Enhanced metanephric specification to functional proximal tubule enables toxicity screening and infectious disease modelling in kidney organoids

Abstract

While pluripotent stem cell-derived kidney organoids are now being used to model renal disease, the proximal nephron remains immature with limited evidence for key functional solute channels. This may reflect early mispatterning of the nephrogenic mesenchyme and/or insufficient maturation. Here we show that enhanced specification to metanephric nephron progenitors results in elongated and radially aligned proximalised nephrons with distinct S1 – S3 proximal tubule cell types. Such PT-enhanced organoids possess improved albumin and organic cation uptake, appropriate KIM-1 upregulation in response to cisplatin, and improved expression of SARS-CoV-2 entry factors resulting in increased viral replication. The striking proximo-distal orientation of nephrons resulted from localized WNT antagonism originating from the organoid stromal core. PT-enhanced organoids represent an improved model to study inherited and acquired proximal tubular disease as well as drug and viral responses.

Introduction

The proximal tubules (PTs) of the kidney represent a highly specialised portion of the nephron performing the bulk of kidney reabsorption and secretion. This occurs via three distinct functional and anatomical segments: the convoluted (S1 and S2) and the straight (S3) segments that traverse the cortico-medullary boundary, with S1 exhibiting the highest capacity for solute, sodium, amino acid, and fluid transport¹. Their unique roles and high metabolic activity render the PTs acutely vulnerable to toxins and metabolic stress². As such, accurately patterned and segmented PTs would represent a critical tool for drug development, toxicology research, and studies of PT dysfunction.

We and others have established robust protocols for the directed differentiation of human pluripotent stem cells to kidney^3,4,5,6,7. While these organoids display a remarkable transcriptional similarity to the developing human kidney^8,9,10,11, their nephron patterning and segmentation remain immature, more closely resembling human trimester 1 fetal tissue⁶. PT maturation and functional segmentation are particularly underdeveloped. Despite possessing nuclear HNF4A (responsible for driving early proximal patterning ref. 12) and apical CUBILIN-MEGALIN complex expression, organoid PTs lack a range of functional solute channels that define each PT subsegment^11,13. Expression levels of the principle water transporting channel, AQP1, the organic anion transporters (OATs), and the organic cation transporters (OCTs) are all low¹³.

Such suboptimal PT maturation may represent inappropriate anteroposterior/mediolateral patterning, suboptimal maintenance of progenitor identity or incomplete maturation. In response to distinct temporospatial signalling, the permanent (metanephric) kidney arises during human embryogenesis as the final of three embryonic excretory organs, developing sequentially from specific rostrocaudal regions of the intermediate mesoderm¹⁴. Metanephric development, commencing during weeks 4−5 of gestation, is preceded by the formation of two more rostral transient organs; the pronephros (human gestation weeks 3–4) and the mesonephros (human gestation week 4–10)¹⁵. While the mammalian pronephros is highly rudimentary, mesonephric nephrons also arise via MET and show similar patterning and segmentation to early metanephric nephron. However, the mesonephros possesses less definitive distal tubule segments and regresses around week 8^16,17,18.

Using fluorescent reporter lines and lineage tracing in human kidney organoids, we have confirmed both the presence of a SIX2⁺ nephron progenitor population and the contribution of these cells to nephrogenesis via MET in kidney organoids^19,20. However, the possibility exists that we are modelling mesonephric rather than metanephric nephrogenesis, potentially contributing to poor PT patterning and maturation (reviewed in ref. 21). It is also possible that suboptimal maintenance of progenitor identity during iPSC differentiation in vitro limits nephron maturation. Several media have been described that are able to support the maintenance of isolated nephron progenitors in vitro^22,23,24,25. While each media contains low levels of canonical WNT activity and FGF2/9, distinct differences in nephron patterning result from the inclusion of a variety of TGFβ superfamily agonists (BMP4, BMP7, Activin A) and antagonists (A83-01, LDN193189), NOTCH inhibition (DAPT), and other growth factors (TGFα, IGF1/2, LIF). The inclusion of LDN193189 (inhibitor of BMP receptor-mediated SMAD1/5/8) supported tubular patterning but not formation of glomeruli²². In contrast, the addition of LIF and either dual-SMAD inhibition (LDN193189 and A83-01) or NOTCH inhibition (DAPT) resulted in the formation of nephrons with podocytes but different nephron morphologies^23,25. Finally, while proximodistal nephron patterning in mouse has previously been shown to be influenced by relative Wnt, Bmp, and Notch signalling in mouse²⁶, these data suggest that distinct nephron progenitor states may show varying competence for different nephron segments, or that distinct SIX2 populations give rise to different regions of the nephron.

Here we show that patterning to a posterior metanephric SIX2⁺ nephron progenitor population by extending the duration of mesodermal patterning, while simultaneously enhancing nephron progenitor expansion, specifies progenitors with improved metanephric identity without influencing anteroposterior/mediolateral patterning. These progenitors form strongly proximalised, elongated, and spatially aligned nephrons. The PTs within these nephrons display distinct segmentation into S1, S2 and S3 cell types, upregulation of key solute channels and transporters, and functional uptake of albumin and organic cations. Treatment with cisplatin upregulates Kidney Injury Marker-1 (KIM-1), while increased expression of key viral entry factors enables improved SARS-CoV-2 infection and replication compared to standard protocols. Notably, striking nephron alignment results from localised WNT antagonism, supporting a role for WNT gradients in human nephron proximodistal patterning. Taken together, this study suggests a requirement for optimal nephron progenitor commitment for appropriate PT identity. Such PT-enhanced kidney organoids represent a model of the human proximal nephrons with likely applications for infectious and genetic disease research, drug development, and nephrotoxicity evaluation.

Results

Delayed nephron induction supports nephron progenitors

As noted previously, optimisation of nephron progenitor maintenance in vitro has been investigated by a range of studies using murine and human pluripotent stem cell-derived nephron progenitors^22,23,25. While all studies reported maintenance of nephron progenitors, variations were evident with respect to the final patterning of resulting nephrons following induction. Given the clear influence that initial differentiation conditions and timing can have on nephron progenitor survival and subsequent nephron patterning, we hypothesised that expanding our nephron progenitor population whilst delaying nephron initiation may create a more metanephric population leading to organoids with improved patterning and PT maturation. We have previously shown that SIX2 expression is not detected until day 10 of pluripotent stem cell differentiation¹⁹. Hence, the initial monolayer differentiation phase was prolonged to between 12–14 days, along with culture in either of two previously defined nephron progenitor (NP) maintenance media, NPSR²³ and CDBLY²⁵ from day 7, which represents the point of intermediate mesoderm commitment^6,27 (Fig. 1a). Compared to control media (TeSR-E6; E6), both NPSR and CDBLY prevented spontaneous epithelialisation of the monolayer (Fig. 1b). However, very little epithelialisation and poor nephron commitment was observed after culture in NPSR (Fig. 1b). In contrast, CDBLY preserved the nephron-forming capacity of the progenitor cells following their formation into a micromass and induction of nephrogenesis with a pulse of canonical WNT signalling) (Fig. 1b). Nephrons of these organoids were also observed to surround a stromal core region that stained positive for markers of kidney stroma MEIS1/2/3 and SNAI2 (SLUG) (Supplementary Fig. 1a)²⁸. Upon prolonged organoid culture (>14 days), portions of this core region formed patches of Alcian blue-positive cartilage (Supplementary Fig. 1B)….