Human nephrogenesis is complete by 34–36-week gestation (WG), with approximately 60% of nephrons forming during the third trimester.1 Human nephron numbers (nephron endowment) range ten-fold, from 0.2 to 2.7 million nephrons per kidney,23–4 and those at the low end are at high risk for hypertension, CKD, and ESKD.56789101112–13 Currently, neonates born at 24 weeks have a survival rate over 60%.14 Because nephrogenesis will progress for no more than 40 days after premature birth,15 preterm infants born between 24 and 30 WG are at the low end of nephron endowment, and at greater risk for CKD and ESKD in adulthood.1617–18 Thus, improved survival of extremely low birth weight and preterm infants will increase the incidence of CKD and ESKD in this population.
In all vertebrates, nephron progenitor cells (NPCs) form nephrons (undergo nephrogenesis), initially through a process associated with bifurcation (branching) of the ureteric bud (UB), referred to here as branching nephrogenesis (BN). In mice, BN persist until postnatal day 3, at which point the progenitor pool differentiates en masse without additional branching and nephrogenesis is completed.19,20 Human nephrogenesis transitions from BN (5–15 WG), to the short arcading phase (15–22 WG), followed by a third period of nephrogenesis (23–36 WG) in the absence of UB branching. The newly formed nephrons in this period connect to an elongating collecting duct, creating lateral “branches.” We will describe this process as lateral branch nephrogenesis (LBN) (Supplemental Figure 1). BN generates around 33,000 nephrons through 15 cycles of a repetitive bifurcation (215) in humans.21 Arcading22,23 and LBN2122–23 combine to create 6–75 nephrons near each of the original 33,000 terminal branch tips to bring the final nephron count between 200,000 and 2.7 million, respectively,21,24 with LBN contributing the vast majority (>60%) of these nephrons during the third trimester. LBN in humans was described morphologically by perinatal autopsy microdissection studies in the 1960s and 1970s,22,23 but no molecular studies were performed given the immense ethical and technical confounders associated with analysis of late-gestation samples. During this critical developmental period, tissue can only be obtained after intrauterine fetal demise or death in the neonatal intensive care unit, resulting in RNA degradation due to the prolonged postmortem interval before preservation.25 Recent in-depth analyses of single-cell RNA sequencing (scRNA-Seq) from early human nephrogenesis have been described but, understandably, the analyses are restricted to time points preceding the neonatal limit of viability (sequencing data up to 17 WG262728–29 and do not capture data relevant to LBN). A recent scRNA-Seq study included human fetal kidneys up to 25 WG,30 however, the entire kidney was dissociated, and the low number of sequenced progenitor cells within the dataset provide limited insight into the nephrogenic zone at that time point. Although these studies established that NPC populations in all stages of human nephrogenesis express SIX2, SIX1, MEOX1, and CITED1,3132–33 the absence of a molecular understanding of LBN hampers the development of therapeutic interventions aimed at improving nephron endowment.
The transition to LBN via arcading may reflect a primate-specific evolutionary adaptation. Indeed, previous morphologic studies have suggested nonhuman primate renal development closely resembles that of humans.3435–36 Nephrogenesis in the rhesus, with material made available to us through ongoing perinatal research collaborative efforts,3738–39 begins late in the first trimester, and continues until the middle of the third trimester, providing a potential model for molecular analyses of LBN.3435–36 We correlated the morphology of rhesus and human in three-dimensional (3D) renderings of late-gestation nephrogenesis, measuring the distance to the last bifurcating branch point and other morphometric criteria to characterize rhesus LBN, and establish it is morphologically similar to late-gestation human archival kidneys. We then performed scRNA-Seq on four preterm rhesus kidneys and single-nucleus RNA-Seq (snRNA-Seq) on one kidney sample to characterize cortical cell populations during LBN. Comparison of the rhesus LBN NPC, the midgestation human NPC, and E14.5, E18.5, and P0 mouse NPC transcriptomes show greater similarity between rhesus NPC and the P0–E18.5 mouse NPC, whereas the human midgestation transcriptome resembled the E14.5–E18.5 mouse NPC. Rhesus NPC and UB-tip markers not reported elsewhere were validated using immunofluorescence and in situ hybridization (RNAScope). We provide an easy-to-use interactive website to enable community access to the molecular data.
All primate procedures were performed at the California National Primate Research Center at University of California Davis, according to the Institutional Animal Care and Use Committee approved protocol 20330 awarded to Drs. Alan Jobe and Claire Chougnet. Kidneys collected in 2015–2016 were bisected and fixed in formalin until 2017, when they were exchanged for storage in PBS with 0.01% sodium azide. Kidneys from 2017 to 2018 were fixed in formalin for 2 days before storage in azide/PBS.37,40,41 Fresh kidneys from control animals euthanized in accordance with the protocol were removed in necropsy and shipped overnight on ice in histidine-tryptophan-ketoglutarate solution utilized for solid organ transplant.42
Human Kidney Studies
All human kidney samples were obtained from the Cincinnati Children’s Hospital Medical Center Biobank after review and approval by the Internal Review Board for use of Discover Together biobank. Consent for storage and future use of biobank tissue was obtained at the time of autopsy. Tissue for biobank storage was fixed in formalin and preserved as formalin-fixed paraffin-embedded (FFPE) blocks for long-term storage. Gestational age was determined by the obstetric/neonatologist documentation taken at the time of autopsy. Kidneys were chosen on the basis of gestational age, low maceration score, and description of grossly normal kidneys in the pathology report. Only kidneys with positive immunofluorescence for SIX1/SIX2 and KRT8 signal were used in this study. For these, the gestational age at death, sex of patient, and postmortem interval is reported in Supplemental Table 1.
Primary antibodies used include SIX1 (128915, 1:500; Cell Signaling Technology), SIX2 (11562–1-AP, 1:1000; Proteintech), KRT8+18 (AB194130, 1:500; Abcam), Lamininß1 (MA5–14657, 1:400; Thermo Fisher Scientific), e-cadherin (618082, 1:400; BD Biosciences). Secondary antibodies include donkey anti-rabbit 488 (711–546–152, 1:250; The Jackson Laboratory), donkey anti-mouse 594 (715–586–151, 1:250; The Jackson Laboratory), donkey anti-guinea pig 647 (AP193SA6, 1:250; Millipore), and goat anti-guinea pig 750 (ab175758; Abcam).
Thick Tissue-section Clearing and Staining
Fixed rhesus tissue was manually sectioned into cross-sectional and cortical 500 µm sections using a razor blade and slicer matrix. Tissue sections were submerged in Quadrol and rocked at 37°C for 24 hours to aid in reducing autofluorescence.43,44 Tissue sections were then embedded in 4% acrylamide hydrogel and heated as previously described.45 The tissue was washed and cleared using 8% SDS in 0.1 M PBS46 solution. Cleared tissue was blocked for 24 hours at room temperature, and rocking in 6% BSA, 0.1% triton, 0.01% sodium azide, with 5% normal donkey serum, followed by incubation in primary antibodies for 4 days, rocked at 37°C (with fresh antibody exchange after 48 hours), followed by incubation in secondary antibody for 4 days, rocked at 37°C. After washing, the kidney tissue was incubated in Refractive Index Matching Solution45 for a minimum of 24 hours at 4°C until the tissue appeared clear by visual inspection. Please see Supplemental Methods for additional detail.
