Skip to main content
NCBI home page
Physiological Genomics logoLink to Physiological Genomics
. 2021 Dec 10;54(2):45–57. doi: 10.1152/physiolgenomics.00102.2021

Kidney cell type-specific changes in the chromatin and transcriptome landscapes following epithelial Hdac1 and Hdac2 knockdown

Kelly A Hyndman 1,✉, David K Crossman 2
PMCID: PMC8791845  PMID: 34890513

graphic file with name pg-00102-2021r01.jpg

Open in a new tab

Keywords: ATAC, histone deacetylase, kidney, single nucleus, transcriptome

Abstract

Recent studies have identified at least 20 different kidney cell types based upon chromatin structure and gene expression. Histone deacetylases (HDACs) are epigenetic transcriptional repressors via deacetylation of histone lysines resulting in inaccessible chromatin. We reported that kidney epithelial HDAC1 and HDAC2 activity is critical for maintaining a healthy kidney and preventing fluid-electrolyte abnormalities. However, to what extent does Hdac1/Hdac2 knockdown affect chromatin structure and subsequent transcript expression in the kidney? To answer this question, we used single nucleus assay for transposase-accessible chromatin-sequencing (snATAC-seq) and snRNA-seq to profile kidney nuclei from male and female, control, and littermate kidney epithelial Hdac1/Hdac2 knockdown mice. Hdac1/Hdac2 knockdown resulted in significant changes in the chromatin structure predominantly within the promoter region of gene loci involved in fluid-electrolyte balance such as the aquaporins, with both increased and decreased accessibility captured. Moreover, Hdac1/Hdac2 knockdown resulted different gene loci being accessible with a corresponding increased transcript number in the kidney, but among all mice only 24%–30% of chromatin accessibility agreed with transcript expression (e.g., open chromatin and increased transcript). To conclude, although chromatin structure does affect transcription, ∼70% of the differentially expressed genes cannot be explained by changes in chromatin accessibility and HDAC1/HDAC2 had a minimal effect on these global patterns. Yet, the genes that are targets of HDAC1 and HDAC2 are critically important for maintaining kidney function.

INTRODUCTION

The mammalian kidney is composed of a diverse array of cell types that perform numerous physiological functions pertinent to maintaining homeostasis. From the vascular structures such as the glomerulus to the different epithelial cells that compose the nephron and to the interstitial cells (e.g., immune, fibroblast, and mesangial), it is perhaps no surprise that with the technological advancement of single-cell and single-nucleus sequencing protocols, at least 20 kidney cell types have been defined (1). Moreover, there are multiomic approaches that permit researchers at the single cell/nucleus level to quantify transcripts, chromatin accessibility, and, likely soon, methylation and proteomes in almost any sample including human biopsies (2). Thus, single cell/nucleus sequencing has expanded our understanding of the diversity of the cellular landscape of the kidney both in a healthy state and in diseased states such as chronic kidney disease, diabetes, and hypertension (3–6).

It is established that gene expression is regulated by various factors. Transcription requires open chromatin, transcription factors, coregulators, and polymerases (7). After transcription, there are a variety of mechanisms that can significantly affect gene expression including interactions with noncoding RNAs such as microRNA (8) or long noncoding (Lnc)RNA (9). In eukaryotic cells, the DNA is condensed and tightly wrapped around histone proteins and in this state called heterochromatin. Histone proteins have numerous well-defined lysine residues that carry a positive charge, bind tightly to the negative DNA, and thereby prevent transcription (10). To transcribe a gene, the interaction between the histone and DNA can be modified by post-translation histone lysine acetylation, which would nullify the charge, leading to opening of the chromatin structure and allowing transcription factors, coregulators, and polymerases access to the gene for transcription. Thus, regulation of lysine acetylation status is a fundamental mechanism for the initiation of transcription.

