Engineering kidney developmental trajectory using culture boundary conditions

Introduction

Branching morphogenesis is a common developmental program for epithelial organs to acquire a tree-like structure. This massively parallelizes their function because branch tips become sites that facilitate organ-specific functions, such as secretion, gas exchange, or filtration¹. In the embryonic kidney, ureteric bud (UB) progenitor cells exchange reciprocal signals with surrounding mesenchyme to build the branched urinary collecting system and the blood-filtering nephrons². Mammalian kidney architecture and the total number of nephrons (nephron endowment) are determined predominantly in utero³. Disruptions to branching morphogenesis result in congenital anomalies, where aberrant tissue-scale organization can hinder adult renal function and negatively impact health⁴. Therefore, it is crucial to understand the factors involved in the spatial patterning of kidney components during development.

Explant culture techniques enable real-time imaging and perturbation of a range of processes including branching morphogenesis and nephron formation⁵. Branching morphogenesis persists in mouse kidneys that are explanted sufficiently early in development (before embryonic day E14) and placed at an air-liquid interface (ALI) on glass, a transwell membrane, or a metal mesh with filter overlay^6,7,8. The failure of explants to develop in suspension has been ascribed to a need for appropriate transport of oxygen or other nutrients⁸. Despite this hypothesis, explant culture also fails in ‘hanging drop’ format, which is thought to have little mass transport limitation⁸. An alternative view arose from the observation that ALI culture is compromised in the presence of a surfactant, suggesting that either mechanical stresses imparted by surface tension and/or kidney-substrate adhesion improve explant morphogenesis⁸. Supporting this, other culture approaches that sandwich the kidney between material interfaces are similarly successful^5,9.

Despite its strengths, ALI-based explant culture presents challenges for questions related to mechanics. While genetic and biochemical regulators of kidney branching have been extensively studied, mechanics is of emerging interest. Tissue mechanics are a key contributor to branching morphogenesis in organs including the lung, salivary gland, mammary gland, and pancreas^10,11,12,13. In the kidney, we previously found that limited organ surface area constrains the ureteric tubule tree geometry, such that branching events generate local mechanical stresses in the surrounding nephrogenic zone^14,15. Current ALI cultures flatten and distort the cortico-medullary structure of the kidney at an air-liquid interface (ALI), which introduces developmental artifacts¹⁶. Historically, Rienhoff first cultured chick metanephros in chicken bouillon hanging drops with a rough description of the 3D morphology¹⁷. Rosines et al. reported sustained kidney development by embedding explants in 3D Matrigel, collagen I, or collagen IV matrices at an ALI¹⁸. However, the relationship between matrix properties and kidney developmental outcomes has not been closely studied.

Although little is known about the effect of matrix properties on in vivo kidney development, explants acquire divergent phenotypes when extracellular matrix (ECM) proteins are added to culture media vs. when bound to the culture substrate in the ALI format¹⁹. Matrix mechanical properties also influence kidney morphogenesis and cell decision-making in organoid models. For example, hydrogel stiffness and viscoelastic properties affect kidney organoid differentiation, final morphology, and proportions of nephron segments^20,21. In vivo, the kidney is embedded within the urogenital system, bounded by the renal interstitium, capsule, renal fascia, and surrounding organs (Fig. 1A, B)²². FOXD1+ stromal progenitors form both the capsule and the outermost stroma adjacent to nephron progenitor niches (the capsule is physically distinct after ~E14)²³. Genetic ablation of these cells leads to severe defects, which are thought to be due to signaling dysregulation^23,24. However, disruption of the capsule may also serve a mechanical role. We sought to develop a live imaging-compatible 3D culture system that would advance the study of in vivo-like branching morphogenesis of the kidney and resolve the influence of mechanical boundary conditions.

**Fig. 1: Hydrogel encapsulation inspired by embryonic microenvironment supports 3D kidney explant culture.**

Inspired by the in vivo urogenital environment, we embedded kidneys in 3D hydrogel domes to mimic their native encapsulation. We demonstrate that embryonic kidneys develop and acquire more in vivo-like morphologies in 3D culture relative to ALI culture. We investigate 3D kidney branching morphogenesis in real time, confirming predictions of tip rotation dynamics and tip-tip spacing inferred from physics-based modeling and fixed kidneys at different ages. Relative to ALI cultures, 3D cultures more faithfully captured in vivo effects of perturbing the GDNF-RET pathway, which drives UB branching. We next show that overall kidney shape and nephron:UB tip ratio (‘nephrogenesis balance’) are influenced by material properties in both ECM-derived and engineered 3D matrices.

Results

3D embedding supports kidney explant culture and development

We first designed a 3D culture system that embeds mouse embryonic kidneys in hydrogel domes. We began with an ECM-derived hydrogel composite consisting of 1:1 collagen I and reconstituted basement membrane (Matrigel). Collagen I is a structural protein that provides tensile strength in the ECM^25,26, and whole E14 embryo staining revealed that collagen I-rich ECM envelops the developing kidney in situ²⁷ (Fig. 1B). Matrigel mainly comprises laminin and collagen IV, two ECM components abundant in embryonic kidneys (Fig. 1A,B)^28,29. Mixing 1 mg/ml collagen I and Matrigel (C + M) at a 1:1 ratio results in 0.5 mg/ml C + M gels that adhered to glass culture substrates, gelled consistently, and had sufficient mechanical integrity to support a consistent 3D geometry during culture up to 5 days (Fig. S1). To contain overlying media, we fabricated and plasma-bonded polydimethylsiloxane (PDMS) rings to coverglass-bottom plates to separate the inner space for explant culture and the outer space for humidity control during gelation (Fig. 1C). Once gelation was complete, culture media was added to the inside of the ring to submerge the hydrogel dome containing the kidney.

Next, we aimed to apply a labeling strategy compatible with time-lapse fluorescence microscopy. Inspired by studies in lung explant systems³⁰, we labeled live kidneys with fluorescently-conjugated EPCAM antibody and PNA lectin (both stains for the ureteric bud, Fig. 1D, E). This approach allowed us to capture time-lapse movies of 3D E12-13 kidney branching using spinning disk confocal microscopy without affecting explant development (Fig. 1D, E, Supplementary Fig. 2, Supplementary Movie 1). After ~3 days (64 h), we compared cultures in 3D C + M to those in ALI by immunofluorescence. Tip duplication rates and final JAG1+ nephron number were equivalent between 3D and ALI cultures (Fig. 1F, Supplementary Fig. 3). Tip localization at the organ surface is a key feature of kidney morphogenesis and is likely necessary to properly localize nephron condensation events to the (outer) cortex relative to the (inner) medulla, contributing to appropriate corticomedullary zonation in the adult organ^31,32. We noted that RET + UB tips were localized organotypically at the kidney surface in 3D and primarily peripherally in ALI, though some more sparsely spaced tips were found internal to the kidney in ALI culture. This is likely due to surface area constraints in flattened ALI cultures, where branching morphogenesis creates more tips than can be geometrically accommodated at the organ periphery¹⁴ (assuming that the top and bottom surfaces of ALI kidneys cannot be considered ‘cortical’³³ such that their areas can be neglected). We next studied geometric distortion during culture by measuring kidney height. While branching morphogenesis proceeded, kidneys in ALI culture flattened to 120 ± 22 µm in height (mean ± S.D.). Kidneys in C + M culture increased their thickness to 310 ± 25 µm from an initial thickness of 240 ± 20 µm for freshly dissected E13 kidneys (Fig. 1D–F). Despite this thickness difference, we found no significant increase in the hypoxia marker Hif1ɑ or downstream inflammatory cytokines in 3D vs. ALI cultures after 3 days (Supplementary Fig. 4). When 3D culture was carried out until 5 days, we observed increased nephron maturity represented by higher areas of positive proximal tubule and podocyte marker expression (Supplementary Fig. 1). Attempts to culture older kidneys (E15) failed in both ALI and 3D formats (Supplementary Fig. 5). Together, these data indicate that kidney embedding in C + M hydrogels retained live imaging capabilities and similar UB tip and nephron numbers to ALI culture, but better approximated in vivo-like tissue thickness and localization of UB tips to a 2D organ surface rather than primarily a 1D organ periphery.