FFPE-sectioned Tissue Immunofluorescence
Rhesus and human FFPE was performed as described in Supplemental Methods. For rhesus tissue fixed in formalin for prolonged periods before arrival at our center, we performed an additional step of incubating slides in 25% Quadrol at 37°C for 24 hours after antigen retrieval to reduce autofluorescence. The slides were blocked and permeabilized in PBS with 0.3% triton with 10% normal donkey serum for 1 hour at RT, followed by incubation overnight in primary antibody at 4°C in humified chamber. Slides were then incubated in secondary antibody for 1 hour at RT, followed by three washes in PBS Tween for 5 minutes each, protected with Prolong Gold Antifade reagent, and covered.
Fluorescence Microscopy of Cleared Tissue
Cleared thick-tissue sections were mounted in a Falcon Tissue Culture Dish within a 1mm thick press-to-seal silicone isolator (Millipore Sigma). The tissue is submerged in Refractive Index Matching Solution45 and covered with a WillCo-dish glass-bottom dish. The 3D images were obtained using a Nikon A1R GaAsP Upright Microscope using the 10× glycerol objective. Next, 500-µm z-stacks were obtained at 512×512 pixels using 2.6 µm step size at 2.51 µm/px. FFPE images were acquired on the Nikon A1 Inverted Confocal Microscope using the 20× objective. The 1024×1024 pixel images were acquired at 1.1 µm step size and 0.61 µm/px.
Quantitative Image Analysis of Cleared Tissue
The 3D tissue z-stack images were processed using Bitplane Imaris 9.3.1 as described in the Supplemental Methods. To quantify the average length to the last branch point in rhesus kidneys, the cross-sectional kidney samples were measured from cortex to base to a maximum depth of 850 µm to standardize depth analysis for all samples. Surfaces for KRT8 and/or CDH1 were created. Branch point depth in 797 ureteric stalks from 126 to 138 DG in the rhesus and 53 stalks in the 32 WG human were individually determined by the investigator. Distances between the last bifurcation point to cortical surface was measured using the measurement tool in Imaris. Niche tips were quantified using “spot detection” of the surfaces rendered ureteric tip using KRT8 and/or CDH1. MATLAB extension was used to calculate the nearest nephrogenic tip using “spot to spot closest distance” function from center of each spot.
Single Molecule Fluorescence In Situ Hybridization Using RNAScope
RNAScope was performed using the Multiplex Fluorescent V2 Assay (Advanced Cell Diagnostics, Inc.) according to the manufacturer’s protocol. Positive controls used were human PPIB (medium abundance) and POL2RA (low abundance). The negative bacterial control (DapB) was used on all tissue samples. Probes used included human SIX1, SHISA8, CACNA1E, PTCHD1, ATP1A4, CCL26, TWIST1, POU3F4, AKR1C1, SAMSN1, and GCNT4. After completion of the manufacturer’s recommendation for RNAScope V2 kit, slides were blocked at room temperature for 1 hour in Tris-buffered saline with 1% BSA, 5% normal goat serum, and incubated in primary antibody guinea pig anti-cytokeratin 8/18 (Abcam) overnight at 4°C, followed by wash and secondary antibody incubation for 1 hour with goat anti-guinea pig 750 (Millipore). Slides were washed and mounted with prolong gold antifade reagent. Images were acquired on the Nikon A1R GaAsP Inverted Confocal Microscope with 20× or 40× objective and laser settings adjusted using positive and negative controls. All subsequent images were acquired with a fixed optical/exposure configuration. The 1024×1024 pixel confocal images were acquired at 1.1 µm step size and 0.61 µm/px (20×), and 0.325 µm step size and 0.16 µm/px (40×). Images of POU3F4 and TWIST1 comparing cortex, cortical-medullary junction, and medulla were acquired using the Nikon Ti-2 SpectraX Widefield Microscope using the 20× objective with all images of the same probe acquired with fixed optical/exposure configurations.
Quantitative Analysis of RNAScope
Regions of interest (ROI) measuring 300×300 pixels were chosen on the basis of whether the 40× image contained NPC, differentiated NPC, or stroma. NPCs were further broken down into NPC 1 (closest to cortical surface), NPC 2 (intermediate), and NPC3 (furthest from cortical surface). At least three ROIs were analyzed per sample of each NPC region. Cropped 300×300 pixel ROI were converted to Imaris files. All display adjustments in Bitplane Imaris 9.3.1 retained the same configuration as initial acquisitions to confirm accurate counts. We used the “spots” algorithm to identify individual transcripts on the basis of an estimated transcript diameter of 0.4 µm and the fluorescence detection threshold at 488, TRITC, and Cy5, respectively, on the basis of negative controls for the respective channels. Algorithm settings remained consistent for both negative control slides and all human tissue samples. Please see Supplemental Methods for additional details.
Mean±SD was determined for the distance from last branch point to cortical surface, distance between niche tips, average number of tips per niche clusters. The median and range for RNAScope transcript counts in varying NPC regions, and plots visualizing young versus old NPC comparisons were calculated using GraphPad Prism 8 software. Mann–Whitney U test was used to compare differences between absolute SHISA8 and SIX1 transcript counts, and SHISA8/SIX1 NPC young and old, as data were not normally distributed. P<0.05 represented statistically significant difference.
Single-Cell Suspension Preparation, scRNA-Seq Procedure, and snRNA-Seq Procedure
Details for all single-cell dissociation are elaborated in the Supplemental Information. In brief, two protocols were used.47 Manually dissected cortical tissue from three donor kidneys were kept on ice and dissociated using cold-active enzyme mixture, enhanced by repeated trituration with a 1 ml pipette and vigorous shaking between each step. Cell suspensions were filtered, red blood cells lysed, and the final pellet was washed before loading into the 10× chromium apparatus. One kidney sample was vortexed in cold collagenase mix to generate a cortically biased cell suspension, and after filtration, processed as above. For snRNA-Seq protocol, the nuclei were isolated from this frozen tissue using a protocol described previously.48
Single-Cell and Trajectory Inference Analyses
The raw scRNA-Seq and snRNA-Seq data (10× Genomics 3’ version 3) were aligned to the Ensembl version 91 reference transcriptome using the Cell Ranger version 3.1.0 workflow and processed downstream using the ICGS249 and cellHarmony50 workflows in the software AltAnalyze. To infer predicted pseudotime trajectories within the full rhesus scRNA-Seq dataset, we applied the software Monocle2, SlingShot,51 and Velocyto52 on the ICGS2 obtained UMAP cell-barcode coordinate plot. Please see Supplemental Methods “Single-Cell Analyses” and “Lineage Trajectory Analyses” for further details, respectively.
