) mouse line as a tool for lineage tracing of Pax7-expressing muscle stem cells.25 To trace the fate of Pax7+ cells in postnatal pituitary gland, we used Pax7CreERp/WT Rosa26LUSEAPm/WT mice for which Pax7+ cells and their progeny are marked with luciferase.25 Immunohistochemistry for luciferase and ACTH was performed on pituitary glands collected 5 weeks after tamoxifen injection (9 weeks old). Luciferase immunoreactive cells were detected not only in the IL but also in the AL (none were detected in the PL). All luciferase+ cells in the IL expressed ACTH (Fig. 2E). Interestingly, some luciferase+ cells in the AL were also ACTH+ (Fig. 2F, arrowhead). Because Pax7 expression is not detected in the AL, these luciferase+ cells of the AL are assumed to have migrated from the IL or its lumen margin.
Basic helix-loop-helix (bHLH) transcriptional factors, which share homology within a basic domain and helix-loop-helix motif, regulate developing organogenesis and postnatal tissue formation. NeuroD1, which is a class B bHLH factor, activates proopiomelanocortin (POMC) transcription by binding to an E-box in the promoter of POMC, thereby regulating corticotroph terminal differentiation.26,27 Myf6, also called MRF4 or Herculin, is expressed in skeletal muscle and is associated with myogenesis as are other MyoD family members such as MyoD, Myf5, and Myogenin.28-31 Here, we demonstrate that Myf6 is expressed in adult pituitary gland. Myf6-expressing (Myf6+) cells in the pituitary gland were located in both the IL and AL but not in the PL (Fig. 3A, arrows), accounting for 56% and 38% of total cells in these lobes, respectively (Fig. 3A). We also confirmed Myf6 gene expression in both mouse and human pituitary gland by RT-PCR (Suppl. Fig. S3A). In contrast, the muscle-specific bHLH transcriptional factors, MyoD and Myogenin, were never detected in pituitary gland (data not shown). Using immunohistochemistry to characterize the Myf6+ cell population in pituitary gland, at least 3 populations of Myf6+ cells were observed based on Myf6 and ACTH expression (Fig. 3B). Because the majority of cells in the IL express Pax7 (Fig. 1), we suspected that Myf6+ cells in the IL could be the same cell population as Pax7+ cells. Immunohistochemistry for Myf6 and Pax7 demonstrated that all of the Myf6+ cells in the IL expressed Pax7 (Fig. 3C). However, 29% of Pax7+ cells were negative for Myf6 (Fig. 3C, arrow and graph), suggesting that a portion of Pax7+ cells in the IL are a different cell population from Myf6+ cells. In addition, we carried out lineage tracing of Myf6+ cells using a newly developed Myf6CreERp/WT mouse line (to be reported elsewhere) in combination with the Rosa26tm(EYFP)Cos/WT reporter to examine whether cellular distribution of the Myf6+ population in the pituitary gland is altered with aging. Surprisingly, Myf6 progeny (YFP+ cells) were restricted to the IL of the pituitary gland in 12-month-old mice yet were never seen in the AL or PL (Fig. 3D). Furthermore, Myf6 progeny in the IL expressed Pax7 and ACTH (Fig. 3D) but not all cells. These results suggest that the Myf6 lineage partially overlaps with the Pax7 lineage but that Myf6 lineage cells may fail to migrate from the IL to the AL with advancing age.
Although the mechanisms underlying human pituitary tumorigenesis are unclear, previous careful studies in the mouse have established that loss of heterozygosity (LOH) of Rb1 causes tumors in the IL with high penetrance, as well as a number of other endocrine tumors.32,33 Biallelic, melanotroph-specific Rb inactivation mediated by the Flp-frt DNA recombination system has also been shown to cause tumorigenesis in the IL at earlier onset than Rb haploinsufficiency.34 So far, however, no reports exist for the consequences of postnatal induction of biallelic Rb deficiency. After the first 30 days of life, melanotroph renewal has been shown to be severely curtailed.35 However, we speculated that tumor susceptibility might still be high in the Pax7+ cells, which we found to be Ki67+ (Fig. 1F,,G).G). In this study using tamoxifen-inducible Pax7-CreER mice, we examined whether tumors developed in the postnatal pituitary gland when Pax7+ cells in the IL undergo biallelic loss of Rb. We also conducted parallel studies using Myf6-Cre (Myf6ICNm/WT) mice, for which Cre-recombinase is expressed in Myf6+ cells from embryogenesis onward25,36 to study if Rb loss in the Myf6+ lineage also results in pituitary tumorigenesis.
