Generation of Atp6v1g3-Cre mice for investigation of intercalated cells and the collecting duct

Abstract

Kidney intercalated cells (ICs) maintain acid-base homeostasis and recent studies have demonstrated that they function in the kidney’s innate defense. To study kidney innate immune function, ICs have been enriched using vacuolar ATPase (V-ATPase) B1 subunit (Atp6v1b1)-Cre (B1-Cre) mice. Although Atp6v1b1 is considered kidney specific, it is expressed in multiple organ systems, both in mice and humans, raising the possibility of off-target effects when using the Cre-lox system. We have recently shown using single-cell RNA sequencing that the gene that codes for the V-ATPase G3 subunit (mouse gene: Atp6v1g3; human gene: ATP6V1G3; protein abbreviation: G3) mRNA is selectively enriched in human kidney ICs. In this study, we generated Atp6v1g3-Cre (G3-Cre) reporter mice using CRISPR/CAS technology and crossed them with Tdtomato^flox/flox mice. The resultant G3-Cre⁺Tdt⁺ progeny was evaluated for kidney specificity in multiple tissues and found to be highly specific to kidney cells with minimal or no expression in other organs evaluated compared with B1-Cre mice. Tdt⁺ cells were flow sorted and were enriched for IC marker genes on RT-PCR analysis. Next, we crossed these mice to ihCD59 mice to generate an IC depletion mouse model (G3-Cre⁺ihCD59^+/+). ICs were depleted in these mice using intermedilysin, which resulted in lower blood pH, suggestive of a distal renal tubular acidosis phenotype. The G3-Cre mice were healthy, bred normally, and produce regular-sized litter. Thus, this new “IC reporter” mice can be a useful tool to study ICs.

NEW & NOTEWORTHY This study details the development, validation, and experimental use of a new mouse model to study the collecting duct and intercalated cells. Kidney intercalated cells are a cell type increasingly recognized to be important in several human diseases including kidney infections, acid-base disorders, and acute kidney injury.

INTRODUCTION

Traditionally, kidney collecting duct intercalated cells (ICs) are considered critical cells in the maintenance of acid-base homeostasis (1, 2). Recent studies have shown that kidney ICs are key innate immune cells for protection against invading uropathogens (3–10). To further study ICs, two novel transgenic mice expressing Cre recombinase or enhanced green fluorescent protein (EGFP) under the control of the 7-kb and 6.5-kb Atp6v1b1 promoter, a gene found in ICs, were generated (B1-Cre and B1-EGFP) by Dr. Raoul Nelson and colleagues (11, 12). We crossed the B1-Cre strain with Tdt mice to generate “IC reporter” mice to enrich ICs (3). Historically, Atp6v1b1 (gene)/V-ATPase-B1 (protein) is considered relatively kidney specific, but cells expressing it were also identified in other organs, including the brain, large intestine, small intestine, uterus, epididymis, and airway cells of the lung (11–13). Thus, a Cre mouse line with increased IC specificity would limit off-target effects in conditional mice generated to evaluate IC function. In 2021, using single-cell RNA sequencing (scRNAseq) of enriched human kidney ICs, we found ATP6V1G3 mRNA to be highly specific to ICs and expressed across IC subtypes (5). Atp6v1g3 (gene)/V-ATPase G3 subunit (protein) has also been noted to be expressed in mouse ICs and therefore represents a target for a more IC-specific Cre mouse line (14–16). Mouse models of deficient ICs have included carbonic anhydrase 2 (Car2) knockout mice, which have reduced IC numbers, and forkhead transcription factor (Foxi1) knockout mice, in which ICs and principal cells (PCs) are replaced with a single cell that has some, but not all, characteristics of both cell types (17–19). These mice with impaired IC formation have extrarenal developmental phenotypes. A model of acquired IC dysfunction would bypass the covariate of developmental phenotypes. In 2008 and 2016, a Cre-inducible human CD59 (ihCD59) model was demonstrated to mediate rapid cell ablation after intermedilysin (ILY) administration and avoid off-target effects (20, 21). The primary objective of this project was to generate a G3-Cre mouse line. The secondary objectives of this project were to compare the G3-Cre and B1-Cre lines and perform a pilot study evaluating use of the G3-Cre with the ihCD59-ILY system to evaluate a model of acquired IC cell depletion.

