2016, 63 (8), 669-680

REVIEW Making pituitary hormone-producing cells in a dish

Hidetaka Suga

Department of Endocrinology and Diabetes, Nagoya University Hospital, Nagoya 466-8550, Japan

Abstract. The hypothalamic-pituitary system is essential for maintaining life and controlling systemic homeostasis. The functional disorder makes patients suffer from various symptoms all their lives. Pluripotent stem cells, such as embryonic stem (ES) cells and induced pluripotent stem (iPS) cells, differentiate into neuroectodermal progenitors when cultured as floating aggregates under serum-free conditions. Recent results have shown that strict removal of exogenous patterning factors during the early differentiation period induces rostral hypothalamic-like progenitors from mouse ES cells. The use of -free, chemically defined medium was critical for this induction. The ES -derived hypothalamic-like progenitors generated rostral-dorsal hypothalamic neurons, in particular magnocellular vasopressinergic neurons. We subsequently reported self-formation of adenohypophysis in three-dimensional floating cultures of mouse ES cells. The ES cell aggregates were stimulated to differentiate into both non-neural head ectoderm and hypothalamic neuroectoderm in adjacent layers. Self-organization of Rathke’s pouch-like structures occurred at the interface of the two epithelia in vitro. Various pituitary endocrine cells including corticotrophs and somatotrophs were subsequently produced from the Rathke’s pouch-like structures. The induced corticotrophs efficiently secreted ACTH in response to CRH. Furthermore, when engrafted in vivo, these cells rescued systemic levels in hypopituitary mice. Our latest study aimed to prepare hypothalamic and pituitary tissues from human pluripotent stem cells. We succeeded in establishing the differentiation method using human ES/iPS cells. The culture method is characterized by replication of stepwise embryonic differentiation. Therefore, these methods could potentially be used as developmental and disease models, as well as for future regenerative medicine.

Key words: , Pituitary, Embryonic stem cells, Differentiation, SFEBq

THE HYPOTHALAMUS and adenohypophysis In addition to somatic stem cells, studies have maintain physiological homeostasis by controlling the focused on embryonic stem (ES) cells and induced . A collection of studies exploring pluripotent stem (iPS) cells. These pluripotent stem their development and function has shown that they are cells exhibit self-renewal properties and pluripotent essential for the regulation of vital functions. However, differentiation abilities. Therefore, they have attracted their regeneration remains largely unclear. attention as a cell source for differentiated tissues in Recently, somatic stem cells have been recognized clinical applications. as a major source for tissue maintenance and regen- eration. Also in the adenohypophysis, the existence of A Need for Hypothalamus and somatic stem cells was reported [1]. Subsequent stud- Adenohypophysis Regenerative Medicine ies have discussed their functions during early postna- tal pituitary maturation [2-6], after pituitary damage The hypothalamus and adenohypophysis are located [7-9], and in pituitary tumorigenesis [10-13]. in adjacent regions, connected with portal vein. They

