Mechanobiology of vertebrate gut morphogenesis

Abstract

Approximately a century after D’Arcy Thompson’s On Growth and Form, there continues to be widespread interest in the biophysical and mathematical basis of morphogenesis. Particularly over the past 20 years, this interest has led to great advances in our understanding of a broad range of processes in embryonic development through a quantitative, mechanically driven framework. Nowhere in vertebrate development is this more apparent than the development of endodermally derived organs. Here, we discuss recent advances in the study of gut development that have emerged primarily from mechanobiology-motivated approaches that span from gut tube morphogenesis and later organogenesis of the respiratory and gastrointestinal systems.

Introduction

The fact that embryonic development occurs by epigenesis, and not simply by magnification of miniature preformed organisms, was solidified when Caspar Friedrich Wolff observed that the embryonic chick intestine arises from folding of an epithelial sheet into a tube that bends and twists into the convoluted adult form [1]. He concluded that physical transformations driven by some “vis essentialis” or vital forces underlie all embryonic development. 250 years later, the study of gut development remains strongly linked to the investigation of how such forces organize cells into complex tissues and organs. Here, we review how recent studies have employed biophysical approaches to advance our understanding of vertebrate gut morphogenesis and how these studies have broadly influenced thinking on the role of physical forces in developmental biology. Because much of the recent progress in this context has focused on the amniote lung and small intestine, our discussion details work in this area, accentuating parallels and important disparities between these and other systems. Both the lung and small intestine function by maximizing surface area across which transport occurs, each relying on distinct morphological adaptations to do so. The lung, derived from the foregut, has a branched, fractal organization, terminating in alveoli that serve as the site of gas transport [2]. The small intestine, derived from the midgut, achieves enhanced nutrient transport via two adaptations: villi and looping. Villi are finger-like projections extending radially into the lumen, while looping refers to the convolution of the intestinal tube into loops that permit proper packing within the body cavity.

Gut Tube Morphogenesis

The gastrointestinal and respiratory systems arise from the gut tube, an embryonic structure consisting of an inner endodermally-derived epithelium surrounded by splanchnic mesoderm-derived mesenchyme. While emergence of the endoderm during gastrulation has been well studied [3,4], far less is understood about subsequent steps of tube formation [5–7]. For example, while it has long been accepted that definitive endoderm displaces visceral (“extraembryonic”) endoderm to the embryo margins during gastrulation, recent mouse studies indicate that definitive and visceral endoderm cells both contribute to the gut tube [8,9]. Our limited understanding of gut tube formation is due in part to the fact that, while aspects of gut organogenesis are well-conserved [10–15], morphogenesis of the tube itself is highly derived among vertebrates. While the amniote gut tube arises through folding of the endodermal sheet [16,17], the zebrafish gut tube forms by condensation of a solid endodermal rod that cavitates to form multiple lumens along the antero-posterior (A-P) axis [18,19]. Osmotic pressure drives inflation and fusion of these lumens through local cellular remodeling [18,20]. Xenopus, on the other hand, forms a gut cavity during gastrulation through involution of definitive endoderm, establishing a narrow gut lumen that expands as the gut tube elongates through radial and longitudinal cell intercalation (Figure 1a) [21,22]. The long-accepted model of tube formation in amniotes, that two invaginations at opposing ends of the embryo migrate toward one another as they internalize the endoderm, was primarily inferred from fate mapping [19,23–26]. However, recent live imaging studies in the chick embryo suggest that the foregut, midgut, and hindgut tubes form through distinct mechanisms. Hindgut formation, for instance, is driven by collective cell movements that arise due to conversion of a morphogenic FGF gradient into an active force gradient (Figure 1b) [17]. Chemo-mechanical modeling revealed a positive feedback mechanism between these signaling and contractile gradients that enables posterior endoderm movement to outpace axis elongation, despite the dependence of both on the same FGF gradient [27,28]. This velocity mismatch drives folding of the hindgut endoderm into a tube. Despite the fundamental importance of gut tube formation for establishing the body plan, it remains severely understudied. This may soon change, however, owing to technical breakthroughs like those in in toto live imaging of early mouse development [29]. Making sense of these extraordinary 4-D datasets will require a multidisciplinary effort relying heavily on biophysical modeling.

Figure 1. Mechanobiology of cell behaviors during gut morphogenesis.