3D cultures more accurately model in vivo morphogenesis

We next focused on live dynamics of tip branching. We found that tip branching proceeded in 3D in embedded cultures, with tips bifurcating and advancing in xy and z (Fig. 2A). During branching in vivo, UB tips become more crowded at the kidney surface, eventually forming semi-crystalline packing after ~E15 in mice and week ~11 in humans^14,15. We wondered if our 3D culture approach would better approximate this close packing of UB tips compared to ALI culture. We quantified the distance between non-sibling tips as a metric for tubule packing. E13 kidneys cultured for ~3 days in C + M hydrogels and size-matched E14 kidneys had similar UB tip-tip distances, while tips in ALI cultures were further away from their neighbors (Fig. 2B). These data indicate that UB tips more closely adopt in vivo-like packing in 3D culture relative to ALI culture.

**Fig. 2: 3D culture models kidney morphogenesis dynamics more closely than ALI cultures.**

We speculated that tissue flattening in ALI culture could account for changes in tip packing by creating branching artifacts that are not representative of in vivo organ growth. Studies of kidneys dissected at different developmental ages have inferred that UB tips rotate and reorganize at the organ surface throughout development^14,15,34,35 (Fig. 2C). Our prior physics-based model also revealed that tip organization in vivo can be explained by physical packing rules in 3D¹⁴. The model performs energy minimization to predict elastic tubule positions within a 3D bounding box and repulsion forces between tips that are inversely proportional to their distances¹⁴ (Fig. 2D). We applied the same model to simulate tip movements in ALI and 3D cultures. To provide a quantitative comparison of tip rearrangement due to packing, we defined a bifurcation angle ϕ and a ‘dihedral’ tip rotation angle θ, similar to ref. ³⁵. We previously quantified ϕ, which becomes smaller over developmental time in vivo^14,36. However, our interest here was to investigate the shorter timescale variation in θ caused by tips accommodating each other due to branching amidst crowding (Fig. 2C). We simulated initial tree geometries for bounding box constraints that were consistent with ALI and 3D culture formats, allowing tubules to adopt energetically favorable positions (see Methods). We then added a new pair of daughter tubules to each existing branch, and measured ϕ and θ before and after a second simulation step (Fig. 2E, Supplementary Fig. 6). We found that the change in θ was significantly higher in the model case simulating 3D culture relative to that simulating ALI, while the change in ϕ was similar between the two. This predicts that kidney flattening in ALI culture constrains tubules to the same plane, leading to a loss of the dihedral rotational degree of freedom.

To validate the geometric model predictions in our experimental system, we projected time-lapse movies of UB tip movements from ALI and 3D cultures onto the xy-plane and annotated UB branching tracks (Fig. 2F). Our analysis revealed that branching tips in ALI culture primarily elongate linearly outward in the radial direction from the center of the kidney as indicated by small rates of change in the directions of their elongation tracks. On the contrary, tips in 3D culture exhibited higher rates of change in their elongation directions due to ongoing repositioning at the explant surface (Fig. 2G, Supplementary Movie 2, 3). This suggests that the linear branching progression in ALI is likely an artifact generated by their limited mobility in the z-direction, essentially confining tip movement/rearrangement to a 2D-like environment. This result validates packing-based tip repositioning that had previously only been inferred from fixed kidneys and geometric modeling. Taken together, the data emphasize that while ALI culture constrains branching and packing to the same xy plane, 3D hydrogel culture retains distinct tip branching and packing planes as in vivo (xz/yz and xy, respectively). This restricts ALI cultures to the study of processes operating at the length scale of individual tips and nephron condensation events, but not to those operating at the scale of groups of tips or the whole organ, including those giving rise to many kidney defects.

Such a capability for capturing larger-scale 3D developmental dynamics in real-time would benefit the study of congenital kidney defects, where the nature or timing of the fault along the developmental trajectory is often obscured. In this vein we applied our 3D culture method to the glial cell-derived neurotrophic factor (GDNF)-REarranged during Transfection (RET) pathway, which is central to kidney branching³⁷. Prior work has established that increased GDNF activation by either ectopic delivery of GDNF or knockout of the signaling antagonist Sprouty (Spry1) leads to more closely spaced branch buds and dilated tips that occasionally merge in ALI culture^37,38 (Fig. 3A). We added GDNF to 3D E12 mouse kidney cultures and compared their morphology to published Spry1^−/− mutants^16,39 (Fig. 3B). GDNF overactivation similarly significantly increased area per tip in 3D^37,38(Fig. 3C). However, we noted that the morphology of the mutant is different in freshly dissected kidneys compared to those explanted and cultured at the ALI. Specifically, Spry1^−/− kidneys in ALI culture show clustered branches with many ‘buried’ tips, while freshly dissected Spry1^−/− kidneys primarily have dilated UB tips rather than clustered branches. Our +GDNF 3D cultures had dilated tips, more similar to the in vivo presentation of the Spry1^−/− mutant compared to its presentation in ALI culture (Fig. 3A, Supplementary Movie 4).

**Fig. 3: 3D culture better approximates the presentation of GDNF-RET signaling defects.**

On the other end of the GDNF-RET axis, loss of RET signaling is detrimental to kidney development, causing renal hypoplasia and agenesis⁴⁰. GDNF-RET inhibition by adding either a GDNF-neutralizing antibody or a MEK inhibitor to block RET-downstream signaling abolishes tip bifurcation in culture, leading to fewer tips in total but retaining persistent tubule elongation^37,41 (Fig. 3D). We confirmed that UB tips in E13 kidneys cultured in ALI format with the selective RET inhibitor (RET-i) selpercatinib⁴² continued to elongate but did not bifurcate, relative to those in littermate control kidneys cultured without RET-i (Fig. 3E). Surprisingly, UB tubules buckled and folded back on themselves while elongating under RET-i conditions in 3D culture, perhaps reflecting the reduced kidney spreading area:volume ratio in 3D culture relative to ALI (Fig. 3E, Supplementary Movie 5). Tubule shape analysis revealed that daughter tubules under RET inhibition acquire higher curvature in 3D culture (Fig. 3F). This morphology resembles a characteristic of renal hypoplasia resulting from retinoic acid signaling inhibition, an upstream regulator of RET⁴³. Together, our experiments with GDNF-RET perturbations indicate that 3D hydrogel culture may more faithfully model morphology changes in response to developmentally relevant defects compared to ALI culture.

ECM-derived matrix composition impacts kidney morphology

In prior work, Sebinger et al. found that ECM proteins affect kidney development in ALI cultures through their composition, concentration, and mode of presentation (e.g. in free solution vs. at the interface between kidney and substrate)¹⁹. This implies that conditions at the boundary of the kidney can affect relatively distant cell and tissue-level behaviors in the ureteric bud and nephrogenic niches at their tips. We hypothesized that one category of conditions here could be mechanical since many ECM proteins have roles in cell adhesion and kidneys fail to develop when cultured in non-adherent ALI conditions (Supplementary Fig. 7). When embedded in fluorescently-labeled collagen 1-containing C + M hydrogels, explants pulled the ECM inward over the culture period, indicating that kidneys actively exert a contractile force on their embedding matrix normal to the interface between the two (Supplementary Movie 6). More generally, we can conceive of a force balance at each point along the interface between forces due to growth, matrix elasticity, tissue contractility, and shear stress due to cell traction (Fig. 4A). Such stresses may feed into morphogenesis through effects on cell migration, differentiation, and proliferation⁴⁴.