To induce Rb loss in the Pax7+ cells in postnatal pituitary gland, Pax7CreERp/WT mice were mated to Rbflox/flox mice,34 and tamoxifen was intraperitoneally injected into Pax7CreERp/WT Rbflox/flox mice at P30. Four weeks after tamoxifen injection, we performed immunohistochemistry for Ki67 and Pax7 on the pituitary glands collected from Pax7CreERp/WT Rbflox/flox and age-matched control mice. About 46% of Pax7+ cells were Ki67 positive in the IL of Pax7CreERp/WT Rbflox/flox mice versus 9% for age-matched controls (Fig. 4A,,B).B). At 3 months after tamoxifen injection, many IL cells were Ki67+ and/or Pax7+, while no AL cells were Ki67+ or Pax7+ (Suppl. Fig. S1A,B). This result indicates that complete loss of Rb in Pax7+ cells of the IL rapidly enhanced proliferation. Although tamoxifen-injected Pax7CreERp/WT Rbflox/flox mice looked healthy for 6 months after injection, almost all Pax7CreERp/WT Rbflox/flox mice died by 8 months after injection in association with an enlarged cranium (Fig. 4C, n = 104). On the other hand, Myf6ICNm/WT Rbflox/flox mice had the same symptoms but only after 12 months of age (Fig. 4C, n = 18). To determine the anatomical location of tumors, we performed coronal slice magnetic resonance imaging (MRI) on coronal Pax7CreERp/WT Rbflox/flox mice (Fig. 4D) demonstrating an infratentorial mass. The presence of a pituitary macroadenoma was confirmed at necropsy for Pax7CreERp/WT Rbflox/flox Rosa26tm(EYFP)Cos/WT mice carrying the green fluorescent protein (GFP) lineage reporter (Fig. 4E). A microCT-based virtual histology scan of a Myf6ICNm/WT Rbflox/flox mouse also demonstrated a cavernous blood-filled pituitary macroadenoma from the cranial base displacing the hypothalamus and nearby structures superiorly (Fig. 4F). Immunohistochemistry of Pax7CreERp/WT Rbflox/flox and Myf6ICNm/WT Rbflox/flox tumors demonstrated ACTH expression in both tumor models (Fig. 4G), while all other hormones such as GH and PRL were negative (data not shown).
Pituitary adenomas are the most common tumors of the adenohypophysis and by definition are nonmetastasizing.37 Pituitary adenomas are classified into several subtypes such as corticotropinomas and prolactinomas based on secretion or nonsecretion of pituitary hormones.6 To characterize tumor-bearing Pax7CreERp/WT Rbflox/flox mice, we compared hormonal gene expression of tumors to the normal pituitary gland by quantitative RT-PCR (qPCR). TSH (β-subunit), GH, FSH (β-subunit), and LH (β-subunit) expression were undetectable in the tumor, while POMC expression was significantly increased in tumor mice (P < 0.05; Fig. 5A). POMC is a prohormone that is processed into biologically active ACTH by prohormone convertase 1/3 (PC1/3) and then into α-melanotropin (αMSH) and β-lipotropin by PC2.38-41 Comparison of ACTH protein expression between the tumor and normal pituitary showed that the ACTH level was significantly increased in tumors (P < 0.05), while the precursor form of ACTH (pro-ACTH) level was decreased (Fig. 5B). These results suggest that ACTH production is accelerated in tumors. On the other hand, however, serum ACTH was undetectable by Western blot analysis for both tumor and normal pituitary. These assays of both tissue and serum suggest that ACTH secretion is inhibited in tumor mice despite increased ACTH protein production (Fig. 5C). Because mice also did not exhibit Cushinoid-like symptoms, these tumors are consistent with silent corticotroph macroadenomas, which are included as nonfunctioning adenomas in humans.
At the cellular level, adenoma cells were uniformly αMSH+, ACTH+, and Pax7+ (Fig. 5D,,E).E). Approximately half of ACTH+ cells were in the proliferating state (data not shown). In addition, primary pituitary tumor cells isolated from Pax7CreERp/WT Rbflox/flox mice also expressed Pax7 as well as POMC (Suppl. Fig. S2A). We speculated that these primary cell cultures may also have value in testing therapeutic agents, as demonstrated by the proof-of-principle experiment showing that viability of primary tumor cells was decreased by short-term treatment with sorafenib, a multikinase inhibitor (Suppl. Fig. S2B).