The enzyme family of histone deacetylases (HDACs) function in the removal of acetyl groups from histone lysines (11). There are 11 HDACs in mammals and they are all expressed in the kidney (12). Class I HDACs (HDAC1, HDAC2, HDAC3, and HDAC8) are predominantly localized to the nucleus and in the kidney HDAC1 and HDAC2 are ubiquitously expressed (13). The central dogma is that HDACs are gene silencers because they restore the positive charge on the lysine through removal of the acetyl group and this would result in tight binding between the histone and DNA. However, a study where HDAC1 and HDAC2 were genetically knockdown from the developing ureteric bud of the kidney, reported that only 0.66% of transcripts were decreased and only 0.55% of transcripts were increased (14). Although only ∼1.2% of genes were significantly affected, HDAC1/HDAC2 knockdown was significant because the kidneys failed to develop normally in these mice (14). Thus, although HDACs regulate gene expression, inhibition of them does not result in global changes in gene expression but the specific genes they affect can have profound effects. But what about in adulthood? Patients take HDAC inhibitors to treat hematological malignancies and neurological disorders (11). Chronic HDAC inhibitor usage results in an increase odds ratio of experiencing a severe fluid-electrolyte disorder (15) and this can be explained in part by inhibition of HDAC1 and HDAC2 activity in the kidney resulting in a downregulation of salt and water transporters/channels (15). What is unknown is whether these kidney transcriptome differences are due to significant changes in the chromatin structure of these genes or other mechanisms that affect gene expression.

The purpose of the current study was to determine to what degree does kidney HDAC1 and HDAC2 knockdown in adulthood affects gene expression through modification of the kidney chromatin landscape. By coupling snRNA-seq and single nucleus assay for transposase-accessible chromatin-sequencing (snATAC-seq) (16) from the same kidney, we determined that kidney HDAC1/HDAC2 regulate a small proportion of chromatin accessibility and gene transcripts, however, the genes they regulate are critical for maintaining a healthy kidney.

METHODS

Animals and Sample Collection

All animal use and welfare adhered to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals following a protocol reviewed and approved by the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham (UAB). Our in-house colony of inducible, kidney epithelial Hdac1/Hdac2 knockdown mice [Hdac1Fl/Fl;Hdac2Fl/Fl; doxycycline-inducible Pax8-reverse tetracycline transactivator (rtTA) and bicistronic Cre (LC-1) hemizygous positive] and littermate controls were used and the genotyping protocol and confirmation of knockdown was previously published (15). The littermate control animal genotype was Hdac1Fl/Fl;Hdac2Fl/Fl;LC-1 hemizygous but Pax8-rtTA negative. For simplicity, we call the kidney epithelial Hdac1/Hdac2 knockdown mice “KO.” Both Hdac1 and Hdac2 were targeted for knockdown because a single allele of either can compensate for the loss of the other alleles in the epithelium (14). All mice were given doxycycline via ad libitum sweetened water (2 mg/mL in 2% sucrose water) provided for 7 days at age 7–8 wk old. This was followed by a 7-day doxycycline washout period where tap water was provided in the cages and euthanasia occurred on the 8th day between 7:00 AM and 8:00 AM. Both male and female mice were used, and genotype and sex were identified as variables in downstream analyses.

For euthanasia and sample collection, the mice were deeply anesthetized with inhaled 2.5% isoflurane, blood collected in a heparinized syringe via cardiac puncture through the diaphragm. Blood chemistry was immediately analyzed with the iSTAT EC8+ cartridge (Supplemental Table S1; see https://doi.org/10.6084/m9.figshare.15141645.v1), and the remainder of the blood centrifuged at 1,000 g for 10 min and plasma collected, snap frozen, and stored at −80°C. Next, the mouse was perfused with 10 mL of sterile 1× phosphate buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH = 7.4) via the heart to help clear the organs of the remaining blood. The kidneys were excised, decapsulated, weighed, and cut in cross sections. One-half of the left kidney was placed in 10% neutral buffered formalin for 24 h at room temperature (22°C) for downstream histological analyses. The other three halves were individually snap frozen and stored −80°C until nuclei isolation.

Histology

After the kidneys were fixed, they were processed and embedded in paraffin by the UAB Animal Resource Program Comparative Pathology Laboratory (Birmingham, AL). Four micron sections of kidney were placed on Superfrost Plus slides (Thermo Fisher Scientific). The slides were deparaffinized with xylenes and rehydrated with a series of 100% ethanol to deionized water washes. The kidney sections were then stained with Gömöri trichrome, following the manufacturer’s instruction (Richard-Allan Scientific, San Diego, CA). The slides were then dehydrated with 70%–100% ethanol (in water), cleared with two changes of fresh xylenes, and mounted with cytoseal 60 and a coverslip. Images of the kidneys were taken with a BX43 microscope fitted with a DP80 camera and CellSens Dimension software (v1.12, Olympus, Tokyo, Japan).

