Deletion of hypoxia-responsive microRNA-210 results in a sex-specific decrease in nephron number

Abstract

Low nephron number results in an increased risk for developing hypertension and chronic kidney disease. Intrauterine growth restriction is associated with a nephron deficit in humans, and is commonly caused by placental insufficiency, which results in fetal hypoxia. The underlying mechanisms by which hypoxia impacts kidney development are poorly understood. microRNA-210 is the most consistently induced microRNA in hypoxia and is known to promote cell survival in a hypoxic environment. In this study, the role of microRNA-210 in kidney development was evaluated using a global microRNA-210 knockout mouse. A male-specific 35% nephron deficit in microRNA-210 knockout mice was observed. Wnt/β-catenin signaling, a pathway crucial for nephron differentiation, was mis-regulated in male kidneys with increased expression of the canonical Wnt target lymphoid enhancer binding factor 1. This coincided with increased expression of caspase-8-associated protein 2, a known microRNA-210 target and apoptosis signal transducer. Together, these data are consistent with a sex-specific requirement for microRNA-210 in kidney development.

Introduction

Human and mouse kidneys contain on average 1 million and 12,500 nephrons (the functional unit of the kidney), respectively (1–3). Nephron number can vary as much as ten-fold amongst individuals and is established prior to birth (4), with a low nephron number being associated with a subsequent increased risk for hypertension and chronic kidney disease (5, 6). Infants born with intrauterine growth restriction (IUGR) have decreased nephron number (7, 8). The underlying mechanisms that result in impaired kidney development in IUGR are largely unknown. However placental insufficiency is a common pregnancy complication that causes fetal hypoxia and IUGR (9, 10), and hypoxia during kidney development can result in abnormalities, including a nephron number deficit (11, 12).

The mammalian kidney develops from reciprocal inductive signaling between the ureteric bud and metanephric mesenchyme (13). Signals from the ureteric bud stimulate condensing of the metanephric mesenchyme into a “cap” of nephron progenitors, which express the transcription factor Sine oculis homeobox homolog 2 (Six2), around the ureteric bud tips (14). Concurrently, signals from the metanephric mesenchyme induce branching of the ureteric bud to generate the collecting duct system (15, 16). As nephrogenesis progresses, a subset of nephron progenitors responds to Wnt9b-mediated signaling from the ureteric bud to promote a mesenchymal-to-epithelial transition to form renal vesicles (expressing lymphoid enhancer binding factor, Lef1), which sequentially develop into comma- and S-shaped body structures as nephron development proceeds (17, 18). In addition, Notch pathway activation is required to downregulate Six2 expression for nephron progenitor differentiation and for nephron segment maturation (19, 20). The proximal end of the developing nephron becomes the glomerulus, while the distal end fuses with the collecting duct, to form the functional nephron (18). Nephron number is partly determined by the balance of nephron progenitor self-renewal versus differentiation; thus, no more nephrons are formed after the nephron progenitor pool has been depleted (4, 21). In mice, about 50% of nephrons are formed postnatally in a burst of nephrogenesis, and nephrogenesis ends by postnatal day 4 (P4) (21, 22). Nephron progenitors reside in a relatively hypoxic environment, and hypoxia-induced signaling is known to be an important factor in regulating their self-renewal and differentiation (23, 24).

Nephron number varies between the sexes, with human females having about 15–17% fewer nephrons than males (25, 26). Despite having fewer nephrons, females are more resistant to kidney injury and disease (27–29). A recent study showed that exposure to prenatal hypoxia resulted in a male-specific nephron deficit that was associated with increased urinary albumin excretion, glomerular hypertrophy and renal fibrosis (11). There is emerging data showing sex-specific differences in expression, but little is known about the functional consequences (30, 31).

miRNAs are small, endogenous non-coding RNAs (~22 nucleotides long) that primarily fine-tune gene expression post-transcriptionally to regulate various biological processes (32). Recent studies have identified miRNAs as important regulators of kidney development, particularly as inhibitors of nephron progenitor apoptosis (33–36). microRNA-210–3p (miR-210) is the most commonly induced hypoxia-sensitive miRNA (37). Hypoxia Inducible Factors (HIFs)—which activate gene expression to promote cell survival in a low oxygen setting and are expressed throughout kidney development (38–40)—bind the miR-210 promoter to upregulate its expression in hypoxia (41, 42). Previously, we have shown in embryonic day 15.5 (E15.5) mice that miR-210–3p, the dominant strand for miR-210, is upregulated in nephron progenitors, compared to whole kidney (43). miR-210–3p has been found to target genes involved in key biological processes associated with hypoxia, such as mitochondrial metabolism, apoptosis, cell cycle regulation, angiogenesis, and DNA damage response (44). Recent studies show that miR-210 protects kidney cells from hypoxia-induced apoptosis (45) and that it is a potential biomarker for acute kidney injury (46) and renal cell carcinoma (47); however, its role in kidney development is unknown.

In this study, we show a male-specific decrease in nephron number in miR-210 knockout (KO) mice. We show that male miR-210 KO kidneys have increased Lef1 expression at P2, suggestive of increased canonical Wnt signaling in developing nephrons. This is associated with increased expression of the known pro-apoptotic miR-210 target, caspase-8-associated protein 2 (Casp8ap2). Together, these data suggest there is a sex-specific requirement for miR-210 in kidney development.