MATERIALS AND METHODS

Study Approval

Murine studies were approved by the Institutional Animal Care and Use Committees at the Indiana University School of Medicine (Protocol Nos. 11333 and 20105) and adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Mice

G3-Cre mice were generated at Albert Einstein College of Medicine, gene modification facility (Bronx, NY), using CRISPR/CAS technology. A CRISPR targeting site, Atp6v1g3 gRNA 76/59, in intron 1 near the 3′ end of exon 1 with acceptable predicted efficiency and specificity was designed. A T2A-Cre-PolyA cassette was inserted into exon1 by CRISPR/homology-directed repair (HDR) strategy. The resulting transcript under the control of the endogenous Atp6v1g3 promoter expresses a short peptide (27 amino acids of ATP6V1G3 N-terminal and 17 amino acids of T2A) and Cre recombinase (Fig. 2A). CRISPR/cas9 gRNA was injected to the zygotes of pronuclei of fertilized eggs and F0 progeny was generated. F0 progeny mice were mated with wild-type (WT) C57BL/6 mice to generate G3-Cre F1 mice. F1 mice were genotyped to identify the 5′-end homologous recombination event of the knockin (KI) allele by PCR with a primer binding to the upstream flanking region of the 5′-homologous arm, Atp6v1g3 CRE 5 Flank F1, and a primer binding to the internal region of the donor DNA, Cre R2. This PCR methodology identifies the 5′-end homologous recombination event of the KI alleles. The sequences of the primer pair were as follows: Atp6v1g3 CRE 5′ Flank F1: GAGCGGAGACTAAAGGAAAG and Cre R2: CGGTGCTAACCAGCGTTTTC, which results in a 1,880-bp G3-Cre KI allele and no WT allele on PCR gel (Fig. 2B, left). These pups were further confirmed for the existence of donor DNA by PCR with primers binding to the internal region of the donor DNA, Atp6v1g3 SF and Cre R2, which resulted in a 770-bp KI allele on the PCR gel, and after careful examination, it was confirmed that the insertion was a KI allele and not a random DNA sequence (Fig. 2B, right). The primer sequence of Atp6v1g3 SF was as follows: Atp6v1g3 SF: GACCCGTATCTCCCAGCTTC. Upon arrival at Indiana University School of Medicine, F1 progeny mice were mated with WT C57BL/6 mice to maintain the colony of G3-Cre mice. Mice were then genotyped by an automated real-time PCR system using tail DNA at Transnetyx (Cardova, TN). B1-Cre mice were kindly provided by Dr. Raoul Nelson (University of Utah, Salt Lake City, UT), in which the 7-kb B1 promoter drives Cre expression in ICs. B1-Cre mice were genotyped “in house” using specific primer sets (3), Cre forward: CATTACCGGTCGATGCAACGAG and Cre reverse: TGCCCCTGTTTCACTATCCAGG. To generate “IC reporter” mice, both B1-Cre and G3-Cre mice were crossed with tdTomato mice (Tdt mice, Stock No. 007909, Jackson Laboratory) and genotyped by automated real-time PCR for G3-Cre expression (Transnetyx). For IC ablation studies, G3-Cre⁺ mice were crossed with ihCD59 mice for 2 generations to generate G3-Cre⁺ihCD59^+/+ mice (21). Mice were genotyped using these primer sets: ihCD59-forward: ACTCATGATGGGAATCCAAGGAGGGTCT; ihCD59-reverse: ACGAATTCTTAGGGATGAAGGCTCCAGGC and WT with SH176: TGGAGGAGGACAAACTGGTCAC; PR387: GTGGGACTGCTTTTTCCAGA. If the mouse was ihCD59 and WT PCR positive, it was determined to be ihCD59^+/− and if the mouse was ihCD59 positive and WT PCR negative, it was determined to be ihCD59^+/+. ICs were depleted in G3-Cre⁺ihCD59^+/+ mice with intraperitoneal administration of ILY at 100 ng/g body wt. Female mice were used in this study.

Figure 2.

Generation and characterization of G3-Cre mice. A: a schematic diagram of murine Atp6v1g3 gene showing 5 exons (brown boxes), untranslated region (UTR) (green box), and the approximate position near exon 1, where T2A-Cre recombinase cassette was inserted with ATP6V1G3 guide RNA (gRNA). B: CRISPR/cas9 gRNA was injected to the zygotes of pronuclei of fertilized eggs, F0 progeny was generated, and pups were genotyped for G3-Cre gene by Flank F1/Cre R2 PCR, resulting in 1,880-bp product on agarose gel (left). Pups with the numbers of 5, 6, 8, 9, and 10 were found to be positive. No band resulted in wild-type B6 (WT) mice shown after lane 10 and B is blank lane. These pups were further confirmed for the existence of donor DNA by PCR with primers binding to the internal region of the donor DNA, Atp6v1g3 SF, and Cre R2, which resulted in 770-bp KI allele (right). A representative G3-Cre⁺Tdt⁺ mouse kidney section stained with intercalated cell subunit markers, V-ATPase E1 (C) and V-ATPase G3 (D), and Tdtomato (anti-RFP antibody) (C) and anti-Tdtomato (D) is shown to confirm the G3-Cre specificity, which showed very strong colocalization. Arrows represent E1- and G3-positive IC cells, which are also Tdt⁺, and the arrowhead represents G3 low expressing cells, which are also Tdt⁺. Magnification, ×40 in C and ×60 in D. KI, knock in.