Nuclei Isolation

Nuclei were isolated from half of a kidney for each group: male control, male KO, female control, and female KO. In 2019, the nuclei were isolated for the snRNA-seq experiment and the details were published in a study by Hyndman et al. (15). In 2020, nuclei were isolated from the other half of the kidney from the same mice following the protocol of Muto et al. (2) for the snATAC-seq experiment. In short, frozen kidneys were chopped in Nuclei EZ Lysis Buffer (Sigma) with 1X halt protease inhibitor cocktail (Thermo Fisher Scientific), transferred to a sterile, RNase/DNase-free tube, and gently homogenized with hand-held pestle mixer (Fisher) on ice. An additional 1 mL of nuclei lysis buffer was gently mixed with the sample by pipetting up and down and the homogenized sample was incubated on ice for 5 min. The homogenate was filtered with a 40-µm pluristrainer (pluriSelect) with an additional 2 mL of lysis buffer. The filtrate was then centrifuged at 500 g for 5 min at 4°C, and the pellet resuspended in 2 mL of nuclei lysis buffer. After a 5 min incubation on ice, the sample was then centrifuged and the pellet resuspension, incubation on ice, and centrifuging cycle completed once more. The final pellet was resuspended in the 10X Genomics Nuclei Buffer (PN-2000153), centrifuged one last time, and the pellet resuspended in 500 µL of the Nuclei Buffer, vacuum filtered with a 5-µm pluristrainer, and taken to the UAB Comprehensive Flow Cytometry Core. At least 4,000 nuclei/sample were loaded in the Chromium Next GEM Chip H Single Cell Kit and the DNA libraries generated with the Chromium Next GEM Single Cell ATAC Library & Gel Bead Kit v1.1 (10X Genomics). Sequencing of these libraries was completed with a NextSeq500 (Illumina) with the help of Dr. Michael Crowley of the UAB Genomics Core Facility. The sequencing depth targeted was 25,000 reads/nucleus, in a paired-end, dual index format which targeted 100 M base pair reads/sample.

snRNA-Seq Bioinformatics Workflow

Our snRNA-seq libraries are available at the National Center for Biotechnology Information Gene Expression Ombnibus (GEO) under accession GSE148354. This data set was originally analyzed with R package Seurat v3.0 and the data were normalized using the SCTransform function (15). In this current study, we reanalyzed the 25,075 high-quality nuclei with the LogNormalization and Harmony batch correction (17) functions RunHarmony using Seurat v4.0.2. This was followed by unsupervised clustering, cell type annotation, and differential expression of genes (DEG) using the DESeq2 function with Seurat. Recent studies have demonstrated that the proximal tubules express sexually dimorphic markers (18). As such, the proximal tubules were annotated based upon the sex of the mouse from which they originated.

snATAC-Seq Bioinformatics Workflow

The fastQ files of the four snATAC-seq libraries were analyzed individually using the CellRanger-ATAC v1.2.0 pipeline. These data files were deposited into GSE181557 (to review GEO accession GSE181557.

A total of 16,545 nuclei with an average 32,340 reads/nucleus (standard deviation (SD) = 5,471) were sequenced (Supplemental Table S2; see https://doi.org/10.6084/m9.figshare.15141660.v1). The median fragments per nucleus was 14,061 (SD = 1,903) and 62% (SD = 2) of the fragments overlapped any target region (Supplemental Table S2). This data set was subsequently merged, integrated, and analyzed with Seurat v4.0.2 and Signac v1.2.0 (19). The peaks.bed files for each library were read using the read.table function and the peaks converted to genomic ranges using the makeGRangesFromDataFrame. This was then combined to create a unified set of peaks to quantify in each library. The metadata (singlecell.csv) were loaded for each sample and low count cells filtered (pass_filters > 500). Next, the fragment files were generated with the “CreateFragmentsObject” using the fragments.tsv.gz and metadata files. From here the count matrixes (FeatureMatrix) and chromatin assays (CreateChromatinAssay) were created using the Mus musculus reference genome (mm10), min.cells = 10, and min.features = 200. Finally, Seurat objects were created using the CreateSeuratObject function. At this point, sex and genotype were annotated in each library and low quality nuclei were further removed based upon the following criteria as recommended by 10X Genomics and published in a study by Muto et al. (2): peak region fragments >2,500 but <25,000, percent reads in peaks >15, blacklist ratio < 0.02, nucleosome ratio < 4, and transcription start site (TSS) enrichment >2. The four Seurat objects were then merged resulting in a final data set of 15,868 high quality nuclei.