**Fig. 4: Matrigel-collagen hydrogels allow kidney rounding and affect nephron:UB tip ratio upon culture.**

We expected that changing collagen concentration would simultaneously affect several material properties including stiffness and adhesion. To begin assaying this, we cultured kidneys in Matrigel with 1.8 mg/ml collagen (1.8 C + M) hydrogels and compared outcomes with those cultured in the Matrigel with 0.5 mg/ml collagen (0.5 C + M) hydrogels used thus far for live imaging experiments (Fig. 4A). The concentration of collagen I gels has been reported to correlate with their stiffness^25,26, predicting that the same would hold in collagen I-Matrigel composite gels. Indeed, microindentation measurements confirmed that 1.8 C + M hydrogels are stiffer than 0.5 C + M hydrogels after swelling in media to mimic the culture environment (Fig. 4B, Supplementary Fig. 8). Aside from stiffness contributions, collagen I contains binding sites for receptors such as integrins that promote cell adhesion²⁶. Therefore, we expected that the adhesion of explants to the surrounding matrix would increase with collagen concentration. Stromal cell migration outward from the explant surface is commonly seen in both ALI and 3D explant cultures (Supplementary Movie 7, 8). In C + M hydrogels, we observed different stromal cell migration patterns in the two collagen concentrations. Stromal cell migration away from the explant was higher in 1.8 C + M gels by day 3, as quantified by the relative stromal cell area beyond the explant boundary (Fig. 4C, Supplementary Fig. 9). We interpreted this as being driven by differences in matrix adhesion and/or stiffness. These data indicate that 1.8 C + M hydrogels are stiffer and potentially more adhesive to stromal cells than 0.5 C + M hydrogels.

To compare developmental outcomes between the two composite hydrogels, we imaged and annotated kidney morphology over 3-day culture. We observed that kidneys converge toward spherical shapes in both conditions, as quantified by progressively decreasing aspect ratio of kidney cross-sectional outlines (Fig. 4D). No significant differences were found in the shape or area of kidneys between the two conditions throughout the culture period (Fig. 4D, Supplementary Fig. 9). To assess branching morphogenesis and nephrogenesis more closely, we quantified the number of UB tips and early nephrons after culture. Both of these are subject to developmental variability (effective gestational age) between littermates at the start of culture^45,46 (Fig. 4E, Supplementary Fig. 10). To correct for this, we quantified the number of nephrons per UB tip as a proxy for nephrogenesis balance in each condition. In developing kidneys in vivo, this ratio is thought to be tightly controlled by a balance between nephron progenitor self-renewal vs differentiation^35,47. We found significant decreases in the nephron:UB tip ratio and tip:tip distance in 1.8 C + M as compared to 0.5 C + M gels (Fig. 4E, F, Supplementary Fig. 11). These data demonstrate that matrix concentration affects UB tip crowding and nephron:UB tip ratio in 3D culture, potentially as a function of either stiffness or adhesion at the kidney-hydrogel boundary.

Stiffness and adhesion tune explant shape and nephrogenesis

Since collagen I and Matrigel are native ECM-derived materials, it is difficult to tune stiffness and adhesion orthogonally. Additionally, their exact compositions are poorly characterized and subject to batch-to-batch variability²⁵. To delineate the effects of stiffness and adhesion on 3D kidney growth, we sought to identify engineered biomaterials that support kidney culture while offering precise control over composition and material properties. Certain engineered 3D hydrogels, namely alginate and puramatrix, have failed to support rat kidney development¹⁸. Hyaluronic acid-based hydrogels and gelatin methacryloyl (GelMA) are widely used in tissue engineering and regenerative medicine applications due to their high biocompatibility and tunability^48,49,50. In pilot experiments, both C + M and acrylated hyaluronic acid (AHA) hydrogels supported kidney culture, while GelMA did not, potentially due to free radicals generated during polymerization (Supplementary Fig. 12). We therefore pursued AHA hydrogels consisting of the AHA polymer, matrix metalloproteinase (MMP)-degradable crosslinkers (MMPc), and covalently-conjugated RGD peptides (Fig. 5A, Supplementary Fig. 13). In this system, changes in crosslinker or RGD ligand concentrations can be used to independently tune hydrogel stiffness and cell adhesion, respectively⁴⁸.

**Fig. 5: Engineered hydrogel stiffness/adhesion controls kidney shape and nephron:tip ratio.**

We first characterized the stiffness of AHA hydrogels over a range of crosslinker concentrations using microindentation. We found that 1.5 mM MMPc was a lower bound for hydrogel integrity over the culture period. Stiffness scaled with the concentration of MMPc present, validating quantitative control over this parameter. AHA gels with 1.5 and 2.25 mM MMPc and 0.5 C + M hydrogels also exhibited similar stiffnesses (Fig. 5B, Supplementary Fig. 14). To confirm orthogonal control over cell adhesion, we plated Madin-Darby canine kidney (MDCK) cells on substrates coated with AHA with 2.25 mM MMPc crosslinker, with and without RGD functionalization. MDCK cells on AHA without RGD did not attach (instead forming spheroids). In contrast, cells successfully adhered and spread on gels containing RGD (Fig. 5C). Cell adhesion to AHA therefore depends on the presence of the RGD functional groups, which did not affect gel stiffness (Fig. 5B). These data indicate orthogonal control over stiffness and adhesion in engineered AHA hydrogels.

To investigate the effect of matrix stiffness on 3D kidney development, we cultured kidneys in AHA + RGD with crosslinkers ranging from 1.5 mM to 4.5 mM (stiffness between ~1 and 20 kPa). Kidneys rounded in the softest hydrogels having 1.5 mM MMPc, similar to those in C + M hydrogels (Fig. 5D). However, kidneys grown in AHAs with ≥2.25 mM MMPc maintained their initial “bean-like” shapes (aspect ratio >1) throughout the culture. We also quantified midplane cross-sectional area increase over time, which showed that kidneys in the softest 1.5 mM AHA gels had a smaller area increase over time relative to those in gels of higher crosslinker concentration (though this effect was not significant, p = 0.074, Fig. 5D). Overall, kidneys maintained a more organotypic elongated shape in AHA + RGD hydrogels as compared to non-adhesive AHA. This may reflect differences in plastic deformation of each gel type in response to tissue-generated stresses that drive rounding. We next quantified UB tips and nephrons in kidneys cultured in AHA + RGD gels with respect to crosslinker concentration (a correlate of stiffness). We noted similar intralitter variability in the number of UB tips and nephrons as seen in C + M cultures (Fig. 5E). However, kidneys achieved a significantly higher nephron:UB tip ratio when cultured in AHA with 2.25 mM crosslinker (stiffness ~2 kPa) relative to softer or stiffer ones (14%, 11.3%, and 26.1% increases relative to 1.5, 3, and 4.5 mM MMPc gels, respectively, Fig. 5E). These data indicate a “Goldilocks” microenvironmental stiffness that maximizes nephron commitment per tip, relative to higher or lower stiffnesses.

We next studied the effect of gel adhesion to kidney morphology and nephrogenesis balance in AHA gels, independent of stiffness. Cell adhesion to the ECM is critical for many biological processes including collective cell movements and tissue-building during development⁴⁴. To probe the role of adhesive boundary conditions, we embedded embryonic kidneys in AHA gels with or without RGD functionalization. Similar to MDCK cells, we observed stromal cells interacting with the matrix in adhesive AHA gels but not in those lacking RGD (Supplementary Fig. 15). Quantification of tissue geometry over the culture period revealed a global rounding effect in kidneys cultured without RGD, similar to the phenotype observed in C + M gels and the softest AHA + RGD gels (Fig. 5F). In contrast, kidneys in +RGD gels retained their initial elongated shapes throughout culture. Kidney midplane area also radically increased in these gels (by ~38%), whereas kidneys in the same gels lacking RGD approximately maintained their initial midplane area (Fig. 5F, Supplementary Fig. 16). We hypothesized that the inability of kidneys to adhere to the 3D matrix would not only adversely affect organ shape and size but also finer-scale development. Indeed, we discovered a significant 10.3% decrease in nephron:UB tip ratio in kidneys cultured without RGD (p = 0.034, Fig. 5G). Taken together, boundary conditions of sufficient stiffness and adhesion are necessary to retain in vivo-like kidney shape and increase in organ size over time. Second, nephrogenesis balance (nephron:UB tip ratio) increases for boundary conditions having a ‘Goldilocks’ stiffness and higher adhesion (Fig. 5H).