To clarify whether the origin of ACTH+ tumor cells in tumor-bearing Pax7CreERp/WT Rbflox/flox mice was the Pax7+ lineage, we performed lineage-tracing experiments using Pax7CreERp/WT Rbflox/flox Rosa26tm(EYFP)Cos/WT mice by immunohistochemistry for ACTH and YFP in tumor sections. In these experiments, all ACTH+ cells in the tumor expressed YFP reporter protein (Fig. 5F), indicating that adenoma cells originated from Pax7-expressing progenitors in the pituitary gland.
To further molecularly characterize tumors, we performed qPCR analysis for some pituitary development- and tumor-associated genes.10,26,41-47 Both NeuroD1 and Tpit (Tbx19) are critical factors for POMC gene transcriptional regulation. NeuroD1 expression was significantly lower than age-matched control mice (P < 0.001), although Tpit expression was almost equal. Other markers except for PC2 were significantly decreased in the tumor (P < 0.001; Fig. 5G). The paucity of Galectin-3 is typical of silent corticotroph adenomas.48 Overexpression of PC2 suggests an acceleration of αMSH production in the tumor because PC2 converts POMC protein to αMSH in melanotroph cells. Thus, our data are most consistent with this murine silent corticotroph adenoma model being an ACTH-producing melanotroph tumor derived from an intermediately differentiated cell of origin within the Pax7 lineage. The possibility that this cell lineage gives rise to silent corticotroph adenomas of the pituitary in higher species is supported by the finding that Pax7+ αMSH+ cells are present in high numbers in the macaque and human intermediate lobe (Suppl. Fig. S3B). The relevance to human pituitary macroadenomas is further supported by the detection of Pax7 expression in different types of human ACTH+ adenomas, including Cushing’s disease corticotropinomas and nonfunctioning corticotropinomas (Suppl. Fig. S3C). Pax7 staining was present in 2 of 4 cases of Cushing adenomas and 5 of 9 cases of silent corticotroph adenomas that we examined.
In these studies, we have defined an unexpected Pax7 lineage for the IL of the mouse and primate pituitary. While these Pax7+ progenitor cells in the postnatal pituitary co-localize with luminal pituitary stem cells and are to an extent proliferative (13% were Ki67+), two thirds of this Pax7+ population express the mature corticotroph/melanotroph marker ACTH. Taken in sum with results of the in vitro differentiation assays of Nestin+ pituitary stem cells, this intermediate lobe Pax7+ lineage has the properties of a committed melanotroph lineage precursor with characteristics between that of a pituitary stem cell and a melanotroph. Interestingly, however, ACTH-expressing cells of this lineage can also be found to have migrated to the anterior lobe by adulthood (9 weeks old).
To our surprise, another muscle-related gene, Myf6 (MRF4, Herculin), was found to have a complementary pituitary cell lineage. Myf6+ cells were localized to both the intermediate and anterior lobes. In many cases, Myf6+ cells in the intermediate lobe were ACTH+ (51%). All Myf6+ cells in the intermediate lobe expressed Pax7, but not all of the Pax7+ cells expressed Myf6. Our studies suggest that the Myf6+ lineage is ontologically downstream of Pax7 for some but perhaps not all cells of the IL.
From the viewpoint of tumorigenesis, both Pax7+ early pituitary progenitors and Myf6+ late pituitary progenitors were found to give rise to ACTH staining (but ACTH nonsecreting) silent corticotroph macroadenomas when the tumor suppressor Rb was homozygously deleted in the first to the second months of adulthood. Tumors themselves also express αMSH, a cleavage product of ACTH, which is characteristic of a melanotroph lineage tumor.