Materials and Methods

Mouse Strains

Wildtype CD-1 time-mated pregnant females were ordered from Charles River Laboratories, Inc. (Wilmington, MA, USA, RRID:MGI:5659424) to collect kidneys from E14.5 and P0 mice. Global microRNA-210 knockout (KO) males (generated as described in (48), on a mixed C57Bl/6J and 129 Elite background) were crossed to female C57Bl/6J wildtype mice from The Jackson Laboratory (RRID:IMSR_JAX:000664; Bar Harbor, ME, USA) to generate heterozygous mice. These heterozygous breeding pairs produced wildtype (control) and miRNA-210 knockout littermates for analysis. Animals were genotyped using genomic DNA isolated from tail clipping by PCR with the following primers: 0.5μM F1 5’-AGACAGGCCTGCTTGTAGGA-3’; 0.5 μM R 5’-TCAGGAGGTGGGTCCTGTAG-3’; and 1μM F2 5’-GGTCACTGCCAGGACTACGT-3’. All animals were housed in the vivarium at the Rangos Research Center at the UPMC Children’s Hospital of Pittsburgh (Pittsburgh, PA, USA) and all animal experiments were carried out in accordance with the policies of the Institutional Animal Care and Use Committee at the University of Pittsburgh.

miRNA Expression

Total RNA from CD-1 E14.5 and P0 kidneys was isolated using the Qiagen miRNeasy Mini Kit (Hilden, Germany). Nephron progenitors were isolated using magnetic-activated cell sorting (MACS) as previously published (49, 50). Briefly, up to 24 CD-1 mouse kidneys were dissected and pooled from an individual litter, then the outer layers of cortical cells were digested into a single cell suspension using a mixture of collagenase A and pancreatin. Cell suspensions were then mixed with magnetic beads biotinylated to α-Itgα8 antibodies (a cell surface protein expressed on nephron progenitors (50); R&D Systems, Minneapolis, MN, USA) using the DSB-X Biotin Protein Labeling Kit (Thermo Fisher Scientific, Grand Island, New York, USA). Bead-bound nephron progenitor cells were immobilized by a DynaMag™-2 Magnet (Thermo Fisher Scientific) then washed, released, and resuspended. Total RNA was isolated from a fraction containing approximately 400,000 cells using the Qiagen miRNeasy Mini Kit. Enrichment of nephron progenitors relative to surrounding cell types was confirmed in each sample by real-time quantitative PCR (qPCR).

To verify the miR-210 knockout model, kidneys from P2 male and female wildtype and knockout littermates were dissected and total RNA was isolated using the Qiagen miRNeasy Mini Kit (Hilden, Germany). U6 snRNA (RT001973), miRNA-210–3p (RT00512), and miRNA-210–5p (RT462444) cDNAs were generated with the TaqMan® MicroRNA Reverse Transcription Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. Expression of mature miRNA was detected by qPCR performed in a 96 well C100 Thermal cycler (Bio-Rad, Hercules, CA, USA) using FAM1 TaqMan® Universal PCR Master Mix with no AmpErase® UNG (Applied Biosystems, Foster City, CA, USA). Expression levels were normalized to the endogenous control (U6 snRNA) and analyzed using the $2^{-\Delta \Delta {\text{C}}_{\text{T}}}$ method (51).

Size Measurements

At the time of sacrifice, kidneys were photographed with a scale bar using a Qimaging QICAM Fast 1394 camera coupled to a Leica M165FC Stereo Microscope, using QCapture software (QImaging, Surrey, British Columbia, Canada, RRID:SCR_014432). ImageJ software (https://imagej.nih.gov/ij/, RRID:SCR_003070) was utilized to determine kidney length in each digital image. Kidney and mouse body masses were measured using a digital platform scale (Summit Series Analytical Balance SI-234, Denver Instrument, Bohemia, NY, USA).

Estimation of Total Glomerular Number

Kidneys were harvested from mice at P30, fixed in 4% paraformaldehyde in phosphate buffered saline (PFA/PBS), embedded in paraffin, and serially sectioned at 4μm. For determination of glomerular number, sections throughout the whole kidney were stained with hematoxylin and eosin (H&E) and a pair of consecutive sections out of every 26 sections was chosen and analyzed using a Walcom drawing tablet (Walcom, Portland, OR, USA) and Stereo Investigator version 9.04 software (RRID:SCR_002526), using a physical fractionator probe (MBF Bioscience, Williston, VT, USA). The total glomerular number (N_glom) was calculated using the following equation:

$${\text{N}}_{\text{glom}}=\frac{{\Sigma \text{Q}}^{-}}{2}\times \frac{\text{A}}{\text{a}}\times \frac{1}{\text{ASF}}$$

where N_glom is the total of glomeruli in the entire kidney, ΣQ⁻ is the total number of glomeruli appearing and disappearing between two consecutive sections, A is the grid size (1000 × 1000), a is the counting frame size (800 × 800), and $\frac{1}{\text{ASF}}$ is the reciprocal of the section-sampling fraction (the number of sections advanced between section pairs) (52). The glomerular counts were performed in a blinded manner.

Serum and Urine Analysis

3-month-old mice were placed in metabolic cages overnight to collect urine. Blood was collected via cardiac puncture at the time of sacrifice into Microtainer® SST™ tubes (BD Biosciences, San Jose, CA, USA). The blood collection tubes were centrifuged to separate out serum. Frozen serum was sent to the Kansas State Veterinary Diagnostic Laboratory (Manhattan, KS, USA) to assay blood urea nitrogen and creatinine levels. Urine albumin/creatinine ratio was assayed using the Exocell Albuwell M and Creatinine Companion kits (Philadelphia, PA, USA) as per manufacturer’s instructions. Sample absorbance was measured using the SpectraMax® i3 plate reader (Molecular Devices, San Jose, CA, USA).

Histology

3-month-old kidneys were fixed in 4% PFA/PBS and embedded in paraffin. For morphological analysis, the Rangos Histology Core at UPMC Children’s Hospital of Pittsburgh sectioned kidneys at 4μm and performed H&E, Period Acid Schiff, and Trichrome staining.