Confirmation of Tdtomato Expression by Flow Cytometry

To determine the specificity of G3-Cre compared with B1-Cre, several tissues from B1-Cre⁺Tdt⁺ and G3-Cre⁺Tdt⁺ “IC reporter” mice were harvested, and a single cell suspension was prepared using Accumax enzymatic digestion solution (Innovative Cell Technologies) with the use of a gentleMACS tissue dissociator (Miltenyi Biotec) tissue-specific program. Cells were labeled with anti-mouse CD45-APC (BioLegend) to gate out immune cells, and total Tdt expression was analyzed using a flow cytometer equipped with a yellow-green or blue laser (Becton Dickinson) at the Indiana University flow cytometry core. Data were analyzed using FlowJo V10 software (Becton Dickinson, Ashland, OR).

Flow Sorting of ICs From the Kidney

Kidneys were harvested from adult (8–10 wk old) B1-Cre⁺Tdt⁺ and G3-Cre⁺Tdt⁺ “IC reporter” mice in cold PBS. Single-cell suspension was prepared with Accumax enzymatic digestion solution (Innovative Cell Technologies) with the help of gentleMACs dissociator (Miltenyi Biotec). In brief, kidneys were placed in a C tube (Miltenyi Biotec) containing 5 mL Accumax solution and rapidly dissociated using the program Lung_02_01 (45 s) and then incubated for 15 min at room temperature with gentle rotation. Cells were again placed on the dissociator and the Spleen_04_01 program (60 s) was used to further dissociate cells, which were again incubated for 15 min at room temperature. After dissociation, cells were gently mixed with pipetting up and down and filtered over 70-µm nylon mesh filter (BD Biosciences). Cells were pelleted by centrifugation at 300 g for 10 min. Red blood cells (RBCs) were lysed with RBC lysing buffer (BioLegend) and filtered and washed again with sterile RPMI1640 media. Dead cells were removed using an Annexin V dead cell removal kit (Stem Cell Technologies). After dead cell removal, cells were washed, and the surface was labeled with an anti-mouse CD45-APC antibody (BioLegend) for 30 min at 4°C to assess Tdt expression on resident immune cells. After surface labeling, ICs were flow sorted after size and singlet gating as CD45⁻ Tdt⁺ (IC) cells at the Indiana University School of Medicine flow cytometry core facility.

Targeted Real-Time PCR

Selected genes found to be upregulated in ICs were validated using real-time PCR in flow-sorted ICs, non-IC kidney cells (NIC), and whole kidney lysates (WKLs). RNA was prepared with RNeasy plus micro kit (Qiagen). Quality was validated with Agilent Bioanalyzer (Agilent) and reverse transcribed to cDNA with high-capacity reverse transcription kit (Applied Biosystems) and amplified with predesigned and validated primer sets (KiCqStart Primers, Sigma) on a 96-well plate real-time PCR CFX system (Bio-Rad) using SYBR Green chemistry. The 2^−ΔΔCt method was used to calculate relative fold change in gene expression. Data were presented as means ± SD from 3 or 4 reporter mice. Primer set sequences were as follows: Gapdh, forward 5′- CTGGAGAAACCTGCCAAGTA-3′ and reverse 5′- TGTTGCTGTAGCCGTATTCA-3′; Atp6v1b1, forward 5′- GAGTTTCTGGACATCAATGG-3′ and reverse 5′- CTGTTCATGACATCAATGGG-3′; Slc4a1, forward 5′- CATGACAGAAAGAGTGTTCC-3′ and reverse 5′- AGGTTAAGCAGAGCTTTTTC-3′; and Slc26a4, forward 5′- ATCCTCTCCATTATCTACACAC-3′ and reverse 5′- GTTCCTTAACAGCCATACAG-3′.