A precedent exists for paired-type homeobox proteins to be expressed in the pituitary gland. While Pax6 is best known as a master regulator of eye development, in the pituitary gland, Pax6 is first expressed in Rathke’s pouch at E10-E12 and acts on the formation of boundary between dorsal somatotroph/lactotroph and ventral thyrotroph/gonadotroph.49,50 Although Pax6 expression is sustained in the postnatal pituitary gland, the precise function of that factor in adults is as yet undetermined.49 Pax7 is another member of the Pax family and expressed in developing muscle and brain, and Pax7 expression is maintained after birth in both tissues.51-54 Here, we show that Pax7 is expressed in adult mouse pituitary gland, consistent with a previous report that Pax7 is expressed in the precursors of the pituitary gland of the zebrafish embryo.55 Taken together with our detection of Pax7 in the macaque pituitary, Pax7 expression in the pituitary gland appears to be well conserved in the macaque and human pituitary as a precursor to a corticotroph/melanotroph population in the intermediate lobe (Suppl. Fig. S3B). A model for the Nestin, Pax7, Myf6 ontogeny of corticotroph and melanotroph cells of the intermediate lobe is given in Figure 6.
We hypothesize that Pax7 may be an important determining factor for melanotroph cells. Raetzman et al.13 have reported that Hes1, which is an inhibitory-type bHLH factor and known as a target of the Notch signaling pathway, is a key factor for melanotroph specification in the developing embryo because absence of Hes1 results in the loss of melanotroph cells as well as AL hyperplasia and increase of somatotroph cells. Because Pax7 directly regulates inhibitory bHLH Id3 transcription in quiescent muscle stem cells,56 we speculate that Pax7 may directly regulate Hes1 gene expression to induce melanotroph specification in the adult pituitary gland. Whether Notch signaling is active in these tumors or could be a target of γ-secretase inhibitors is a topic of ongoing investigation.
While both Pax7 and Myf6 lineages give rise to silent corticotroph macroadenomas, the shorter latency and higher frequency of tumors in the Pax7 lineage suggest that the potential cell of origin of these tumors in mammals is more likely to be the intermediately differentiated Pax7+/ACTH+ cell than a more differentiated Myf6+/ACTH+ cell, although both are possible. Whether a Nestin+ pituitary tumor stem cell gives rise to adenomas, particularly pituitary corticotropinomas, with any more or less susceptibility than Pax7+ cells remains to be tested. We showed that Pax7 was expressed in half or more of human corticotropinomas tested. While phenotype does not imply cell of origin (because marker expression can be lost or gained opportunistically by tumors), the expression of Pax7 in corticotropinomas does imply Pax7 may play an ongoing functional role. This potential role will be the topic of future investigation.
Multiple prior reports establish that Rb heterozygous loss in mice results in a spectrum of tumors, which include thyroid C cell carcinomas and neuroendocrine tumors as well as pituitary intermediate lobe tumors.32,35,57,58 In contrast to these studies, complete (homozygous), lineage-restricted loss of Rb in Pax7-expressing and Myf6-expressing cells rarely if ever resulted in tumorigenesis for another tissue (e.g., skeletal muscle or thyroid gland). Although mutation of the RB1 gene has not been identified in human pituitary adenomas, hypermethylation at the CpG island of the RB1 promoter region is frequent in tumor cells, resulting in loss of RB protein expression.59,60 In addition, LOH at 13q, the locus of the RB1 gene, has been identified in human pituitary adenomas.61 These reports suggest that Rb is a tumor suppressor in human pituitary, although further studies are needed. To address potential cooperative or disease-modifier mutations, we have also generated Pax7CreERp/WT Rbflox/flox mice with homozygous p53 mutation to determine if additional mutations could accelerate pituitary tumorigenesis. Tumor initiation and survival rate in the pituitary gland were not significantly different from mice without p53 mutation (data not shown). The role of other potential modifiers in this lineage remains to be explored.
Therapeutic targets for corticotroph adenoma, including SCA, are still unclear. As proof of principle, we have obtained preliminary results revealing the Raf/ERK signaling pathway as a potential therapeutic target for corticotroph adenoma in vitro (Suppl. Fig. S2B), with in vivo preclinical experiments now in progress.
Perhaps the most exciting finding from these studies is the correlation to the human disease. Indeed, by immunohistochemistry, we have found Cushing’s disease and nonfunctioning adenomas to express Pax7 (Suppl. Fig. S3C). Since the 1930s, ACTH+ αMSH+ (basophilic) cells of the pituitary have been observed to divide, exhibit hyperplasia, and even invade the posterior lobe of the pituitary.62,63 ACTH+ microadenomas of the IL also account for 10% of ACTH-producing adenomas of the pituitary.64 In what may be a related lineage, large ACTH+ tumors can silently arise from the intermediate lobe and displace the brain superiorly.62 These so-called silent corticotroph macroadenomas may in fact be better described as silent melanotroph macroadenomas. Previous related studies support this concept, in that 40% of ACTH+ microadenomas of the IL are immunoreactive for αMSH.64 We speculate that the Pax7+ lineage may also be responsible for the frequent occurrence of asymptomatic microadenomas found incidentally at autopsy,64,65 but additional studies are needed to confirm that assertion. By having defined the cell of origin for this disease and having as a result created a reproducible model, the opportunity now exists to test new, molecularly targeted therapies to serve as adjunct to surgery and/or to replace the role of radiation in treating this disease. Furthermore, our results of lineage tracing and tumorigenesis draw attention to the intermediate lobe, a region of the human pituitary bearing closer scientific interest as a reservoir of pituitary progenitor cells.