Immunostaining

Kidneys were fixed in 4% PFA/PBS, embedded in paraffin, and sectioned at 4μm. Sections were deparaffinized, rehydrated, and permeabilized in PBS with 0.1% Tween-20 (PBS-T). Then antigen retrieval was performed by boiling in 10mM sodium citrate pH 6.0 buffer or Trilogy (Cell Marque, Rocklin, CA, USA). Sections to be stained using the Tyramide Signal Amplification Kit (PerkinElmer, Waltham, Ma, USA) were treated with 3% H₂O₂ to inhibit endogenous peroxidase activity. Sections were then blocked with either 3% bovine serum albumin (BSA) or 5% normal donkey serum (NDS) in PBS-T. Sections were incubated overnight with primary antibody, washed with PBS-T, incubated with secondary antibody, washed again with PBS-T, incubated with 1:5000 4′,6-diamidino-2-phenylindole, washed with PBS-T, and then mounted in Fluoro Gel with DABCO™ (Electron Microscopy Science, Hatfield, PA, USA). Primary antibodies were visualized either by staining with fluorescence-conjugated secondary antibodies or by horseradish peroxidase-conjugated antibodies followed by TSA-Plus Cyanine 5 or Fluorescein antibodies, as per the manufacturer’s instructions. Some sections were co-stained with fluorescein- or rhodamine-labeled Dolichos Biflorus Agglutinin (DBA; 1:100), both purchased from Vector Laboratories (Burlingame, CA, USA), to visualize the ureteric bud epithelium / collecting duct system. Immunostaining was visualized with a Leica DM2500 microscope and photographed with a Leica DFC 7000T camera using LAS X software (Buffalo Grove, IL, USA, RRID:SCR_013673). The list of antibodies and their dilutions used is shown in Table 1.

Table 1.

List of antibodies used for immunostaining

	Dilution for Immuno-fluorescence	Dilution for Western Blot	Company	Animal Source
Primary Antibodies
α-β-Act	-	1:5000	Cell Signaling Technologies	Rabbit
α-c-Casp3	1:100	-	Cell Signaling Technologies	Rabbit
α-c-Casp8 p18	-	1:1000	Santa Cruz Biotechnology	Rabbit
α-β-Cat active	1:100	-	Cell Signaling Technologies	Rabbit
α-β-Cat total	1:80	-	Cell Signaling Technologies	Rabbit
α-Emcn	1:50	-	Santa Cruz Biotechnology	Monoclonal Rat
α-Jag1	1:100	1:1000	Cell Signaling Technologies	Rabbit
α-Lef1	1:100	1:1000	Cell Signaling Technologies	Rabbit
α-Ncam	1:100	-	Sigma Aldrich	Rabbit
α-pHH3	1:100	-	Sigma Aldrich	Rabbit
α-Six2	1:100	1:1200	Proteintech	Rabbit
α-β-Tub	-	1:1000	Sigma Aldrich	Monoclonal Mouse
α-Wt1	1:100	-	Thermo Fisher Scientific	Rabbit
Secondary Antibodies
α-Mouse-488	1:100	-	Jackson ImmunoResearch Laboratories	Donkey
α-Mouse-594	1:100	-	Jackson ImmunoResearch Laboratories	Donkey
α-Mouse-HRP	1:100	1:3000	Cell Signaling Technologies	Horse
α-Rabbit-488	1:100	-	Jackson ImmunoResearch Laboratories	Donkey
α-Rabbit-594	1:100	-	Jackson ImmunoResearch Laboratories	Donkey
α-Rabbit-HRP	1:100	1:4000 or 1:8000 for α-Six2	Sigma Aldrich	Goat

To estimate the number of ureteric branch tips, kidney sections stained with DBA were imaged around the whole kidney section (9–14 images per section) and the number of DBA-positive ureteric tips was quantified per image using Image J. This semi-quantitative analysis was performed in a blinded manner.

To estimate the number of apoptotic and proliferating nephron progenitor cells, kidney sections were stained for c-Casp3 or pHH3 expression, respectively. Nephron progenitors were visualized by co-staining with anti-Six2 antibodies. Stained sections were imaged around the whole kidney section (9–14 images per section) and the percentage of c-Casp3- and pHH3-positive nephron progenitor cells per image was quantified using ImageJ. This semi-quantitative analysis was performed in a blinded manner. To estimate the number of apoptotic cells in early developing nephrons, kidney sections were co-stained for Jag1 expression and the quantification of c-Casp3- and Jag1-double positive cells was determined using the same method.

qPCR for mRNA Expression

Total RNA was isolated using the miRNeasy Mini Kit (Qiagen), as per the manufacturer’s instructions. cDNA was synthesized from total RNA using Superscript™ III First-Strand Synthesis System (Thermo Fisher Scientific), as per the manufacturer’s instructions. qPCR was performed in a 96 well C100 Thermal Cycler (Bio-Rad) using Sso Advanced SYBR Green Master Mix (Thermo Fisher Scientific). The list of primers used in these experiments is shown in Table 2. Expression levels were normalized to that of the endogenous control Actb and analyzed using the $2^{-\Delta \Delta C_T}$ method (51).

Table 2.

qPCR primer sequences and product sizes

Gene	Forward	Reverse	Product Size (bp)
Actb	GGCTGTATTCCCCTCCATCG	CCAGTTGGTAACAATGCCATGT	154
Casp3	TGACTTCCTGTATGCTTACT	TTGCCACCTTCCTGTTAA	161
Casp8ap2	GAGCTTCCGTCTCAGGACAAA	GCCGTAATGTTTCACGTCATTC	135
Cited1	GTCTCCAGGTCTTACCACCGA	GCAGAGATGGCCACGTGTAT	155
Efna3	GTTCACCATGTACAGCACGTA	GGAACAGCTCCAATCAGCA	143
Fgf8	GCTAATTGCCAAGAGCAACG	GGTAGTTGAGGAACTCGAAGC	245
Hey1	AAGTCGCCAGTAAGTCAG	GTTCGTAATCACTACCTCAATT	174
Jag1	GTCTCCAGGTCTTACCACCGA	GCAGAGATGGCCACGTGTAT	144
Lef1	AGCTTGTTGAAACCCCAGAC	TTTTTGGAAGTCGGCGCTTG	160
Lhx1	CTACATCATAGACGAGAACAAG	TCATTACTACCACCTTCCTTAT	198
Ret	ACACTCAGCACTCCTCTA	AGCATTCTCAGCCACATAA	224
Six2	GCAGGACTCCATACTCAA	GATACCGAGCAGACCATT	215
Sox9	AAGGAAGGAAGGAAGGAAG	AGGCACAGTGAATGTTCTA	201
Vegfr2	GAGAGGTGCTGCTTAGAT	GAGAGTAGAGTCAACACATTC	164
Wnt4	TGGGAAGGTGGTGACACAAG	TGACCACTGGAAGCCCTGT	166