Immunofluorescence

Deidentified human kidney tissue was obtained from the Cooperative Human Tissue Network (CHTN), Midwest Division (Columbus, OH), as previously described (22, 23). The use of CHTN samples was reviewed by the Indiana University Institutional Review Board and was exempt (Protocol No. 1802253259), because no identifying information was included in patient samples and permissions and consents for human sample collection were from the CHTN and its member institutions. The use of human tissue expression patterns was contextualized with images obtained from the Human Protein Atlas (https://www.proteinatlas.org) (24, 25). To localize ICs in G3-Cre “IC reporter” mice, WT C57BL/6 mice and human kidney biopsies, 7-µm thick paraffin embedded kidney tissue sections were stained with polyclonal chicken anti-V-ATPase E1 (1:500 dilution) (Sigma, St. Louis, MO, Cat. No. GW22284F) antibody or rabbit polyclonal anti-V-ATPase G3 antibody (Cat. No. ab122012, 1:200 dilution) (Abcam, Boston, MA). To localize PCs in mice and humans, goat polyclonal anti-aquaporin 2 (AQP2) antibody (1:200 dilution) (Santa Cruz Biotechnology, Dallas, TX, Cat. No. SC-9882) was used. To localize the distal convoluted tubule (DCT), calbindin D-28k (1:500 dilution) (Swant AG, Burgdorf, Switzerland, Cat. No. CB-38a) was used. For Tdt visualization in G3-Cre⁺Tdt⁺ mice, anti-RFP rabbit antibody (1:200 dilution) (Cat. No. 600-401-379, Rockland Antibodies, Limerick, PA) and goat polyclonal Tdtomato antibody (1:500 dilution) (Cat. No. TA150129, OriGene, Rockville, MD) were used. To visualize Cre in mice, anti-Cre recombinase antibody (1:500 dilution) (Cat. No. ab216262, Abcam) was used. To visualize V-ATPase G3 and Cre, anti-rabbit AF488 (1:600 final dilution), to visualize V-ATPase E1, anti-chicken Cy3 (1:600 final dilution), and to visualize RFP/Tdtomato, anti-rabbit Cy3 and anti-goat Cy3 (1:600 final dilution) secondary antibodies were used (Jackson ImmunoResearch). Sections were visualized using a Keyence BZ-9000 microscope (Keyence, Osaka, Japan).

IC Ablation and Blood Chemistry

G3-Cre⁺ mice were crossed with ihCD59^+/+ mice for two generations to generate G3-Cre⁺ihCD59^+/+ mice. IC ablation was achieved by intraperitoneal injection of two doses of ILY administration at 100 ng/g body wt with the last dose at 18 h before blood chemistry. For control, WT C57BL/6 (B6) mice were given the same dose of ILY. Blood chemistries were performed the next morning using the iSTAT system (Abbott Point of Care) with EC8+ cartridges (REF: 03P79-25, Abbott) to measure blood glucose, sodium, hemoglobin, pH, bicarbonate, potassium, chloride, and anion GAP (AnGAP) levels.

IC Image Stitch and Cell Count

To quantify the extent of IC ablation in G3-Cre⁺ihCD59^+/+ mice compared with WT B6 mice, kidney sections stained with V-ATPase E1 were acquired on the BZ-X810 microscope (Keyence, Osaka, Japan) available at the Indiana University O’Brien Center for Advanced Microscopic Analysis (Indianapolis, IN), using a ×20 objective. Mosaic images were generated using automatic stitching in BZ-X800 software with threshold adjusted manually to include V-ATPase E1+ (Cy3) cells, which allowed us to eliminate nonspecific cells. ICs were counted in whole kidney image stitch sections using BZ-X800 software hybrid cell count function (Keyence, Osaka, Japan). Total cell counts were reported from each mouse kidney. After acquisition, red color images were pseudo color converted on LAS X software (Leica Microsystems Inc, Illinois) using under/over glow color palette to visualize the image stitch.

Data Analysis

Targeted PCR data were analyzed, and graphs were prepared using Graph Pad Prism. Differences between groups were compared with unpaired t test (two groups) when datasets were normally distributed and Mann–Whitney test if they were not normally distributed. For comparison of three groups, one-way analysis of variance (ANOVA) was used. Significance was assigned for a P value <0.05. Images were prepared using Adobe Photoshop (Adobe Software, San Jose, CA) and BZII Analyzer (Keyence) software with processing limited to black balance and optimization of brightness and contrast. The graphical abstract was created with BioRender (Toronto, ON, Canada).

