An emerging role for tissue plasticity in developmental precision Open Access

A prevalent feature of embryonic development is the robustness of tissue morphogenesis, where reproducible morphology emerges in a tissue from the same species. Robustness emerges despite the presence of noise emerging from the inherent stochasticity in signaling, molecular interactions and gene expression, in addition, to genetic and environmental variation in a population. To understand the mechanisms conferring robustness in a system, precise quantitative measurements of the entity being probed needs to be undertaken after careful consideration of the different sources of noise and variation [1]. Robustness in general is said to be conferred by several mechanisms including the expression of genes with redundant functions, through multiple signaling pathways together contributing towards a common functional output, also known as distributed robustness [2], gene organization in the chromosome [3], as well as through intricate network topologies including gene regulatory networks [4] that comprises feedback and feed-forward loops [5, 6]. Over the last decade, tissue mechanics has added a new chapter in the context of emergence of precise and reproducible tissue morphologies [7–13].

In this mini-review, I will discuss the role of tissue mechanics in developmental precision with a particular focus on the emergence of left–right symmetry in somites followed by directing attention to a few examples where mechanics has been demonstrated to buffer intrinsic noise in the context of shape and size regulation. Left–right symmetry, also known as bilateral symmetry, is a prevalent feature across most animals where similar tissue shapes and sizes form within the same individual, for example in paired tissues such as somites, ears and eyes that develop on the left and right sides of the embryo. Studying how precision emerges on the two sides in the same individual is beneficial as genetic and environmental variation can be safely ignored likely enabling the discovery of processes that buffer unavoidable intrinsic noise in the system.

Somites, the precursors of most of the adult muscle and skeletal system, are left–right symmetric tissues that emerge during embryonic development. Somite morphogenesis has been extensively studied across species and the underlying mechanisms driving somite formation in a rhythmic and sequential manner are being increasingly parsed out in detail. Briefly, somite formation is driven by a segmentation clock that comprises gene transcriptional oscillations of the Hes/Her gene family resulting from negative feedback loops [14–16] and synchronization of this transcriptional state among a local group of cells via Delta–Notch signaling that depends on a combination of positive and negative feedback loops [17]. At the tissue-scale, this is manifested as a wave of transcription that traverses from the posterior of the embryo to the presumptive position where the next somite pair forms in the anterior. The exact mechanism that converts the temporal information carried by the waves to a spatial information that sets the initial size of a somite are, however, still hotly debated, with roles for an intrinsic timer that regulates the clock [18], Wnt [19] and retinoic acid morphogen gradients [20, 21] as well as clock-mediated periodic inhibition of FGF signaling [22] being proposed.

Somites are three-dimensional units of groups of epithelial and mesenchymal cells [23]. The size of a somite, largely described only in the form of its length along the anteroposterior (AP) axis (also known as AP length), has a stereotypical pattern along the body axis. In all species examined, the initial AP length in the anterior somites are small, the trunk somites are relatively bigger, followed by a gradual decrease in length towards the posterior somites [24, 25]. Many of the above-mentioned processes that drive somite formation have an effect on the initial AP length. A slowdown [26] or increasing the speed of the clock [27] leads to a corresponding increase or decrease in AP length, respectively, and similarly, inhibition or activation of the various signaling gradients [19, 28, 29] and synchronization processes [30] results in a change in AP length. Although not studied in detail, the precision with which the initial AP length of a somite is set is, therefore, likely to depend on the network topology and the strength of interactions between these different processes.

Importantly, even in an unperturbed system, i.e. in wild type, there still exists substantial variability in the initial AP length of somites as quantified recently in zebrafish embryos [8] (Figure 1A–C). This intrinsic variability can be considered to be a measure of the collective noise in the upstream signaling and transcriptional processes that drive somite formation. Zebrafish has 32 somites along the body axis and the variability in somites 1–8 were analyzed in this study and discussed in detail below. The initial AP length of many somites varied by 15–20% compared with the mean across embryos and this variability was observed even for the same somite ordinal number across the left and right sides indicating the presence of initial asymmetries. Although not explicitly mentioned, a similar extent of variability has been previously observed in a few studies across species [19, 24, 25, 29]. It has been previously reported that a consistent 6–7% increase in somite length along the body axis can lead to lasting effects in the musculoskeletal system with a reduced number of skeletal elements [26]. A much larger left–right difference in somite lengths observed in wild-type embryos [8] could, therefore, potentially lead to left–right asymmetries in adults, but such asymmetries persisting from the embryonic stage into adulthood have never been reported in wild type.