The snATAC-seq data set was normalized with term-frequency inverse-document-frequency (TFIDF), dimensional reduction performed with the singular value decomposition (SVD), and uniform manifold approximation and projection (UMAP). To correct for potential batch effects, the method Harmony was applied to the data set (RunHarmony). A gene activity matrix was generated with the GeneActivity function, and the data set processed again with TFIDF, SVD, and UMAP. Clustering was performed with FindNeighbors and FindCluster functions and cluster annotation was performed with two methods. The first cluster annotation application was to identify markers with the FindAllMarkers function and manually annotate clusters based upon published cell-specific markers. The second cluster annotation was performed using the FindTransferAnchors, which transfers labels based upon the cluster annotation of the snRNA-seq data set.

Differential accessibility of chromatin (DAC) between cell types or between genotypes within a cell type were calculated using the FindMarkers function for peaks detected in a minimum of 20% of cells, with a log2-fold-change (log2FC) cutoff of 0.25, and an Bonferroni-adjusted P value of <0.05. Correlations between DEGs and DACs were performed for the control and KO mice. Because there are often multiple DACs within one gene, we opted to keep the highest Log2FC per gene per cluster. Correlations were performed with GraphPad Prism (v9.1.2) between the DEG log2FC vs the DAC log2FC. A χ2 test with 3 degrees of freedom was used to test for statistical differences between the control and KO mice in each of the four quadrants. Gene ontology (GO) and pathway analyses with Kyoto Encyclopedia of Genes and Genomes (KEGG) were performed with the genes from these quadrants using the Database for Annotation, Visualization and Integrated Discovery v. 6.8 (20).

Tn5 integration was determined globally within the genome and specifically in each cluster for all mice using the code and methods reported by Muto et al. (2). Gene coverage plots and snATAC-seq and snRNA-seq expression plots were generated using Signac and Seurat. All code used in the study analyses can be found at github (https://github.com/hyndmank/snATAC_Hyndman).

Quantitative Real-Time PCR

Follow-up validation using quantitative real time PCR was performed from an additional cohort of mice (control males n = 5, females n = 4; KO males n = 4, females n = 4). Plasma chemistries and histological analyses of these mice were previously published (15) and matched the phenotype of the mice used in the seq analyses. RNA was isolated from a half of kidney with the Zymo Direct-zol RNA kit. One microgram of RNA was reverse transcribed into cDNA with the VILO-SSVI kit (Thermo Fisher Scientific) and SYBR green based quantitative real-time PCR (qPCR) performed as previously reported (21). Primer sequences were from the Primer Bank (22) and are listed in Supplemental Table S3 (see https://doi.org/10.6084/m9.figshare.15141663.v1).

RESULTS

Demographics of the Mice

The kidneys from the same mice were used in the snRNA-seq and snATAC-seq analyses. Previously, it was reported from a large cohort of male and female mice that kidney epithelial Hdac1/Hdac2 knockdown (KO) in adulthood resulted in a serious fluid-electrolyte disorder and kidney damage compared with littermate control mice (15). We confirmed that the KO male and female mice used in the sequencing experiments had elevated plasma sodium, a higher kidney/body mass ratio (Supplemental Table S1), and kidney damage that was assessed by histological analysis (Fig. 1, A and B, and Supplemental Fig. S1; all Supplemental Figures are available at https://doi.org/10.6084/m9.figshare.15173679.v1). The male KO mouse had a loss of the proximal tubule brush border and mild interstitial fibrosis that was not observed in the control male mouse (Fig. 1A and Supplemental Fig. S1, A and B). The female KO mouse had a mild reduction in brush border integrity compared with the female control or male KO (Fig. 1B and Supplemental Fig. S1, C and D); however, there was no obvious interstitial fibrosis in the female KO (Fig. 1B and Supplemental Fig. S1, C and D). Thus, the mice used in the sequencing experiments are representative of the previously reported phenotype of this colony (15).

Figure 1.

Figure 1.

