Abstract
Many organs, such as the gut or the spine are formed through folding of an epithelium. Whereas genetic regulation of epithelium folding has been investigated extensively, the nature of the mechanical forces driving this process remain largely unknown. Here we show that monolayers of identical cells proliferating on the inner surface of elastic spherical shells can spontaneously fold. By measuring the elastic deformation of the shell we inferred the forces acting within the monolayer. Using analytical and numerical theories at different scales, we found that the compressive stresses arising within the cell monolayer through proliferation quantitatively account for the shape of folds observed in experiments. Our study shows that forces arising from epithelium growth are sufficient to drive folding by buckling.
One Sentence Summary Epithelial cells proliferating in an artificial, elastic hollow sphere accumulate compressive stresses that drive inward buckling.
Main Text
Epithelium folding is essential for the formation of many organs, such as the gut during gastrulation and the central nervous system during neurulation (1–2). There are three major mechanisms driving invaginations: convergent extension, apical constriction, and cellular flows (1–3–4), which contribution to epithelial folding is however difficult to evaluate in vivo. Alternatively, theoretical studies proposed that cell proliferation in confined geometries induce epithelium folding through buckling (5–8). Buckling is a bending instability occurring in elastic material under compressive forces (9).
To test if forces generated by proliferation can induce spontaneous buckling of epithelia, we undertook an in vitro approach and studied the proliferation of epithelial cells confined in a spherical shell. The spherical geometry presents several advantages over others: it is the one of early embryos, without boundaries, such that all cells experience the same environment. Specifically, we encapsulated MDCK-II cells in hollow alginate spheres, hereon referred to as capsules. To form them, we used a 3D-printed microfluidic device to generate three-layered droplets, which outer layer made of alginate undergoes gelation when falling in 100mM CaCl2 solution (Fig 1A, Fig S1 and methods)(10–11). The inner surface of the capsules is coated with a 3-4 μm thick layer of Matrigel to which cells adhere (Fig 1B)(10).
Encapsulated cells were imaged using 3D time-lapse confocal microscopy. Time zero corresponds to the start of imaging, 24h after capsules formation (Methods). Initially, MDCK-II cells were sparsely distributed on the capsule’s inner surface. Through proliferation, cells first formed clusters, which then merged into a monolayer (Fig 1C and Fig S2). Monolayers reached confluency at 8.8±0.8 hours (mean±SEM, as in the rest of the text, unless noted, N=54). Strikingly, monolayers folded after 14.5±0.8 hours (N=54) in approx. 80-90% of the capsules. In this process, a portion of the monolayer detached from the alginate shell and bent inwards (Fig 1D and Fig S3, SI Movie1). Proliferation was unaffected by confluency or folding, as the cell number increased linearly with a rate of 3.6±0.1 cells per hour (N=54) during 55 hours (Fig 1E). We concluded that a cell monolayer proliferating inside a spherical shell can fold similarly to embryonic epithelia during development.
To understand the driving mechanism of folding, we first checked that confinement was essential. Capsules were dissolved using alginase prior to folding and no monolayer folding was observed (Fig S3). Furthermore, as the apical side faced the interior of capsules (Fig 1F), apical constriction would rather oppose than promote folding in our system (4–7–12–13). In addition, no large-scale collective cellular flows were observed (Fig 1G). These observations show that neither apical constriction nor cellular flows can cause folding in our system. We then tested the last possibility that folding occurred through buckling.
Buckling is driven by compressive stresses in the monolayer. Since alginate is elastic, stresses caused by proliferating cells can be inferred from its deformations. Notably, proliferation of the monolayer generated pressure resulting in capsule wall thinning (Fig 2A). Detecting contours of capsules from confocal images (Fig 2D & 2C), the averaged wall thickness with time was measured for four different alginate concentrations (1, 1.5, 2, and 2.5%, see methods, Fig S4 and Fig 2D). Capsules without cells kept constant wall thickness (Fig S4). We used empty capsules to measure the Young modulus of alginate by atomic force microscopy (AFM) (methods and Fig 2E)(14). From the wall thickness and the Young modulus, the pressure exerted on capsules by cells over time can be calculated (methods and Fig 2F). For all alginate concentrations, the pressure increased during approx. 55h, reaching a plateau around 300-400 Pa. We concluded that compressive stresses corresponding to pressures below 300 Pa had no significant impact on the proliferation rate (Fig 1E).