Submitted May 8, 2016; Accepted May 10, 2016 as EJ16-0232 Otp, orthopedia homeobox; Released online in J-STAGE as advance publication May 28, 2016 Brn2, Pou3f2 POU domain, class 3, transcription factor 2; Correspondence to: Hidetaka Suga, Department of Endocrinology Pitx, paired-like homeodomain; and Diabetes, Nagoya University Hospital, 65 Tsurumai-cho, Lhx3, LIM homeobox protein 3; Showa-ku, Nagoya 466-8550, Japan. Tbx19, T-box 19; E-mail: [email protected] DAPT, N-[N-(3,5-Difluorophenacetyl-L-alanyl)]-(S)-phenylglycine Abbreviations: Rax, retina and anterior neural fold homeobox; t-butyl ester; Six3, sine oculis-related homeobox 3; Pit1, Pou1f1 POU domain, class 1, transcription factor 1; Vax1, ventral anterior homeobox 1; BIO, (2’Z,3’E)-6-Bromoindirubin-3′-oxime ©The Japan Endocrine Society 670 Suga coordinate as the center for the endocrine systems. In Three-dimensional ES Cell Culture case of their dysfunction, patients suffer from vari- ous systemic symptoms. Current treatment consists Organ formation during embryogenesis consists of of hormone replacement therapy, but various factors complicated processes that involve various local inter- can complicate its proper dose in some cases. Drug actions between different tissues or cells. Despite this administration cannot precisely mimic the circadian complexity, organogenesis can be modeled in vitro. or stress-induced changes of hormone requirements. Our colleagues established a three-dimensional cul- For example, we have reported that some patients with ture method for ES cells called “serum-free culture of central diabetes insipidus show unstable serum Na lev- embryoid body-like aggregates with quick re-aggre- els, resulting in a poor prognosis [14]. This instability gation (SFEBq)” [23, 24]. The culture method is seems to be caused by the lack of positive and negative quite simple. First of all, the quality of maintenance control systems, which is characteristic of hormone- for undifferentiated ES cells is quite important. For producing cells, in the conventional hormone-replace- SFEBq culture, maintained ES cells are dissociated to ment therapy. As for hypopituitarism, it has been single cells in trypsin or something similar. The cells reported that adrenal crisis occurs in a substantial pro- are then quickly aggregated using low-cell adhesion portion of hypopituitary patients, and adrenal crisis- 96-well plates in differentiation medium suitable for associated mortality is not negligible, even in educated each differentiation purpose (Fig. 1). patients [15]. Furthermore, ACTH-dependent adre- This culture method is appropriate for induction nal insufficiency, as well as high-dose hydrocortisone of various ectodermal derivatives from ES cells. In treatment, serves as a predictor for acromegaly-associ- SEFBq cultures, the ES cell aggregates exhibit self- ated mortality [16, 17]. Taken together, there are many organization [25] and spontaneous formation of a prospects for pituitary regenerative medicine. highly ordered structure or patterning. This float- ing culture has revealed intrinsic programs that drive Mouse Embryonic Stem Cells locally autonomous modes of organogenesis and homeostasis. Using the SFEBq method, mesence- The establishment of mouse pluripotent ES cells phalic dopamine neurons [26, 27], cortex neurons [24, significantly contributed to the advancement of biol- 28, 29], the optic cup [22, 30, 31], cerebellar neurons ogy and medicine. In 1981, Evans and Kaufman suc- [32], and hippocampal neurons [33] have been gener- cessfully established mouse ES cells from the inner ated from mouse and human ES cells. cell mass of mouse blastocyst-stage embryos [18]. Various knockout and knock-in mice have been estab- Induction of Hypothalamic Neurons lished using genetically modified ES cells, which from Mouse ES Cells have contributed to our understanding of gene func- tions [19, 20]. Using SFEBq cultures, hypothalamic neurons, such There are two reasons for the use of mouse ES cells as vasopressin-positive neurons, have been induced as a first step, rather than human ES cells, in our recent from mouse ES cells [34]. The differentiation occurs studies. One reason is the short developmental period; efficiently when the ES cell aggregates are cultured the duration of mouse fetal development is about 20 in growth factor-free, chemically defined medium days, which is much shorter than the 300 days of human (gfCDM). Strict removal of exogenous patterning development. Therefore, mouse ES cells are suitable for factors during early differentiation steps induces effi- establishing novel differentiation methods with numer- cient generation of rostral hypothalamic-like progeni- ous trial-and-error processes. Another reason is the sim- tors (Rax(+)/Six3(+)/Vax1(+); these combinations are ilarity between mouse and human cells. For example, characteristic for hypothalamic precursors) in mouse the retinal differentiation method from human ES cells ES cell aggregates. The use of gfCDM is critical. [21] was established based on a previous report using For example, even the presence of exogenous insulin, mouse ES cells [22]. Although the two inducing cul- which is commonly used in cell culture, strongly inhib- ture methods considerably differ, their key principles are its differentiation via the Akt-dependent pathway. The similar. The fundamental processes of mouse ES cells ES cell-derived hypothalamic progenitors generate appear to be applicable to human ES cells. Otp(+)/Brn2(+) neuronal precursors (characteristic of In vitro induction of pituitary 671

Fig. 1 SFEBq method Dissociated ES/iPS cells are distributed into the low-cell-adhesive well plate. Cells are quickly aggregated. Using inducing signals in the culture medium, the aggregate differentiates into ectodermal tissue. rostral-dorsal hypothalamic neurons) and subsequent PRL, TSH, LH, and FSH. The posterior pituitary magnocellular vasopressinergic neurons that release gland consists of axons and terminals of hypothalamic vasopressin upon stimulation. Additionally, differen- neurons, i.e., vasopressin and oxytocin neurons. The tiation markers of rostral-“ventral” hypothalamic pre- development of the adenohypophysis is a complex pro- cursors and neurons have been induced from ES cell- cess. During early development, the adenohypophysis derived Rax(+) progenitors by treatment with Sonic anlage originates as a placode in the non-neural ecto- Hedgehog (Shh). derm adjacent to the anterior neural plate (Fig. 2a). Thus, in the absence of exogenous growth factors Both the adenohypophysis placode and hypothalamic in the medium, ES cell-derived neuroectodermal cells anlage interact with each other. Accordingly, the thick- spontaneously differentiated into rostral (particularly ened placode invaginates and subsequently detaches rostral-dorsal) hypothalamic-like progenitors, which from the oral ectoderm to form a hollowed vesicle generate characteristic hypothalamic neuroendocrine termed “Rathke’s pouch” [35] (Fig. 2b). The molecu- neurons in a stepwise fashion, as observed in vivo. lar nature of this local inductive interaction during this These findings indicated that, instead of the addition of initial phase of pituitary formation has been intensively inductive signals, minimization of exogenous pattern- investigated, but still remains elusive. Among them, ing signaling played a key role in rostral hypothalamic FGF and BMP signals appear to be involved as impor- specification of neural progenitors derived from plu- tant factors [36, 37]. ripotent cells. This work also showed that the default fate of mouse ES cells is the rostral hypothalamus [34]. Two-layer Formation in Vitro is the First Step of Adenohypohysis Differentiation Embryology Based on the knowledge of embryology, we next We next tried to establish an in vitro differentia- tried to establish an in vitro differentiation method for tion method for the . The key of our the anterior pituitary [38]. It is known that Rathke’s SFEBq method is replicating the embryonic differenti- pouch is formed as a result of interactions between the ation environment. Therefore, the developmental biol- hypothalamus and neighboring oral ectoderm [35]. To ogy of pituitary is important. recapitulate these embryonic pituitary developmental The adenohypophysis, which corresponds to the processes, we co-induced these two tissues simultane- anterior pituitary gland, contains several types of endo- ously within one ES cell aggregate. crine cells that secrete factors including ACTH, GH, Previous results have shown hypothalamic differ- 672 Suga