Open in a new tab

Representative histological images of kidney sections stained with Gomori’s trichrome from the mice used in this study. A: high-magnification images of the cortex, outer medulla, and inner medulla from the male control and male kidney epithelial Hdac1/Hdac2 knockdown mice (KO). B: high-magnification images from female control and KO mice. Scale bar, 20 µm.

snRNA-Seq and snATAC-Seq Kidney Clusters and Annotation

In the snRNA-seq data set the 25,075 nuclei (12,546 from the control mice and 12,529 from the KO mice) formed 26 clusters (Supplemental Table S4; see https://doi.org/10.6084/m9.figshare.15141669.v1). Cluster annotation based upon conserved markers previously reported (2, 3, 5, 6, 18) revealed all of the expected cell types in the kidney including endothelial, immune, stromal, and epithelial cells (Fig. 2A and Supplemental Table S4). Clear sexual dimorphisms in the proximal tubules were evident (Fig. 2, A and E, and Supplemental Table S4). Cluster PT5, which shares conserved gene markers with proximal tubules, was highly enriched in the KO mice as we previously reported (15). Cluster PT shared markers for both the S1 and S2 segment and was expressed by both sexes and genotypes (Supplemental Table S4). A limitation to this snRNA-seq data set was that it was not possible to annotate the proximal tubules to their segment specific nomenclature of S1, S2, or S3.

Figure 2.

Figure 2.

Open in a new tab

Unsupervised clustering of nuclei from the single nucleus (sn)RNA-sequencing (seq) (A) and snATAC-seq (B) experiments from control and kidney specific Hdac1/Hdac2 knockdown mice. Sixteen clusters in the snRNA-seq and 20 clusters in the snATAC-seq were annotated based upon published gene markers. Note that the proximal tubules display significant sex-specific clusters (denoted male = m, female = f). The snATAC-seq was used to identify S1, S2, and S3 segments of the proximal tubule, but this level of resolution was not obtained in the snRNA-seq. Typical cell-specific markers in each cluster for the snRNA-seq (C) and snATAC-seq (D) data sets. Unsupervised clustering of the nuclei grouped by sex with females in pink and males in blue: the snRNA-seq (E) and snATAC-seq (F) data sets. c, cortical; CDPC, collecting duct principal cell; DCT, distal convoluted tubule; EC, endothelial cell; IC, intercalated cell; m, medullary; PT, proximal tubule; TAL, thick ascending limb.

With the snATAC-seq data set the 15,868 nuclei (8,016 from the control mice and 7,852 from the KO mice) formed 26 clusters (Fig. 2B and Supplemental Table S4). All major cell types expected in the kidney were identified (Fig. 2B). Moreover, there were clear sexual dimorphisms in the chromatin accessibility (Tn5 insertion) of the proximal tubules (Fig. 2, B and F). Similar to the snRNA-seq data set, cluster PT5 was enriched in only the KO mice regardless of sex, and cluster PT was found in all mice. However, contrary to the snRNA-seq, the proximal tubules were clearly identifiable as the S1, S2, and S3 segments (Fig. 2B). This level of annotation was also reported for human kidney proximal tubule cells where the snATAC-seq could differentiate the tubules into S1–S3 while the snRNA-seq was less reliable (2). In addition, the snATAC-seq clusters were annotated informatically based upon the differentially expressed genes in the snRNA-seq data set (Supplemental Fig. S2, A and B) and further confirmed our manually clustered annotation. Finally, we overlaid the snATAC-seq and snRNA-seq data sets and demonstrated overlap within the clusters from these different technologies (Supplemental Fig. S2C).

Examples of conserved snRNA-seq genes and snATAC-seq accessible genes are reported in Fig. 2, C and D, respectively. Genetic coverage plots of chromatin accessibility (frequency of Tn5 insertion) from the TSS to the 3′-untranslated region (UTR) for cell specific genes are presented in Supplemental Fig. S3. For example, there are increased chromatin peaks along the phosphoenolpyruvate carboxykinase 1 (Pck1) gene in all proximal tubule clusters, while the principal cell marker, Aqp2, is enhanced in the annotated cluster collecting duct principal cell (CDPC) (Supplemental Fig. S3).