Rhesus Nephrogenesis Cessation Occurs During the Third Trimester
The gestational period of the rhesus lasts 165 days gestation (DG) and proceeds in three trimesters. Stereology determined that 1100 nephrons are present per kidney at 80 DG (middle of the second trimester), 32,000 at 100 DG (late second trimester), 130,000 at 120 DG (early third trimester), and approximately 350,000 nephrons near term.34 The third trimester in the rhesus34 corresponds to 26–40 WG in humans. The 3D renderings of third-trimester rhesus macaque nephrogenic niches from 35 formalin-fixed, cleared, and acrylamide-embedded rhesus samples spanning 124–147 DG (Supplemental Table 2) were used to determine the timing of nephrogenesis cessation and the presence of lateral branches (see Methods for further details). Some of the samples were exposed in utero to LPS and Escherichia coli infections, or to steroid administration, mimicking some perinatal human exposures in utero. The tissue was immunostained with SIX2 to identify NPC, cytokeratin 8/18 (KRT8) to identify the collecting ducts, and e-cadherin (CDH1) to visualize epithelia. Lamininß1, a basement membrane marker, identified SIX1/2+ cells undergoing mesenchymal to epithelial transition5354–55 (Supplemental Figure 1, A and B). It is important to note recent studies have questioned the specificity of Krt8 to the collecting duct as it has been found in both UB-derived epithelium and NPC-derived distal nephron.30,56 NPC clusters condense from 126 to 134 DG, and are depleted by 136 DG (Figure 1, A5). The bulbous ureteric tip (Figure 1, A1 and B1) matures into a thin branch similar to the lateral branches lower in the stalk, marking the end of both nephrogenesis and the ureteric tip (Figure 1, A5, B5, and C5, yellow arrow). From these analyses, we determined that SIX2 immunoreactivity in the rhesus disappeared between 136 and approximately 147 DG, becoming sparse after 138 DG. We sorted the animals by perinatal exposure and observed a trend (without statistical significance) toward earlier differentiation and cessation in those exposed to perinatal infection or steroid exposure, common exposures for premature infants, compared with control animals.
Rhesus and Human Display Similar Morphology of LBN
To compare the rhesus nephrogenic niche to late-gestation human kidney, we used six human samples from our institutional biobank on the basis of preservation of SIX1/2 immunofluorescence (Supplemental Table 2). Samples ranged from 16 to 32 WG to include early– and late–second-trimester and third-trimester samples. As no previous study has visualized third-trimester human kidney 3D morphology, we processed one 32 WG thick sample for 3D renderings under the same conditions applied to the rhesus tissue (see Methods; Figure 1, A and B). As shown in Figure 1, A6, B6, D, and F, the 32-WG human kidney closely resembles the organization of rosettes, NPC distribution, and lateral branches of the 126–131 DG rhesus kidney, with lateral branch points observed along single ureteric stalks at regular intervals (Figure 1, B and E, Supplemental Movies 1 and 2). As previously described by Lindström et al.27 the ureteric tip is bulbous and capped with asymmetric SIX2+ NPCs (Figure 1, D and E). Bilateral and unilateral CDH1+, KRT8− lateral branches are seen along the length of the ureteric stalk in both rhesus and humans. As mentioned above, a lateral branch may be in fact part of the distal nephron invading the elongating duct,30,56 but this distinction will not be addressed further here. The SIX2+ NPC and associated UB tips form rosette-like patterns throughout the cortex in both human and rhesus (Figure 1F), indicating this process is consistent among primates. Thus, the 3D morphology and markers of the primate late-gestation kidney are distinct from the mouse.
To confirm the rhesus third-trimester kidney is engaged in LBN, we identified the last bifurcating branch point visualized to a depth of 850 µm during late-gestation rhesus kidney development. A total of 797 ureteric stalks from 126 to 138 DG were analyzed. Only 4.3% of all stalks in the rhesus, and 3.8% in the human, contained a bifurcation (marking the last branching event) at that depth (Figure 2, A–C, blue arrow, Supplemental Data 1). The last branchpoint was distributed randomly among unbranched ureteric stalks (Figure 2C, blue arrow). These data suggest the majority of UB tips transition to LBN almost synchronously throughout the cortex, with minimal variation, confirming late-gestation human and rhesus kidneys proceed along the same developmental trajectory,22,23,26,28 consistent with a developmentally coordinated molecular mechanism.
Finally, we investigated the distances between ureteric tips and the number of tips per nephrogenic niche cluster (Figure 2, D–G, Supplemental Data 1). The average minimum distance between tips is 130.9 µm (±30.68 µm) at 124 DG, decreasing by 19% to 106.3 µm (±33.31 µm) at 132 DG, followed by a 28% increase in distance to 147.7 µm (±45.83 µm) at 136 DG with two to nine tips residing in each radially symmetric rosette-like cluster (Figure 2, F and G, Supplemental Data 1). We hypothesize this sequential increase could be attributed to the maturation and lengthening of the ureteric tip after nephrogenesis ends. The mean tip number per cluster was 4.43±1.20 at 124 DG, increasing modestly to a mean of 4.87±1.01 at 137 DG. This increase is small relative to the midgestation human kidney seen by Lindström et al.26,28 likely due to minimal bifurcating branching events at late gestation.
The same analysis was also performed on a 32-week human sample, with an average tip distance of 123±33.89 µm and 5.44±0.99 tips per cluster (Figure 2, E and G). This suggests that after 23 WG27 or 124 DG, a few new tips may be added via rare bifurcating events, but the overwhelming majority of nephrons are added via LBN that does not increase UB tip numbers. These data solidify our conclusion that morphologically, the third-trimester rhesus kidney is resembles the 32-week human kidney.
sc and snRNA-Seq Reveals Novel Progenitor Populations During Lateral Branch Nephrogenesis
To investigate the transcriptome of NPC undergoing LBN, scRNA-Seq was performed on cold protease-dissociated cortical cells from four kidneys obtained from four different control fetal rhesus age 129–131 DG, the time point that most closely resembled the human 32 WG.47 We applied the unsupervised analysis workflow (ICGS2)49 to identify distinct cell populations from the combined cortical scRNA-Seq data. For these analyses, we parameterized the software to stringently exclude cell-cycle effects, which have been previously shown to hinder the separation of self-renewing from other populations of NPCs, during cluster discovery.49 This analysis revealed 37 transcriptionally distinct cell clusters from 23,608 cells, with no strong evidence of batch/donor effects (Figure 3, A and B, Supplemental Tables 3–5). To obtain possible cell-type identities of the cell clusters, we analyzed the top 100 ICGS2 significant marker genes using a large, assembled database of over 2300 human scRNA-Seq cell-type marker sets, including human fetal kidney progenitors from three independent studies,26,29,30 using the GO-Elite57 algorithm run from AltAnalyze (Supplemental Figures 2 and 3A, Supplemental Table 6). These data identified the expected continuum of nephron cell states, spanning diverse epithelial cell types (Figure 3A, Supplemental Table 7).