Western Blot

Kidneys were harvested from P0 and P2 pups and dissociated in RIPA buffer (20mM Tris-HCl pH7.5; 150 mM NaCl; 1% Triton X-100; 1% sodium deoxycholate, and 1% SDS) using the Sonic Dismembrator Model 100 (Fisher Scientific, Hampton, NH, USA). The protein concentration of extracts was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific), as per manufacturer’s instructions, and the SpectraMax® i3 plate reader (Molecular Devices). 10μg or 20μg from each sample was run on a reducing 8%, 10%, or 12% SDS-PAGE gel or on a 4–20% Mini-PROTEAN® TGX™ Precast protein gel (Bio-Rad) and blotted to an ImmunoBlot polyvinylidene difluoride membrane (Bio-Rad) using a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). The list of antibodies used is shown in Table 1. The signals were developed on CL-Xposure™ Film (Thermo Fisher Scientific) using either the Pierce™ ECL Western Blotting Substrate (Thermo Fisher Scientific), Pierce™ ECL Plus Western Blotting Substrate (Thermo Fisher Scientific), or SuperSignal™ West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific). Densitometric analysis of bands was performed using Image J software.

in situ Hybridization

This assay was performed as described in (53). Kidneys were harvested from P2 pups and fixed in 4% PFA/PBS overnight at 4°C. The kidneys were rinsed with PBS, incubated with 30% sucrose/PBS overnight at 4°C, embedded in optimal cutting temperature compound (OCT) (Scigen Scientific Inc., Carson, CA, USA), and stored at −80°C. The OCT-embedded kidneys were sectioned at 10–12μm using the Microm HM 550 cryo-tome (Thermo Fisher Scientific) and affixed to Fisherbrand® Superfrost Plus Microscope Slides, Precleaned (Thermo Fischer Scientific). The sections were fixed in fresh 4% PFA/PBS for 10min, rinsed with PBS, and permeabilized with 15μg/mL Proteinase K for 10min. The sections were rinsed with PBS and refixed with 4% PFA/PBS for 5min. The sections were acetylated for 10min in a solution of 1.5% Triethanolamine, 0.02M hydrochloric acid, and 0.375% acetic anhydride. The sections were rinsed with PBS and blocked with hybridization buffer (50% formamide, 1.3X standard saline citrate (SSC) pH 4.5, 5mM EDTA pH 8.0, 50μg/mL yeast tRNA, 0.2% Tween-20, 0.5% CHAPS, 100μg/mL Heparin) for 2hr. The Wnt9b Digoxigenin-labelled probe (prepared as in (54) with forward primer 5’-GTCTTTGCCAAGTCTGCCTC-3’ and reverse primer 5’-CGATGTTAATACGACTCACTATAGGGGC-3’) was diluted 1:250 in hybridization buffer and heated at 80°C for 5min, before incubating the sections with the probe solution at 68°C overnight. The sections were rinsed with 0.2X SSC and NTT (0.15M NaCl, 0.1M Tris pH 7.5, 0.1% Tween-20) and then blocked with 5% heat inactivated sheep serum in 2% blocking reagent (Roche, Basel, Switzerland)/NTT for 2hr. Sections were then incubated with 1:2500 anti-Digoxigenin-AP antibody (Roche) in 1% heat inactivated sheep serum in 2% blocking reagent/NTT at 4°C overnight. Sections were rinsed with NTT and NTTML (0.15M NaCl, 0.1M Tris pH 9.5, 0.1% Tween-20, 50mM MgCl₂, and 2mM Levamisole) then incubated with BM Purple (Roche) at room temperature until color developed (~48hrs). The sections were rinsed with PBS, fixed with 4%PFA/PBS with 0.2% glutaraldehyde for 1hr, rinsed, and mounted in Cytoseal™ 280 (Richard-Allan Scientific, San Diego, CA, USA). The sections were visualized with a Leica DM2500 microscope and photographed with a Leica DFC 7000T camera using LAS X software.

Statistical Analysis

All experiments were performed with at least three biological replicates collected from multiple litters. Mann-Whitney U test and two-way ANOVA with Tukey correction were used to determine statistical significance where applicable: *P ≤ 0.05; **P ≤ 0.01; and ***P ≤ 0.001. All statistical analyses were performed using Prism 8 software package (GraphPad Software, Inc., La Jolla, CA, USA; RRID:SCR_002798).

Results

Expression of miR-210 throughout kidney development

To begin our investigation of the role of miR-210 in kidney development, we first evaluated its expression during nephrogenesis. We isolated whole kidneys and nephron progenitors (using a partial digestion, followed by MACS with antibodies against the nephron progenitor cell surface marker Itga8, as has been published previously (49, 50)) from E14.5 and P0 CD1 WT mice. qPCR analysis showed minimal expression of miR-210–3p at E14.5 (early nephrogenesis), and increased expression at P0 in both whole kidneys and nephron progenitors (Figure 1A,B). As expected, the expression of miR-210–5p was undetectable (data not shown), since it is the passenger strand for miR-210 (44). To verify loss of miR-210–3p in the miR-210 global knockout (KO) mouse model, we performed qPCR analysis of miR-210–3p expression in P2 wildtype (WT) and KO male and female kidneys and found no expression of miR-210–3p in KO kidneys (Figure 1C). Of note, it has previously been shown that there were no gross differences in embryonic and placental growth of progeny from miR-210 HET matings, suggesting that miR-210 is likely dispensable for fetal growth (48).