Discussion

We developed a 3D kidney culture technique for the study of morphogenesis dynamics and the effect of matrix material boundary conditions on development. 3D embedding in ECM-derived composite hydrogel rescued in vivo-like branching as compared to traditional flattened culture at air-liquid interfaces, validated with a computational model. Changing the ECM-gel concentration altered the UB tip distribution at the organ surface. By introducing engineered AHA hydrogels that allow independent control of stiffness and adhesion, we discovered that both factors contribute to the overall 3D organ shape as well as nephrogenesis balance in culture.

Kidney culture techniques have evolved from metal grids supporting cloth/filter materials supplied with inductive signals from the spinal cord^6,7,17,51 to commercially available transwells and other formats imposing ALI conditions^8,9. However, outcomes in ALI culture can differ from those in vivo. For example, BMP7 inhibition in ALI culture leads to rare UB tip collision events, while knockout of Bmp7 in vivo does not affect tip:tip distance^16,52. Similarly, HGF signaling promotes kidney branching in culture, while Hgf^−/− mouse embryos show normal ureteric bud organization^53,54. These discrepancies motivate improved culture approaches, particularly to avoid gross tissue flattening that occurs in ALI culture. We present a 3D culture approach that better recapitulates in vivo morphogenesis. Beyond this study, it will be informative to reexamine known ALI culture artifacts in the 3D format and to revisit mouse models of kidney defects, many of which are poorly understood at the level of live cell behavior.

Our previous work showed that without an appropriately timed transition in UB tip orientation, tip crowding and repulsion at the organ surface cause a ‘buried tips’ phenotype, where some tips are displaced to deeper tissue layers¹⁴. In this study, we observed rare buried tip-cap mesenchyme niches in C + M gels but not in AHA gels (Supplementary Fig. 17). Buried tips are predicted to occur when the available organ surface area per niche falls below a minimal threshold. In the case of C + M gels, the shift in the shape of kidneys from oblong to spherical would reduce their surface area:volume ratio, consistent with the emergence of buried tips. A similar logic follows for ALI culture, where buried tips (relative to the strip of circumferential organ area) are common. This underscores the importance of appropriate control over global organ shape for the kidney to maintain correct positioning of UB tips at the surface during branching.

Our data show that stiffness and adhesion boundary conditions both affect nephron endowment per UB tip in explants. The data complement reports that stiffness and viscoelasticity affects stem cell-derived nephron-lineage kidney organoid differentiation^20,21, even though the spatial organization and reciprocal signaling between cells in organoids do not closely resemble that of UB and nephron lineages in vivo^55,56. This suggests a cell-autonomous component to interpretation of the mechanical microenvironment. Further work is needed to understand other potential effects of mechanical boundary conditions, for example in setting the length and time scales over which endogenous stresses associated with branching morphogenesis or nephron condensation persist^15,57. The connection between these factors and nephron progenitor renewal vs. differentiation balance remains an open question⁵⁸.

One area warranting future investigation are alternative materials for 3D embedding that could confer broader engineering control over physical properties and bioorthogonality. Systems with higher or lower resemblance to the native ECM properties of the kidney microenvironment are open to further study, as well as the effect of viscoelasticity^10,20. While we focused on hyaluronic acid hydrogels for their biocompatibility and potential for controlled mechanical modification, HA itself is a biologically active ECM component in mesoderm-derived tissues including the kidney. During kidney development, HA is mainly deposited by nephron progenitors around the UB tips⁵⁹. HA regulates both UB branching and nephrogenesis in a concentration and molecular weight-dependent manner when supplied in media^60,61. Though our HA gels are crosslinked and do not penetrate UB tip niches, unreacted HA chains or those liberated by crosslink degradation could diffuse into kidney tissue. However, we expect minimal confounding effects since the HA used to synthesize our gels has a molecular weight of 70 kDa, similar to a 64 kDa soluble HA condition that previously yielded comparable phenotypes to untreated controls in ALI culture⁶¹. As a preliminary test we performed ALI culture in the presence of uncrosslinked AHA and found differences in tip and nephron numbers but not nephrogenesis balance after culture (Supplementary Fig. 18). However, this simulated a maximum possible concentration expected if AHA gels had fully degraded in the 3D case (0.15%), despite AHA degradation typically being observed on the weeks timescale rather than days^62,63. This will require further study to resolve; biologically inert materials such as poly(ethylene glycol) (PEG)-based hydrogels are potential alternatives⁶⁴.

Our work motivates continued investigation of the role of the force balance at the kidney periphery to its global and local organization. Appropriate patterning of tensile, shear, and pressure forces may contribute to new strategies for rational control over morphogenetic modules such as tubule elongation, branching, and cell fluxes in the ureteric bud, cap mesenchyme, and stroma. Such a capability would address engineering barriers to setting nephron condensation, connectivity with the ureteric bud, and corticomedullary patterning toward synthetic kidney replacement tissues.

Methods

Ethical statement

All mouse experiments followed National Institutes of Health (NIH) guidelines and were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Pennsylvania.

Mouse strain and microdissection

E12-15 embryos were collected from wild-type timed pregnant CD-1 mice (Charles River Laboratories, RRID:IMSR_CRL:022). Mice were housed in standard ventilated cages (single occupancy) in a conventional rodent facility with free access to water and food and kept on a 12 h light cycle. Pregnant breeders were 8–10 weeks old upon receipt. Embryo age was confirmed by limb staging⁶⁵. Kidneys as well as whole urogenital tissues were microdissected from mouse embryos and kept on ice before experiments.

Culture plate fabrication

50 ml of 10:1 (base:crosslinker) polydimethylsiloxane (PDMS) elastomer (Sylgard 184) was cast onto a 15 cm petri dish and baked at 60 °C overnight to form a sheet. The cured PDMS was punched using a 0.625 “ hole punch and a 12 mm biopsy punch to create PDMS rings. The PDMS was then plasma-bonded to glass-bottom 12-well plates (P12-1.5H-N, Cellvis) in a plasma oven (HPT100, Princeton Scientific) and baked at 60 °C for >1 h to enhance bonding. The plates were sterilized before use and the outer space between the PDMS and the well was filled with PBS with 1X Pen-Strep.

Live antibody labeling

To live label the explants, kidneys were incubated at 37 °C in fluorophore-conjugated antibodies for 2 h and washed once before culture. Live labels were anti-CD326/EPCAM with FITC (11-5791-82, Invitrogen) and Lectin PNA from Arachis hypogaea with Alexa Fluor 647 (L32640, Invitrogen) with 1:250 dilution in culture media. In long-term time-lapse experiments, cultured explants were labeled again by adding fluorescent antibodies in media for 1 h in the incubator and washed before resuming imaging.

Explant culture in 3D hydrogel and ALI

3D culture

Dissected kidneys were washed in PBS and transferred to PDMS rings with a pre-cut gelatin-coated p20 pipette tip and excess liquid was removed. 10 µl of hydrogel solution was added to create a droplet and suspend the kidney. The culture plate was incubated at 37 °C for 30 min for gelation. After gelling, DMEM media with 10% fetal bovine serum (FBS, MT35-010-CV, Corning) and 1× pen/strep (100 IU/mL penicillin, and 100 μg/mL streptomycin, 100× stock, 15140122, Invitrogen) was added to the PDMS ring.

ALI culture

ALI culture experiments were primarily done in transwells (3460, Corning) by placing the kidneys on top of a filter membrane with 0.4 µm pores and supplying culture media underneath. The time-lapse movies at ALI were acquired using low-volume dishes modified from Sebinger et al.⁸. Briefly, kidneys were placed in the middle of the PDMS rings and a low volume of media (~80 µl) was added to generate the air-liquid interface.

Collagen and Matrigel composite hydrogel

Rat tail collagen I (354236, Corning) was neutralized on ice with 1 M NaOH, 10X PBS and DI water to the desired concentration at pH 7. For collagen labeling, NHS-ester(Succinimidyl Ester) with Alexa 555 (A20009, Invitrogen) was added to the collagen stock at 1:1000 the day before. Collagen was neutralized with 1 part labeled collagen and 2 parts unlabeled collagen. Human ESC-qualified Matrigel (354277 lot 17823003, Corning) was also labeled using the same technique. Neutralized collagen at 1 mg/ml or 3.6 mg/ml and Matrigel are then mixed at 1:1 ratio to make the 0.5 and 1.8 C + M composite gels, respectively.

