It is difficult to recapitulate organ development in vitro, especially when interactions between tissues are essential. Nonetheless, researchers have now achieved this for the pituitary gland. See Article p.57
Suga et al.1 report in vitro generation of a functional embryonic organ from mouse embryonic stem cells on page 57 of this issue. They have induced formation of Rathke's pouch — a precursor structure to the pituitary gland — from which hormone-secreting cells can be made to differentiate. This work is a notable advance towards studying pituitary development and for better management of pituitary-hormone deficits through regenerative medicine. It also illustrates that inductive processes — whereby tissues interact to form organs — can be triggered in a controlled and largely autonomous manner, which bodes well for the prospect of developing even more complex organs in the lab.
In vertebrates, the pituitary — under the control of the hypothalamus in the brain — regulates essential physiological processes, including growth, puberty and reproduction. In humans, poor functioning of this organ can be congenital or acquired later in life, and affects one in 3,000–4,000 live births. Although hormone-replacement therapy can be effective in these patients, it has its disadvantages, not least that it cannot reproduce normal hormone-secretion patterns2.
In the early embryo, the tissues that will form the hypothalamus and Rathke's pouch are adjacent, because they form from the most anterior (rostral) neuroepithelium and the ectoderm cell layers, respectively (Box 1). Although both structures are then displaced posteriorly, they remain in intimate contact: the presumptive hypothalamus induces formation of Rathke's pouch in the developing embryo and regulates pituitary-hormone production after birth3.
Differentiation of rostral hypothalamus from embryonic stem-cell aggregates has been reported previously4. Some of the same authors reasoned1 that induction of an even more anterior feature in these aggregates might allow for the concomitant development of rostral ectoderm along with neuroepithelium — a prerequisite for pituitary development. To see whether this was the case, they obtained floating aggregates in which an external layer of rostral-like ectoderm covered an inner layer of hypothalamus-like neuroepithelium. They then treated the aggregates with activators of the signalling molecule Sonic Hedgehog, a protein that is required for proper development of Rathke's pouch in mice5. Suga et al. observed spectacular formation of invaginated hollowed vesicles that were strikingly reminiscent of Rathke's pouch in vivo, not only in shape and location, but also in the molecular markers they expressed (see Fig. 1 of the paper1).
It is not surprising that factors required for Rathke's-pouch formation in vivo (including Sonic Hedgehog, bone morphogenetic proteins and fibroblast growth factors) are also needed for pouch formation in vitro. What was unexpected, however, was such a degree of organization in this simple culture system. In vivo, morphogenetic movements (development of the head folds and neural-tube closure) affect the final location of Rathke's pouch. Yet, in Suga and colleagues' experimental system, similar movements, at least for the pouches, were recapitulated outside the 'head' context. It is also noteworthy that, in contrast to the in vivo situation, these aggregates did not contain mesodermal or neural-crest tissues, despite data suggesting that at least the mesoderm6 is required for Rathke's-pouch development.
The authors1 also find that inhibition of the Notch signalling pathway — which is involved in cell-fate decisions and, in many systems, in progenitor maintenance versus differentiation — is sufficient to induce differentiation of corticotrophs, the endocrine cells of the anterior pituitary lobe that secrete adrenocorticotropic hormone. This observation nicely recapitulates previous in vivo data7,8 obtained from Notch mutant mice. Corticotrophs can also form spontaneously ex vivo, when Rathke's pouch is dissected out and maintained in culture9,10. As all endocrine cell types begin to differentiate during the same period of development11, ex vivo and in vitro corticotroph differentiation probably reflects a permissive context for this particular cell type at this early time point, rather than a corticotroph-specific inductive event. Indeed, Suga et al. report differentiation, albeit less efficient, of all other anterior-lobe endocrine cell types. This is consistent with in vitro differentiation of all endocrine cell types from cellular spheres derived from adult pituitary stem and/or progenitor cells2.
The authors convincingly show that their in vitro-generated corticotroph-containing aggregates are functional when grafted into mice in which the pituitary has been ablated. A previous study12 also showed that progenitor cells from adult pituitary could differentiate in vivo, in vascularized microchambers. However, low cell yield and low differentiation efficiency were limiting factors for the use of cells obtained in this way in regenerative medicine. By contrast, Suga et al. had access to an unlimited supply of cells to obtain endocrine cells in vitro. Theoretically, this means that endocrine cells may also be obtainable from patient-derived induced pluripotent stem cells, which would offer a real improvement for the management of pituitary-hormone deficits.
More work is required to understand similarities and differences between in vivo and in vitro development of Rathke's pouch, and to improve endocrine-cell differentiation — perhaps by including supporting tissues and vasculature, as the authors1 suggest. Obtaining adult cell types from embryonic stem cells in vitro has proved problematic13. It should therefore also be investigated whether the embryonic-like pituitary cells that Suga et al. generated are fully mature.
Within the mature anterior pituitary lobe, the various cell types are linked together in complex networks that are thought to help coordinate hormone release in response to cues from the hypothalamus or the periphery14. As part of this response, cell mobility and the organization of these networks can change15. Is any of this mimicked in Suga and co-authors' in vitro-derived pituitaries? Plasticity of the pituitary — whereby the relative proportions of different hormone-secreting cell types can change in response to events such as puberty, pregnancy and lactation — may also depend, in part, on populations of stem cells and their progenitor cells within the gland2. It will be interesting to see whether these cells are present in pituitaries generated in the lab. If so, it may be easier to study these cells' origin and behaviour in vitro.
Many mutations can cause pituitary defects. But it is not always clear whether these are intrinsic to specific cell types and aspects of pituitary development or due to secondary constraints, imposed by abnormalities elsewhere that compromise inductive interactions. Suga and co-workers' method should allow the question of how the various pituitary defects develop to be addressed: it can be combined with the powerful technique of live-tissue imaging, which is difficult to apply to developing tissues in vivo.
In recent years, the field of tissue engineering has made considerable progress, especially with the use of natural and artificial matrices that can be seeded with tissue-specific stem cells to create various organs16,17. But perhaps — as this paper1 shows — early embryonic tissues already know what to do. Indeed, it has long been known that teratoma tumours, derived from pluripotent stem cells, can contain a complex mixture of tissues (such as teeth, skin and gut tissue), albeit in a haphazard manner. What Suga et al. have achieved is to harness this ability to obtain a robust and directed system. Could it also work with similar types of inductive event, such as those that lead to more complex organs, including the lungs, liver and pancreas18?
About this article
Endocrine Pathology (2013)