Figure 1. Deletion of miR-210 results in a sex-specific nephron deficit.

(A-B) qPCR analysis of miR-210–3p expression at E14.5 and P0 in (A) whole kidneys and (B) nephron progenitors (n = 3 pooled embryonic kidneys per litter). (C) qPCR analysis of P2 WT and KO male and female kidneys for miR-210–3p expression (n = 3 mice per genotype and sex). (D-G) H&E staining of P30 miR-210 (D, F) WT and (E, G) KO (D-E) male and (F-G) female kidneys. Black arrowheads indicate glomeruli. (H) Estimated total glomerular number for P30 WT and KO male and female kidneys (n = 5 mice per genotype and sex). (I) Kidney to body mass ratio of P30 WT and KO male and female kidneys (n ≥ 4 mice per genotype and sex). Error bars ± SEM, *P ≤ 0.05, ** P ≤ 0.01, ***P ≤ 0.001, (A-C) Mann-Whitney U test, (H-I) two-way ANOVA with Tukey’s multiple comparisons test.

Global deletion of miR-210 results in decreased nephron number

To determine if miR-210 KO mice have abnormal kidney development, we evaluated histological sections from postnatal day 30 (P30) kidneys. There were no gross morphologic differences between WT and KO male and female tissues (Figure 1D–G). We then used the “gold standard” physical disector/fractionator combination method (55) to estimate glomerular number in WT and KO littermates. We found that not only do the male KO mice have approximately 35% fewer nephrons than WT males, but also that both WT and KO females have approximately 27% fewer nephrons than WT males (Figure 1H). This decrease in nephron number is not due to an overall decrease in body or kidney size in females (Figure 1I, Supplementary Figure 1). Analysis of 3-month-old animals demonstrated that male and female WT and KO mice have normal tissue histology, kidney to body mass ratio, and kidney function (Supplementary Figure 2). This observed decrease in nephron number could be due to several factors, including 1) altered nephron progenitor differentiation, 2) early termination of nephrogenesis, 3) decreased nephron progenitor proliferation, or 4) increased apoptosis.

Kidney development appears normal at P0 in KO mice

In order to identify the onset of the nephron deficit, we first investigated kidney development at P0, when miR-210 is normally expressed, and before the burst of nephrogenesis that occurs from P1 to P3 (21, 22). H&E staining showed that P0 kidneys were histologically normal in both WT and KO males (Figure 2A,B) and females (Supplementary Figure 3A,B). Then we performed immunofluorescent staining for the nephron progenitor marker Six2 (14), and nephron progenitor and developing nephron structure marker (renal vesicle, comma-shaped body, and S-shaped body) Ncam (56). Both of these were normally expressed in WT and KO males (Figure 2C,D) and females (Supplementary Figure 3C,D). Wnt/β-catenin signaling is essential for normal induction of nephron progenitors into the renal vesicle, and disruption of this pathway results in abnormal kidney development (57). Immunofluorescent staining for its downstream effector, Lef1 (58), was normal in both male (Figure 2E,F) and female P0 kidneys (Supplementary Figure 3E,F). Further, immunofluorescent staining for Jag1, the ligand for Notch signaling (which is also essential for normal nephrogenesis (20)), appears normal in male (Figure 2G,H) and female kidneys (Supplementary Figure 3G,H). Western blot analysis for Jag1 and Six2 expression in male WT and KO P0 kidneys showed no differences in expression (Figure 2 I–K). Together, these data suggest that neither male nor female kidney development exhibit significant changes by miR-210 deletion at P0.

Figure 2. Normal nephrogenesis in P0 miR-210 KO male kidneys.

(A-B) H&E staining of P0 (A) WT and (B) KO male kidneys. Black arrowheads indicate glomeruli. (C-D) Immunofluorescent staining of nephron progenitor marker Six2 (red) and nephron progenitor and developing nephron structures marker Ncam (green) in (C) WT and (D) KO kidneys. (E-F) Immunofluorescent staining of Wnt/β-catenin signaling marker Lef1 (red) and ureteric epithelium marker DBA (green) in (E) WT and (F) KO kidneys. (G-H) Immunofluorescent staining of Six2 (red) and Notch signaling ligand Jag1 (green) in (G) WT and (H) KO kidneys. (I-K) Western blot analysis comparing levels of (J) Jag1 and (K) Six2 normalized to β-Act loading control (n = 4 mice per genotype). Error bars ± SEM, Mann-Whitney U test.

Differentiation is affected by P2 in KO male kidneys

We next investigated kidney development at P2. Histological evaluation demonstrated no gross morphological differences in P2 kidney sections from WT and KO male kidneys (Figure 3A,B). Similarly, immunofluorescent staining for Six2 and Ncam demonstrated no significant difference in nephron progenitors and developing nephron structures in WT and KO male kidneys (Figure 3C,D). However, qPCR analysis of the renal vesicle markers, Wnt4, Fgf8, and Lhx1 showed increased expression of Lhx1 in KO kidneys (Figure 3I). Interestingly, Lhx1 is thought to function downstream of Fgf8 and Wnt4 as nephron progenitors form pre-tubular aggregates and then renal vesicles (18, 58, 59). Furthermore, activation of Wnt/β-catenin signaling initiates the commitment of nephron progenitors to nephron formation (58), so we evaluated the expression of the canonical Wnt target, Lef1 (60). Although immunostaining for Lef1 showed no difference in the spatial expression pattern in developing nephron structures between WT and KO kidneys (Figure 3E,F), there was increased expression of Lef1 mRNA (Figure 3I) and protein (Figure 3J,K) levels in KO kidneys. This increase in Lef1 expression does not appear to be due to increased expression in the peri-ureteric bud stroma (Supplemental Figure 4A,B). Furthermore, in situ hybridization for Wnt9b (which is critical in inducing nephron formation) showed a normal ureteric bud tip expression pattern in both WT and KO male kidneys (Supplemental Figure 4C,D), and immunostaining for total and active β-catenin (Supplementary Figure 5) showed no overt spatial differences in their expression patterns between male WT and KO kidneys. Together, this suggests that there is increased Wnt/β-catenin signaling in P2 KO male kidneys, which has been associated with impaired nephron formation (61), although the spatial localization of components of this signaling pathway are preserved.