Tip rotation model

The tip rotation model was based on a physics-based form-finding model built in Rhino (v.7, Robert McNeel & Associates) using the Grasshopper Kangaroo2 package (Daniel Piker). It models trees of elastic edges with self-repelling branch nodes enveloped within an elastic bounding domain. Quantitative principles of this software environment are detailed in ref. ¹⁴ We used the SphereCollide goal to model node mutual repulsion at a radius of 80 µm (mean tip-tip distance for E14 kidneys from Fig. 2B). We used the Length goal to model tree edge elasticity. Branching nodes below either the grandparent generation (3D) or the parent generation (ALI) were fixed to prevent unrealistic global rotation of tree families as seen in our time-lapse movies. We ran simulated trees within a bounding box of fixed volume at different heights to mimic the 3D and ALI environments. We estimated the initial tree (three sub-trees with two generations of branches) and bounding box geometries from our culture data shown in Figs. 1D and 2. The model was run to equilibrium before adding new tips in the same direction as parents using a customized MATLAB (2024) script. Two new tips were assigned to each terminal end with a 60-degree bifurcation angle either 90 degrees rotated from their parent plane (3D) or along the direction of their parents in the xy plane (ALI). The 3D bounding box was also expanded assuming equal increase in x, y, and z directions based on spherical growth in our 3D culture. The model was run again and ϕ and θ angles were measured both before and after this step. Detailed model parameters are included in Supplementary Fig. 3.

GDNF-RET signaling perturbations

GDNF overactivation was achieved by supplying 100 ng/ml recombinant GDNF (212-GD, R&D systems) in culture media at day 0 of 3D 0.5 C + M gel culture. Similarly, RET inhibition was carried out through adding Selpercatinib (76-961-0, Tocris) at 100 nM in the culture media. Selpercatinib was stored as a 20 mM stock solution in DMSO at −20 °C. Kidneys were cultured in exogenous factors till culture end point.

AHA hydrogel

Synthesis of acrylated hyaluronic acid was based on previous methods⁴⁸. Briefly, AHA modification was performed through the reaction of acrylic anhydride (Sigma Aldrich) with sodium hyaluronate (HA, MW = 70 kDa, Lifecore Biomedical) at a pH 9–10 for 3 h at RT in DI water. AHA was purified via dialysis, lyophilized and modification was confirmed using ¹H NMR (Bruker Neo 400). AHA of 26% modification was confirmed by ¹H NMR, and used for all experiments (Supplementary Fig. 10, analysis performed in TopSpin). To form hydrogels, RGD (GCGYGRGDSPG, Genescript) and MMP-degradable peptides (GCNSVPMSMRGGSNCG, Genescript) were conjugated to AHA via Michael addition between thiols on cysteines and acrylates during gelation. The reaction was carried out at 37 °C for 30 min in pH 8.5 Media 199 (11150067, Gibco) with 10% FBS and 1% Penstrep. 3% w/v AHA was used for all studies. For the varying crosslinker experiments, all AHAs were functionalized with 1 mM RGD.

GelMA hydrogel

To make GelMA, 5% (w/v) GelMA (A-0460, Allevi) and 0.05% (w/v) lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) were dissolved in DMEM. Kidney explants were mixed into precursor solution and exposed to UV light (365 nm, 15 mW/cm²) in culture plates for 5 min. Crosslinked hydrogels were rinsed 3× with PBS to remove free radicals prior to culture over time.

Hydrogel microindentation

Hydrogel preparation

100 µl hydrogel solutions were added to PDMS molds and incubated at 37 °C for 30 min to generate r = 4 mm cylinders. To mimic the osmotic environment in culture, formed hydrogels are then swelled in culture media overnight in the incubator before taking measurements. AHAs with varying crosslinker concentrations were all fabricated without RGD for microindentation, except for the with and without RGD comparison.

Microindentation

Microindentation was performed as described by Prahl et al.¹⁵. The system consists of a stepper motor attached to a μN-resolution tensiometer that includes a 255 µm cylindrical 30 gauge (AWG) SAE 316 L stainless steel wire and a microbalance. The indenter was lowered at a rate of 12.5 µm/s to indent the sample while force, time, and displacement data were recorded. The force-displacement relationship was found to behave as a Hookean spring with force linearly related to displacement, and the spring constant was calibrated using a glass substrate before each measurement.

2D AHA MDCK cell adhesion assay

MDCK cell line production

The MDCK-II cell line stably co-expressing Lifeact-GFP and H2B-mRuby2 was generated by viral transduction using pTK92_Lifeact-GFP (#46356, Addgene) and pLentiPGK Hygro DEST H2B-mRuby2 (#90236, Addgene) transfer vectors. Lifeact-GFP retroviral particles were produced by co-transfecting 7 × 10⁵ Lenti-X 293T cells (#632180, Takara Bio) in a 6-well plate with 1.25 µg transfer vector, 0.5 µg pCMV-VSV-G (#8454, Addgene), and 0.75 µg pCMV-gag/pol using calcium phosphate. H2B-mRuby2 lentivirus was similarly produced by co-transfecting 1.5 µg transfer plasmid along with 1.33 µg pCMV-deltaR9.81 (#12263, Addgene) and 0.17 µg pMD2.G (#12259, Addgene). For calcium phosphate transfection, plasmids were diluted in 1× HEPES buffered saline (HEBS, 2×: 50 mM HEPES, 280 mM NaCl, 1.5 mM Na₂HPO₄, pH 7.05) followed by dropwise addition of 2.5 M calcium chloride (CaCl₂) to 150 mM, and incubated for 2 min before adding to cells. Media was replaced the next day. Viral supernatants were collected after 2 days and 1 ml supernatant was used to infect 10⁵ MDCK-II cells in the presence of 8 µg/ml polybrene (#TR-1003-G, EMD Millipore). Stable cell lines were expanded and enriched for fluorescent marker expression using a BD FACSAria™ III cell sorter (BD Biosciences) and periodically exposed to 1 µg/ml puromycin (#A11138-03, Life Technologies) or 200 µg/ml hygromycin B (#10843555001, Roche) to remove non-expressing cells. Lenti-X 293T cells were cultured in DMEM supplemented with 10% FBS and pen/strep, with antibiotics removed during transfection.

2D AHA gel plating

50 µl 3% AHA solution with 2.25 mM MMP-degradable crosslinker and with or without 1 mM RGD peptides was added to 96-well glass-bottom plates and crosslinked as described above. MDCK-II Lifeact-GFP H2B-mRuby cells were lifted with 0.25% trypsin-EDTA (2530056, Corning) and plated on AHA with and without RGD at the same density. The MDCKs on AHA were cultured in minimum essential medium (MEM, Earle’s salts and L-glutamine, MT10-010-CV, Corning) supplemented with 10% FBS and 1x pen/strep overnight. Cell adhesion was quantified on the second day by manually annotating attached cells and floating cells in each well after washing.

Immunofluorescence and optical clearing

Immunofluorescence staining and imaging was performed as previously described¹⁴. Kidney explants were fixed in 4% paraformaldehyde (J19943.K2, Thermo Scientific), washed twice in 1× DPBS-glycine and once in DPBS. Blocking solution was made by adding 10% donkey serum to an IF wash solution composed of 1 mg/ml bovine serum albumin, 0.2% (v/v) Triton X-100, 0.041%(v/v) Tween-20 in 1X DPBS. Fixed samples were blocked overnight and then incubated in primary antibody in the same blocking solution at 4 °C for >2 days depending on their age. A full list of antibodies is provided in Supplementary Table 1. Samples were washed in three exchanges of 1X PBS at least 1 h each and then incubated with secondary antibodies in blocking solution at 4 °C, all with 1:300 dilution in 10% donkey-serum. Stained samples were then cleared using for 2 days in ScaleA2 (4 M urea + 0.1% Triton X-100 + 10% glycerol) followed by 2 days in ScaleB4 (8 M urea + 0.1% Triton X-100)⁶⁶, and imaged in ScaleA2.