Among these populations, we note that CITED1 was the principal defining marker for a cluster of 316 predicted self-renewing naïve NPCs (c25), marked also by expression of previously identified NPC markers (MEOX1, EYA1, and TMEM100). This cluster was most similar to human 16–17 WG NPCs26,29 (Supplemental Tables 8 and 9). No apparent bias was observed in this cluster among any one of the four different rhesus fetal kidneys (ranging from 52 to 102 cells per donor; Figure 3C, Supplemental Figure 4B). Adjacent clusters (c29, c30) resembled primed NPCs. When designated as the origin cluster for the trajectory prediction software SlingShot, c25 was predicted to contain a lineage branch point going to (1) podocyte, (2) epithelial, or (3) NPC in-cycle (c29)/Cap mesenchyme (c30) (Figure 3B, Supplemental Figure 4A).
To determine whether any cortex populations were missing or underestimated in our scRNA-Seq database, we performed snRNA-Seq48 on frozen tissue from the 129-DG cortex sample analyzed by scRNA-Seq. This sample yielded 5972 nuclei, corresponding to 29 ICGS2 clusters (Figure 3D, Supplemental Tables 10 and 11). We annotated these cell populations against the same reference single-cell marker database and found similar alignments to human kidney fetal cell populations as to our scRNA-Seq (Supplemental Figures 4, C and D, Supplemental Table 12). In this dataset, we found a single cluster (c26), with a near identical GO-Elite enrichment profile to that the presumptive self-renewing, naïve NPCs scRNA-Seq cluster (c25) (Supplemental Table 13). As an orthogonal method to assess the similarity of cells between the scRNA-Seq and snRNA-Seq, we used the label projection tool, cellHarmony, to identify nearest neighbor cells between these two datasets.50 Using the cortical scRNA-Seq dataset as a reference, we were able to identify the large majority of similar cell populations (Figure 3E, Supplemental Figure 5, A–C, Supplemental Table 10). However, cellHarmony identified two nuclei clusters that were missing (unclassified) from the scRNA-Seq in this comparison; c18, predicted to be neurons, and c26 (self-renewing NPCs, marked by EYA1, MEOX1, CITED1, and TMEM100) containing many nuclei transcribing unspliced pre-mRNA transcripts that were retained in the nucleus before cytoplasmic export (Figure 3F, Supplemental Figure 3D).We note some overlap of single-cell cluster 25 and 26 compared with single-nucleus cluster 26 (Figure 3D). This is likely attributed to the fact that snRNA-Seq resulted in only 29 ICGS2 clusters, whereas scRNA-Seq resulted in 37 clusters. Similar to previous reports,26,29SIX1/2 transcripts, weakly detected in the NPC by scRNA-Seq or snRNA-Seq, are present as confirmed by RNAScope and immunofluorescence.
To determine if primate late-gestation NPC subsets differ from early-gestation NPCs, we next compared the rhesus NPC and human NPC gene sets to mouse NPC transcriptome58(preprint) analyzed at different developmental stages (Figure 3G), and to scRNA-Seq data from human NPCs that identified two self-renewing NPC clusters (CRABP2 and SLC15A1) at (16–17GA).26,29 Late-gestation rhesus NPC markers more closely aligned to late-gestation murine NPCs, whereas human early second-trimester (16–17 WG) NPCs (CRABP2+) aligned more closely to midgestation murine NPCs, consistent with an age-dependent shift in the primate NPC transcriptome. To assess the developmental similarity of the rhesus and human NPCs, we performed a cross-species integrative analysis of all 16–17 WG and the rhesus single-cell/nucleus profiles (Supplemental Figure 4, Supplemental Tables 14–16). The late-gestation rhesus progenitor markers and early-gestation human markers26 show overlap within the same clusters, yet overall, there is a species-specific and/or temporal shift differentiating human and rhesus cells (Supplemental Figure 4, B and D).
Validation of Transcripts Expressed During Late-Gestation Primate Development
Analysis of these rhesus and human scRNA-Seq datasets identified several differentially expressed transcripts (>1.2 fold) increased in late-gestation rhesus NPC versus non-NPC (scRNA-Seq and snRNA-Seq) that were not identified in midgestation human NPC, compared with non-NPC cell populations (http://altanalyze.org/LateGestationNephrogenesis.html). These genes include CACNA1E, CCL26, and PTCHD1 in both datasets and ATP1A4 and SHISA8 in scRNA-Seq data. Note that snRNA-Seq highlights transcripts actively being transcribed, whereas scRNA-Seq reflects mRNA transcripts already present.
Having validated the use of RNAScope on human archival material (Supplemental Figure 5), we applied this technique to several of these rhesus LBN NPC-enriched markers. CACNA1E was detected at low levels in the NPC and differentiated NPC (Supplemental Data 2). PTCHD1 was noted at low levels in the NPC, and ATP1A4 and CCL26 were not detected with existing human probes. By contrast, SHISA8 is abundantly expressed in a cortical/medullary gradient within the SIX1+ progenitor population (Figure 4, A and B, Supplemental Data 2), increasing in NPC located furthest from the cortex (Figure 4, B and C; NPC1 marks the cortical NPC, NPC2 is intermediate and NPC3 are the deepest). Cells in the NPC3 zone were occasionally positive for the apical marker ZO-1 (Supplemental Figure 1E) but not fully epithelialized (CDH1, Lamininß1 negative, Supplemental Figure 1, C and D). We first assessed the absolute transcript counts of SIX1 and SHISA8 within the three NPC regions. We found there was no statistically significant difference in SIX1. By contrast, we did find a significant difference between NPC1 versus NPC3 (P<0.0001) and NPC2 versus NPC3 (P=0.02) in the older gestation samples only, and notably a significant difference between NPC3 young and old samples (P=0.03). Due to concern for intersample variability between NPC niches, we also assessed the ratio of SHISA8/SIX1 to normalize for these potential differences. Higher SHISA8/SIX1 ratio in NPC1 versus NPC3 was significant in all gestational ages (P=0.007 at 16–17 WG; P<0.0001 at 26–27 WG), but SHISA8/SIX1 ratio in NPC2 versus NPC3 only reached significance in the 26–27 WG (P<0.0001). Interestingly, old NPC3 have a significantly higher SHISA8/SIX1 ratio than young NPC3 (P=0.04). These data could suggest the overall abundance of SHISA8 and SHISA8 relative to SIX1 increases with gestational age as the NPC population becomes primed for differentiation, consistent with the tipping point model.59
Tip and Stalk Populations Identified in the Ureteric System During LBN
The UB may play a role in the transition from BN to LBN. Two UB clusters (c11 and c13) were identified in our scRNA-Seq on the basis of known marker gene expression. To delineate subsets of UB tip and stalk cells, we subclustered these 1350 cells after removal of a single doublet-cell cluster (DoubletDecon) into two tip and four stalk clusters (Figure 5A, Supplemental Tables 17–19). UMAP visualization of these clusters suggests a lineage transition starting from c3/c9 (tip) through c7, with a subset of stalk c1 cells positioned adjacent to cells in the tip. The same trajectory was also inferred by pseudotemporal ordering with Monocle2 (Supplemental Table 20). The markers TWIST1 and POU3F4 were highly expressed broadly throughout the tip (c3/c9), overlapping with prior defined UB markers. c3 was characterized by AKR1C1, SAMSN1, and GCNT4 (Figure 5, B and C, Supplemental Figure 5), and c9 by AKR1B1, C2orf40, and EHF (Supplemental Figure 5). RNAScope probes for hTWIST1 and hPOU3F4 (UB tip) or hAKR1C1, hSAMSN1, and hGCNT4 (C3 early-tip specific) found that POU3F4 was asymmetrically expressed across the UB with strong cortical bias (Figure 6, Supplemental Figure 6, B and C). TWIST1 was expressed in both the UB and stroma (Supplemental Figure 6, D and E), but absent from the cortical-medullary junction or the medulla (Supplemental Figure 6, F and G). AKR1C1, SAMSN1, and GCNT4 were not detectable. These data suggest novel UB markers can be expressed in a more restricted manner, suggesting a spatial and temporal heterogeneity. However, a much larger sample needs to be analyzed to determine if any UB transcripts contribute to the BN to LBN transition.