(a) Convergent extension (left), mediated by radial and longitudinal intercalation (right, paired black arrow heads), simultaneously lengthens and thins the Xenopus gut tube, facilitating luminal expansion. (b) In chick hindgut tube formation, collective cell movements are driven by graded cell contractility (orange) in response to a morphogenic gradient of FGF (blue). These graded contractile forces draw neighboring passive cells (grey) toward higher FGF, thereby inducing those cells to contract as well and propagating a cascade of contraction and compaction that shifts the entire cell sheet. An example recruited cell is marked red. (c) Asymmetric migration (indicated by black arrows at right) of lateral plate mesoderm (orange) displaces the endodermal rod (grey) to the left, generating a leftward loop in the primitive zebrafish gut. (d) In amniotes, ECM remodeling facilitated by hyaluronan stabilization (orange) asymmetrically expands the right half of the mesentery (right, black arrow), resulting in leftward gut tube tilting. The tilting is enforced by mesenchymal condensation in the right with increased N-Cadherin (blue). (Abbreviations: gt, gut tube; FGF, fibroblast growth factor; en, endoderm; dm, dorsal mesentery; LPM, lateral plate mesoderm; N-Cad, N-Cadherin; HA, hyaluronan; D, dorsal; V, ventral; L, left; R, right.)

Organogenesis

A-P patterning of the gut begins before tube morphogenesis is complete [6,30,31]. This is due in part to reciprocal signaling interactions between the endodermally-derived epithelium and mesodermally-derived mesenchyme of the gut, which continue throughout later organogenesis [30,32,33]. In the past decade, it has become apparent that the mechanical interactions between the epithelium and mesenchyme are equally important for morphogenesis. Because molecular aspects of gut patterning have been described elsewhere [5,6], here we primarily discuss the mechanobiology of gut morphogenesis.

Respiratory Tract

The lungs begin as a simple evagination of the foregut tube, which then undergoes a series of remarkably stereotyped fractal branching events to generate complex alveolar networks in organized lobes [2,34]. Traditionally, efforts have primarily focused on identifying the molecular antecedents of these patterns [35], but there has been a recent shift towards understanding how forces coordinate airway branching morphogenesis. Distinct but mechanically analogous mechanisms are responsible for branch initiation and extension in the mouse and chick airway. In mice, both domain branching (lateral budding) and terminal bifurcation (tip splitting) result from epithelial growth against the constraint of contracting smooth muscle (Figure 2a) [34,36]. Therefore, branching requires spatiotemporal coordination between epithelial growth and smooth muscle patterning, generating the mechanical interplay that leads to characteristic thinning and elongation of new branch buds. This coordination is not necessary for branch initiation itself, but is essential for stereotyped branching [36]. Explant culture of mouse airway epithelium suggests that mechanical instability arising from proliferative growth and differential stiffness between the epithelium and undifferentiated mesenchyme is sufficient to initiate branching via viscoelastic buckling [37]. Analogously, bud elongation and thinning in chick also appears to be mediated by dynamic spatial restriction and mesenchymal reorganization, though here the constraint is provided by mechanoresponsive basement membrane thinning and mesenchymal fluidization at branch tips [38]. Branch initiation in chick is driven by intrinsic forces in the epithelium alone, with local apical constriction leading to branch budding without the proliferative growth required in mouse [39]. It is unclear whether extracellular matrix (ECM) remodeling also contributes to smooth muscle patterning. Investigating this potential interaction may aid in synthesizing findings from chick and mouse into a unifying theory of amniote branching.

Figure 2. Mechanical constraints direct form during tissue growth.

(a) Domain branching (left) and terminal bifurcation (right) of mouse airway occur by similar mechanisms. Differentiation of contractile smooth muscle surrounding the growing airway epithelium (grey) provides a spatial constraint (orange) to direct bud thinning and elongation (black arrow heads). (b) Higher pressure (orange arrows) in the lumen relative to the environment drives uniform luminal expansion if unconstrained (top), as in zebrafish gut tube, or epithelial protrusion and extension if constrained (bottom), as in mouse lung branching. Greater pressure differences accelerate rates of bud extension (bottom right). (c) The dorsal mesentery serves as a mechanical constraint (orange) on the elongation of the gut tube (grey), progressively buckling the straight tube (top) into loops (bottom) in proportion to differences in length and stiffness between the tissues. (d) Intestinal villi in chick form from luminal folds in the gut tube (right) generated by mechanical buckling as the luminal tissue grows against the surrounding smooth muscle constraint (orange). (Abbreviations: P, pressure; gt, gut tube; dm, dorsal mesentery).