Microscopy

Confocal imaging was performed using a Nikon Ti2-E microscope equipped with a CSU-W1 spinning disk (Yokogawa), a white light LED, laser illumination (100 mW 405, 488, and 561 nm lasers and a 75 mW 640 nm laser), a Prime BSI sCMOS camera (Photometrics) or a ORCA-Fusion camera (Hamamatsu), motorized stage, 4×/0.2 NA and 10×/0.25 NA lenses (Nikon), and a humidified stage top environmental chamber (OkoLabs). For time-lapse imaging, xy-positions were defined in the Nikon Elements software and stacks were collected using 15 µm step size every hour with the 10× objective lens. To image the cleared samples, kidneys were immersed in ScaleA2 solution in glass-bottom plates and imaged using either the 4× or 10× lens.

Kidney morphology analysis

Tip rotation analysis: Time-lapse movies were processed to remove background and speckles and correct for photobleaching over time in FIJI (version 2.16.0). Maximum projection of the live EPCAM channel was used to annotate tip movements with TrackMate⁶⁷. The segmented tips and tracks were manually filtered to remove misidentified objects. The mean directional change rate was computed by calculating the angle between two consecutive track segments (two time frames) and averaging the angles over the entire track. Kidney structure annotation: UB tips and nephrons were manually annotated from confocal stacks in FIJI using the point tool. UB tips were identified as EPCAM/ECAD+ terminal tips in live samples and RET+ tips in fixed samples. Early nephrons were identified by JAG1 expression. Nephron per tip ratio was calculated for each kidney. UB tip-tip distance: For tip-tip distance analysis, neighbors were defined as RET + UB tips that are at least two generations away (not sharing the same grandparent node) and presented in the same z-plane. The distance was obtained by measuring the distance between the lumen of one tip to its neighbor’s. Kidney cross-sectional shape analysis: The shapes of cultured kidneys were segmented from the maximum projection of the brightfield channel from confocal z-stacks. The outlines were manually traced and measured using the freehand tool and the shape descriptors in FIJI and smoothed by 10% for visual presentation in Figs. 4 and 5. The measurements are defined as: Aspect ratio: major_axis/minor_axis based on the ellipsoid fit of the segmented organ outline. 1 indicates a perfect circle. % area increase: (final area-initial area)*100%/initial area. Initial area was obtained on day 1 and final area was obtained on day 3. Stroma migration area: The stroma migration outlined was obtained by thresholding the brightfield images. The migration area was calculated by subtracting the organ outline area from the total migration outline. Tip area: RET+ tips were manually segmented using the freehand tool. Tubule curvature: Only tubules parallel to the imaging plane were selected and manually outlined by tracing the lumen up until the bifurcation point with the segmented line tool. The line ROIs were exported into the FIJI plugin “Kappa” to measure their curvatures⁶⁸.

Quantitative real-time PCR

Kidneys were manually dissected out of the gel or from the membrane on day 3 of culture and cut into pieces using dissection forceps. One freshly dissected E17 kidney was also collected as a control. RNA isolation and reverse transcription were performed using the quick RNA miniprep kit (R1053, Zymo) and high-capacity RNA-to-cDNA kit (4387406, Applied Biosystems). Expression of target genes was quantified by SYBR green master mix (A2574, Applied Biosystems) with a real-time qPCR instrument (7300, Applied Biosystems). The Hif1α expression was normalized to the Gapdh gene within each sample. Relative expression was then normalized to the E17 kidney expression. Primer sequences are included in Supplementary Table 2.

Cytokine measurement

Conditioned kidney culture media was collected from both ALI culture on transwell and 3D 0.5 C + M culture on day 2. Mouse cytokine array panel (ARY006, R&D Systems) was used to quantify the expression of 40 cytokines (full list in Supplementary Table 3) following the manufacturer’s instructions. For each panel, 1.5 mL conditioned media was used by pooling from 4 cultured kidneys from each group.

Soluble AHA in ALI culture

0.15% AHA macromer was dissolved in culture media without any crosslinker or RGD. E13 kidneys were cultured on transwells either with basal media or 0.15% soluble AHA. Cultured kidneys were fixed on day 3 for structure quantification.

Statistics and reproducibility

Statistical details for each experiment, e.g., sample size (n), litters, and type of statistical test can be found in the figure legends. Normality test was performed on all datasets before statistical analysis in Prism 10 (GraphPad). Outliers were removed using nonlinear regression (ROUT) method with false discovery rate at 1%. Unpaired two-tailed t-test was used for all comparisons between two conditions. One-way ANOVA with correction for multiple comparisons using Tukey’s test was used for all analyses for three or more conditions within one variable unless noted. Two-way ANOVA with correction for multiple comparisons within one variable was used for tests with two variables. Statistical significance denoted as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All the data generated or analyzed during this study are included within this article, its Supplementary Information and its Source Data. Raw image stacks are available upon request due to prohibitive file sizes. All raw image requests should be sent to Alex Hughes at ajhughes@seas.upenn.edu. The expected response time is one week. Once access is granted, the data will be available online for one month. Source data are available at https://github.com/AriazyH/3D-culture-rhino-matlab⁶⁹. Source data are provided with this paper.

Code availability

References

Paramore, S. V., Goodwin, K. & Nelson, C. M. How to build an epithelial tree. Phys. Biol. 19, 061002 (2022).
Short, K. M. & Smyth, I. M. The contribution of branching morphogenesis to kidney development and disease. Nat. Rev. Nephrol. 12, 754–767 (2016).

CAS PubMed Google Scholar
McMahon, A. P. Development of the Mammalian Kidney. Curr. Top. Dev. Biol. 117, 31–64 (2016).

PubMed PubMed Central Google Scholar
Houat, A. P. et al. Congenital anomalies of the upper urinary tract: a comprehensive review. Radiographics 41, 462–486 (2021).

PubMed Google Scholar
Lawlor, K. T. et al. Nephron progenitor commitment is a stochastic process influenced by cell migration. Elife 8, e41156 (2019).
Trowell, O. A. A modified technique for organ culture in vitro. Exp. Cell Res. 6, 246–248 (1954).

CAS PubMed Google Scholar
Saxen, L., Vainio, T. & Toivonen, S. Effect of polyoma virus on mouse kidney rudiment in vitro. J. Natl Cancer Inst. 29, 597–631 (1962).

CAS PubMed Google Scholar
Sebinger, D. D. R. et al. A novel, low-volume method for organ culture of embryonic kidneys that allows development of cortico-medullary anatomical organization. PLoS One 5, e10550 (2010).

PubMed PubMed Central Google Scholar
Saarela, U. et al. Novel fixed z-direction (FiZD) kidney primordia and an organoid culture system for time-lapse confocal imaging. Development 144, 1113–1117 (2017).

CAS PubMed PubMed Central Google Scholar
Varner, V. D., Gleghorn, J. P., Miller, E., Radisky, D. C. & Nelson, C. M. Mechanically patterning the embryonic airway epithelium. Proc. Natl Acad. Sci. USA 112, 9230–9235 (2015).

CAS PubMed PubMed Central Google Scholar
Shih, H. P., Panlasigui, D., Cirulli, V. & Sander, M. ECM signaling regulates collective cellular dynamics to control pancreas branching morphogenesis. Cell Rep. 14, 169–179 (2016).

CAS PubMed Google Scholar
Buchmann, B. et al. Mechanical plasticity of collagen directs branch elongation in human mammary gland organoids. Nat. Commun. 12, 2759 (2021).

CAS PubMed PubMed Central Google Scholar
Wang, S., Matsumoto, K., Lish, S. R., Cartagena-Rivera, A. X. & Yamada, K. M. Budding epithelial morphogenesis driven by cell-matrix versus cell-cell adhesion. Cell 184, 3702–3716.e30 (2021).

CAS PubMed PubMed Central Google Scholar
Prahl, L. S., Viola, J. M., Liu, J. & Hughes, A. J. The developing murine kidney actively negotiates geometric packing conflicts to avoid defects. Dev. Cell 58, 110–120.e5 (2023).