RESULTS

ATP6V1B1 Is Widely Expressed in Human and Mouse Tissues Including Kidney ICs

Historically, V-ATPase subunits were described as kidney specific or ubiquitous with V-ATPase B1 being well described as kidney specific. To study the function of kidney ICs, B1-Cre mice (gift of Dr. Raoul Nelson) were crossed with Tdt^flox/flox mice to generate B1-Cre⁺Tdt⁺ “IC reporter” mice to enrich ICs. In flow sorting, we noticed that along with ICs, kidney CD45⁺ immune cells robustly expressed Tdt including CD45⁻Tdt⁺ (IC) cells (Supplemental Fig. S1, bottom). For IC flow sorting, CD45⁺Tdt⁺ contaminating immune cells were efficiently gated out to enrich renal ICs (Tdt⁺CD45⁻), which expressed all the markers specific to ICs as determined by RT-PCR or bulk RNA sequencing (3, 4). Kidney CD45⁺ immune cell Tdt expression led us to speculate whether V-ATPase B1 was widely expressed in other tissues. We first analyzed the freely available Human Protein Atlas portal (www.proteinatlas.org) for V-ATPASE-subunit B1 expression (https://www.proteinatlas.org/ENSG00000116039-ATP6V1B1/tissue) and found that indeed V-ATPase B1 protein and RNA are expressed in various human organs (Supplemental Fig. S2). To confirm the Human Protein Atlas finding, we performed Tdt expression analysis in B1-Cre⁺Tdt⁺ mouse tissue single-cell suspensions by flow cytometry. Tdt was expressed in tissues assessed, such as the brain, heart, liver, lung, small intestine, spleen, and kidney, at varying levels (Supplemental Fig. S3A, top, and Supplemental Fig. S3B).

ATP6V1G3 Is a Highly Specific Marker for Kidney ICs

The Human Protein Atlas tissue expression profile of ATP6V1G3 is mostly specific to the kidney (https://www.proteinatlas.org/ENSG00000151418-ATP6V1G3/tissue) (Supplemental Fig. S4). These data were confirmed in our human kidney CKIT⁺ IC scRNAseq experiment, which revealed ATP6V1G3 mRNA to be highly enriched in α-IC (A-IC) subsets (5). A study of mouse kidney scRNAseq also found that Atp6v1g3 mRNA is highly enriched in A-ICs (14). To confirm these findings, we immunolabeled human and C57BL/6 mouse kidney sections with anti-ATP6V1G3 (to label collecting duct cells) and anti-AQP2 (to label PCs) antibodies and colabeled with V-ATPase subunit E1 (ATP6V1E1) in the mouse kidney and found AQP2⁻ATP6V1G3⁺ ICs as well as a unique cell population of AQP2⁺ATP6V1G3⁺ cells (Fig. 1, A and B). Our previous scRNAseq study of human c-KIT enriched cells also showed AQP2⁺G3⁺ cells (5). Recent studies by Gao et al. have shown that AQP2⁺ progenitor cells (AP) express both AQP2 and ATP6V1B1 and ATP6V1B2 (26, 27). It is likely that AP cells may also express ATP6V1G3.

Figure 1.

Immunofluorescence staining of a wild-type C57BL/6 mice and human kidney showing collecting duct staining. A: V-ATPase G3 staining (green, arrows) to locate intercalated cells (ICs) and AQP2 staining (red) to locate principal cells (PCs). In both mouse and human collecting duct, G3⁺AQP2⁻ and several unique AQP2⁺G3⁺ cells were present. B: murine collecting duct ICs stained with V-ATPase E1 (green) and a subunit of ICs colocalized with V-ATPase G3 (red) IC subunit staining. Arrows in A represent AQP2⁻G3⁺ and arrowhead represents AQP2⁺G3⁺ cells. Arrows in B represent V-ATPase E1⁺V-ATPase G3⁺ cells. Magnification ×40. *Green autofluorescent red blood cells (RBCs) in the human kidney section.

Tdt Expression in G3-Cre⁺Tdt⁺ Mice Primarily Localizes to ICs

Based on the observations that ATP6V1G3 is highly kidney specific and enriched in ICs, we generated ATP6V1G3-Cre mice using CRISPR/CAS methodology as described in methods (Fig. 2, A and B). F0 progeny mice were bred with WT C57BL/6 mice and genotyped by RT-PCR to maintain a colony of mice. To confirm the specificity of G3-Cre, mice were bred with Tdt mice to generate G3-Cre⁺Tdt⁺ mice and various tissue single-cell suspensions were analyzed for Tdt expression by flow cytometry. Tdt was robustly expressed in the kidney while no Tdt expression was found in the brain, heart, liver, and spleen and little expression in lung cells (Supplemental Fig. S3A, bottom, and Supplemental Fig. S3C). Lung Tdt expression was further confirmed by immunofluorescence staining of G3-Cre⁺Tdt⁺ mouse lung tissue section, which showed few low V-ATPase G3 cells expressing robust Tdt and few high V-ATPase G3 expressing cells with low Tdt expression (Supplemental Fig. S5). Cre specificity in G3-Cre⁺Tdt⁺ was confirmed by immunofluorescence staining of kidney sections with anti-RFP antibody and colabeling with IC-specific V-ATPase E1 subunit antibody, which showed complete colocalization (Fig. 2C). Similarly, direct anti-Tdtomato antibody staining colocalized with V-ATPase G3, expressed in collecting duct cells (Fig. 2D and Supplemental Fig. S6). We also confirmed Cre recombinase protein expression in G3-Cre⁺Tdt⁺ mouse kidneys. Cre was expressed in cortical ICs with complete colocalization with Tdt in the cortex, whereas in the medulla, there were few ICs that expressed low-level Cre and strong Tdt, and a few ICs had Cre but no Tdt expression (Supplemental Fig. S7). For IC flow sorting experiments, G3-Cre mice were crossed with Tdt mice to generate G3-Cre⁺Tdt⁺ mice. Tdt⁺ cells were flow sorted and analyzed for enrichment of intercalated cell-specific genes by RT-PCR and found to be highly enriched for IC-specific genes (Fig. 3).