Recent studies in mouse have also begun to resolve potential mechanical regulators of lung epithelial proliferation. Strikingly, microfluidic control of transluminal pressure modulates rates of branching and transcriptional maturation in embryonic lung explants (Figure 2b), without altering the branching pattern [40]. Pressure-induced epithelial tension also regulates proliferation directionality (Figure 3a) [41], and luminal pressure is upstream of alveolus differentiation and functional maturation [42]. Interestingly, deletion of the mechanoresponsive transcriptional coactivator YAP reduces epithelial tension and causes formation of distal cysts rather than organized branches [43], suggesting that transluminal pressure and epithelial tension are transduced through mechanosensitive transcription.

Figure 3. Mechanical forces instruct growth and organization.

(a) Tension (right) in the airway epithelium drives alignment of cell long axes along the direction of stretch (orange arrows), orienting mitotic spindles and subsequent cell division along this axis (black arrow). Cells in a relaxed tissue (left) will be less likely to align and will therefore divide in more random directions. (b) Passive, static tension in the first forming smooth muscle layer (orange arrowheads, center) is generated by radial growth of the intestinal tube, driving circumferential alignment of this layer. Longitudinal alignment of the second smooth muscle layer (bottom) results from cyclic deformations generated by peristaltic contractions (orange arrowheads) of the preceding circumferential layer. (Abbreviations: SM, smooth muscle.).

Although this physically-driven view of branch pattering is compelling, it remains to be fully reconciled with prior signaling-focused studies of branch morphogenesis [44–47]. Efforts to understand the signals that regulate differences in apical constriction, buckling, and constrained growth in chick and mouse lungs may lead to a precise mechanistic connection between sequential branch initiation and elongation events.

Gastrointestinal Tract

The diverse functions of the gastrointestinal tract arise during organogenesis through regional symmetry breaking events that shape the organs and determine their relative placement within the body. One of the earliest such events is a tilting or lateral looping of the initially straight gut tube. While this event is conserved across vertebrates [48–50], the underlying mechanobiology is not. In zebrafish, ECM remodeling facilitates asymmetric migration of lateral plate mesoderm that physically displaces the endodermal rod leftward (Figure 1c), setting the eventual position of the esophagus, intestinal bud, and liver [50,51]. In amniotes, midgut laterality is established by asymmetries in the dorsal mesentery, a stalk of tissue connecting the gut tube to the dorsal body wall [52]. ECM expansion in the right and N-Cadherin-mediated mesenchymal condensation in the left tilt the tube leftward (Figure 1d) [49,53]. Despite disparate cell behaviors driving laterality in zebrafish and amniotes, both are downstream of left-specific Nodal/Pitx2 expression and rely on localized ECM remodeling. In Xenopus, Pitx2 drives rightward directionality of the tube’s first symmetry breaking bend [48]. Interestingly, the right expansion of ECM in amniote mesentery arises primarily from hyaluronan stabilization, independent of Pitx2-dependent mesenchymal condensation on the left [53], demonstrating a right-specific symmetry breaking mechanism and the potential for important left-right crosstalk during tilting. These studies demonstrate the utility of the gut as a key system for elucidating the links between axis specification and downstream cell behaviors as effectors of morphogenesis.

Subsequent to tilting, the amniote small intestine transforms from an approximately straight tube into one consisting of many regular loops. This is attributed to a mechanical instability arising from rapid elongation of the tube against the constraint of the attached dorsal mesentery, resulting in compressive forces that buckle the tube into loops (Figure 2c) [54]. Loop morphology is conserved for a given species, and can be accurately predicted with a handful of experimentally accessible physical parameters, including differential growth rate and stiffness. While these relations describe the mechanics of looping, less is known of the underlying mechanobiology. Recently, bone morphogenic protein (BMP) signaling was identified as a key regulator of intestinal buckling, modulating differential growth through negative regulation of mesentery elongation [55]. However, the cellular mechanisms of organ-scale buckling remain largely unknown. Interestingly, torturous looping also occurs on the whole-gut tube scale in Xenopus as the tube elongates within the A-P confines of the tadpole’s body, strikingly reminiscent of amniote midgut looping [19,48]. It is compelling to consider that analogous buckling mechanics might be at work in this and similar systems where differential growth or asymmetric thickening is observed in hand with bulging or folded morphologies, as in the frog stomach and liver bud [56,57].