How do embryos then buffer such variability to ensure minimal long-lasting effects? Interestingly, by following the same somites through development, it was observed that the variability in somite length decreased over time, with AP lengths getting adjusted towards a precise target length even within the same embryo [8]. As a consequence of this adjustment, AP lengths on the two sides become similar over time, thus ensuring increased left–right symmetry (Figure 1B,C). This observation of precision in tissue size emerging over time is not exclusive to somites and has been previously reported in left–right otic vesicle (precursors of the ear) development in zebrafish, where a decrease in variability of tissue size from 16 hours post fertilization (hpf) to 48 hpf was shown to result in increased symmetry in otic vesicle morphology [31]. Notably, in both cases, the correction mechanism was shown to be independent of processes that drive upstream morphogenesis of the respective tissues. In the case of otic vesicle, perturbation of FGF signaling, an important regulator of initial ear size, led to increased asymmetries but still the embryos recovered to attain symmetry by 48 hpf. Similarly, in somites, despite mild perturbations to the clock and the synchronization process, which led to a different initial size and increased variability, respectively, length adjustments still proceeded normally [8]. Furthermore, adjustment continued unperturbed on one side even in the total absence of contralateral somites suggesting that somites attain a target length independently on the two sides [8], a phenomenon that is also well appreciated in left–right otic vesicle [31, 32] and limb bud development [33]. In early embryos, it, therefore, appears that minimal cross-talk is required between the two sides to result in tissues of precise sizes.

What drives the size control of a somite on each side? During the time period of AP length adjustment, the volume of a somite remained unchanged, which directly pointed to the role of shape changes in a somite [8]. Accordingly, it was observed that a change in AP length was highly correlated to a corresponding change in mediolateral (ML) length (Figure 1B). To test whether neighboring tissues play a role in this shape transition or if this is entirely intrinsic to a somite, ablation experiments were performed. Ablating the presomitic mesoderm immediately posterior to the most-recently-formed somite resulted in a bulging of the nearest somite boundary (Figure 1D). This suggested the presence of significant contact stresses along the AP axis, which when relieved by ablation resulted in somite shape changes (Figure 1D). A similar observation was made in chick embryos, where somites bulged into regions devoid of neighboring somites that were mechanically removed [34]. In contrast, ablating the lateral plate mesoderm, a tissue that is present lateral to somites, did not result in any somite shape changes, suggesting relatively lesser constraint on somites laterally. A consequence of this could be readily observed in the shape of the lateral somite boundary, which tends to be curved, in contrast with the relatively straighter somite boundaries abutting other tissues (Figure 1). In addition, this provides a possible explanation for the ML length majorly buffering adjustments in AP length, as this dimension is relatively free to change shape given the lesser constraint, particularly in anterior somites. Whether the overlying ectoderm, which abuts the lateral plate mesoderm further away from the somites, also plays a role is yet to be tested.

Although not mapped out in detail, as the shape of a somite changed, concomitantly, cells within anterior somites were observed to undergo rearrangements akin to a diffusive process [8]. It is important to note here that shape changes in anterior somites immediately after formation were probed in this study when minimal cell shape changes are observed. The mechanism by which precision emerges could be different, however, in posterior somites, where upon somite formation, cells within a somite in addition undergo substantial cell shape changes [35]. Together, these observations indicate that somites attain a precise AP length that stems from a cross-talk, which remains to be characterized in detail, between external contact stresses from neighboring tissues and internal active forces that drive cell rearrangements. Given a set size for the entire presomitic mesoderm, it was also proposed that adjustments of length will automatically lead to positional symmetry of somites [8]. However, what ensures a similar size of the presomitic mesoderm on the two sides is unknown, although at later somite stages, the symmetric flow of cells into the presomitic mesoderm could play a role [36].

To understand shape changes in a tissue quantitatively, it is imperative to be cognizant on the material property of a tissue. Of late, it appears that cells and tissues can be considered as yield stress materials [37–41]. The yield stress of a tissue corresponds to the maximum mechanical stress it can withstand in a solid-like state, before cells within the tissue start to rearrange and flow thereby permanently changing the shape of a tissue. This regime characterized by a permanent change in the shape of a material (in this context, read as ‘tissue’) even when the stress is relieved represents the plasticity of a material [42].