Figure 3.

Flow-sorted ICs from Atp6v1g3-Cre⁺Tdt⁺ mice are enriched in IC-specific mRNAs. ICs and nonintercalated cells (NICs) were enriched by flow sorting. For control, whole kidney lysate (WKL) cDNA was used. ICs were found to be enriched for IC-specific mRNA markers including Atp6v1b1 (A), Slc4a1 (B), and Slc26a4 (C) compared with NIC and WKL. Data are represented as a scatter dot plot of flow-sorted cells from 3 individual mice and the horizontal line represents the median. Each dot represents an individual mouse. The statistical test used was one-way ANOVA with **P < 0.01 and *P < 0.05 but >0.01. IC, intercalated cell; ns, not significant.

G3-Cre Is a Suitable Strain for IC Ablation

We further evaluated G3-Cre reporter mice by determining the functional consequences of specifically ablating ICs in vivo. To this end, we crossed G3-Cre mice with Cre-inducible ihCD59 mice for two generations to generate G3-Cre⁺ihCD59^+/+ mice (Fig. 4A). ILY, a toxin secreted by Streptococcus intermedius, binds exclusively to hCD59 and, after binding, forms toxin pores that lyse the cells within seconds (20). G3-Cre⁺ihCD59^+/+ mice were given two doses of 100 ng/g body wt ILY intraperitoneally and compared with untreated WT B6 mice. Blood chemistry was performed using the iSTAT system. Blood pH was significantly lower (7.338 ± 0.030 in WT vs. 7.222 ± 0.021 in G3-Cre⁺ihCD59^+/+, P = 0.0013) and potassium levels were significantly higher (4.120 ± 0.258 in WT vs. 4.900 ± 0.264 in G3-Cre⁺ihCD59^+/+, P = 0.0064), with bicarbonate trending lower (20.14 ± 0.9503 in WT vs. 19.20 ± 0.3606 in G3-Cre⁺ihCD59^+/+, P = 0.1602) in IC-ablated mice compared with WT control mice, suggestive of a distal renal tubular acidosis (dRTA) phenotype (Fig. 4, B–E). Kidneys were harvested and stained for IC-specific V-ATPase subunit marker V-ATPase E1 (Fig. 4, F and G) and counted. Mean IC cell count in G3-Cre⁺ihCD59^+/+ mice was 888 ± 177 versus 2,159 ± 183 in WT B6 mice (P < 0.05) (Fig. 4H). The IC cell area (91 ± 95 µm² WT vs. 61 ± 48 µm² G3-Cre⁺ihCD59^+/+, P < 0.0001) and cell perimeter (31 ± 20 µm WT vs. 24 ± 12 µm G3-Cre⁺ihCD59^+/+, P < 0.0001) were also significantly reduced in IC-ablated mice (Supplemental Fig. S8).

Figure 4.

G3-Cre mice are suitable for IC ablation in vivo. A: a representative PCR gel showing genotyping for generation of G3-Cre⁺ihCD59^+/+ mice. Lane 1 and lane 4 mice are iCD59^+/− (both WT and ihCD59 band) mice, whereas lanes 2 and 3 are G3-Cre⁺ihCD59^+/+ mice (only ihCD59 band). B–E: WT B6 and G3-Cre⁺ihCD59^+/+ mice were intraperitoneally treated with 2 doses of ILY, and blood chemistry analysis was performed the next morning using the iSTAT system. Blood pH was significantly lower (B), and potassium was higher in G3-Cre⁺ihCD59^+/+ mice (D). F and G: representative confocal image stitch of the V-ATPase E1-stained whole kidney section from WT B6 (F) and ihCD59^+/+ mice treated with ILY (G). F and G: after image stitches were generated of the kidney section, images were pseudoconverted on Leica LASX software using under/over glow color palette; most likely, the red cells are autofluorescent proximal tubule cells, whereas the white cells are ICs. H: scatter dot plot of whole kidney image stitch IC cell count. In A, M represents the DNA ladder. Each dot represents an individual mouse. The horizontal line in B–E and H is the median. An area from F and G was digitally enlarged to show V-ATPase E1 (ICs)-stained cells with arrows. The statistical test used was a 2-tailed unpaired t test with *P < 0.05, **P < 0.01, and ns = nonsignificant. IC, intercalated cell; ILY, intermedilysin; WT, wild type.