Tn5 Insertion within the Mouse Genome

Next, we determined the percent Tn5 insertion along the genome and within each cluster of nuclei for each genotype and separated by sex. For all the mice, the greatest proportion of Tn5 insertion was within 1 kb (kilobase) of the TSS of the promoter (Fig. 3, A–D). Between the female control and KO mice, there were no statistically significant differences among the Tn5 integration at the whole genome level (Fig. 3, A and B, χ2 = 11.05, df = 8, P = 0.2). However, between the male control and male KO there was significantly greater Tn5 integration in the promoter region (<1 kbp) of the control mouse and greater integration in the distal intergenic regions (Fig. 3, C and D, χ2 = 89.05, df = 8, P < 0.0001). At the individual cluster level, there was significantly greater Tn5 integration in the PT cluster of the KO mice compared to controls regardless of sex (female χ2 = 88.67, df = 5, P < 0.0001; male χ2 = 59.18, df = 5, P < 0.0001; Fig. 3, E–H). This PT cluster expresses markers of both the S1 and S2 segment and is highly expressed in the control mice compared with the KO (Supplemental Table S4).

Figure 3.

Figure 3.

Open in a new tab

Tn5 integration in the genomes of the control and kidney specific Hdac1/Hdac2 knockdown mice (KO) separated by sex. The percent Tn5 integration at the whole genome level of the control female mouse (A), KO female mouse (B), male control mouse (C), male KO mouse (D). The distribution of Tn5 integration was significantly different between male control and KO (χ2 = 89.05, df = 8, P < 0.0001). Tn5 integration within each cluster for a female control (E), male control (F), female KO (G), male KO (H). KO mice had significantly different Tn5 integration in cluster PT (*female χ2 = 88.67, df = 5 P < 0.0001; #male χ2 = 59.18, df =5, P < 0.0001). c, cortical; CDPC, collecting duct principal cell; DCT, distal convoluted tubule; EC, endothelial cell; f, female; IC, intercalated cell; m, male; m, medullary; PT, proximal tubule; TAL, thick ascending limb.

Differentially Expressed Genes and Differentially Accessible Chromatin between Control and KO Mice

Unsupervised clustering of the nuclei between the control and KO mice resulted in some distinctions among the clusters (Fig. 4, A and B). In both the snRNA-seq (Fig. 4A) and snATAC-seq (Fig. 4B) there were clear distinctions between the PTs of control and KO mice. Nephron segments distal to the PTs had fewer differences (greater overlaps) as shown in the unsupervised clustering depicted in Fig. 4, A and B.

Figure 4.

Figure 4.

Open in a new tab

Unsupervised clustering of nuclei from Fig. 2 demonstrating clear genotype specific differences in the clusters of the snRNA-seq (A) and snATAC-seq (B) data sets. Nuclei from control mice are in pink and from kidney specific Hdac1/Hdac2 knockdown mice (KO) in blue. Correlations between the log2-fold change (log2FC) of cluster specific differentially accessible chromatin peaks (DAC) and differentially expressed genes (DEG) from the snRNA-seq data set: genes from the control mice (C) and genes from the KO mice (D). Numbers in the quadrants represent the gene count. E: summary of the number of genes either unique (nonoverlapping) or conserved (overlapping) between control (pink) and KO (blue) mice in each quadrant. CDPC, collecting duct principal cell; DCT, distal convoluted tubule; EC, endothelial cell; IC, intercalated cell; Podo., podocyte; TAL, thick ascending limb.

