Dual-origin vascularization of kidney organoids on the chick CAM

What the work claims

This is a methods-plus-mechanism primary study from the Vainio group at the University of Oulu. It makes two linked claims. First, the vasculature of the developing mouse kidney has a dual origin: intrinsic endothelial progenitors identified as CD146-positive cells and more mature CD31-positive endothelial cells build most of the renal and glomerular endothelium (vasculogenesis), while endothelial cells from the graft host contribute additional vessels (angiogenesis).¹

Second, a practical one: if you strip embryonic kidney tissue of every CD31-positive and CD146-positive cell, the reaggregated organoid forms glomerular structures with normal nephrogenic potential but no vessels of its own, and when grafted onto the vascularized chicken chorioallantoic membrane it is vascularized exclusively by host endothelial cells, which reach the glomerular tufts and carry circulating host blood cells.¹

How it works

The technical advance is a mini reservoir for the classic chorioallantoic membrane (CAM) assay. Traditional CAM grafting drops an organ rudiment onto the membrane of a fertilized chicken egg and accepts that many grafts fail to attach, float, or dry out. Here, commercial 12-well inserts are trimmed into 2 mm-high rings with a 0.4 micrometer transparent membrane; the E11.5 mouse kidney sits on the inverted insert, which is pressed onto the CAM and filled with medium. Graft survival rises from roughly 62 percent to roughly 97 percent (n = 4 biological replicates, P < 0.05 by Welch's t-test), and in a side-by-side series 8 of 8 improved-assay grafts vascularized versus 1 of 10 by the traditional method.¹

Origin was resolved with two complementary lineage systems. In Tie1Cre;mT/mG mice, endothelial-lineage cells fluoresce green: after seven days on the CAM, glomerular tufts consist mostly of mouse-derived green endothelial cells. In the reciprocal setup, kidneys are grafted onto CAMs from ubiquitously green fluorescent chickens: chimeric vessels containing both mouse CD31-positive and chicken-positive endothelium appear, confirming host contribution.¹

The depletion experiment is the load-bearing result. Dissociated E11.5 kidney cells were FACS-sorted to remove CD31-positive, CD146-positive, or both populations, then reaggregated. Sorted organoids verified by immunostaining and by qPCR (reduced Tie1, Tie2, VEGFR2 and VE-cadherin) formed no endothelial network in static organ culture, yet stereomicroscopy showed nephrogenesis proceeding normally in all of them. Grafted onto the CAM, the double-depleted organoids developed glomeruli at s-shaped-body to tuft-stage maturity containing exclusively chicken endothelial cells, with wheat-germ-agglutinin labeling confirming blood flow through the vessels.¹

Where a skeptic should push

The most load-bearing assumption is that FACS depletion removed endothelial progenitors cleanly. The authors' own data qualify it: qPCR showed not only endothelial genes but also Gata3, a ureteric-bud epithelial marker, reduced in the depleted fraction, and rare residual CD31-positive cells appeared outside glomeruli in some grafted organoids, which the authors attribute to incomplete depletion or differentiation from a double-negative precursor. "Endothelial depletion" may therefore be partially "endothelial plus some epithelial depletion," and a hidden progenitor reservoir is explicitly left open in the discussion.¹

Second, the functional bar is deliberately low. Transmission electron microscopy shows chicken erythrocytes inside glomerular capillaries next to podocyte-like cells with junctions near the basement membrane, which demonstrates perfusion, not filtration. No glomerular barrier selectivity, no podocyte foot-process maturity benchmarked against adult kidney, no filtration assay of any kind appears in the paper.¹

Third, the model system is a mouse-embryo, chicken-egg chimera. The CAM is an embryonic, immunotolerant, tumor-like vascular bed. It is an excellent window onto developmental mechanism and a poor proxy for what a human iPSC kidney organoid will do under the kidney capsule, on a chip, or in a patient. The paper also settles a forty-year-old dispute with Sariola's 1983 claim that grafted kidney vessels are exclusively avian, but one alternative the authors themselves raise is uncomfortable: endothelial cells may simply die in grafts before host vessels arrive, so "dual origin" could partly reflect differential survival rather than a programmed developmental program.¹

Finally, scale: survival quantification rests on n = 4 biological replicates per condition, and the depletion-to-grafting pipeline is artisanal. None of this invalidates the lineage findings, but it caps how far each number should travel.

Kidney organoid vascularization and drug screening

For organoid models of human organs, the transferable message is about where the bottleneck sits. The field's standard complaint is that iPSC kidney organoids build nephrons but no vessels. This paper shows, in the embryonic setting, that the failure is environmental, not intrinsic: the kidney primordium carries competent endothelial progenitors that survive and assemble glomerular endothelium the moment they are given a perfused bed, and even a completely avascular organoid can recruit host endothelium into its glomeruli. Whatever static dish culture is missing, it is not the tissue's capacity to vascularize.¹

That reframes two drug-discovery workhorses. Nephrotoxicity screening on avascular kidney organoids is measuring the response of an organoid that has never experienced flow: shear stress, endothelial crosstalk, and oxygen gradients are all absent, so toxicity thresholds learned there may not transfer to perfused tissue. The depletion-reconstitution design here is, almost by accident, a screening assay: reaggregate defined endothelial-free kidney organoids, expose them to candidate pro-vascular factors or endothelial co-cultures, and graft to read out recruitment and perfusion. That is a blueprint for screening vascularization strategies that current static protocols cannot run.¹

The threat is over-translation. A result obtained in an embryonic mouse-to-chicken xenograft will be cited as evidence that human kidney organoids "can be vascularized," and grant language will run ahead of the fact that no human iPSC-derived tissue, no filtration function, and no adult-equivalent maturation appears anywhere in this study. The CAM itself cannot be the endpoint for therapeutic work: it is a developmental imaging platform, not a transplant bed or a screening format.

The bottom line

Established: with a perfused host bed, embryonic mouse kidney tissue and even endothelial-depleted kidney organoids vascularize, glomeruli acquire flow-carrying endothelium of partly donor and partly host origin, and the mini-reservoir CAM assay makes that process visible and quantifiable. Not established: that the same progenitor logic operates in human iPSC kidney organoids, or that any grafted organoid in this system filters blood. The confirming experiment is a human organoid version of the depletion-reconstitution assay with a filtration readout; the breaking experiment would be evidence that human kidney organoid endothelium is entirely host-recruited with no intrinsic progenitor contribution at all.