Given the profound implication of low nephron numbers to human health and longevity, there is an unmet need to characterize the primate-specific process of generating most nephrons and sustaining it in those born prematurely. The direct study of third-trimester neonatal human kidney RNA is not feasible, as tissue can only be obtained after intrauterine fetal demise or death in the neonatal intensive care unit, resulting in RNA degradation during the postmortem interval.25 Therefore, we leveraged two existing resources—the human bioarchive at Cincinnati Children’s Hospital Medical Center (CCHMC) used for validation with RNAScope and ongoing study generating both archival and fresh material in rhesus—to demonstrate the suitability of the rhesus as a bridge for molecular study of late-gestation human nephrogenesis. We established that, despite its reduced size, the rhesus kidney engages in lateral branching in a manner highly reminiscent of human development down to the fine details of cortical substructure. At the molecular level, the small sample size was sufficient to show that aging is associated with a different molecular signature in primate NPCs, similar to the rodent.59 As NPC age, they transition from a BN transcriptome to one that may support or promote LBN and later, differentiation. Alternatively, age-dependent changes in the stroma, the UB, or the crosstalk between these niche components sustains LBN.
Although our UB tip sample was small, we did validate two unexpected findings related to TWIST1 and POU3F4. TWIST1, a driver of epithelia to mesenchyme transition, is expressed after renal insults from CKD and inflammation.60TWIST1 is expressed broadly in the epithelia (collecting duct, proximal tubule, and distal tubule) of renal transplants.61 It is faintly expressed in the intercalated B cells of cortical collecting duct in the mouse62 and human,63 but its role in ureteric development is not known. POU3F4 is present in the principal cells of the collecting duct of adult kidney transplants,61 and is enriched in the B intercalated cells and principal cells of adult kidney snRNA-Seq paired with snATAC-seq data.64(preprint) Whereas POU3F4 was enriched in the tip in our expression data, we found a broad, cortically biased expression domain that included both tip and stalk. It is possible that a larger sample would have identified this transcript in stalk cluster. POU3F4 is a transcription factor involved in neural cell commitment and differentiation,65 previously identified in the developing kidney of the Xenopus, and localized to the distal and connecting tubule in the mature pronephric tubules. POU3F4 is present in the principal cells of the collecting duct of adult kidney transplants,61 and is enriched in the B intercalated cells and principal cells of adult kidney snRNA-Seq paired with snATAC-Seq data.64(preprint) Although POU3F4 was enriched in the tip in our expression data, we found a broad, cortically biased expression domain by RNAScope that included both tip and stalk. It is possible a larger sample would have identified this transcript in stalk clusters.
One transcript increasing with age in rhesus and humans is SHISA8, known also as CKAMP29 and ORF26. SHISA8 is a ubiquitously expressed membrane protein that is enriched in the brain-controlling AMPA-type glutamate receptors.66,67SHISA8 was recently identified in ciliated cell progenitors in the developing murine airway.68 Given that ciliopathies underly numerous kidney pathogenesis, it may be interesting to study the effect of SHISA8 in the context of cystic kidney disease.
In summary, our analysis is the first to molecularly analyze LBN in a nonhuman primate model and establish it as a bona fide bridge for molecular study of late-gestation human nephrogenesis. An increase in sample size will enable identification of LBN transcriptome and help formulate mechanistic hypotheses when coupled with validation studies on human archival material. Nonetheless, our preliminary molecular findings do support the tipping point model of accumulating age-dependent changes within the nephrogenic niche during late-gestation renal development. The data created in this study are available in an interactive website (http://altanalyze.org/LateGestationNephrogenesis.html) to serve as a resource for all researchers interested in human and primate nephrogenesis. We are convinced the study of late-gestation primate kidney development will prove critical to decreasing CKD risk in infants born very prematurely, and this study takes the first step toward a mechanism-based therapeutic intervention.
R. Kopan reports receiving research funding from the National Institutes of Health (NIH); reports having patents and inventions with Washington University; reports being a scientific advisor or member of the Developmental Cell Editorial Board; and having other interests/relationships as a member in the American Association for the Advancement of Science. All remaining authors have nothing to disclose.
Funding for this project was provided by the CCHMCPediatric Cell Atlas Center to N. Salomonis and R. Kopan and NIHRO1 DK106225 to R. Kopan. M. Schuh was funded by Pediatric Scientist Development Program (4K12HD000850-32/Cincinnati Children’s Research Foundation), the K12/Child Health Research Career Development Award (5K12HD028827-28), and the CCHMC P50 Pediatric Center of Excellence Pilot and Feasibility Funding (P50DK096418). Rhesus kidneys were slavaged from animals studied with support from the Bill and Melinda Gates Foundation (OPP 1132910) (Dr. Jobe, PI), NIHU01 ES029234 (Dr. Chougnet, PI), a Burroughs-Wellcome Fund award (Dr. Chougnet, PI), and a CCHMCPerinatal Infection and Inflammation Collaborative Grant (Dr. Chougnet, PI).
Data Sharing Statement
The raw scRNA-Seq and snRNA-Seq are deposited in GEO (GSE158304) and the aligned and processed data are deposited in Synapse (syn22647742). A publicly available interactive gene-browser was created for all analyzed scRNA-Seq datasets at: http://altanalyze.org/LateGestationNephrogenesis.html.
