Direct Isolation and Characterization of Human Nephron Progenitors

Abstract Mature nephrons originate from a small population of uninduced nephrogenic progenitor cells (NPs) within the cap mesenchyme. These cells are characterized by the coexpression of SIX2 and CITED1. Many studies on mouse models as well as on human pluripotent stem cells have advanced our knowledge of NPs, but very little is known about this population in humans, since it is exhausted before birth and strategies for its direct isolation are still limited. Here we report an efficient protocol for direct isolation of human NPs without genetic manipulation or stepwise induction procedures. With the use of RNA‐labeling probes, we isolated SIX2+CITED1+ cells from human fetal kidney for the first time. We confirmed their nephrogenic state by gene profiling and evaluated their nephrogenic capabilities in giving rise to mature renal cells. We also evaluated the ability to culture these cells without complete loss of SIX2 and CITED1 expression over time. In addition to defining the gene profile of human NPs, this in vitro system facilitates studies of human renal development and provides a novel tool for renal regeneration and bioengineering purposes. Stem Cells Translational Medicine 2017;6:419–433

hAF-derived populations comes as no surprise, since hAKPC-P (from which the SIX2 + CITED1 + cells are isolated) were previously enriched for a renal phenotype (Fig. 5) by selection for CD24, OBcadherin and podocalyxin [2].

Acquisition of hFK and hAF samples, single cell suspension and fibroblast obtainment
For the current study, 3 samples of hAF (17.1, 17.8, 18.0 GA) with normal male karyotypes and normal fetal ultrasounds (kindly donated by Dr R. Habibian, Labcorp, Monrovia, CA) were collected by amniocentesis. Written or verbal consent was not required since samples of AF were not identified and information obtained about the samples was limited to karyotype and fetal health status (45 CFR 46.102). Based on these facts, and after a detailed review, the requirement for an approval was waived by the CHLA IRB/CCI committee. From each sample, human kidney progenitor cells (hAKPC-P) were isolated as previously described [2]. A total of 14 hFK (around 17 GA) were obtained from CHLA Bank Tissue and used for all the experiments.
Tissue collection for this study was approved by the Institutional Review Boards of both Children's Hospital Los Angeles and the University of Southern California. Tissue donors provided informed consent, and no Identifying Health information was collected. After digestion with 0.05% collagenase I (BD Biosciences) at 37°C for 90 minutes and elimination of erythrocytes by Blood Lysis kit (Miltenyi Biotech), single cell suspension from hFK were obtained. Mouse amniotic fluid was collected as reported by our laboratory [2][3][4] following standard procedures approved by CHLA IACUC. Human Lung fibroblasts were purchased from LifeLine Cell Technology and expanded with Fibrolife Media (LifeLine Cell Technology) in tissue culture dishes for up to 5 passages.
RNA extraction was performed immediately after FACS (passage 0) by using the Qiagen RNeasy Micro Kit following manufacturer recommendations. Low input amount of RNA was converted to cDNA using the Clontech SMARTer V3 kit. RNA amounts were estimated following analysis on an Agilent Bioanalyzer Pico RNA chip. The ERCC standards were added, with appropriate dilutions being calculated based on the estimated RNA concentrations. The Clontech protocol for cDNA production was followed according to manufacturer's instructions. Amplification cycles were estimated based on input amounts of RNA. cDNA was again visualized by Bioanalyzer to confirm size and amounts, then sonicated on a Covaris S2 according to Clontech recommended conditions. DNA libraries were constructed using the Kapa Hyper prep kit and NextFlex adapters (Bioo Scientific). Libraries were visualized by bioanalyzer analysis and quantified by qPCR (Kapa library quantification kit) prior to application on an illumina NextSeq 500. RNA sequencing short reads were analyzed with FastQC (bioinformatics.babraham.ac.uk, 2015) and were of acceptable quality. Adapter sequences and sequences with low Phred quality scores were removed with Trimmomatic [5]. Reads were aligned to the Gencode version 22 human genome [6]corresponding to the GRCh38.p2 human genome (Genome Reference Consortium) and supplemented with sequences for the External RNA Controls Consortium synthetic RNA controls (ERCC ExFold, Ambion) [7]. Data from the mouse single cell analysis [8], were downloaded in the SRA format of the NCBI GEO database, accession GSE59127, processed similarly to the human samples and aligned to the Gencode version M6 mouse genome corresponding to the GRm38.p3 mouse genome (Genome Reference Consortium). All reads were aligned using the RNA-star short read aligner with the ENCODE recommended parameters [9]. Read counts per transcript were obtained using the HTSeq-count python script [10]. Reads per kilobase per million mapped reads (RPKM) were generated using the edgeR [11] R/Bioconductor software package [12]. Relative log expression graphs and principle component graphs were generated using the plotting functions of the EDASeq R/Bioconductor software package [13]. Differential gene expression was analyzed using the ERCC ExFold probes with the Remove Unwanted Variation R/Bioconductor software package [14] combined with edgeR [15].
GO enrichment analysis was performed using the GOstats R/Bioconductor software [16].
Comparisons between human and mouse were facilitated by the use of the EnsemblCompara web service of Ensembl biomart which allowed correspondence between mouse and human genes and a metric for homology [17]. Clustering and plotting of heatmaps was performed with the R software package 'gplots'. Directed acyclic graphs with data generated from the package GOstats were plotted using the 'Rgraphviz' software package. Smear plots of RNA seq data were plotted with the ggplot2 software package [18] enhanced with the RColorBrewer package.
For the purpose of comparing RNA-seq data between human and mice, single cell analysis were stratified by their expression for SIX2, CITED1 and FOXD1. Specifically, cells were grouped as Enrichment for NP genes from hAKPC-P using negative selection for induced/differentiative markers As suggested by Brown et al. [1] we used a modified (no use of anti-erythroid marker, since cells in culture were already depleted for this cell type) negative selection for enrichment of NP markers. Following trypsinization of hAKPC-P, cells were blocked with 1X human IgG for 15 minutes and stained with CD140a, CD326, CD105 antibodies for 1 hour on ice. After washing steps, cells were sorted using a FacsAria flow cytometer (BD Biosciences) and expanded for further analysis. RNA extraction and PCR analysis were performed as reported in the Experimental Procedure section in the main Manuscript.

Clones generation
Clones from both populations were obtained by limiting dilution immediately after sorting Dissociation/reaggregation Assays hFK cells were mixed in a 10:1 ratio with either hAF-or hFK-derived SIX2 + /CITED1 + cells, at passage 5 after selection, previously labeled with CM-DiI (Invitrogen) following standard protocols [19]. Cells were transferred onto polycarbonate membrane (3 µm pore size) at the airliquid surface in DMEM growth medium in a 24 well plate for 7 days. After 7 days of culture, the kidney explants were fixed with 4% PFA. CM-Dil-labeled cells were visualized by immunofluorescence microscopy after immunostaining.

Wnt9B induction, 3D collagen experiments and podocyte induction
Induction of hFK cells toward differentiation was performed by adding Wnt9b (0.4 μg/ml) and BMP7 (0.05 μg/ml) to the culture media for 7 days. Cells were then harvested, fixed and flow cytometry analysis to evaluate expression of SIX2 and CITED1 was performed as previously described.
Induction into tubular-like cells was performed by seeding the cells at passage 5 after selection into a three dimensional collagen layer using the EMD-Millipore 3D collagen assay kit, following the manufacturer instructions. Cells were placed into 24 well plates and cultured for up to 21 days with RPMI 1640, 10% ES-FBS and 1% penicillin-streptomycin. Podocyte induction was performed as previously published [2]. Briefly