Figure 3. Increased Wnt/β-catenin signaling in P2 miR-210 KO male kidneys.

(A-B) H&E staining of P2 (A) WT and (B) KO male kidneys. Black arrowheads indicate glomeruli. (C-D) Immunofluorescent staining of nephron progenitor marker Six2 (red) and nephron progenitor and developing nephron structure marker Ncam (green) in (C) WT and (D) KO kidneys. (E-F) Immunofluorescent staining of Wnt/β-catenin signaling marker Lef1 (red) and ureteric epithelium marker DBA (green) in (E) WT and (F) KO kidneys. (G-H) Immunofluorescent staining of Notch signaling ligand Jag1 (red) and DBA (green) in (G) WT and (H) KO kidneys. (I) qPCR analysis of expression of differentiation markers in WT and KO kidneys (n = 6 mice per genotype). (J-K) Western blot analysis of Lef1 expression normalized to loading control β-Act (n = 4 mice per genotype). (L-M) Western blot analysis of Jag1 expression normalized to loading control β-Tub (n = 4 mice per genotype). Error bars ± SEM, Mann-Whitney U test, *P ≤ 0.05.

Since Notch signaling is required for priming nephron progenitors for differentiation (20), we analyzed its activation in P2 kidneys. Immunostaining for the Notch ligand Jag1 showed no difference in its spatial expression in renal vesicles between WT and KO male kidneys (Figure 3G,H); however, Western blot analysis of Jag1 showed decreased expression in KO kidneys (Figure 3L,M). qPCR for Jag1 and Notch signaling downstream marker Hey1 showed no differences in expression (Figure 3I). Together, these data suggest that while there is less Notch ligand expression, the downstream activation of Notch signaling is not affected by deletion of miR-210 in the whole kidney.

Importantly, analysis of female P2 WT and KO kidneys showed no gross histological differences (Figure 4A,B). Immunostaining for Six2, Ncam, Lef1, and Jag1 showed normal expression in female WT and KO kidneys (Figure 4C–H), and qPCR analysis of expression of Lhx1 and Lef1 showed no difference between P2 WT and KO female kidneys (Figure 4I). Thus, the effect of global miR-210 deletion is predominantly in the development of the male kidney.

Figure 4. Normal nephrogenesis in P2 miR-210 KO female kidneys.

(A-B) H&E staining of P2 (A) WT and (B) KO female kidneys. Black arrowheads indicate glomeruli. (C-D) Immunofluorescent staining of nephron progenitor marker Six2 (red) and nephron progenitor and developing nephron structures marker Ncam (green) in (C) WT and (D) KO kidneys. (E-F) Immunofluorescent staining of Wnt/β-catenin signaling marker Lef1 (red) and ureteric epithelium marker DBA (green) in (E) WT and (F) KO kidneys. (G-H) Immunofluorescent staining of Notch signaling ligand Jag1 (red) and DBA (green) in (G) WT and (H) KO kidneys. (I) qPCR analysis of expression of differentiation markers in WT and KO kidneys (n = 6 mice per genotype). Error bars ± SEM, Mann-Whitney U test.

Nephrogenesis timing is not affected by miR-210 KO

To investigate if the nephron deficit is the result of a premature end to nephrogenesis in miR-210 KO mice, we evaluated kidneys at P3 and P4 (when nephrogenesis ceases in wildtype kidneys (21, 22)). H&E staining showed normal histology at P3 (Figure 5A,B) and P4 (Figure 5C,D) in WT and KO male mice. At P3, Six2 and Ncam were expressed in a normal pattern in both WT and KO male mice (Figure 5E,F). In addition, there was normal expression of Lef1 (Figure 5I,J) and Jag1 (Figure 5M,N). At P4, there was no Six2 expression in both WT and KO male mice (Figure 5G,H), as expected (22), and developing nephron structures still had normal expression of Lef1 (Figure 5K,L) and Jag1 (Figure 5O,P). Together, this shows that KO male mice have no gross differences in the timing of nephrogenesis cessation compared to WT.

Figure 5. Nephrogenesis ends by P4 in miR-210 WT and KO male kidneys.

(A-D) H&E staining of (A-B) P3 and (C-D) P4 (A, C) WT and (B, D) KO male kidneys. Black arrowheads indicate glomeruli. (E-H) Immunofluorescent staining of nephron progenitor marker Six2 (red) and nephron progenitor and developing nephron structures marker Ncam (green) in (E-F) P3 and (G-H) P4 (E, G) WT and (F, H) KO kidneys. (I-L) Immunofluorescent staining of Wnt/β-catenin signaling marker Lef1 (red) and ureteric epithelium marker DBA (green) in (I-J) P3 and (K-L) P4 (I, K) WT and (J, L) KO kidneys. (M-P) Immunofluorescent staining of Notch signaling ligand Jag1 (red) and DBA (green) in (M-N) P3 and (O-P) P4 (M, O) WT and (N, P) KO kidneys.

Development of other compartments is not affected by miR-210 KO

Since decreased ureteric branching can also result in decreased nephron number, we measured expression of the ureteric tip markers, Ret and Sox9 (62), and found no difference between WT and KO in male (Supplementary Figure 4E) and female P2 kidneys (Supplementary Figure 4G). To determine if there were fewer branches of the ureteric bud, we counted the number of DBA-positive ureteric bud tips. There was no difference between WT and KO male (Supplementary Figure 4F) and female kidneys (Supplementary Figure 4H). Thus, miR-210 deletion does not have a significant effect on ureteric development.

miR-210 is a well-known promoter of angiogenesis (63), and since vascular development in the kidney is necessary for normal overall kidney development (24), we performed co-immunostaining for the endothelial cell marker Emcn (64) and the nephron progenitor and podocyte marker Wt1 (65) in P2 WT and KO male and female kidneys. There were no differences in the spatial expression pattern (Supplemental Figure 6A–L). qPCR analysis of endothelial cell marker Vegfr2 (66) and the anti-angiogenic miR-210–3p target Efna3 (67) showed no significant difference in expression in male (Supplemental Figure 6M) and female kidneys (Supplementary Figure 6N). This suggests that the development of renal vasculature is not markedly affected by loss of miR-210.