The authors thank Dr. Alan Jobe and Dr. Claire Chougnet, and the California National Primate Center for donating the rhesus kidneys used for this study. The Gene Expression Core at CCHMC produced the 10x data. This study used samples, data, and/or services from the Discover Together Biobank at Cincinnati Children’s Research Foundation. The authors thank Dr. Mathew Kofron and the Confocal Imaging Core staff for their assistance in 3D imaging. The authors dedicate this work to the patients and families who contributed to the Discover Together Biobank to make this study possible.
Meredith P. Schuh and Raphael Kopan conceptualized and planned the experiments, analyzed data, and supervised Lyan Alkhudairy. Meredith P. Schuh and Lyan Alkhudairy performed image acquisition and 3D reconstructions. Andrew Potter and S. Steven Potter performed single-cell dissociation of the rhesus tissue. Kashish Chetal, Kairavee Thakkar, and Nathan Salomonis performed the bioinformatics analyses in collaboration with Meredith P. Schuh and Raphael Kopan. Meredith P. Schuh, Nathan Salomonis, and Raphael Kopan assembled figures, wrote the manuscript, and incorporated input from all authors.
This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020101459/-/DCSupplemental.
Supplemental Figure 1. Undifferentiated NPCs are noted in both rhesus and human samples analyzed in this study.
Supplemental Figure 2. Annotation and alignment of scRNA-Seq and snRNA-Seq clusters.
Supplemental Figure 3. Common kidney progenitor populations in cortex single cells and nuclei.
Supplemental Figure 4. Maturation associated differences in nephron progenitors across species.
Supplemental Figure 5. Previous work supporting RNA and protein stability in human archival material.
Supplemental Figure 6. RNAScope validation of UB markers POU3F4 and TWIST1.
Supplemental Table 1. Gestational Age, sex, and post-mortem interval (PMI) of human archival biobank kidney samples.
Supplemental Table 2. Developmental Distribution of Rhesus Macaque Kidneys.
Supplemental Table 3. 10x Genomics Quality Control Metrics - Rhesus cortex sc/snRNA-Seq.
Supplemental Table 4. Rhesus Cortex scRNA-Seq (4 animals) - Cell-to-Cluster Associations.
Supplemental Table 5. Rhesus Cortex scRNA-Seq (4 animals) - Marker Genes per Cluster.
Supplemental Table 6. .Human Marker Genes for Prior Published single-cell Studies.
Supplemental Table 7. GO-Elite Enrichment Results - Rhesus Cortex scRNA-Seq (AltAnalyze single-cell compendium).
Supplemental Table 8. GO-Elite Enrichment Results - Rhesus Cortex scRNA-Seq (cross-species Kidney compendium).
Supplemental Table 9 GO-Elite Enrichment Results - Rhesus Cortex scRNA-Seq (cross-species Kidney compendium) Heatmap.
Supplemental Table 10 Rhesus Cortex snRNA-Seq (1 animal) - Cell-to-Cluster Associations.
Supplemental Table 11. Rhesus Cortex snRNA-Seq (1 animal) - Marker Genes per Cluster.
Supplemental Table 12. GO-Elite Enrichment Results - Rhesus Cortex snRNA-Seq (cross-species Kidney compendium).
Supplemental Table 13.GO-Elite Enrichment Results - Rhesus Cortex snRNA-Seq versus Rhesus Cortex scRNA-Seq.
Supplemental Table 14. Cross-Species Cortex scRNA-Seq/snRNA-Seq - Seurat3 - Cell-to-Cluster Associations.
Supplemental Table 15. Cross-Species Cortex scRNA-Seq/snRNA-Seq - Seurat3 Marker Genes per Cluster.
Supplemental Table 16. GO-Elite Enrichment Results - Cross-Species Cortex scRNA-Seq/snRNA-Seq Heatmap.
Supplemental Table 17. DoubletDecon Cell Doublet Predictions - Rhesus scRNA-Seq (37 cell clusters).
Supplemental Table 18. Rhesus Cortex scRNA-Seq (4 animals) - Cell-to-Cluster Associations.
Supplemental Table 19. Rhesus Cortex scRNA-Seq (4 animals) - Marker Genes per Cluster.
Supplemental Table 20. Monocle2 Pseudotime Analysis Results - Rhesus scRNA-Seq.
Supplemental Data 1. Rhesus morphology.
Supplemental Data 2. RNAScope.
Supplemental Movie 1. Human 32 WG SIX2 CDH1 KRT8.
Supplemental Movie 2. Rhesus 129DG SIX2 CDH1 KRT8.
1. Hinchliffe SA, Sargent PH, Howard CV, Chan YF, van Velzen D: Human intrauterine renal growth expressed in absolute number of glomeruli assessed by the disector method and Cavalieri principle. Lab Invest 64: 777–784, 1991 2046329
2. Nyengaard JR, Bendtsen TF: Glomerular number and size in relation to age, kidney weight, and body surface in normal man. Anat Rec 232: 194–201, 1992 1546799
3. Hoy WE, Hughson MD, Bertram JF, Douglas-Denton R, Amann K: Nephron number, hypertension, renal disease, and renal failure. J Am Soc Nephrol 16: 2557–2564, 2005 16049069
4. Bertram JF, Douglas-Denton RN, Diouf B, Hughson MD, Hoy WE: Human nephron number: Implications for health and disease. Pediatr Nephrol 26: 1529–1533, 2011 21604189
5. Barker DJ: The intrauterine environment and adult cardiovascular disease. Ciba Found Symp 156: 3–10, discussion 10–16, 1991 1855414
6. Barker DJ, Eriksson JG, Forsén T, Osmond C: Fetal origins of adult disease: Strength of effects and biological basis. Int J Epidemiol 31: 1235–1239, 2002 12540728
7. Barker DJP, Bagby SP: Developmental antecedents of cardiovascular disease: A historical perspective. J Am Soc Nephrol 16: 2537–2544, 2005 16049070
8. Barker DJ: The fetal and infant origins of disease. Eur J Clin Invest 25: 457–463, 1995 7556362
9. Barker DJP, Bagby SP, Hanson MA: Mechanisms of disease: In utero programming in the pathogenesis of hypertension. Nat Clin Pract Nephrol 2: 700–707, 2006 17124527
10. Barker DJ, Bull AR, Osmond C, Simmonds SJ: Fetal and placental size and risk of hypertension in adult life. BMJ 301: 259–262, 1990 2390618
11. Barker DJ, Osmond C, Golding J, Kuh D, Wadsworth ME: Growth in utero, blood pressure in childhood and adult life, and mortality from cardiovascular disease. BMJ 298: 564–567, 1989 2495113
12. Lackland DT, Bendall HE, Osmond C, Egan BM, Barker DJ: Low birth weights contribute to high rates of early-onset chronic renal failure in the Southeastern United States. Arch Intern Med 160: 1472–1476, 2000 10826460
