Human Amniotic Membrane as a Novel Scaffold for Inducible Pluripotent Stem Cell-derived Kidney Organoids

Human inducible pluripotent stem cells (hiPSCs) can be differentiated into kidney organoids1 that could be used for engineering replacement kidneys. However, delivery approaches need to be further explored. Here, we describe how hiPSC-derived kidney organoids can be differentiated on decellularized human amniotic membrane (dhAM). We envision that an elastic dhAM could be useful as a scaffold for the implantation of kidney organoids into the peritoneum, where dynamic volumetric changes are expected. We investigated the effects of axial stretch on the kidney organoids on the dhAM and observed that gentle stretch does not disrupt tubules and that tubular elongation can be induced. The need for kidney tissue engineering is significant. Millions of patients worldwide have end-stage renal disease, and for many, transplantation is the best treatment option. However, there is a severe shortage of donor kidneys available. Previous research has demonstrated that kidney organoid vasculature can be perfused after implantation in the kidney subcapsular space.2 However, filtrate outflow tracts are lacking in such a system. Furthermore, in end-stage renal disease patients, significant renal fibrosis or even cystic disease would impair successful vascular anastomoses and perfusion if implanted in this manner. Alternative heterotopic implantation strategies should be considered like peritoneal implantations where filtrate could be more readily removed (Figure 1A).Figure 1.: dhAM as a scaffold for differentiation of kidney organoids. A: Concept of kidney organoid implantation into the peritoneum using dhAM as a biologic scaffold. Left panel illustrates normal peritoneum showing mesothelial cells overlying the normally vascularized submesothelial layer. Lumen side is indicated. Middle panel shows implantation of dhAM with loosely attached differentiated kidney organoids (purple) coated with a biodegradable hydrogel-like fibrin to support vascularization of the kidney organoids from the submesothelial layer. Right panel shows that once vascularized, the kidney organoids will generate filtrate that should separate organoids from the dhAM. Filtrate could be drainable by a peritoneal dialysis-like catheter or a tissue-engineered ureter-like construct. For clarity, structures are not drawn to scale. B: Multiphoton microscopy image of kidney organoid differentiated on dhAM. Inset shows closer view of complex tubular structures within. Images are pseudo-colored with H&E. C: Immunofluorescence of kidney organoid differentiated on dhAM. Lotus tetragonolobus lectin (green-proximal tubule), anti-E-cadherin (white-distal tubule), and anti-CD31 (red-vascular endothelial cells) were used for staining. Scale bars are (B) 100 and (C) 50 µm. dhAM, decellularized human amniotic membrane; H&E, hematoxylin and eosin.A structural scaffold upon which kidney organoids could be delivered would have significant utility for heterotopic implantation. The scaffold should be biocompatible with both the host and kidney organoids. The material should be supportive of kidney organoid differentiation. It should be durable, not immunogenic, and easily manipulated for surgical procedures. Thus far, only decellularized kidney tissue, fibrin/collagen, and silk have been investigated as scaffolds for kidney organoids.3 However, these biomaterials require complex cellular localizations or lack the elasticity to accommodate volume changes that would occur with large-volume filtrate production. Here, we report that dhAM is an elastic biomaterial that could serve as kidney organoid scaffold. AM is the innermost layer of placental membranes and is composed of a single layer of epithelial cells and a basement membrane composed of collagen I, -III, -IV, laminin, and fibronectin.4 It has remarkable durability and elasticity, as it must protect and accommodate a growing fetus. hAM is readily attained postpartum and has been used for numerous surgical and tissue engineering applications, including burn treatment, wound healing, vaginal reconstruction, and corneal ulcerations.4–6 In previous preclinical bladder reconstruction surgeries, AM implants were vascularized at 7 days, had host smooth muscle cell recellularization by 21 days and lasted for at least 42 days in a rat.7 For kidney organoid differentiation, hiPSC-derived kidney organoids were generated as described previously with APEL2 Medium (Stemcell Technologies) supplemented with 5% protein-free hybridoma medium 2 and antimycotic-antibiotic (Life Technologies).1 hiPSCs (designated Y6 at the Yale Stem Cell Center) were generated from neonatal skin fibroblasts using integration-free Sendai virus. Decellularization of the hAM was carried out as previously described.4 iPSCs cell aggregates were pelleted directly onto dhAMs saturated with differentiation media to form kidney organoids. Kidney organoids could be readily differentiated on the dhAM (Figure 1, B and C). Organoids were loosely attached, and the side of the dhAM did not affect attachment or differentiation. The striking durability and elasticity of AM could be beneficial for transplantation of kidney organoids into the peritoneum, where stretch of the biomaterial scaffold after implantation and volume changes due to filtrate outflow from the tubules would be anticipated. To test the effects of stretch on the kidney organoids attached to the dhAM, we built a stretch device to facilitate kidney organoid differentiation and application of uniaxial dhAM stretch (Figure 2A). The dhAMs were held by adjustable clamps, and organoids were loaded onto the center of the membrane. A glass well beneath the dhAM contained media to supply organoids with nutrients and to keep the membrane saturated. We investigated how acute stretch would affect kidney organoids and also how long-term stretch might affect the morphology of the tubules within the organoids. For acute stretch (within 2 minutes), dhAM was stretched 33% beyond the initial length. Long-term stretch was initiated on day 10 of differentiation (after tubule formation) and continued daily for 10 days total to achieve the same degree of elongation. To assess tubular morphology, we used multiphoton microscopy as previously described8 to achieve the necessary depth of visualization. Z-stacks were reconstructed using Fiji9 with 3D Roi plugin10 (Figure 2B). We observed that application of acute and long-term stretch (10 days) increased the elongation of the tubules (Figure 2C) but did not damage them. In addition, tubular volumes were increased significantly when stretch was applied over a prolonged period (Figure 2D). These data suggest that tubular morphology itself can be modified by the application of mechanical forces onto an elastic substrate.Figure 2.: Tubular effects of axial stretch of dhAM and kidney organoids. A: Stretch device schematic and titanium version used for experiments. B: Examples of 3D tubular reconstructions after multiphoton microscopy of experimental groups: control (no stretch), acute, and long-term stretch applications. Scale bars for horizontal axes are 80 µm. C: Elongation and (D) tubular volumes of control, acute, and long-term (10 days) stretch of kidney organoids. Statistical analyses were unpaired Welch’s t-tests performed with Prism 7.04 (GraphPad) software. dhAM, decellularized human amniotic membrane.We believe that the dhAMs are a promising scaffold for studying effects of mechanical forces on human kidney organoids in vitro and can help facilitate the implantation of kidney organoids into the peritoneum or other heterotopic locations. In vivo, dhAMs would likely experience multiaxial stretch with volumetric changes. We would anticipate that such multiaxial stretch would also increase tubular volume; however, tubular elongation may be limited. In terms of tissue engineering, dhAM organoid-containing sheets could be used as a patch, be readily modified into tubes after rolling, or could be used to form saclike structures. Fortunately, the normal complement of nephrons may not be necessary to achieve therapeutic benefit, as even 10–15% of renal function might be sufficient to maintain patients off dialysis. Future studies will investigate how dhAMs and kidney organoids react to implantation in the peritoneum of an animal model system. In particular, we will focus on the durability of the dhAM, organoid, and host responses. Critical questions will be whether vasculature of the kidney organoids is perfused, can filtrate form and accumulate, and whether the host can re-epithelialize the implanted dhAM. If re-epithelialization occurs, we anticipate deposition of host extracellular matrix. Of note, dhAM has not been approved by the FDA for purposes described here, and experiments shown represent investigational work only.