To determine if the genes used to define the clusters within these data sets had accessible chromatin structure and thus were considered “open” correlated with an increase in transcript number, or if the gene was inaccessible (considered closed) and correlated with a decrease transcript number, correlations between the log2FC for a gene in the snATAC-seq and snRNA-seq data sets were performed. These analyses were performed on the control and KO samples independently to determine if knock down of Hdac1/Hdac2 affected the number of open and closed genes. The control mice had a total of 5,811 (31%) gene specific DEGs and DACs in agreement out of the 18,585 cluster-specific DEGs (absolute log2FC > 0.25, adjusted P value < 0.05). Of these genes, 3,206 genes were considered open and expressed and 2,036 genes were closed and not highly expressed (Fig. 4C). Within each cluster, the proportion of DEGs that agreed with the DAC was on average 0.27 with the minimum proportion of 0.09 in the EC2 cluster and maximum proportion of 0.46 for the podocytes (Supplemental Table S5; see https://doi.org/10.6084/m9.figshare.15141672.v1). Of these DEGs and DACs only 9.7% did not agree (e.g., were considered open in the snATAC-seq but were not highly expressed in the snRNA-seq; Fig. 4C). Similar results were found in DACs and DEGs from the KO mice (Fig. 4D). There were 4,226 DEGs that had a corresponding DAC out of 17,378 (24%). Of these, 2,543 were considered open and expressed and 1,202 were closed and not highly expressed (Fig. 4D). Only 11% of the genes did not agree. Within each cluster, the proportion of DEGs that agreed with DACs was on average 0.27 with a minimum proportion of 0.07 in cluster EC2 and a maximum proportion of 0.75 in the male proximal tubules (Supplemental Table S5). The number of expected genes in each quadrant of Fig. 4, C and D was statistically different than observed for both the control and KO mice [degrees of freedom (df) = 3, χ2 = 50.21, P < 0.0001). This was driven by a greater than expected number of genes in the positive DAC and positive DEG quadrants of the KO mice and greater than expected number of genes in the negative DAC and negative DEG quadrants of the control mice (Fig. 4, C and D). When comparing the genes within each quadrant between the control and KO mice, there are genes that are conserved between the genotypes (Fig. 4E). However, there are many genes that are unique to each genotype (Fig. 4E). To determine if there are significant pathways that are associated with these genes we performed gene ontology (GO) analyses. In the positive DAC and positive DEG quadrants of Fig. 4C, control mice have enrichment in metabolic processes (GO:008152) and fatty acid metabolism (CO:006631), whereas the KO mice (Fig. 4D) have genes enriched in cell adhesion (GO:007155) and transforming growth factor beta receptor signaling (GO:007179) (Supplemental Table S6; see https://doi.org/10.6084/m9.figshare.16811488.v1). When comparing the genes in the DAC and DEG negative quadrants (closed genes), again the control mice are enriched in genes in metabolic processes (GO:008152) but the KO mice are enriched in pathways such as ρ signaling (GO:0035023) and actin filament bundle assembly (GO:0051017). In the other quadrants (e.g., positive DAC but negative DEG), there were very few statistically significant pathways suggesting individual gene differences and not enrichment of pathways. All of the GO and KEGG analyses are reported in Supplemental Table S6.

Next, we examined and validated specific genes of interest to determine if there was agreement between the DACs and DEGs. The kidney specific Hdac1/Hdac2 KO mice have a severe fluid-electrolyte disorder and present with kidney injury (15). The elevated plasma sodium in the KO mouse was associated with fewer aquaporins (water channels) expressed along the nephron compared to the control mouse (15; Supplemental Table S7; see https://doi.org/10.6084/m9.figshare.15141678.v1; Supplemental Table S8; see https://doi.org/10.6084/m9.figshare.15141687.v1). Figure 5 shows the proximal tubule and thin limb of Henle expressed water channel, Aqp1, and the CDPC aquaporins, Aqp2 and Aqp4. A coverage plot of the Aqp1 gene (chromosome 1, position 55336432-5534855 in the mouse genome) depicting the Tn5 insertion peaks is presented in Fig. 5A. Control mice have higher Tn5 insertion peaks along the Aqp1 gene, especially near the first exon, compared to the KO mouse (Fig. 5A). This Tn5 insertion can be quantified and DACs calculated within each cluster. The DAC log2FC for PTs from male mice was for example in mPTS1 1.07 (P = 8.61 × 10−4) and female mice fPTS1 0.96 (P = 2.83 × 10−51), thus in both sexes control mice had greater Tn5 insertion in the Aqp1 gene than KO mice (Fig. 5B and Supplemental Table S8). This correlated with significantly greater Aqp1 RNA expression in the control mice (Fig. 5C and Supplemental Table S7) and was further validated with a larger cohort of male and female control and KO mice (Fig. 5D).

Figure 5.

Figure 5.

Open in a new tab

Aquaporin (Aqp) chromatin accessibility and RNA expression from the snATAC-seq and snRNA-seq data sets with validation in a lager cohort of mice with quantitative real-time PCR. A–D: Aqp1 is highly expressed in the proximal tubule (PT). A: coverage plots of the frequency of Tn5 insertion in the Aqp1 gene. Nuclei from control mice are presented in pink and from the kidney specific Hdac1/Hdac2 knockdown (KO) in blue. B: violin plots of quantification of the Tn5 insertion along the Aqp1 gene for male (m) and female (f) PTs. The PT cluster was identified in both sexes. Dots represent individual nuclei. C: Aqp1 RNA expression from the snRNA-seq data set plotted as a violin plot with dots representing individual nuclei for the PTs. D: validation with cDNA derived from whole kidney RNA from female (closed symbols) and male (open symbols) control and KO mice. n = 4 or 5 per sex per genotype, individual mice plotted. P values are reported from an unpaired, two-tailed, Student’s t test. E–H: Aqp2 is highly expressed in the collecting duct principal cell (CDPC). Coverage plot (E) and quantification of the Tn5 insertion (F). RNA expression (G) and validation with qPCR (H). I–L: Aqp4 is also highly expressed in the CDPC. Coverage plot (I) and quantification of the Tn5 insertion (J). RNA expression (K) and validation with qPCR (L).

