Building additional complexity to in vitro-derived intestinal tissues

Gastrointestinal (GI) disorders affect up to 25% of the US population. Common intestinal disorders include malabsorption, irritable bowel syndrome and fecal incontinence. Some GI disorders such as Hirschsprung's disease have a genetic basis and are associated with an absence or paucity of enteric nerves. Current treatment plans for GI disorders range from changes in diet to bowel resection, and there are very few drugs available that target the primary deficiencies in intestinal function such as controlled peristalsis. While animal models can recapitulate the broad range of intestinal pathologies of the GI tract, they are intrinsically complicated and of low throughput. Several in vitro systems have been established, and these range from epithelial enteroids to more complex organoids, which contain most intestinal cell types. One of the more complex organoid systems was derived from adult mouse intestines and contains functional enteric nerves and smooth muscle capable of peristalsis. Establishing an equivalent human intestinal system is challenging due to limited access and variable quality of human intestinal tissues. However, owing to recent advances, it is possible to differentiate human induced and embryonic pluripotent stem cells, collectively called pluripotent stem cells, into human intestinal organoids (HIOs) in vitro. Although HIOs contain a significant degree of epithelial and mesenchymal complexity, they lack enteric nerves and thus are unable to model the peristaltic movements of the gut. The goal of this review is to discuss approaches to generate complex in vitro systems that can be used to more comprehensively model common intestinal pathologies. New and more biologically complete human models of the intestine would allow for unprecedented studies of the cellular and molecular basis of normal and pathological gut function. Furthermore, fully functional HIOs could serve as a platform for preclinical drug studies to model absorption and efficacy.

turns over approximately every 7 days. Th e renewal of the epithelium is driven by intestinal stem cells that reside in the crypts of Lieberkühn at the base of the villi. Th e submucosal layers contain smooth muscle myocytes, fi broblasts, subepithelial myofi broblasts, as well as enteric ganglia. Th ere is a rudimentary understanding of how the complex architecture and cellular diversity of the intestine arise during embryonic development, which has helped eff orts towards directing diff erentiation of human induced and embryonic pluripotent stem cells, collectively called pluripotent stem cells (PSCs), into the intestine. However, there is still much to be learned.

Intestinal development
Development of the intestine can be broadly subdivided into several steps including endoderm formation, midgut and hindgut specifi cation, gut tube morphogenesis, assembly of mesenchyme, colonization by neural crest cells, crypt-villus morphogenesis and cytodiff erentiation [1] (Figure 2a). Several important signaling pathways have been identifi ed as being required for directing these early stages of intestinal development in a broad range of vertebrate species, including birds, frogs and mice [2][3][4]. For example, Wnt and fi broblast growth factor signaling pathways direct endoderm into a midgut and hindgut fate, and this process is required for subsequent intestinal development. Moreover, the synergistic activity of both pathways is involved in the initiation of gut tube morphogenesis and formation of the gut mesenchyme [5]. Inhibition of either of these pathways results in abnormal intestinal morpho genesis and the loss of expression of posterior/hindgut markers, including Cdx genes. Once formed, the simple cuboidal epithelium of the hindgut and the surrounding mesenchyme undergo a series of reciprocal signaling events resulting in formation of a polarized columnar epithelium containing villi, a proliferative progenitor zone, and distinct intestinal lineages [5,6]. Along with development of the epithelium, there is a coordinated development of gut mesenchyme into the submucosal and smooth muscle layers. As discussed below, vagal neural crest cells that give rise to the enteric nervous system (ENS) migrate ventrally, undergo extensive proli fera tion, and incorporate into the developing gut shortly after gut tube formation [7,8] ( Figure 2a).