All animal procedures were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Texas Health Science Center at San Antonio (UTHSCSA). Mice were a mixed strain of SV/J129, C57Bl/6, and FVB. Conditional Rb mice were obtained from the Mouse Models of Human Cancer Consortium (NCI-Frederick, NIH). Tamoxifen-inducible Pax7-CreER and tamoxifen-independent Myf6-Cre mice were previously described.25,36 The generation and description of Myf6CreERp/WT mice will be presented as a separate study (Hosoyama et al., in preparation). Detailed genotyping protocols and animal imaging procedures are given in supplementary methods. Frozen macaque pituitary samples were provided by WFG and DLM as routinely harvested necropsy materials from the Oregon National Primate Research Center (Beaverton, OR).
To induce Cre in Pax7+ and Myf6+ cells, 200 µL of 10 mg/mL tamoxifen (Sigma-Aldrich, St. Louis, MO) suspended in corn oil was injected intraperitoneally into 4-week-old Pax7CreERp/WT Rbflox/flox, Pax7CreERp/WT Rbflox/flox Rosa26tm(EYFP)Cos, Pax7CreERp/WT Rbflox/flox Rosa26LUSEAPm/WT, Pax7CreERp/WT Rosa26LUSEAPm/WT, or Myf6CreERp/WT Rosa26tm(EYFP)Cos mice once a day for 5 days at a dose of 2 mg/20 g body weight/day. Rosa26tm(EYFP)Cos and Rosa26LUSEAPm/WT Cre/LoxP reporter mice that express eYFP or luciferase, respectively, following Cre expression have been described previously.25,66,67 Mice were euthanized by CO2 asphyxiation at 5 weeks (Pax7-CreER) and 11 months (Myf6-CreER) after tamoxifen injection, respectively, and pituitary glands were harvested for immunohistochemistry for luciferase or eYFP as a lineage-tracing marker.
Total RNA was isolated from pituitary gland and skeletal muscle of 8-week-old wild-type mice using TRIzol reagent (Promega, Madison, WI). Single-strand cDNA was synthesized from 1 µg of total RNA using the RevertAid™ M-MuLV Reverse Transcriptase (Fermentas Life Sciences, Burlington, Ontario, Canada) according to the manufacturer’s protocol. The primer set spanning exon 5 to 6 of mouse Pax7 was 5′-GCA CAG AGG ACC AAG CTC AC-3′ and 5′-TGG TGG TGG GGT AGG TAG AG-3′ (189 bp). Cycling conditions were as follows: 95°C for 3 minutes, 35 cycles of 95°C for 30 seconds, 58°C for 45 seconds, and 72°C for 1 minute, followed by 72°C for 7 minutes. PGK was used as an internal control. PGK primers were described previously.68 For qPCR, total RNA was isolated from tumor-bearing and age-matched normal pituitary glands. Expression of 6 neuroendocrine genes and 8 pituitary development and tumorigenesis-associated genes were compared between age-matched control and tumor mice by quantitative RT-PCR on a 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA). All primer sequences are described in Supplemental Table S1 and Table S2.
Histology was performed as previously described.68 The electron microscopy protocol is described in supplemental methods. For immunohistochemistry, dissected pituitary glands were fixed in 10% formalin for 30 minutes at room temperature (RT) and soaked in sucrose solution at 4°C. Tissues were then embedded in OCT compound (Sakura Finetek, Torrance, CA). For immunohistochemistry of frozen pituitary tissue sections (7 µm thick), staining was performed using the M.O.M. immunodetection kit (Vector Laboratories, Burlingame, CA).