DISCUSSION

ICs are in the collecting duct of the kidney and are traditionally known for regulation of acid-base homeostasis. ICs express vacuolar-ATPase (V-ATPases), which is an ATP-driven pump known as a proton pumping rotary nanomotor (28, 29) and are involved in many physiological processes (30). V-ATPases are historically described as kidney specific or ubiquitous. In recent years, understanding of IC function has included the innate defense against invading uropathogens (3, 6–9). ICs are located in the distal kidney tubule segment as the collecting duct and would be among the first responders against invading uropathogens. To study ICs, two novel mouse models, namely B1-Cre and B1-EGFP, were produced by the Dr. Raoul Nelson research group (11, 12). To study IC gene expression, we previously generated B1-Cre “IC reporter” mice. ICs were sorted from these mice using Tdtomato fluorescence in the kidney, and whole transcriptome analysis was performed (3, 4), although ICs could be flow sorted from the B1-Cre mouse kidney after gating out CD45⁺ cells (which expressed robust Tdt) as CD45⁻Tdt⁺ with relative high purity. Robust Tdt expression on resident CD45⁺ immune cells and in a range of tissues matches the human V-ATPase B1 expression profile shown in the Human Protein Atlas (31) (https://www.proteinatlas.org). Later, we also sequenced human c-KIT-enriched ICs and identified innate immune functions and gene expressions in this cell type (5). To study the role of innate IC genes and their role in protection against uropathogens, an IC-specific gene knockout mouse would be helpful. Since V-ATPase B1 is expressed in multiple tissues, including immune cells, this strain has limitations regarding the study of genes of interest in a kidney IC-specific manner. In our human IC scRNAseq data, we found the ATP6V1G3 gene to be largely kidney IC specific, Atp6v1g3 has also been shown to be largely IC specific in mice (5, 14). Based on these observations, we generated G3-Cre mice and crossed them with Tdt mice and compared G3-Cre⁺Tdt⁺ with B1-Cre⁺Tdt⁺ mouse tissue Tdt expression by flow cytometry. Tdt expression using G3-Cre mice was more specific to the kidney compared with B1-cre mice, in which Tdt expression was found to be expressed in various tissues. ICs sorted from G3-Cre mice were highly enriched for IC-specific genes. In our pilot study, the G3-Cre mouse model was found to be suitable for IC ablation to determine the cellular functions and regeneration of ICs when crossed with ihCD59 mice (20, 21). IC ablation using the G3-Cre model resulted in a dRTA phenotype and represents a unique mouse model of acquired IC dysfunction that is more specific to ICs than pharmacologically induced disruption of acid-base regulation with carbonic anhydrase inhibitors (32). Blood chemistry in terms of glucose, sodium, and hemoglobin was not significantly different (Supplemental Fig. S9) but blood pH and potassium were different, suggestive of a more kidney-specific phenotype in G3-Cre mice compared with WT B6 mice after ILY administration. In addition, serum bicarbonate trended lower as well (P = 0.160). This study does have limitations. First, the percentage of ICs was lower when the Atp6v1g3 Cre mouse was used for enrichment versus when the B1-Cre mouse was used. Potential explanations for this include that B1-Cre is “leakier” with some off-target Cre expression, that Atp6v1b1 is expressed in cells other than ICs or that not all ICs identified using the G3-Cre mice did not label all inner medullary collecting duct ICs, or a combination of some of these explanations. The leakiness of B1-Cre mice is also supported by the fact that mRNA expression of Atp6v1b1 is similar to Atp6v1g3 in WT B6 mice in different tissues assessed (Supplemental Fig. S10). It may also be speculated that the G3 subunit is less abundantly expressed than the B1 subunit in V-ATPase assembly, resulting in lower rates of IC detection, and/or that G3-Cre promoter activity is different in different IC locations. We also found AQP2⁺ cells that were also G3⁺ (Fig. 1). The most likely explanation for this is that these represent hybrid PC-IC cells. Another possibility is that G3-Cre is not 100% IC specific. The hybrid cell explanation is most likely because we and others have previously found that hybrid IC-PC cells account for 5% of V-ATPase E1 cells of the mouse kidney (4, 33) and identified V-ATPase