Building a better system: engineering additional complexity into human intestinal organoids
Low-throughput, limited genetic diversity, and species diff erences are recognized limitations for the use of animal models to study gastrointestinal (GI) disease. Recent advances in the understanding of human intestinal stem cells now allows for the derivation of human intestinal epithelial enteroids from patient biopsies [9]. Human enteroids have been cultured from both the small and large intestines and contain all major epithelial cell types. However, enteroids do not contain any additional cell and tissue types such as smooth muscle, supporting fi broblasts, endothelial cells, or enteric nerves. While there are more complex organoid systems derived from adult mouse intestines [10], it would be advantageous to have a more functional human intestinal organoid system as a screening platform to monitor drug absorption and the impact on intestinal motility. Brugmann and Wells Stem Cell Research & Therapy 2013, 4(Suppl 1):S1 http://stemcellres.com/content/4/S1/S1 Generation of more biologically complex human intestinal tissue has been accomplished through the directed diff erentiation of human PSCs (Figure 2b) [5,11]. Th e process is initiated using the Nodal-related protein activin, which directs diff erentiation of human PSCs into defi nitive endoderm [12,13]. Synergistic activity of the fi broblast growth factor and Wnt signaling pathways was then used to promote a posterior gut tube fate [2][3][4], to induce gut tube morphogenesis, and to promote growth of the intestinal mesenchyme. Th e resulting gut tube spheroids were strik ingly similar to the gut tube of an embryonic day 9 mouse embryo, consisting of a CDX2expressing cuboidal epithelium surrounded by a CDX2expressing mesen chyme. Growth of gut tube spheroids in three-dimen sional conditions that favor intestinal growth [14] resulted in the effi cient production of human intestinal organoids (HIOs) that have both secretory and absorptive function. Moreover, the cellular diversity and architecture was strikingly similar to that of the developing gut. Th e epithelium contained crypt and villus-like structures as well as all of the cell types normally found in the gut. Th e mesenchyme underwent diff er en tiation into stratifi ed layers. Some layers expressed smooth muscle markers whereas others expressed markers of fi broblasts and subepithelial myofi broblasts, both found in the submucosal layer. Furthermore, the HIOs generated via these methods were capable of basic intestinal function, including absorption of amino acids and secretion of mucus, and were able to be passaged in vitro for over 1 year, resulting in 50,000-fold expansion.
Despite the signifi cant level of complexity, PSC-derived HIOs lack both an enteric nervous system and a vascular system. Th is is probably due to the absence of vascular and ENS precursors, specifi cally neural crest stem cells. PSCs can be directed to diff erentiate into vascular and neural crest stem cells [15][16][17][18][19], raising the intriguing possibility that additional tissue complexity can be engineered into HIOs by addition of neural and/or vascular precursors in a contrived manner to recapitulate the normal develop ment of the intestinal vasculature and ENS.

Knowledge gaps
Th e ENS is a division of the autonomic nervous system and consists of a nerve plexus that innervates the GI system allowing for peristaltic contractions. Several GI disorders, such as Hirschsprung's disease, are due to an absence or paucity of enteric nerves [20,21]. Th e ENS is derived from a specialized cell population called the neural crest. Th e neural crest is a multipotent population of cells that derive from the dorsal neural tube and give rise to a myriad of cell types depending on their anteriorposterior position in the embryo [22]. Anterior/cranial neural crest cells give rise to neurons, bone, and cartilage of the head, whereas posterior/trunk neural crest cells give rise to components of the peripheral nervous system including the ENS. Embryonic studies of cranial neural crest stem cell (NCSC) development have paved the way for recent protocols to direct PSC diff erentiation into anterior/cranial NCSCs [15,23]. However, there are no published methods to direct the diff erentiation of PSCs into posterior/trunk NCSCs. Th is lack may be due to the fact that trunk neural crest development, in particular formation of the ENS, is less understood than development of cranial NCSC-derived structures.
Th ere are several reported methods to generate anterior/cranial NCSCs. In one case, ectoderm and neural ectoderm are derived from human ESCs by cellular aggregation or growth in media devoid of signaling molecules that might favor endoderm and mesoderm formation. Th ese conditions result in the formation of neural progenitor cells capable of being further directed into various neural derivatives. For example, exposing neural progenitor cells to factors, such as bone morphogenetic protein, that promote dorsal neural tube fate promotes the formation of cells that have NCSC properties and express NCSC markers, such as p75 [23]. In several reports, human ESC-derived NCSCs were shown to be multi potent and competent to diff erentiate into an array of NCSC-derived tissues in vitro and in vivo [15,23,24]. A similar approach should be possi ble to generate more posterior/trunk NCSCs, possibly by manipulating anterior-posterior patterning pathways such as retinoic acid or Wnt [25]. Subsequently, it should be possible to integrate human trunk NCSCs into develop ing intestinal organoids, at a time that approximates normal ENS formation, in attempts to generate HIOs with enteric nerves.
Given our current knowledge of the HIO system, we have identifi ed two major avenues for ongoing research. First, we will need to generate HIOs with more tissue/ biological complexity and functionality. For example, it should be possible to integrate enteric nerves and a capillary plexus into the HIO system. A more functional HIO would be able to transport luminal factors, such as drugs, across the epithelium and deliver them to an integrated capillary network. Th e impact of drugs on the ENS and peristalsis could also be evaluated. One approach to building a more complex HIO might be to incorporate vascular precursors and neural crest cells into developing HIOs. Second, as more systems-based approaches are developed, it would be advantageous to integrate HIOs into a microfl uidics platform that would allow for precise administration of drugs and other compounds, as well as monitoring of drug absorption and drug impact on intestinal function. Moreover, with an integrated network of microorgan systems on a device, it should be possible to measure the impact of intestinal function on drug bioavailability and the eff ects on other organ systems. Th e ability to perform this type of testing on a high-throughput scale, with PSC lines from a broad genetic background, could signifi cantly improve our predictions of drug toxicity and effi cacy in clinical trials.