Fig. 2 Diagram of mouse pituitary development a: Dorsal view of neural plate and placodes. b: Sagittal view of pituitary embryogenesis. (from Suga et al. (2011) Nature 480: 57-62. [38] modified.) entiation from mouse ES cells [34]. Mouse ES cells [38]. Quantitative polymerase chain reaction analyses can be induced to differentiate into hypothalamic revealed significantly higher internal BMP4 expres- cells when cultured as floating aggregates using the sion in LCA aggregates [38]. Moreover, Koehler et SFEBq method with gfCDM. Therefore, the present al. succeeded in differentiating the otic placode (Fig. study used some technical modifications to co-induce 2a) [39], which belongs to the head and oral ectoderm, oral ectodermal differentiation in addition to hypotha- following BMP treatment of mouse ES cells, which lamic differentiation. supports the reliability and robustness of this strategy. We attempted to slightly shift the positional infor- Our recent study showed that very low concentrations mation so that the oral ectoderm co-existed with of exogenous BMP4 treatment facilitated differentia- hypothalamic tissues [38]. As shown in Fig. 2a, the tion into non-neural ectoderms, which contained not oral ectoderm is generated from the rostral and mid- only pituitary primordium, but also dental germs [40]. line region adjacent to the hypothalamic resion in Taken together, appropriate BMP4 signal appears to be the mouse embryo. Therefore, the rostral and mid- important for head ectoderm induction [41-43]. line shifting information was theoretically relevant for mouse ES cell aggregates in the SFEBq culture. We Self-formation of Rathke’s Pouch tested many culture conditions known to affect early ectodermal patterning. We finally identified two condi- In the developing embryo, Rathke’s pouch forms at tions that efficiently induced oral ectoderm. One con- the midline of the head ectoderm. Shh is expressed dition was the addition of bone morphogenetic protein in the ventral diencephalon and oral ectoderm but is 4 (BMP4). However, treatment with 0.5 µM BMP4 excluded from the invaginating Rathke’s pouch [35, strongly inhibited hypothalamic neuron differentia- 44]. Rathke’s pouch receives Shh signals from neigh- tion, instead of inducing oral ectodermal differentia- boring tissues in vivo, and Shh is known to provide tion. The other condition was high-density cell aggre- positional information to adjust towards the midline gation (10,000 cells per aggregate instead of 3,000 in [35]. Therefore, we added smoothened agonist (SAG) SFEBq culture), which we refer to as large cell aggre- as a strong Shh signal to the differentiation medium of gation (LCA) (Fig. 3a). In the LCA culture, both the mouse ES cell aggregates in vitro. On day 13, mul- oral ectoderm (Pitx1/2+) and hypothalamic tissues co- tiple oval structures formed in the SAG-treated LCA existed within one aggregate (Fig. 3b). SFEBq aggregates (Fig. 3c). The vesicles were situ- LCA culture allows for the formation of oral ecto- ated between the oral ectoderm and hypothalamic neu- derm on the surface of mouse ES cell rons. Lim3 (formal gene name is Lhx3) expression aggregates, as well as hypothalamic neural tissue in indicated that the vesicles had similar characteristics the inner layer adjacent to the oral ectoderm (Fig. 3b). to Rathke’s pouch. These Lim3+ tissues appeared as Treatment with a BMP4 antagonist, dorsomorphin, has a thick epithelium on the surface, which then invagi- been shown to suppress the generation of oral ectoderm nated and finally formed hollowed vesicles. The length In vitro induction of pituitary 673