The anti-Pax7 antibody (Developmental Studies Hybridoma Bank [DSHB]) was used at concentration of 1:50 for frozen sections or 1:5 for formalin-fixed 3-µM paraffin sections. The anti-Ki-67 (Thermo Scientific, Waltham, MA), anti-ACTH (Abcam, Cambridge, MA), and anti-Myf6 (GenWay Biotech, San Diego, CA) antibodies were used at concentration of 1:100. The anti-GFP (Chemicon, Temecula, CA), anti-ACTH (DAKO, Glostrup, Denmark), anti-luciferase (Abcam) and anti-Nestin (Abcam) antibodies were used at concentration of 1:500. The anti-αMSH antibody (Peninsula Laboratories, Torrance, CA) was used at concentration of 1:250. AlexaFluor 594 conjugated anti-mouse IgG, AlexaFluor 488 conjugated anti-chicken IgG, and AlexaFluor 594 conjugated anti-rabbit IgG (Invitrogen, Carlsbad, CA) were used as secondary antibodies. Human tumor ACTH staining was performed at the Cleveland Clinic pathology laboratory. Anti-ACTH (DAKO), anti-TSH β-subunit (DAKO), anti-FSH β-subunit (C10; DAKO), anti-GH (DAKO), anti-Prolactin (DAKO), and anti-LH (C93; DAKO) antibodies were used at concentration of 1:1000, 1:8000, 1:200, 1:2500, 1:800, and 1:320, respectively.
Cells were fixed with 4% paraformaldehyde (PFA)/phosphate-buffered saline (PBS) for 20 minutes at RT. After washing with PBS, cells were soaked in 0.1% Triton-X/PBS for 15 minutes at RT and blocked with 5% normal goat serum (NGS)/PBS for 60 minutes at RT to inhibit nonspecific binding of antibodies. Primary antibodies were applied, and cells were incubated overnight at 4°C. The primary antibodies anti-Pax7 (DSHB), anti-Nestin (Abcam), and anti-ACTH (Abcam) were used at titers of 1:50, 1:200, and 1:400, respectively.
To isolate pituitary stem cells, 4-week-old mice were sacrificed and pituitary glands were dissected. Cell isolation was performed as previously described.14 Cells were plated on poly-L-lysine and fibronectin-coated culture plates and stained with anti-Nestin (Abcam) and anti-Pax7 (DSHB) antibodies at day 0 and day 3 after plating. To induce terminal differentiation of isolated pituitary stem cells, cells were incubated in differentiation media for 7 days.14 At day 14 after plating, cells were stained with anti-ACTH (Abcam), Nestin, and Pax7 antibodies as described above.
Protein samples were collected from tumor-bearing pituitary and age-matched control pituitary glands. For comparison of blood ACTH level, serum samples were collected from 2 to 4 pm from tumor mice and age-matched control mice. Then, 20 µg of protein was applied on acrylamide gel and transferred to a PVDF membrane. Membrane was soaked in 5% skim milk/PBS for 60 minutes at RT. Anti-ACTH antibody (1:5000, Abcam) was used as primary antibody. For phospho-proteins, anti-ERK1/2 (Cell Signaling Technology, Beverly, MA) and anti-phospho-ERK1/2 (Cell Signaling Technology) antibodies were used at a concentration of 1:1000.
Kaplan-Meier survival analysis was generated using Systat12 (Creation Engine, Inc., Mountain View, CA). For comparison between wild-type and tumor assay measurements, we used the Student t test.
The anti-Pax7 monoclonal antibody was obtained from the Developmental Studies Hybridoma Bank developed under the NICHD and maintained by the University of Iowa. These studies were funded in part by a grant from the National Brain Tumor Society. The authors thank Z. David Sharp, Damon Herbert, and Kris Vogel for thoughtful comments in the preparation of this manuscript. They also thank David Rodriguez for graphic design assistance, David M. Weinstein of the Scientific Computing and Imaging Institute for assistance in rendering virtual histology images using SCIRun™ (freeware made available by NIH/NCRR 5P41RR012553), and James McMahon for assistance with electron microscopy.
The authors declared no potential conflicts of interests with respect to the authorship and/or publication of this article except to state that CK is co-founder of Numira Biosciences, which commercializes MicroCT-based Virtual Histology.
This study was supported by the start-up funds to CK from UTHSCSA. This work was also partially supported by grants from the National Pediatric Brain Tumor Society and the St. Baldrick’s Foundation to CK.
Supplementary material for this article is available on the Genes & Cancer Web site at http://ganc.sagepub.com/supplemental.