What makes a tissue plastic? In the context of somites, as mentioned above, during the time period of somite shape change, cells within a somite were observed to rearrange with a diffusive dynamics over the time-scale of length adjustment [8] (Figure 2C). One can, therefore, argue that tissues have to be in a fluid-like state to exhibit plasticity, i.e. lasting shape changes. In contrast, recent experiments by injection of ferrofluid droplets in zebrafish embryos to directly measure cellular and supracellular stresses in the anterior presomitic mesoderm just before somites form, have shown the tissue to behave as a yield stress material with a solid-like state [37] (Figure 2B). One possible explanation for the observed difference in [8] is that once a somite forms and an extracellular matrix encapsulates it (Figure 2A), sufficient stresses are generated to drive the somite tissue beyond its yield stress and, therefore, towards a fluid-like state. Accordingly, an isolated somite was observed to round up within 30 min of explanting the tissue (Figure 2D), signifying the role of surface tension in generating sufficient stresses to fluidize the tissue [8]. Furthermore, perturbations to molecules implicated in surface tension slowed down the observed shape changes validating this claim. However, direct measurements of somite surface tension are yet to be performed, which will be crucial to resolve the debate on the material state of the somite tissue leading to the observed shape changes. The detailed molecular activities that drive cell and tissue shape changes as well as a general overview of tissue mechanics will, however, not be discussed here and I will instead refer to extensive reviews previously written on these topics [43–53].

In general, tissues can also change shape without accompanying cell rearrangements, where changes in tissue shape can be directly correlated with concomitant changes in cell shapes. For example, during Drosophila ventral furrow formation, local transient cell shape changes, known as apical constriction, driven by the apical actomyosin cortex, lead to an out-of-plane bending of the tissue [54]. Notably, even though no cell rearrangements have been reported in this system, the resulting tissue deformation seems to be dependent on a viscous response of the cytoplasm to the force generated apically [55]. Thus, irrespective of the precise material property of the tissue, a fluid-like state of the tissue seems to be important for mediating shape changes. Other examples where a fluid-like state of tissues have shown to be important for tissue shape changes include the spreading of the zebrafish blastoderm before gastrulation [56, 57], epithelial gap closure during insect gastrulation [58], tissue flows during avian gastrulation [59], elongation of the zebrafish tail bud [36, 60], convergence–extension during germ band elongation in Drosophila [61] and Kupffer’s vesicle morphogenesis in zebrafish [62]. Tissue fluidization in fact is important for various other contexts of morphogenesis such as the rotational motion of connected cells in the Drosophila eye [63], wound healing [64] and collective cell migration [65–67]. Most of the above-mentioned shape transitions emerge from regulated changes in gene expression patterns that in turn drive shape changes in tissues. For example, the initial shape changes important for apical constriction are driven by spatiotemporally regulated expression of specific transcription factors [54]. Similarly, zebrafish somites also change from a coarsely cuboidal to a chevron shape [35, 68] over time, which is regulated by a coordination of FGF and Shh signaling pathways [69]. In contrast, the subtle shape changes that were observed in [8] purely emerge as a response to a particular mechanical environment given an initial state and can be considered to be a case of self-organization.

Notably, as mentioned in the previous section, changes in somite shape are important to ensure increased precision and symmetry. The prevailing idea here is that the upstream segmentation clock-related processes set a crude initial AP length, which gets fine-tuned by tissue mechanics to result in increased precision. In other words, tissue mechanics serves to buffer collective noise emerging from transcriptional and signaling processes, an emerging theme across diverse contexts. For example, the cephalic furrow, an invagination that forms all along the dorsal–ventral axis in gastrulating Drosophila embryos, is known to form in a straight manner without the appearance of any kinks or branching. This precise formation of the furrow is dependent on tissue-scale mechanical coupling [70] that buffers spatiotemporal heterogeneities in furrow initiation as well as the inherent noise in generating contractile forces [9]. Similarly, mechanical forces ensure the straightness of other morphological boundaries such as the dorsoventral compartment boundary in the Drosophila wing disc [71], parasegmental boundaries in early Drosophila embryos [72] and the notochord-presomitic mesoderm boundary in zebrafish embryos [73].