Importantly, at the time of folding (approximately 14.5h), cell proliferation rate remained constant. Upon folding, the pressure was between 50 and 100 Pa for 1.5, 2 and 2.5% alginate concentration (Fig 2G). For 1%, the force exerted by cells onto the alginate shell was estimated by Hooke’s law from the deflection of the capsule and its stiffness (Fig 2H, SI Movie 2). The capsule stiffness (0.03±0.003 N/m, N=44, Fig 2K) was measured using an FT-S100 indenter (FemtoTools, Buchs, Switzerland, see methods, Fig S4 and Fig 2J). The average deflection force was 1.8±0.2 μN (N=20). Dividing this value by the invagination area, we obtained a pressure of 128±6 Pa (N=20) (Fig 2L), of the same order than for other alginate concentrations (Fig 2G). Our results show that monolayer folding occurred at a pressure of approx. 100 Pa, which is independent of alginate stiffness.
Next, we determined the corresponding compressive stress within the monolayer. To this end, we used a continuum theoretical description, without cellular details. The cell monolayer is described as a circular elastic ring, reproducing the geometry found on equatorial confocal planes (Fig 3A). The so-called cellular ring has a radius r if it is not confined. Two elastic parameters characterize its mechanics: the bending rigidity 𝒦 and the compressional rigidity λ (SI). Confinement is accounted for by restricting the cellular ring to a circular domain of radius R. For r>R, confinement is achieved by a harmonic spring of constant k. Hence, the total energy is a combination of the bending, compression, and confinement energies (SI). From now on, we consider the realistic limit 𝒦/λR2≪1 and λ/kR2≪1, meaning that compression is energetically favored over bending and that the confinement is rigid (SI).
This system exhibits a first order buckling transition controlled by the excess strain Δɛ=ΔL/2πR (15–16), with the excess length ΔL=2π(r-R) (Fig 3B and SI). The threshold excess strain at the transition can be deduced by comparing the energies in the two limiting cases of a compressed unbuckled ring (Fig 3C) and of an uncompressed buckled ring (Fig 3D). The buckling transition occurs when the buckled ring energy equals the compressed ring energy, leading to a threshold excess strain Δɛc~(𝒦/λR2)3/5 that depends on both the tissue material properties and the confinement geometry.
From this, the pressure at buckling can be determined. Below the threshold (Δɛ<Δɛc), the cell ring compression is stabilized by a uniform pressure P exerted by the confinement (Fig 3C). The energy associated with this pressure is ~PR2Δɛ and equals the ring compression energy ΔE~λRΔɛ2, yielding P~λΔɛ/R (Fig 3E). Hence, the pressure at the buckling transition Pbuckling, when Δɛ~Δɛc, depends only on the material properties of the monolayer and on the capsule geometry, but not on the capsule stiffness k (Fig 3F), in agreement with our experimental observations (Fig 2G,K). Above the threshold (Δɛ>Δɛc), the buckled ring is stabilized by a uniform pressure PB on the undeformed ring segment (Fig 3D and SI). PB can be expressed in terms of α and δ and independent of λ (16) (SI). α, δ, and PB can be experimentally measured in capsules with a single fold (Fig 3G). The dependence of α with δ fulfils theoretical prediction with no other free parameter (Fig 3H), supporting that the folded monolayer’s shape emerges from buckling.
We then wondered how single cell properties could participate to the buckling transition. To this end, we numerically analyzed the cell monolayer dynamics inside alginate capsules using a 2D vertex model (6–8–17–18). When simulations start, cells are characterized by a resting area A0 = 300μm2and a resting edge length L0, with (L0)2 = A0 (SI and Fig 4A). Deviations from these values are penalized by harmonic spring energy terms with constants K and ks for the area and the length, respectively (Fig 4A and SI) (17–19). In addition, large bending deformations of the monolayer are penalized by a harmonic spring energy term with constant cb (SI). K, the cell elasticity, can be estimated by K~λ/A0~109 N/m3. As in the continuum theory, the monolayer is confined to a circular domain of radius R by a spring constant k = 0.06N/m (SI).