Fig. 3 In vitro differentiation into anterior pituitary from mouse ES cells a: Diagram of SFEBq. b: Two-layer formation in LCA aggregates. c: Self-formation of Rathke’s pouches. d: Subsequent generation of ACTH+ cells. (from Suga et al. (2011) Nature 480: 57-62. [38] modified.) ** P < 0.01. of the major axis was about 200 µm, which is almost thelium on the surface of the ES cell aggregate, which equal to the size of the embryonic Rathke’s pouch. is reminiscent of the Vax1 knockout mouse [45]. A sec- Interactions between oral ectoderm and hypotha- ond Rathke’s pouch develops in addition to the ortho- lamic neurons appear to be critically important for in topic anlage in the Vax1 knockout mouse. Ectopic vitro induction of Rathke’s pouch. Neither isolated expression of FGF10, which is expressed in the infun- surface ectoderm alone, nor isolated hypothalamic dibulum and implicated in pituitary induction, is also tissues alone, formed Lim3+ pouches. Only in cases detected in the hypothalamic neuroepithelium over- where the two divided components are re-assembled, lying the second pouch. Thus, Vax1 likely limits the Lim3+ expression recovered to some extent [38]. hypothalamic neuroepithelium area that generates pitu- These findings demonstrate self-formation of itary-inducing signals. Indeed, Vax1 expression in vivo Rathke’s pouch in mouse ES cell aggregates. It has is eliminated near the infundibulum, which has induc- also been shown that Rathke’s pouch forms even with- ing activity for pituitary development. In the mouse out mesenchymal cells, because this model contains ES aggregates used for pituitary differentiation in the only ectodermal cells. present study, Vax1-positive cells did not exist in the Interestingly, a single aggregate often contains sev- hypothalamic area. Conversely, Wataya’s aggregate eral pouches, whereas there is usually only one pouch for hypothalamic differentiation [34] has been shown in the embryo [38]. This finding suggests that sev- to contain Vax1-positive cells. We speculate that pre- eral morphogenetic fields for pituitary placodes can be cise positioning in the hypothalamus slightly shifts as a independently generated within the oral ectoderm epi- result of BMP4 and Shh signals. 674 Suga

Differentiation into Hormone-producing peripheral blood. ACTH secretion from the pituitary Endocrine Cells gland is negatively regulated by the downstream glu- cocorticoid hormone in vivo. Consistent with this con- During pituitary development in the embryo, Lim3+ trol principle, in vitro ACTH secretion as a result of pituitary progenitors commit to several lineages [46], CRH stimulation was suppressed by glucocorticoid i.e., corticotroph, somatotroph, lactotroph, thyrotroph, pre-treatment (Fig. 4b). gonadotroph, and melanotroph lineages. Among them, Similar to in vivo endocrine systems, these data the ACTH-producing corticotroph lineage expresses demonstrate that mouse ES cell-derived ACTH+ cells the transcription factor Tbx19 prior to ACTH expres- respond to both positive and negative regulators. sion. It is known that Notch signaling inhibits Tbx19 These hormonal responses to surrounding regulators expression [47-49]. Therefore, we evaluated the effect are indispensable for homeostasis. For this reason, the of the Notch inhibitor DAPT. As a result, DAPT treat- generation of anterior pituitary tissue that retains regu- ment increased Tbx19 expression in SAG-treated LCA latory hormonal control in vitro is an important step for SFEBq aggregates. A substantial number of ACTH+ the development of cell transplantation therapies for cells appeared in the Tbx19+ lesion (Fig. 3d). Without pituitary diseases. Furthermore, we suggest that the DAPT treatment, corticotroph differentiation efficiency endocrine organoid formed in this three-dimensional was decreased, and other lineages were not detected. culture condition might better reflect thein vivo micro- Previous reports have shown that canonical Wnt sig- environment. Such approaches may be beneficial for naling promotes Pit1 expression [50-52]. Consistent producing other functionally mature endocrine tissues. with this finding, treatment with the Wnt agonist BIO increased Pit1 expression, resulting in subsequent GH+ Effect of Transplantation and PRL+ cell differentiation. into Hypophysectomized Model Animals Head mesenchyme has been suggested to promote pituitary development in vivo [53]. Therefore, we Finally, we evaluated the transplantation effect of applied conditioned medium from PA6 stromal cells to the induced ACTH+ cells. Because of technical dif- SAG-treated LCA SFEBq aggregates. As a result, we ficulties, we chose ectopic transplantation into the successfully induced LH-positive, FSH-positive, and kidney subcapsule (Fig. 4c), instead of orthotopic TSH-positive cells. Further investigation is necessary transplantation into the sella turcica. At 1 week after to identify factors in the PA6-conditioned medium. transplantation, blood ACTH levels were slightly, but Lim3 is essential for these hormone-producing lin- significantly, increased. CRH loading induced a sub- eages. Knockdown of Lim3 inhibited subsequent dif- stantial elevation in blood ACTH levels (Fig. 4d). ferentiation into hormone-producing cells, which sup- The downstream glucocorticoid hormone corticoste- ports altered pituitary development in Lim3 knockout rone was also significantly increased, indicating that mice [54]. ACTH from the graft sufficiently induced the down- These results demonstrate the competence of ES stream hormone (Fig. 4d). cell-derived pituitary progenitors to generate multiple Even without CRH loading, the basal levels of endocrine lineages in vitro. ACTH were higher. Importantly, corticosterone lev- els were also increased, suggesting that partial recov- Functionality of Induced ACTH+ Cells ery of blood ACTH has a moderate, but biologically significant, effect (note that ED50 of the ACTH recep- Positive and negative regulations by exogenous tor MC2R for glucocorticoid production is around stimuli are characteristic for endocrine cells. To inves- 9 pg/mL) [55]. In accordance with this, the treated tigate in vitro functionality, we induced ACTH+ cells hypophysectomized mice displayed higher spontane- for evaluation because they are most efficiently gener- ous locomotor activities and survived significantly lon- ated using the SAG-treated LCA SFEBq method. ger (Fig. 4e). Although CRH, which is secreted from After 10 min of stimulation by CRH, substantial the hypothalamus, should be diluted in the peripheral amounts of ACTH were secreted from SAG-treated site, mESC-derived pituitary tissues rescued survival LCA SFEBq aggregates in vitro (Fig. 4a). The secreted and spontaneous activities, suggesting that basal secre- ACTH concentration was similar to levels in mouse tion from these tissues was sufficient for those effects. In vitro induction of pituitary 675