Thus far, we have primarily focused on how endoderm-derived organs are shaped through the action of purely physical forces. Intestinal villi, however, provide an important example of how tissue deformations inform patterning as well. In chick, sequentially differentiating circumferential and longitudinal smooth muscle layers spatially constrain the growing epithelium, progressively buckling it first into longitudinal ridges, then tight parallel zigzags, and finally driving growth of individual villi (“villification“) (Figure 2d) [58]. This final villification step is a consequence of buckling-induced changes in the mesenchymal distribution of endodermally-secreted Hedgehog ligands, creating BMP-responsive pockets that ultimately restricts intestinal stem cells to the base of villi, where crypts form postnatally [59]. Recent work suggests that LGR5+ stem cells residing in adult crypts arise from diverse, non-LGR5+ embryonic populations, suggesting that once the basic pattern of villi is established, there is continued remodeling as villi fuse and separate [60]. An alternative to the buckling model has been proposed in mouse, where intermediate buckling morphologies are not observed. Specifically, BMP pathway ligands and antagonists set up Turing patterns that place villi through reciprocal signaling between epithelium and mesenchyme [61–63]. Whether Turing patterns play a similar role in chick is unclear. Together, these studies illustrate the need to account for the bidirectional relationship between morphogenic patterning and tissue mechanics, which may themselves be under morphogenic control.

Contractile smooth muscle layers provide a prime example of this interplay between mechanobiology and patterning. In chick small intestine, radial growth generates static stretch that circumferentially aligns the initial inner layer of smooth muscle, which through cyclic contractions generates dynamic stretch that longitudinally aligns the later outer layer (Figure 3b) [10]. While smooth muscle orientation is mechanically set, its placement is determined by inhibitory Hedgehog-induced BMP signaling and by local suppression of this signaling by enteric neurons and preceding smooth muscle [10]. YAP signaling serves to refine this pattern by inhibiting muscle differentiation, suggesting that mechanoresponsive feedback may also be involved [64]. Interestingly, similar signaling, timing, and smooth muscle organization in fish suggests that these morphogenic mechanics may extend to non-terrestrial vertebrates [65–67].

Conclusions

There are many important parallels between the mechanobiology of lung and intestinal morphogenesis, speaking to broader themes and considerations for the study of mechanobiology in other developmental contexts. For example, lung branching, intestinal looping, and villification all rely on heterogeneities in tissue stiffness and growth rates, generating forces through constrained growth that are accommodated through buckling. This suggests that mechanical properties must be tightly controlled at the molecular and genetic level to ensure stereotyped morphogenesis. However, the means by which developmental signals specify tissue stiffness remains an open question.

In addition, lung branching and villification occur by surprisingly distinct mechanisms in birds and mammals, despite the vital role of these organs and the high degree of conservation in amniote development in general. This apparent disparity may suggest an incomplete view of morphogenesis in either species. In Xenopus, like chick, longitudinal ridges form from a smooth epithelium before kinking into zigzags [58]. These zigzags do not progress further into villi, providing an evolutionary snapshot of villi development in chick. Further, mice deficient in laminin develop zigzags, an intermediate step in the chick embryo that is not observed in wildtype [68]. Full villification most likely requires both the action of patterning cues [69] and mechanical deformations [58]. Similarly, lung branching has been separately attributed to molecular [70] and mechanical [34,37] mechanisms. Bridging between molecular and mechanical influences may lead to a unifying model of morphogenesis that describes both the development of these organs and how remarkably diverse morphologies can be achieved across species by evolutionary modulation of tissue mechanics. While chemo-mechanical coupling in morphogenesis has been extensively studied in theoretical work [71], less progress has been made toward understanding these interactions experimentally. A unifying theory of morphogenesis that accounts for the incredible morphological diversity across vertebrates will require such integration of mechanical and molecular aspects of development in place of the present, oftentimes dichotomous view.

Looking forward, parallel advances in several fields may lead to many great advances in the mechanobiology of vertebrate morphogenesis. A growing toolkit for force and stiffness measurement in vivo [72–74], breakthroughs in the speed and resolution of live imaging [29,75], single-cell omics, and nearly boundless computational power have made a whole new class of problems accessible and tractable for the first time. Cell reprogramming and the emergence of organoid cultures [76,77] further advances this accessibility and enables the study of human specific morphogenesis. As these technologies converge on vertebrate morphogenesis, the field is poised for rapid and exciting advances in understanding the “vis essentialis” of gut development.