Nephron progenitor proliferation is unaffected by miR-210 KO

Since miR-210–3p has been shown to regulate cell cycle progression (44), immunostaining for the proliferation marker phospho-histone H3 (pHH3) was performed and demonstrated no significant differences in the percentage of proliferating Six2-positive nephron progenitors in P2 WT and KO male (Figure 6A–C) and female (Supplementary Figure 7A–C) P2 kidneys. Western blot analysis of Six2 showed no difference in expression in male P2 kidneys (Figure 6J,K). Consistent with this finding, qPCR analysis of the nephron progenitor markers Cited1 (68) and Six2 showed no difference in expression between WT and KO male (Figure 6L) and female P2 kidneys (Supplementary Figure 7D). Together, these data suggest that deletion of miR-210 does not affect the self-renewal capacity of nephron progenitors.

Figure 6. Increased expression of Casp8ap2 in P2 miR-210 KO male kidneys.

(A-B) Immunofluorescent staining for nephron progenitor marker Six2 (green) and proliferation marker pHH3 (red) in P2 (A) WT and (B) KO male kidneys. (C) Analysis of the percentage of Six2 and pHH3 co-positive cells (n = 4 mice per genotype). (D-E) Immunofluorescent staining for nephron progenitor marker Six2 (green) and pro-apoptotic marker cleaved-Caspase3 (c-Casp3; red) in P2 (D) WT and (E) KO male kidneys. White arrowheads indicate c-Casp3-positive cells at site of differentiation. Inset shows magnified view of apoptotic cells. (F) Analysis of the percentage of Six2 and c-Casp3 co-positive cells (n = 4 mice per genotype). (G-H) Immunofluorescent staining for developing nephron structures marker Jag1 (green) and c-Casp3 (red) in P2 (G) WT and (H) KO male kidneys. White arrowheads indicate c-Casp3-positive cells at site of differentiation. Inset shows magnified view of apoptotic cells. (I) Analysis of the percentage of Jag1 and c-Casp3 co-positive cells (n = 4 mice per genotype). (J-K) Western blot analysis of Six2 levels normalized to loading control β-Act in WT and KO kidneys (n = 4 mice per genotype). (L) qPCR analysis of expression of self-renewing nephron progenitor cell marker Cited1 and Six2 (n = 6 mice per genotype). (M-O) Western blot analysis of apoptosis marker Casp8 levels, both the (N) inactive pro-Casp8 and (O) activate p18 forms, normalized to loading control β-Act in WT and KO kidneys (n = 4 mice per genotype). (P) qPCR analysis of expression of pro-apoptotic marker and miR-210 target gene Casp8ap2 and pro-apoptotic Casp3 (n = 6 mice per genotype). Error bars ± SEM, Mann-Whitney U test, *P ≤ 0.05.

Expression of Casp8ap2 is increased in P2 male KO kidneys

Since miR-210–3p is known to target several pro-apoptotic genes (44), we tested whether apoptosis may be dysregulated in miR-210 KO kidneys. We co-immunostained for cleaved-caspase 3 (c-Casp3), an intra-cellular signal for apoptosis (69), and the nephron progenitor marker Six2 at P2. There was no difference in the percentage of apoptotic Six2-positive nephron progenitors in WT and KO male (Figure 6D–F) and female kidneys (Supplementary Figure 7E–G). Interestingly, it appeared that there were more apoptotic cells at the bottom of the cap mesenchyme—which are poised for differentiation—in KO male kidneys (Figure 6E). Some of these cells are Six2- and c-Casp-3 co-positive, and some are Six2-negative and c-Casp3-positive but are located just below the edge of the cap mesenchyme where renal vesicles are forming (Figure 6E). To determine if there were more apoptotic cells in early differentiating nephrons, we co-immunostained for c-Casp3 and Jag1 (Figure 6G,H). There was no statistically significant difference in the percentage of c-Casp3- and Jag1-double positive cells in developing nephron structures between male WT and KO kidneys (Figure 6I).

To assay apoptosis in the whole kidney, we first performed qPCR analysis of the pro-apoptosis markers Casp3 and Casp8ap2, the latter of which is also a miR-210–3p target (70, 71). Casp3 is unaffected by miR-210 deletion, which is unsurprising since it is regulated at the protein level, while Casp8ap2 expression is increased in male KO kidneys (Figure 6P). While Western blot analysis of Casp8, which is activated by Casp8ap2 to in turn activate Casp3 (72), did not show altered expression of the active c-Casp8 p18, it did show increased expression of its full-length protein pro-Casp8 in male KO kidneys (Figure 6M–O), which binds Casp8ap2 at the Fas receptor (72). qPCR analysis of Casp8ap2 and Casp3 showed no difference in expression between WT and KO female kidneys (Supplementary Figure 7H). Together, these data suggest that miR-210 deletion results in increased expression of the miR-210 target gene, Casp8ap2, specifically in the developing male kidney.

Discussion

Understanding the mechanisms that determine nephron number are important in defining risk factors for kidney disease. In this study, we show a male-specific decrease in nephron number in miR-210 knockout mice. The male miR-210 knockout kidneys have increased Lef1 expression at P2, consistent with increased canonical Wnt signaling in developing nephrons, which has previously been shown to result in a nephron deficit (61). This is associated with increased expression of the known miR-210 target, caspase-8-associated protein 2, suggesting that increased apoptosis contributes to the observed nephron deficit. Together, these data are consistent with a sex-specific requirement for miR-210 in the determination of nephron number during kidney development.