Fig. 4 Functional tests of mouse ES-derived ACTH+ cells a: In vitro release from mouse ES-derived ACTH+ cells. “F” = glucocorticoid pretreatment. Among the releasing factors, CRH most efficiently induces ACTH secretion. b: Negative feedback test. Pretreatment with hydrocortisone suppresses CRH- stimulated ACTH secretion from aggregates. c: In vivo functional test by ectopic transplantation. All mice, except for the WT mice, received a hypophysectomy; hypopituitarism was confirmed by CRH loading. “S+D+” = SAG- and DAPT-treated aggregates. “S-D-” = no SAG or DAPT treatment. d: Blood ACTH and subsequent release of corticosterone. e: Improved activity and survival. (from Suga et al. (2011) Nature 480: 57-62. [38] modified.) (The values shown on graphs represent mean ± s.e.m. * P < 0.05; ** P < 0.01; *** P < 0.001) 676 Suga

These findings showed that induced ACTH+ cells pouch formed by mouse ES cells, which was in accor- derived from mouse ES cells acted as endocrine tis- dance with the size difference between human and sues, and that regenerative medicine for pituitary dys- mouse embryonic Rathke’s pouches. These pituitary function is feasible. placodes subsequently differentiated into pituitary hor- mone-producing cells. All types of pituitary hormone- Adaptation to Human ES/iPS Cell Culture producing cells were identified (Fig. 5b). Among them, we confirmed that the human ES-derived corti- The recovery from pituitary functional disorder cotroph responded normally to releasing and feedback is an important issue for medical studies because the signals. Electron microscopy revealed secretory gran- anterior pituitary has poor potential for regeneration. ules stored in the cytoplasm of these cells (Fig. 5c). Because some pituitary dysfunctions cannot be solely For both mouse and human ES cells, SFEB culture treated by drugs [14-16], regenerative therapy employ- is a favorable method that can generate functional pitu- ing stem cells should be considered as a new form of itary cells. Future studies will confirm whether human therapeutic intervention. Our SFEBq method [38] iPS cells can differentiate into pituitary cells using the induces pituitary cells that can auto-regulate hormonal same culture methods. secretion and respond to changing circumstances. The application of this culture method to human ES cells Future Perspectives is necessary for clinical purposes. However, poor sur- vival of human ES cells in SFEB culture might limit There are two primary uses for human ES/iPS the use of these cells for future medical applications. cell-derived pituitary cells. One is the human model Our colleagues found that a selective Rho-associated of development or disease. Results from our study kinase (ROCK) inhibitor, Y-27632, markedly dimin- showed that the present culture methods recapitulated ished dissociation-induced apoptosis of human ES embryogenesis, suggesting that it could be used in the cells and enabled the cells to form aggregates in SFEB area of developmental biology. In terms of diseases culture [56]. Using this fundamentally important dis- due to gene mutations, tissues derived from disease- covery, we attempted to adapt our pituitary-differen- specific iPS cells can be used for therapy screenings in tiating culture method for human ES cell culture. We a human disease model. have recently established the differentiation method The second major use for human ES/iPS cell-derived into corticotrophs and somatotrophs from human ES pituitary cells is for regenerative medicine. Although cells [57]. stem cell-based therapeutics provide high expectations The characteristics for differentiating pluripotent for the treatment of diabetes mellitus, the use of regen- stem cells into pituitary cells were as follows; erative medicine for hypothalamus-hypophyseal dys- - simultaneous induction of neighboring hypotha- functions has received little attention. lamic neuroectoderm and oral ectodermal tissue, simi- ES cell-derived ACTH-producing cells function lar to embryo. even after ectopic transplantation. This finding raises - self-formation of pituitary anlage (Rathke’s pouch) the possibility of relatively simple grafting of artificial as a result of interaction between those two layers. ES/iPS cell-derived pituitary tissues into a peripheral - generation of multiple endocrine lineages from site. These cells can function effectively if hormone Lim3(+) pituitary progenitors. secretion can be extrinsically controlled by releasing - functionality confirmation as endocrine tissue. factors or small molecule agonists. However, ectopic Combining these approaches, we designed a cell cul- transplantation is not perfect, because physiological ture scheme for human ES cells. CRH released from the hypothalamus does not directly Our results demonstrated that the anterior pituitary affect these grafts. Orthotopic transplantation of hor- self-forms in vitro following co-induction of the hypo- mone-producing cells that are controlled by positive thalamic and oral ectoderm (Fig. 5a). The juxtaposition and negative regulators is one of the future candidates of these tissues facilitated the formation of the pituitary for complete therapy. placode, and their features were consistent with char- In future studies, it will be challenging to recapitu- acteristics of Rathke’s pouch in vivo. The human ES late an entire anterior pituitary gland that contains all cell-derived Rathke’s pouch was much larger than the endocrine components in three-dimensional cultures of In vitro induction of pituitary 677