13. Sayer AA, Cooper C, Evans JR, Rauf A, Wormald RPL, Osmond C, et al.: Are rates of ageing determined in utero? Age Ageing 27: 579–583, 1998 12675097
14. Stoll BJ, Hansen NI, Bell EF, Walsh MC, Carlo WA, Shankaran S, et al.; Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network: Trends in care practices, morbidity, and mortality of extremely preterm neonates, 1993-2012. JAMA 314: 1039–1051, 2015 26348753
15. Rodriguez MM, Gomez A, Abitbol C, Chandar J, Montané B, Zilleruelo G: Comparative renal histomorphometry: A case study of oligonephropathy of prematurity. Pediatr Nephrol 20: 945–949, 2005 15856326
16. White SL, Perkovic V, Cass A, Chang CL, Poulter NR, Spector T, et al.: Is low birth weight an antecedent of CKD in later life? A systematic review of observational studies. Am J Kidney Dis 54: 248–261, 2009 19339091
17. Vikse BE, Irgens LM, Leivestad T, Hallan S, Iversen BM: Low birth weight increases risk for end-stage renal disease. J Am Soc Nephrol 19: 151–157, 2008 18057216
18. Crump C, Sundquist J, Winkleby MA, Sundquist K: Preterm birth and risk of chronic kidney disease from childhood into mid-adulthood: National cohort study. BMJ 365: l1346, 2019 31043374
19. Rumballe BA, Georgas KM, Combes AN, Ju AL, Gilbert T, Little MH: Nephron formation adopts a novel spatial topology at cessation of nephrogenesis. Dev Biol 360: 110–122, 2011 21963425
20. Hartman HA, Lai HL, Patterson LT: Cessation of renal morphogenesis in mice. Dev Biol 310: 379–387, 2007 17826763
21. al-Awqati Q, Goldberg MR: Architectural patterns in branching morphogenesis in the kidney. Kidney Int 54: 1832–1842, 1998 9853247
22. Osathanondh V, Potter EL: Development of human kidney as shown by microdissection. V. Development of vascular pattern of glomerulus. Arch Pathol 82: 403–411, 1966 5923463
23. Potter EL: Normal and Abnormal Development of the Kidney, Chicago, Year Book Medical Publishers, 1972, pp 1–305
24. Moritz KM, Wintour EM, Black MJ, Bertram JF, Caruana G: Factors influencing mammalian kidney development: Implications for health in adult life. Adv Anat Embryol Cell Biol 196: 1–78, 2008 18637522
25. Carithers LJ, Ardlie K, Barcus M, Branton PA, Britton A, Buia SA, et al.; GTEx Consortium: A novel approach to high-quality postmortem tissue procurement: The GTEx project. Biopreserv Biobank 13: 311–319, 2015 26484571
26. Lindström NO, Guo J, Kim AD, Tran T, Guo Q, De Sena Brandine G, et al.: Conserved and divergent features of mesenchymal progenitor cell types within the cortical nephrogenic niche of the human and mouse kidney. J Am Soc Nephrol 29: 806–824, 2018 29449449
27. Lindström NO, McMahon JA, Guo J, Tran T, Guo Q, Rutledge E, et al.: Conserved and divergent features of human and mouse kidney organogenesis. J Am Soc Nephrol 29: 785–805, 2018 29449453
28. Lindström NO, Tran T, Guo J, Rutledge E, Parvez RK, Thornton ME, et al.: Conserved and divergent molecular and anatomic features of human and mouse nephron patterning. J Am Soc Nephrol 29: 825–840, 2018 29449451
29. Lindström NO, De Sena Brandine G, Tran T, Ransick A, Suh G, Guo J, et al.: Progressive recruitment of mesenchymal progenitors reveals a time-dependent process of cell fate acquisition in mouse and human nephrogenesis. Dev Cell 45: 651–660.e4, 2018 29870722
30. Wang P, Chen Y, Yong J, Cui Y, Wang R, Wen L, et al.: Dissecting the global dynamic molecular profiles of human fetal kidney development by single-cell RNA sequencing. Cell Rep 24: 3554–3567.e3, 2018 30257215
31. Kobayashi A, Valerius MT, Mugford JW, Carroll TJ, Self M, Oliver G, et al.: Six2 defines and regulates a multipotent self-renewing nephron progenitor population throughout mammalian kidney development. Cell Stem Cell 3: 169–181, 2008
32. Boyle S, Misfeldt A, Chandler KJ, Deal KK, Southard-Smith EM, Mortlock DP, et al.: Fate mapping using Cited1-CreERT2 mice demonstrates that the cap mesenchyme contains self-renewing progenitor cells and gives rise exclusively to nephronic epithelia. Dev Biol 313: 234–245, 2008
33. Brunskill EW, Aronow BJ, Georgas K, Rumballe B, Valerius MT, Aronow J, et al.: Atlas of gene expression in the developing kidney at microanatomic resolution [published correction appears in Dev Cell
16: 482, 2009 10.1016/j.devcel.2008.09.007]. Dev Cell 15: 781–791, 2008 19000842
34. Batchelder CA, Keyser JL, Lee CC, Tarantal AF: Characterization of growth, glomerular number, and tubular proteins in the developing rhesus monkey kidney. Anat Rec (Hoboken) 296: 1747–1757, 2013 23997038
35. Batchelder CA, Lee CC, Martinez ML, Tarantal AF: Ontogeny of the kidney and renal developmental markers in the rhesus monkey (Macaca mulatta
). Anat Rec (Hoboken) 293: 1971–1983, 2010 20818613
36. Batchelder CA, Lee CC, Matsell DG, Yoder MC, Tarantal AF: Renal ontogeny in the rhesus monkey (Macaca mulatta
) and directed differentiation of human embryonic stem cells towards kidney precursors. Differentiation 78: 45–56, 2009 19500897
37. Schmidt AF, Kannan PS, Bridges JP, Filuta A, Lipps D, Kemp M, et al.: Dosing and formulation of antenatal corticosteroids for fetal lung maturation and gene expression in rhesus macaques. Sci Rep 9: 9039, 2019 31227752
38. Fisher AL, Sangkhae V, Presicce P, Chougnet CA, Jobe AH, Kallapur SG, et al.: Fetal and amniotic fluid iron homeostasis in healthy and complicated murine, macaque, and human pregnancy. JCI Insight 5: e135321, 2020 31990688
39. Jain VG, Kong F, Kallapur SG, Presicce P, Senthamaraikannnan P, Cappelletti M, et al.: IRAK1 is a critical mediator of inflammation-induced preterm birth. J Immunol 204: 2651–2660, 2020 32238461
40. Cappelletti M, Presicce P, Lawson MJ, Chaturvedi V, Stankiewicz TE, Vanoni S, et al.: Type I interferons regulate susceptibility to inflammation-induced preterm birth. JCI Insight 2: e91288, 2017