The sex-specific difference in nephron number in wild-type animals in our model is consistent with prior observations in both humans and mice that females have fewer nephrons compared to males (11, 26). Indeed, it has been reported that there are differences in sex-related changes in nephron number in different inbred mouse strains (3). However, our observation of a significant nephron deficit in male miR-210 knockout mice, but not in females, is the first description of a sex-specific functional role for a miRNA in kidney development, to our knowledge. Deletion of miR-210 was found to have no significant effect on mouse placental development during both normoxic and hypoxic pregnancies (48). However, this study did not analyze the sexes separately. In humans, female placentas from overweight and obese mothers had increased miR-210 expression compared to male placentas (73). miR-210 has also been identified as a potential biomarker for pre-eclampsia (74). Recently, a sex-specific role for a miRNA in adult kidneys was identified, where females with a miR-146b-5p deletion had exacerbated renal hypertrophy after 5/6 nephrectomy in rats (75). Thus, it is possible that in an injurious setting (e.g. hypertension, hypoxia) miR-210 responds in a sexually dimorphic manner.

Interestingly, a sex-specific difference in nephron number similar to that of miR-210 deletion was observed in a model of intrauterine hypoxia (which would be predicted to induce miR-210 expression), where hypoxia-exposed males had about 25% fewer nephrons than untreated males (11). Prolonged prenatal hypoxia has also been shown to result in a male-specific disruption of collecting duct patterning through altered Wnt/β-catenin and retinoic acid signaling, which resulted in a urine concentrating defect (76). Taken together, this raises the question of whether both over- and under-expression of miR-210 could result in a nephron deficit in a sex-specific manner. For example, both over- and under-activity of β-catenin signaling in nephron progenitors results in fewer nephrons due to an inability to either proceed to the renal vesicle stage, or undergo mesenchymal to epithelial transition, respectively (61).

Canonical Wnt signaling is necessary for the commitment of nephron progenitors to undergo a mesenchymal to epithelial transition and form the renal vesicle (77). Lef1 forms a transcriptional complex with Tcf that inhibits expression of Wnt pathway genes until β-catenin is activated, translocated to the nucleus, and complexes with Lef1/Tcf, to increase expression of target genes (78). It has previously been shown that miR-210–3p target Tcf7l2 (79) and that increased expression of miR-210–3p results in down-regulation of Ctnnb1, the β-catenin transcript (80). In the miR-210 KO kidneys, we see increased expression of Lef1, a marker of increased β-catenin activity, and overexpression of β-catenin has previously been shown to result in early depletion of the nephron progenitor pool (61). Together, this suggests that during nephrogenesis miR-210–3p fine-tunes Wnt/β-catenin activation to promote normal nephron number formation.

Furthermore, deletion of miR-210 results in the increased expression of its pro-apoptotic target, Casp8ap2. Other studies have shown that miR-210 targets Casp8ap2 for degradation to promote stem cell survival after ischemic preconditioning (70) and to protect human umbilical vein endothelial cells against oxidative stress (71). Under normal conditions, Casp8ap2 promotes S-phase progression and histone biosynthesis (81–83). Upon Fas receptor activation, Casp8ap2 translocates to the cytoplasm to promote apoptosis through the Fas receptor death-inducing signaling complex (DISC) (72, 84) as well as through activating the mitochondrial apoptotic pathway (85). Proper regulation of apoptosis is essential for normal kidney development (86), and part of this regulation is carried out by miRNAs (33, 36). We show a male-specific increase of Casp8ap2 in miR-210 KO kidneys, which has previously been reported have increased expression in males compared to females in developing mouse lung tissue (87). Our data suggest that loss of miR-210-mediated regulation of the pro-apoptotic Casp8ap2 transcript contributes to a sex-specific decrease in nephron number; although it remains unclear in which developmental compartment this occurs.

In summary, we have shown that deletion of the hypoxia-induced miR-210 results in a sex-specific nephron deficit due to dysregulation of Wnt/β-catenin signaling and apoptosis in developing male kidneys. This demonstrates how loss of a single miRNA can have a significant impact on nephrogenesis. Future studies will evaluate whether miR-210 deletion results in a predisposition to hypertension or renal disease with aging, high salt diet or angiotensin-induced hypertension, as seen in other studies regarding the prenatal programming of hypertension (88).

Nonstandard Abbreviations

β-Act

Actb = Beta-actin

Casp3

Caspase-3

c-Casp3

Cleaved Caspase-3

Casp8

Caspase-8

Casp8ap2

Caspase-8 associated protein 2

Cited1

Cbp/P300 interacting transactivator with Glu/Asp rich carboxy-terminal domain 1

DBA

Dolichos biflorus agglutinin

Efna3

Ephrin A3

Emcn

Endomucin

Fgf8

Fibroblast growth factor 8

H&E

Hematoxylin & eosin

Hey1

Hes related family bHLH transcription factor with YRPW motif 1

pHH3

Phosphorylated histone H3

Itgα8

Integrin subunit alpha 8

IUGR

Intrauterine growth restriction

Jag1

Jagged1

KO

Knockout

Lef1

Lymphoid enhancer-binding factor 1

Lhx1

LIM homeobox 1

MACS

Magnetic-activated cell sorting

miR / miRNA

microRNA

Ncam

Neural cell adhesion molecule

OCT

Optimal cutting temperature compound

PBS

Phosphate buffered saline

PBS-T

Phosphate buffered saline with 0.1% Tween-20

PFA

Paraformaldehyde

qPCR

real-time quantitative PCR

Ret

Ret proto-oncogene

Six2

Sine oculis homeobox homolog 2

snRNA

small nucleolar RNA

Sox9

SRY-box 9

β-Tub

Beta-tubulin

Vegfr2

Vascular endothelial growth factor receptor 2

Wnt4

Wnt family member 4

Wnt9b

Wnt family member 9B

WT

Wildtype

Wt1

Wilms tumor 1