Fig. 5 Human ES cells culture a: Recapitulation of Rathke’s pouch formation. b: Differentiation into multiple lineages. c: Secretory granules characteristic of endocrine cells. 678 Suga human ES or iPS cells and to use such artificial pitu- Disclosure itary tissues for orthotopic transplantation into the sella of a large mammal. To achieve this long-term goal, The author does not have any potential conflicts of further studies are needed before pituitary regenerative interest associated with this research. medicine can be directly transferred to clinical use.

References

1. Chen J, Hersmus N, Van Duppen V, Caesens P, Denef Romero S, Senra A, et al. (2012) Craniopharyngiomas C, et al. (2005) The adult pituitary contains a cell popu- express embryonic stem cell markers (SOX2, OCT4, lation displaying stem/progenitor cell and early embry- KLF4, and SOX9) as pituitary stem cells but do not onic characteristics. Endocrinology 146: 3985-3998. coexpress RET/GFRA3 receptors. J Clin Endocrinol 2. Fauquier T, Guérineau NC, McKinney RA, Bauer K, Metab 97: E80-87. Mollard P (2001) Folliculostellate cell network: a route 13. Li H, Collado M, Villasante A, Matheu A, Lynch CJ, for long-distance communication in the anterior pitu- et al. (2012) p27(Kip1) directly represses Sox2 during itary. Proc Natl Acad Sci USA 98: 8891-8896. embryonic stem cell differentiation. Cell Stem Cell 11: 3. Kikuchi M, Yatabe M, Kouki T, Fujiwara K, Takigami 845-852. S, et al. (2007) Changes in E- and N-cadherin expres- 14. Arima H, Wakabayashi T, Nagatani T, Fujii M, sion in developing rat adenohypophysis. Anat Rec Hirakawa A, et al. (2014) Adipsia increases risk of (Hoboken) 290: 486-490. death in patients with central diabetes insipidus. Endocr 4. Chen J, Gremeaux L, Fu Q, Liekens D, Van Laere S, J 61: 143-148. et al. (2009) Pituitary progenitor cells tracked down by 15. Hahner S, Spinnler C, Fassnacht M, Burger-Stritt S, Lang side population dissection. Stem Cells 27: 1182-1195. K, et al. (2015) High incidence of adrenal crisis in edu- 5. Gremeaux L, Fu Q, Chen J, Vankelecom H (2012) cated patients with chronic adrenal insufficiency: a pro- Activated phenotype of the pituitary stem/progenitor spective study. J Clin Endocrinol Metab 100: 407-416. cell compartment during the early-postnatal maturation 16. Sherlock M, Reulen RC, Alonso AA, Ayuk J, Clayton phase of the gland. Stem Cells Dev 21: 801-813. RN, et al. (2009) ACTH deficiency, higher doses of 6. Mollard P, Hodson DJ, Lafont C, Rizzoti K, Drouin J hydrocortisone replacement, and radiotherapy are inde- (2012) A tridimensional view of pituitary development pendent predictors of mortality in patients with acro- and function. Trends Endocrinol Metab 23: 261-269. megaly. J Clin Endocrinol Metab 94: 4216-4223. 7. Luque RM, Lin Q, Córdoba-Chacón J, Subbaiah PV, 17. Ben-Shlomo A (2010) Pituitary gland: predictors of Buch T, et al. (2011) Metabolic impact of adult-onset, acromegaly-associated mortality. Nat Rev Endocrinol isolated, growth hormone deficiency (AOiGHD) due 6: 67-69. to destruction of pituitary somatotropes. PLoS One 6: 18. Evans MJ, Kaufman MH (1981) Establishment in cul- e15767. ture of pluripotential cells from mouse embryos. Nature 8. Fu Q, Gremeaux L, Luque RM, Liekens D, Chen J, et 292: 154-156. al. (2012) The adult pituitary shows stem/progenitor 19. Bernstein A, Breitman M (1989) Genetic ablation in cell activation in response to injury and is capable of transgenic mice. Mol Biol Med 6: 523-530. regeneration. Endocrinology 153: 3224-3235. 20. Babinet C, Cohen-Tannoudji M (2001) Genome engi- 9. Langlais D, Couture C, Kmita M, Drouin J (2013) Adult neering via homologous recombination in mouse pituitary cell maintenance: lineage-specific contribution embryonic stem (ES) cells: an amazingly versatile tool of self-duplication. Mol Endocrinol 27: 1103-1112. for the study of mammalian biology. An Acad Bras 10. Gaston-Massuet C, Andoniadou CL, Signore M, Cienc 73: 365-383. Jayakody SA, Charolidi N, et al. (2011) Increased 21. Nakano T, Ando S, Takata N, Kawada M, Muguruma K, Wingless (Wnt) signaling in pituitary progenitor/stem et al. (2012) Self-formation of optic cups and storable cells gives rise to pituitary tumors in mice and humans. stratified neural retina from human ESCs. Cell Stem Proc Natl Acad Sci USA 108: 11482-11487. Cell 10: 771-785. 11. Andoniadou CL, Gaston-Massuet C, Reddy R, 22. Eiraku M, Takata N, Ishibashi H, Kawada M, Sakakura Schneider RP, Blasco MA, et al. (2012) Identification E, et al. (2011) Self-organizing optic-cup morphogen- of novel pathways involved in the pathogenesis of esis in three-dimensional culture. Nature 472: 51-56. human adamantinomatous craniopharyngioma. Acta 23. Watanabe K, Kamiya D, Nishiyama A, Katayama T, Neuropathol 124: 259-271. Nozaki S, et al. (2005) Directed differentiation of tel- 12. Garcia-Lavandeira M, Saez C, Diaz-Rodriguez E, Perez- encephalic precursors from embryonic stem cells. Nat In vitro induction of pituitary 679