CAS PubMed PubMed Central Google Scholar
Prahl, L. S. et al. Jamming of nephron-forming niches in the developing mouse kidney creates cyclical mechanical stresses. Nat. Mater. 23, 1582–1591 (2024).
Lefevre, J. G. et al. Branching morphogenesis in the developing kidney is governed by rules that pattern the ureteric tree. Development 144, 4377–4385 (2017).

CAS PubMed Google Scholar
Rienhoff, W. F. Development and growth of the metanephros or permanent kidney in chick embryos (eight to ten days’ incubation). Johns. Hopkins Hospital Bull. 33, 392–405 (1922).

Google Scholar
Rosines, E. et al. Constructing kidney-like tissues from cells based on programs for organ development: toward a method of in vitro tissue engineering of the kidney. Tissue Eng. Part A 16, 2441–2455 (2010).

CAS PubMed PubMed Central Google Scholar
Sebinger, D. D. R., Ofenbauer, A., Gruber, P., Malik, S. & Werner, C. ECM modulated early kidney development in embryonic organ culture. Biomaterials 34, 6670–6682 (2013).

CAS PubMed Google Scholar
Nerger, B. A. et al. 3D hydrogel encapsulation regulates nephrogenesis in kidney organoids. Adv. Mater. 36, e2308325 (2024).

PubMed PubMed Central Google Scholar
Garreta, E. et al. Fine tuning the extracellular environment accelerates the derivation of kidney organoids from human pluripotent stem cells. Nat. Mater. 18, 397–405 (2019).

CAS PubMed PubMed Central Google Scholar
Kobayashi, A. et al. Identification of a multipotent self-renewing stromal progenitor population during mammalian kidney organogenesis. Stem Cell Rep. 3, 650–662 (2014).

CAS Google Scholar
Levinson, R. S. et al. Foxd1-dependent signals control cellularity in the renal capsule, a structure required for normal renal development. Development 132, 529–539 (2005).

CAS PubMed Google Scholar
Hum, S., Rymer, C., Schaefer, C., Bushnell, D. & Sims-Lucas, S. Ablation of the renal stroma defines its critical role in nephron progenitor and vasculature patterning. PLoS One 9, e88400 (2014).

PubMed PubMed Central Google Scholar
Frantz, C., Stewart, K. M. & Weaver, V. M. The extracellular matrix at a glance. J. Cell Sci. 123, 4195–4200 (2010).

CAS PubMed PubMed Central Google Scholar
Tang, V. W. Collagen, stiffness, and adhesion: the evolutionary basis of vertebrate mechanobiology. Mol. Biol. Cell 31, 1823–1834 (2020).

CAS PubMed PubMed Central Google Scholar
Timin, G. & Milinkovitch, M. C. High-resolution confocal and light-sheet imaging of collagen 3D network architecture in very large samples. iScience 26, 106452 (2023).

CAS PubMed PubMed Central Google Scholar
Lipp, S. N., Jacobson, K. R., Hains, D. S., Schwarderer, A. L. & Calve, S. 3D mapping reveals a complex and transient interstitial matrix during Murine kidney development. J. Am. Soc. Nephrol. 32, 1649–1665 (2021).

CAS PubMed PubMed Central Google Scholar
Hughes, C. S., Postovit, L. M. & Lajoie, G. A. Matrigel: a complex protein mixture required for optimal growth of cell culture. Proteomics 10, 1886–1890 (2010).

CAS PubMed Google Scholar
Akram, K. M. et al. Live imaging of alveologenesis in precision-cut lung slices reveals dynamic epithelial cell behaviour. Nat. Commun. 10, 1178 (2019).

PubMed PubMed Central Google Scholar
Little, M. H., Combes, A. N. & Takasato, M. Understanding kidney morphogenesis to guide renal tissue regeneration. Nat. Rev. Nephrol. 12, 624–635 (2016).

CAS PubMed Google Scholar
Yu, W. et al. Simple rules determine distinct patterns of branching morphogenesis. Cell Syst. 9, 221–227 (2019).

CAS PubMed PubMed Central Google Scholar
Chang, C.-H. & Davies, J. A. In developing mouse kidneys, orientation of loop of Henle growth is adaptive and guided by long-range cues from medullary collecting ducts. J. Anat. 235, 262–270 (2019).

PubMed PubMed Central Google Scholar
al-Awqati, Q. & Goldberg, M. R. Architectural patterns in branching morphogenesis in the kidney. Kidney Int. 54, 1832–1842 (1998).

CAS PubMed Google Scholar
Short, K. M. et al. Global quantification of tissue dynamics in the developing mouse kidney. Dev. Cell 29, 188–202 (2014).

CAS PubMed Google Scholar
Lindström, N. O., Chang, C.-H., Valerius, M. T., Hohenstein, P. & Davies, J. A. Node retraction during patterning of the urinary collecting duct system. J. Anat. 226, 13–21 (2015).

PubMed Google Scholar
Vega, Q. C., Worby, C. A., Lechner, M. S., Dixon, J. E. & Dressler, G. R. Glial cell line-derived neurotrophic factor activates the receptor tyrosine kinase RET and promotes kidney morphogenesis. Proc. Natl Acad. Sci. USA 93, 10657–10661 (1996).

CAS PubMed PubMed Central Google Scholar
Basson, M. A. et al. Branching morphogenesis of the ureteric epithelium during kidney development is coordinated by the opposing functions of GDNF and Sprouty1. Dev. Biol. 299, 466–477 (2006).

CAS PubMed Google Scholar
Basson, M. A. et al. Sprouty1 is a critical regulator of GDNF/RET-mediated kidney induction. Dev. Cell 8, 229–239 (2005).

CAS PubMed Google Scholar
Skinner, M. A., Safford, S. D., Reeves, J. G., Jackson, M. E. & Freemerman, A. J. Renal aplasia in humans is associated with RET mutations. Am. J. Hum. Genet. 82, 344–351 (2008).

CAS PubMed PubMed Central Google Scholar
Fisher, C. E., Michael, L., Barnett, M. W. & Davies, J. A. Erk MAP kinase regulates branching morphogenesis in the developing mouse kidney. Development 128, 4329–4338 (2001).

CAS PubMed Google Scholar
Oliveira, L. C. B. & Mulligan, L. M. Selpercatinib: first approved selective RET inhibitor. Cell 186, 1517 (2023).

CAS PubMed Google Scholar
Rosselot, C. et al. Non-cell-autonomous retinoid signaling is crucial for renal development. Development 137, 283–292 (2010).

CAS PubMed PubMed Central Google Scholar
Wu, D., Yamada, K. M. & Wang, S. Tissue morphogenesis through dynamic cell and matrix interactions. Annu. Rev. Cell Dev. Biol. 39, 123–144 (2023).

CAS PubMed PubMed Central Google Scholar
Allen, E. & MacDowell, E. C. Variation in mouse embryos of 8 days gestation. Anat. Rec. 77, 165–173 (1940).

Google Scholar
Gupta, I. R., Lapointe, M. & Yu, O. H. Morphogenesis during mouse embryonic kidney explant culture. Kidney Int. 63, 365–376 (2003).

PubMed Google Scholar
Cebrian, C., Asai, N., D’Agati, V. & Costantini, F. The number of fetal nephron progenitor cells limits ureteric branching and adult nephron endowment. Cell Rep. 7, 127–137 (2014).

CAS PubMed PubMed Central Google Scholar
Xu, K. L. et al. Microinterfaces in biopolymer-based bicontinuous hydrogels guide rapid 3D cell migration. Nat. Commun. 15, 2766 (2024).

CAS PubMed PubMed Central Google Scholar
Highley, C. B., Prestwich, G. D. & Burdick, J. A. Recent advances in hyaluronic acid hydrogels for biomedical applications. Curr. Opin. Biotechnol. 40, 35–40 (2016).

CAS PubMed Google Scholar
Yue, K. et al. Synthesis, properties, and biomedical applications of gelatin methacryloyl (GelMA) hydrogels. Biomaterials 73, 254–271 (2015).

CAS PubMed PubMed Central Google Scholar
Grobstein, C. Inductive epitheliomesenchymal interaction in cultured organ rudiments of the mouse. Science 118, 52–55 (1953).