Tissue mechanics is also important for maintaining precise shapes in the presence of heterogeneities in a developing organ. For example, sepal growth in Arabidopsis flowers is driven by cell proliferation and if the position of the dividing cells and the rate at which they divide are spatially heterogeneous, which do occur, sepal shapes can get deformed over time. However, during the growth phase, sepals are known to largely maintain their shape and this was shown to depend on spatiotemporal averaging of cellular growth [74] as well as on a stress-feedback mechanism, where dividing cells are mechanically shielded by a ring of microtubules spread across the immediate neighbors [10], ensuring robust maintenance of shapes.

Finally, tissue mechanics is important for robust contralateral cell and pattern matching during development. This has been demonstrated in Drosophila embryos in two different contexts. One, during dorsal closure, spatiotemporal regulation of cytoskeletal tension drives precise matching of segmentation gene expression patterns across the fusing epithelial flanks [75]. And, two, during heart morphogenesis, differential expression of adhesion proteins ensure precise cell matching by modulating filopodia binding dynamics in contralateral cardioblasts [76]. Taken together, across diverse systems, the mechanical properties of cells and tissues play a vital role in ensuring robustness during morphogenesis.

While it is readily apparent that a specific tissue forms with reproducible shapes and sizes across individuals from the same species, we are just beginning to understand how precision emerges in tissue morphology during embryo development. Given that noise is inevitable in biological systems, to understand the emergence of precision, mechanisms that buffer noise, therefore, need to be studied in detail. Studying transcription and signaling network topologies that suppress noise is important, but at the same time, the role of mechanics in ensuring increased precision should also be taken into account as illustrated through several examples mentioned in this review.

One aspect I have conveniently ignored here is the importance of scaling in achieving a precise size in a tissue relative to the overall body size, which is referred to as morphological allometry [77]. In several contexts, during the growth phase of a tissue, morphogen gradients have in fact been shown to scale with tissue size suggesting that a tissue can read its own size as it grows [78–80]. However, the details of the underlying mechanism is unknown and it is also yet to be clearly established if mechanics plays a role in regulating scaling properties during development particularly in the context of precision.

In general, to understand the role of mechanics, quantitative analysis of shape changes in a developing tissue need to be performed. This is possible only by performing long time-lapse imaging of developing tissues with techniques that lead to negligible phototoxicity such as light-sheet microscopy [81] and coupling such imaging modules with advanced image analysis algorithms. Furthermore, in addition to existing methodologies [82] for measuring mechanical properties such as ferrofluid droplets [83] and actuators [84], newer tools need to be developed that can allow probing the mechanical state of tissues directly in vivo. These tools need to be standardized in multiple systems and made widely available that can be utilized in a user-friendly manner, which will enable comparisons of mechanical properties across diverse tissues using the same tools. Importantly, such quantitative efforts need to be coupled with the development of mechanical models to provide a deeper understanding of mechanisms that ensure the precision of tissue shapes and sizes.

• It is well known that investigating the mechanical properties of cells and tissues is vital for understanding tissue morphogenesis. It is now increasingly appreciated that tissue mechanics is not just important to mediate shape changes but also to ensure precision in the final shape and size a tissue attains.

• To emerge in a precise manner, intrinsic noise in biological processes need to be buffered. Over the next few years by studying more tissues from the perspective of precision, particularly in the context of tissue plasticity and mechanics, the similarity and diversity of noise buffering mechanisms can be better laid out.

• An in-depth understanding of how reproducible shapes emerge is likely to have a huge impact on tissue engineering as well as in generating reproducible tissues on a chip from patient-derived samples, an important bottle-neck to solve in the quest towards personalized medicine.

The authors declare that there are no competing interests associated with the manuscript.

I would like to thank Andrew Oates, Marko Popovic, Swarali Dighe and Siddhartha Bardhan for comments on the review. S.R.N. acknowledges financial support from the Department of Atomic Energy (DAE), Govt. of India (Project Identification no. RTI4003, DAE OM no. 1303/2/2019/R&D-II/DAE/2079 dated 11.02.2020) and the Max Planck Society Partner Group program (M.P.G.A MOZG0010).

AP

anteroposterior

hpf

hours post fertilization

ML

mediolateral