G3⁺AQP2⁺ hybrid cells on scRNAseq of the human kidney cells (5). It has also been shown that AQP2⁺ cells can give rise to ICs (34); therefore, it is likely that in our data, AQP2⁺G3⁺ cells are transitional cells. Magnetic or flow sorting does not result in pure cell populations, but rather relative enrichment, and we were able to demonstrate robust enrichment using the G3-Cre mice. Miller et al. (12) showed that V-ATPase B1 subunit promoter-driven expression of EGFP may extend beyond ICs in the kidney in the connecting tubule segment, which expresses calbindin (12). Our V-ATPase B1 and TdTomato double labeling identified “leakiness” of Cre expression. We confirmed findings of Miller et al. and compared it with G3-Cre mice, which showed Tdt expression in mostly calbindin-low-expressing cells, whereas B1-Cre mice had several Tdt-expressing cells with high calbindin positive cells (DCT) (Supplemental Fig. S11). Although recent studies by Brown et al., Frische et al., and Loffing et al. (35–37) have shown that V-ATPase B1 is also expressed by DCTs that are not ICs but express sodium-chloride cotransporter (NCC) and epithelial sodium channel (ENaC), hence the Tdt expression in B1-Cre⁺Tdt⁺ mice, DCT cells may also be true V-ATPase B1 expression, not leakiness. Interestingly, there was no direct overlap between our Atp6v1g3 mRNA and V-ATPase G3 subunit protein expression. However, this does mirror what is reported for human ATP6V1G3 in the Human Protein Atlas (https://www.proteinatlas.org/ENSG00000151418-ATP6V1G3/tissue) (31). We did find some Atp6v1g3/V-ATPase G3 subunit expression in the lung, albeit lower than lung Atp6v1b1/V-ATPase B1 subunit expression, potentially reflecting some minor expression differences between human and mice. We were unable to collect urine from the mice when we collected the blood samples. In future studies, urine pH, urine electrolytes, or both will help to further define the renal tubular acidosis status. Finally, we did not obtain 100% ablation of ICs using the ihCD59/ILY system. This could be due to the dose or administration needing to be further optimized or intrinsic resistance to complete ablation by ICs or inherent property of G3-Cre mice that does not label all ICs and hence cannot be ablated. For the ILY ablation experiment, we used ATP6V1E1 (E1) as an IC marker as has been done by others (38, 39). However, E1 has been reported to be more ubiquitous than G3 (40). It is possible that some of the nonablated labeled cells are E1⁺G3⁻ cells, perhaps in the proximal tubule or in the loop of Henle (30). Nonetheless, we appeared to have a partial IC ablation in the G3-Cre mouse kidneys, a finding that supported the low serum pH and cell counting. To show that we labeled B1-Cre⁺Tdt⁺ and G3-Cre⁺Tdt⁺ mice with Tdtomato antibody, we compared them with WT B6 mice labeled with V-ATPase B1 and V-ATPase G3 antibody and performed whole kidney image stitching. These data showed that G3-Cre⁺Tdt⁺ mice do not label all the ICs in the inner medulla (Supplemental Fig. S12). More mice are needed to confirm this finding, although we did demonstrate that they developed a phenotype consistent with a hyperkalemic voltage-dependent dRTA. Specifically, following IC ablation, serum pH was significantly lower, potassium was significantly higher, and serum bicarbonate trended lower. Thus, this partial IC ablation may more closely reflect acquired IC damage seen in conditions such as obstructive uropathy (41). Taken together, the G3-Cre mouse is a novel strain that can be used to enrich kidney ICs and ablate genes of interest in a largely kidney IC-specific manner while limiting the confounding factor of nonspecific deletion in nonkidney tissues and other immune cells.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S12: https://doi.org/10.6084/m9.figshare.24029907.v1.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01DK106286 (to A.L.S. and D.S.H.) and by the Eli Lily Foundation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

V.S., D.S.H., and A.L.S. conceived and designed research; V.S., S.A., and J.C. performed experiments; V.S., S.A., S.Z., D.S.H., and A.L.S. analyzed data; V.S., S.Z., D.S.H., and A.L.S. interpreted results of experiments; V.S. and A.L.S. prepared figures; V.S. drafted manuscript; V.S., S.Z., J.C., X.Q., D.S.H., and A.L.S. edited and revised manuscript; V.S., D.S.H., and A.L.S. approved final version of manuscript.