41. Presicce P, Park C-W, Senthamaraikannan P, Bhattacharyya S, Jackson C, Kong F, et al.: IL-1 signaling mediates intrauterine inflammation and chorio-decidua neutrophil recruitment and activation. JCI Insight 3: e98306, 2018 29563340
42. Erhard J, Lange R, Scherer R, Kox WJ, Bretschneider HJ, Gebhard MM, et al.: Comparison of histidine-tryptophan-ketoglutarate (HTK) solution versus University of Wisconsin (UW) solution for organ preservation in human liver transplantation: A prospective, randomized study. Transpl Int 7: 177–181, 1994 8060466
43. Tainaka K, Kubota SI, Suyama TQ, Susaki EA, Perrin D, Ukai-Tadenuma M, et al.: Whole-body imaging with single-cell resolution by tissue decolorization. Cell 159: 911–924, 2014 25417165
44. Jing D, Zhang S, Luo W, Gao X, Men Y, Ma C, et al.: Tissue clearing of both hard and soft tissue organs with the PEGASOS method. Cell Res 28: 803–818, 2018 29844583
45. Lee E, Choi J, Jo Y, Kim JY, Jang YJ, Lee HM, et al.: ACT-PRESTO: Rapid and consistent tissue clearing and labeling method for 3-dimensional (3D) imaging. Sci Rep 6: 18631, 2016 26750588
46. Yang B, Treweek JB, Kulkarni RP, Deverman BE, Chen CK, Lubeck E, et al.: Single-cell phenotyping within transparent intact tissue through whole-body clearing. Cell 158: 945–958, 2014 25088144
47. Adam M, Potter AS, Potter SS: Psychrophilic proteases dramatically reduce single-cell RNA-seq artifacts: A molecular atlas of kidney development. Development 144: 3625–3632, 2017 28851704
48. Wu H, Kirita Y, Donnelly EL, Humphreys BD: Advantages of single-nucleus over single-cell RNA sequencing of adult kidney: Rare cell types and novel cell states revealed in fibrosis. J Am Soc Nephrol 30: 23–32, 2019 30510133
49. Venkatasubramanian M, Chetal K, Schnell DJ, Atluri G, Salomonis N: Resolving single-cell heterogeneity from hundreds of thousands of cells through sequential hybrid clustering and NMF. Bioinformatics 36: 3773–3780, 2020 32207533
50. DePasquale EAK, Schnell D, Dexheimer P, Ferchen K, Hay S, Chetal K, et al.: cellHarmony: cell-level matching and holistic comparison of single-cell transcriptomes. Nucleic Acids Res 47: e138, 2019 31529053
51. Street K, Risso D, Fletcher RB, Das D, Ngai J, Yosef N, et al.: Slingshot: Cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genomics 19: 477, 2018 29914354
52. La Manno G, Soldatov R, Zeisel A, Braun E, Hochgerner H, Petukhov V, et al.: RNA velocity of single cells. Nature 560: 494–498, 2018 30089906
53. Miner JH, Sanes JR: Molecular and functional defects in kidneys of mice lacking collagen alpha 3(IV): Implications for Alport syndrome. J Cell Biol 135: 1403–1413, 1996 8947561
54. Miner JH, Sanes JR: Collagen IV alpha 3, alpha 4, and alpha 5 chains in rodent basal laminae: Sequence, distribution, association with laminins, and developmental switches. J Cell Biol 127: 879–891, 1994 7962065
55. Ekblom P: Formation of basement membranes in the embryonic kidney: An immunohistological study. J Cell Biol 91: 1–10, 1981 6795211
56. Combes AN, Phipson B, Lawlor KT, Dorison A, Patrick R, Zappia L, et al.: Single cell analysis of the developing mouse kidney provides deeper insight into marker gene expression and ligand-receptor crosstalk. Development 146: dev178673, 2019 31118232
57. Zambon AC, Gaj S, Ho I, Hanspers K, Vranizan K, Evelo CT, et al.: GO-Elite: A flexible solution for pathway and ontology over-representation. Bioinformatics 28: 2209–2210, 2012 22743224
58. Brunskill E, Jarmas A, Chaturvedi P, Kopan R: Age-dependent changes in the progenitor translatome coordinated in part by Tsc1 increase perception of signaling inputs to end nephrogenesis. bioRxiv. 10.1101/2020.03.12.989343 (Preprint posted March 12, 2020)
59. Chen S, Brunskill EW, Potter SS, Dexheimer PJ, Salomonis N, Aronow BJ, et al.: Intrinsic age-dependent changes and cell-cell contacts regulate nephron progenitor lifespan. Dev Cell 35: 49–62, 2015 26460946
60. Ning X, Zhang K, Wu Q, Liu M, Sun S: Emerging role of Twist1 in fibrotic diseases. J Cell Mol Med 22: 1383–1391, 2018 29314610
61. Wu H, Malone AF, Donnelly EL, Kirita Y, Uchimura K, Ramakrishnan SM, et al.: Single-cell transcriptomics of a human kidney allograft biopsy specimen defines a diverse inflammatory response. J Am Soc Nephrol 29: 2069–2080, 2018 29980650
62. Ransick A, Lindström NO, Liu J, Zhu Q, Guo JJ, Alvarado GF, et al.: Single-cell profiling reveals sex, lineage, and regional diversity in the mouse kidney. Dev Cell 51: 399–413.e7, 2019 31689386
63. Wu H, Uchimura K, Donnelly EL, Kirita Y, Morris SA, Humphreys BD: Comparative analysis and refinement of human PSC-derived kidney organoid differentiation with single-cell transcriptomics. Cell Stem Cell 23: 869–881.e8, 2018 30449713
64. Muto Y, Wilson PC, Wu H, Waikar SS, Humphreys BD: Single cell transcriptional and chromatin accessibility profiling redefine cellular heterogeneity in the adult human kidney. bioRxiv. 10.1101/2020.06.14.151167 (Preprint posted June 15, 2020)
65. Cosse-Etchepare C, Gervi I, Buisson I, Formery L, Schubert M, Riou J-F, et al.: Pou3f transcription factor expression during embryonic development highlights distinct pou3f3 and pou3f4 localization in the Xenopus laevis
kidney. Int J Dev Biol 62: 325–333, 2018 29877572
66. Farrow P, Khodosevich K, Sapir Y, Schulmann A, Aslam M, Stern-Bach Y, et al.: Auxiliary subunits of the CKAMP family differentially modulate AMPA receptor properties. Elife 4: e09693, 2015 26623514
67. Poole LJ, Zong J-C, Ciufo DM, Alcendor DJ, Cannon JS, Ambinder R, et al.: Comparison of genetic variability at multiple loci across the genomes of the major subtypes of Kaposi’s sarcoma-associated herpesvirus reveals evidence for recombination and for two distinct types of open reading frame K15 alleles at the right-hand end. J Virol 73: 6646–6660, 1999 10400762
68. He M, Wu B, Ye W, Le DD, Sinclair AW, Padovano V, et al.: Chloride channels regulate differentiation and barrier functions of the mammalian airway. Elife 9: e53085, 2020 32286221