CAS PubMed Google Scholar
Davies, J. A., Hohenstein, P., Chang, C.-H. & Berry, R. A self-avoidance mechanism in patterning of the urinary collecting duct tree. BMC Dev. Biol. 14, 35 (2014).

PubMed PubMed Central Google Scholar
Schmidt, C. et al. Scatter factor/hepatocyte growth factor is essential for liver development. Nature 373, 699–702 (1995).

CAS PubMed Google Scholar
van Adelsberg, J. et al. Activation of hepatocyte growth factor (HGF) by endogenous HGF activator is required for metanephric kidney morphogenesis in vitro. J. Biol. Chem. 276, 15099–15106 (2001).

PubMed Google Scholar
Howden, S. E. et al. Plasticity of distal nephron epithelia from human kidney organoids enables the induction of ureteric tip and stalk. Cell Stem Cell 28, 671–684.e6 (2021).

CAS PubMed Google Scholar
Porter, C. M., Qian, G. C., Grindel, S. H. & Hughes, A. J. Highly parallel production of designer organoids by mosaic patterning of progenitors. Cell Syst. 15, 649–661.e9 (2024).

CAS PubMed PubMed Central Google Scholar
Liu, J., Prahl, L. S., Huang, A. Z. & Hughes, A. J. Measurement of adhesion and traction of cells at high yield reveals an energetic ratchet operating during nephron condensation. Proc. Natl Acad. Sci. USA 121, e2404586121 (2024).

CAS PubMed PubMed Central Google Scholar
Davis, S. N. et al. Nephron progenitors rhythmically alternate between renewal and differentiation in synchrony with kidney branching morphogenesis. bioRxiv, https://doi.org/10.1101/2023.11.21.568157 (2023).
Rabelink, T. J., Wang, G., van der Vlag, J. & van den Berg, B. M. The roles of hyaluronan in kidney development, physiology and disease. Nat. Rev. Nephrol. 20, 822–832 (2024).
Pohl, M., Sakurai, H., Stuart, R. O. & Nigam, S. K. Role of hyaluronan and CD44 in in vitro branching morphogenesis of ureteric bud cells. Dev. Biol. 224, 312–325 (2000).

CAS PubMed Google Scholar
Rosines, E., Schmidt, H. J. & Nigam, S. K. The effect of hyaluronic acid size and concentration on branching morphogenesis and tubule differentiation in developing kidney culture systems: potential applications to engineering of renal tissues. Biomaterials 28, 4806–4817 (2007).

CAS PubMed PubMed Central Google Scholar
Khetan, S. & Burdick, J. A. Patterning network structure to spatially control cellular remodeling and stem cell fate within 3-dimensional hydrogels. Biomaterials 31, 8228–8234 (2010).

CAS PubMed Google Scholar
Wade, R. J., Bassin, E. J., Rodell, C. B. & Burdick, J. A. Protease-degradable electrospun fibrous hydrogels. Nat. Commun. 6, 6639 (2015).

CAS PubMed Google Scholar
Xu, K. L., Mauck, R. L. & Burdick, J. A. Modeling development using hydrogels. Development 150, dev201527 (2023).

CAS PubMed PubMed Central Google Scholar
Wanek, N., Muneoka, K., Holler-Dinsmore, G., Burton, R. & Bryant, S. V. A staging system for mouse limb development. J. Exp. Zool. 249, 41–49 (1989).

CAS PubMed Google Scholar
Hama, H. et al. Scale: a chemical approach for fluorescence imaging and reconstruction of transparent mouse brain. Nat. Neurosci. 14, 1481–1488 (2011).

CAS PubMed Google Scholar
Tinevez, J.-Y. et al. TrackMate: An open and extensible platform for single-particle tracking. Methods 115, 80–90 (2017).

CAS PubMed Google Scholar
Mary, H. & Brouhard, G. J. Kappa (κ): analysis of Curvature in Biological Image Data using B-splines, https://doi.org/10.37473/dac/10.1101/852772 (2019).
Huang, A. Z. Engineering kidney developmental trajectory using culture boundary conditions. Github, https://doi.org/10.5281/ZENODO.15708471 (2025).
Li, W., Hartwig, S. & Rosenblum, N. D. Developmental origins and functions of stromal cells in the normal and diseased mammalian kidney: origins and functions of kidney stromal cells. Dev. Dyn. 243, 853–863 (2014).

PubMed Google Scholar

Download references

Acknowledgements

We thank past and current members of the Hughes lab for their help and discussions, especially J. Viola, J. Liu, and S. Grindel. We also thank Y. Zhang and D. Huang for advice on biomaterials, and X. Luo for help with the Rhino model. We are grateful to P. Mollenkopf and P. Janmey for training and assistance with material characterization. We thank G. Timin and M. Milinkovitch for sending image files from ref. ²⁷. Cell sorting was performed on a BD FACSAria Fusion maintained by the Penn Cytomics and Cell Sorting Resource Laboratory and obtained in part through NIH grant S10 1S10OD026986. This work was supported by the trainee pilot award from Center for Engineering MechanoBiology (CEMB), an NSF Science and Technology Center, under grant agreement CMMI: 15-48571 (A.Z.H), NIH F32 fellowship DK126385 and Penn Center for Soft & Living Matter fellowship (L.S.P.), NIH F30 fellowship F30AG074508 (K.X.), NSF CAREER award 2047271 (A.J.H.), NIH NIDDK R01DK132296 (A.J.H.), NSF MRSEC award DMR-2309034 (A.J.H.), and Pennsylvania Department of Health HRFF CURE 585499 (A.J.H.).

Author information

Authors and Affiliations

Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, USA

Aria Zheyuan Huang, Louis S. Prahl, Karen Xu, Robert L. Mauck, Jason A. Burdick & Alex J. Hughes
Bioengineering Graduate Group, University of Pennsylvania, Philadelphia, PA, USA

Aria Zheyuan Huang, Karen Xu, Robert L. Mauck & Alex J. Hughes
Center for Engineering Mechanobiology, University of Pennsylvania, Philadelphia, PA, USA

Aria Zheyuan Huang, Karen Xu & Robert L. Mauck
Center for Soft and Living Matter, University of Pennsylvania, Philadelphia, PA, USA

Louis S. Prahl & Alex J. Hughes
McKay Orthopedic Research Laboratory, Department of Orthopaedic Surgery, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Karen Xu & Robert L. Mauck
Translational Musculoskeletal Research Center, Corporal Michael J. Crescenz VA Medical Center, Philadelphia, PA, USA

Karen Xu & Robert L. Mauck
BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, CO, USA

Jason A. Burdick
Department of Cell & Developmental Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Alex J. Hughes
Cell and Molecular Biology Graduate Group, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Alex J. Hughes
Institute for Regenerative Medicine, University of Pennsylvania, Philadelphia, PA, USA

Alex J. Hughes

Authors

Aria Zheyuan Huang
Louis S. Prahl
Karen Xu
Robert L. Mauck
Jason A. Burdick
Alex J. Hughes

Contributions

A.Z.H. and A.J.H. designed the project. A.Z.H. performed explant/cell line experiments and hydrogel characterization with assistance from L.S.P. Data analysis was by A.Z.H. The engineered hydrogel study was carried out by A.Z.H. and K.X. R.L.M. and J.A.B. assisted with the interpretation of the engineered hydrogel data. Writing was by A.Z.H. and A.J.H. Reviewing and editing was by A.Z.H., L.S.P., K.X., R.L.M, J.A.B. and A.J.H. All authors discussed results, assisted in data interpretation, and provided input on the paper.

Corresponding author

Correspondence to Alex J. Hughes.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Ian Smyth, who co-reviewed with Kieran Short, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Source data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Huang, A.Z., Prahl, L.S., Xu, K. et al. Engineering kidney developmental trajectory using culture boundary conditions. Nat Commun 16, 7829 (2025). https://doi.org/10.1038/s41467-025-63197-5

Download citation

Received: 28 January 2025
Accepted: 12 August 2025
Published: 22 August 2025
DOI: https://doi.org/10.1038/s41467-025-63197-5