Neurosci 8: 288-296. induction from the diencephalon. Development 125: 24. Eiraku M, Watanabe K, Matsuo-Takasaki M, Kawada 4835-4840. M, Yonemura S, et al. (2008) Self-organized formation 37. Brinkmeier ML, Potok MA, Davis SW, Camper SA of polarized cortical tissues from ESCs and its active (2007) TCF4 deficiency expands ventral diencephalon manipulation by extrinsic signals. Cell Stem Cell 3: signaling and increases induction of pituitary progeni- 519-532. tors. Dev Biol 311: 396-407. 25. Sasai Y, Eiraku M, Suga H (2012) In vitro organogen- 38. Suga H, Kadoshima T, Minaguchi M, Ohgushi M, Soen esis in three dimensions: self-organising stem cells. M, et al. (2011) Self-formation of functional adeno- Development 139: 4111-4121. hypophysis in three-dimensional culture. Nature 480: 26. Kawasaki H, Suemori H, Mizuseki K, Watanabe K, 57-62. Urano F, et al. (2002) Generation of dopaminergic neu- 39. Koehler KR, Mikosz AM, Molosh AI, Patel D, rons and pigmented epithelia from primate ES cells by Hashino E (2013) Generation of inner ear sensory epi- stromal cell-derived inducing activity. Proc Natl Acad thelia from pluripotent stem cells in 3D culture. Nature Sci USA 99: 1580-1585. 500: 217-221. 27. Morizane A, Takahashi J, Shinoyama M, Ideguchi M, 40. Ochiai H, Suga H, Yamada T, Sakakibara M, Kasai T, et Takagi Y, et al. (2006) Generation of graftable dopa- al. (2015) BMP4 and FGF strongly induce differentia- minergic neuron progenitors from mouse ES cells by tion of mouse ES cells into oral ectoderm. Stem Cell Res a combination of coculture and neurosphere methods. J 15: 290-298. Neurosci Res 83: 1015-1027. 41. Wilson PA, Hemmati-Brivanlou A (1995) Induction 28. Danjo T, Eiraku M, Muguruma K, Watanabe K, Kawada of epidermis and inhibition of neural fate by Bmp-4. M, et al. (2011) Subregional specification of embryonic Nature 376: 331-333. stem cell-derived ventral telencephalic tissues by timed 42. Basch ML, Bronner-Fraser M (2006) Neural crest and combinatory treatment with extrinsic signals. J inducing signals. Adv Exp Med Biol 589: 24-31. Neurosci 31: 1919-1933. 43. Davis SW, Camper SA (2007) Noggin regulates Bmp4 29. Kadoshima T, Sakaguchi H, Nakano T, Soen M, Ando activity during pituitary induction. Dev Biol 305: S, et al. (2013) Self-organization of axial polarity, 145-160. inside-out layer pattern, and species-specific progeni- 44. Wang Y, Martin JF, Bai CB (2010) Direct and indirect tor dynamics in human ES cell-derived neocortex. Proc requirements of Shh/Gli signaling in early pituitary Natl Acad Sci USA 110: 20284-20289. development. Dev Biol 348: 199-209. 30. Ikeda H, Osakada F, Watanabe K, Mizuseki K, 45. Bharti K, Gasper M, Bertuzzi S, Arnheiter H (2011) Haraguchi T, et al. (2005) Generation of Rx+/Pax6+ Lack of the ventral anterior homeodomain transcription neural retinal precursors from embryonic stem cells. factor VAX1 leads to induction of a second pituitary. Proc Natl Acad Sci USA 102: 11331-11336. Development 138: 873-878. 31. Osakada F, Ikeda H, Mandai M, Wataya T, Watanabe K, 46. Davis SW, Mortensen AH, Camper SA (2011) et al. (2008) Toward the generation of rod and cone pho- Birthdating studies reshape models for pituitary gland toreceptors from mouse, monkey and human embryonic cell specification. Dev Biol 352: 215-227. stem cells. Nat Biotechnol 26: 215-224. 47. Lamolet B, Pulichino AM, Lamonerie T, Gauthier Y, 32. Muguruma K, Nishiyama A, Ono Y, Miyawaki H, Brue T, et al. (2001) A pituitary cell-restricted T box fac- Mizuhara E, et al. (2010) Ontogeny-recapitulating tor, Tpit, activates POMC transcription in cooperation generation and tissue integration of ES cell-derived with Pitx homeoproteins. Cell 104: 849-859. Purkinje cells. Nat Neurosci 13: 1171-1180. 48. Zhu X, Zhang J, Tollkuhn J, Ohsawa R, Bresnick EH, 33. Sakaguchi H, Kadoshima T, Soen M, Narii N, Ishida Y, et al. (2006) Sustained Notch signaling in progenitors et al. (2015) Generation of functional hippocampal neu- is required for sequential emergence of distinct cell lin- rons from self-organizing human embryonic stem cell- eages during organogenesis. Genes Dev 20: 2739-2753. derived dorsomedial telencephalic tissue. Nat Commun 49. Kita A, Imayoshi I, Hojo M, Kitagawa M, Kokubu H, 6: 8896. et al. (2007) Hes1 and Hes5 control the progenitor pool, 34. Wataya T, Ando S, Muguruma K, Ikeda H, Watanabe intermediate lobe specification, and posterior lobe for- K, et al. (2008) Minimization of exogenous signals in mation in the pituitary development. Mol Endocrinol ES cell culture induces rostral hypothalamic differentia- 21: 1458-1466. tion. Proc Natl Acd Sci USA 105: 11796-11801. 50. DiMattia GE, Rhodes SJ, Krones A, Carrière C, 35. Zhu X, Gleiberman AS, Rosenfeld MG (2007) Molecular O’Connell S, et al. (1997) The Pit-1 gene is regulated physiology of pituitary development: signaling and tran- by distinct early and late pituitary-specific enhancers. scriptional networks. Physiol Rev 87: 933-963. Dev Biol 182: 180-190. 36. Takuma N, Sheng HZ, Furuta Y, Ward JM, Sharma K, 51. Olson LE, Tollkuhn J, Scafoglio C, Krones A, Zhang et al. (1998) Formation of Rathke’s pouch requires dual J, et al. (2006) Homeodomain-mediated beta-catenin- 680 Suga