To simulate proliferation, cells stochastically divide with a probability pmitosis (SI). Simulations started with 40 cells, similar to the cell number in a confocal section at confluency, and ended when the cell number doubled. Simulations reproduced folding (Fig 4B, SI Movie 3). The monolayer pressure grew continuously over time before and after buckling as in experiments (Fig 4C). However, the continuum theory predicts that pressure after buckling decreases as the excess strain increases (SI). Thus, proliferation accounts for the pressure dynamics in experiments.
Strikingly, we found that simulation folds acquired budded shapes unlike to experiments, owing to cell flows towards the fold’s neck (Fig 4D). We reasoned that in experiments, cell adhesion to the Matrigel acts as an effective friction (Fig 1G). We found that a friction force FNS~5.10−3μN in simulations prevents cell lateral displacements on the capsule’s inner surface and fold necks stayed wide, resembling experimental shapes (Figs 4E-1D, SI Movie 4).
We next wondered which cell parameters set the number positions of folds (Fig 4F-H). In experiments, two folds is the most frequent case (50%) followed by one fold (20%) (Fig 4G). This is different from the continuum theory, where equilibrium shapes feature a single fold (SI). In case of two folds, angles between folds were between 150° and 180°, whereas in case of three folds, they were between 90° and 120° (Fig 4H). In simulations, we search for parameter values that reproduce these experimental observations. A shape diagram (SI and Fig S9) shows that along the line ks/cb = 3.3 10−8, the fold number (Fig 4I) and angle distributions (Fig 4J) were similar to experiments (Fig 4G,H). Moreover, the shape of folds obtained in these simulations matched the ones observed in experiments and continuum theory, as seen from the relation between α and δ (Fig 4K-3H). To find realistic values of ks and cb we used parameter sets for which the pressure at buckling would match experimental values (Fig 2F and 2G), yielding to ks~3. 10−3N/m and cb~9 μN.μm with buckling occurred between 90 and 180 Pa (SI). It shows that, in addition to proliferation and friction, cell mechanics played essential roles in controlling shape, number and distribution of monolayer folds.
In this study, we show that an epithelium growing under spherical confinement buckles due to the compressive stresses arising from cell proliferation. The pressure necessary for epithelium buckling is at least five times lower than that required to hinder cell proliferation. Our study adds to the mechanical processes described in vitro to play a role in tissue folding through proliferation, from mesenchymal constriction (20) to constrained heterogeneous growth (21–22). It also strengthens the concept that unique physical properties of cells are required during morphogenesis, like superelasticity during growth of epithelium domes (23–24), or liquid to solid phase transition during elongation of the fish embryo (25). Our in vitro findings establish a buckling mechanism that could participate in many epithelium folding events occurring during embryogenesis.
Funding
AR and BC acknowledge funding from the SystemsX EpiPhysX consortium. AR acknowledges funding from Human Frontier Science Program Young Investigator Grant RGY0076/2009-C, the Swiss National Fund for Research Grants N°31003A_130520, N°31003A_149975 and N°31003A_173087, and the European Research Council Consolidator Grant N° 311536. IDM and AR acknowledge funding from Secrétariat d’Etat à la Recherche et à l’Innovation grant agreement REF-1131-52107. I.D.M., S.A., A.R. and J.G. acknowledge funding from the EU Horizon2020 Marie Sklodowska-Curie ITN “BIOPOL” (grant agreement No 641639).
Author contributions
A.T. and A.R. designed the project; A.T. performed all experiments and image analyzes, with exception of cell proliferation measurements; I.D.M. helped with capsule formation, 4D confocal imaging, and performed cell proliferation measurements. K.A. and P.N. designed and fabricated microfluidic devices, and optimized the encapsulation technology. A.M. and B.C. designed numerical simulations, and A.M. performed them. C.B-M. and K.K. developed continuum theory. S.A. and J.G. measured Young’s moduli of capsules by AFM. A.T., I.D.M., C.B-M, A.M., B.C., K.K. and A.R. analyzed the results, and wrote the paper, with editions from other co-authors.
Acknowledgements
Authors thank Oresti Malaspinas for his useful insights into the project.