dependent switching events dictate cell-lineage deter- lineages by the LIM homeobox gene Lhx3. Science mination. Cell 125: 593-605. 272: 1004-1007. 52. Sornson MW, Wu W, Dasen JS, Flynn SE, Norman 55. S Melmed (2010) The Pituitary (3rd) Academic DJ, et al. (1996) Pituitary lineage determination by the Press: 61. Prophet of Pit-1 homeodomain factor defective in Ames 56. Watanabe K, Ueno M, Kamiya D, Nishiyama A, dwarfism. Nature 384: 327-333. Matsumura M, et al. (2007) A ROCK inhibitor permits 53. Gleiberman AS, Fedtsova NG, Rosenfeld MG (1999) survival of dissociated human embryonic stem cells. Tissue interactions in the induction of anterior pituitary: Nat Biotechnol 25: 681-686. role of the ventral diencephalon, mesenchyme, and 57. Ozone C, Suga H, Eiraku M, Kadoshima T, Yonemura notochord. Dev Biol 213: 340-353. S, et al. (2016) Functional anterior pituitary generated in 54. Sheng HZ, Zhadanov AB, Mosinger B Jr, Fujii T, self-organizing culture of human embryonic stem cells. Bertuzzi S, et al. (1996) Specification of pituitary cell Nat Commun 7: 10351.