Next, we examined the principal cell water channels Aqp2 and Aqp4. Similar to Aqp1, Aqp2 Tn5 insertion (chromosome 15, position 99577556–99585545) was significantly greater in the control mice than KO (log2FC = 1.05, P = 0.002) (Fig. 5, E and F and Supplemental Table S8). This was associated with a significant DEG log2FC of 1.28 (P = 0.0002) (Fig. 5G and Supplemental Table S7) and validated with qPCR in the whole kidney (Fig. 5H). However, kidney specific Hdac1/Hdac2 knockdown did not significantly affect all aquaporins, as the principal cell Aqp4 (chromosome 18, position 15389394–15403684) was not differentially accessible (Fig. 5, E and F and Supplemental Table S8), nor was it differentially expressed (Log2FC = 0.82, P = 1, Fig. 5, G and H and Supplemental Table S7).

Last, we examined the relationship between DACs and DEG for two highly significant genes in the PT clusters: Target of Nesh-SH3 (Abi3bp, also called Tarsh) and acyl-coenzyme synthetase (Acsm2). Abi3bp encodes a newly identified extracellular matrix protein (23). Tn5 was inserted in the Abi3bp gene of only 1%–2% of PTs from control mice but was inserted in 20%–30% of PTs from KO male mice (mPTS3 log2FC = −2.10, P = 3.59 × 10−29) and female mice (fPTS1 log2FC = −1.3, P = 3.06 × 10−17; Fig. 6, A and B, and Supplemental Table S8). In the stromal cell clusters from control and KO mice Tn5 insertion was observed but it was not statistically different between the genotypes (Fig. 6B). Similar results are presented from the snRNA-seq, where only 1%–2% of PTs from control mice expressed Abi3bp compared with 24%–33% of PTs from KO male mice (log2FC = −0.5, P = 3.40 × 10−64) and female mice (log2FC = −0.44, P = 2.64 × 10−10) (Fig. 6C and Supplemental Table S7). Whole kidney Abi3bp expression was twofold greater in the KO mice compared with control (Fig. 6D).

Figure 6.

Figure 6.

Open in a new tab

Examples of differentially accessible chromatin and differentially expressed genes from the snATAC-seq and snRNA-seq data sets with validation in a lager cohort of mice with quantitative real-time PCR. A–D: Abi3bp is predominantly expressed in the stromal cells but has increased Tn5 insertion and RNA expression in the kidney specific Hdac1/Hdac2 knockdown mice (KO). A: Tn5 chromatin insertion coverage plots. Nuclei from control mice are presented in pink and from the KO mice in blue. B: violin plots of quantification of the Tn5 integration along the Abi3bp gene for stromal cells and male (m) and female (f) proximal tubules (PTs). The PT cluster was identified in both sexes, and the PT5 cluster is predominantly from the KO mice. Dots represent individual nuclei. C: Abi3bp RNA expression from the snRNA-seq data set plotted as a violin plot with dots representing individual nuclei. D: validation with cDNA derived from whole kidney RNA from female (closed symbols) and male (open symbols) control and KO mice. n = 4 or 5 per sex per genotype, individual mice plotted. P values are reported from an unpaired, two-tailed, Student’s t-test. E–H: Acsm2 is one of the greatest differentially accessible and expressed genes in the control mice compared with KO. Coverage plot (E) and Tn5 insertion (F). G: RNA expression. H: validation with qPCR.

The Ascm2 gene had the greatest log2FC DEG between control and KO for all PT clusters, regardless of sex (mPT = 3.46, P = 0; fPT = 1.50, P = 0; PT = 3.13, P = 1.56 × 10−178; PT5 = 2.4, P = 2.46 × 10−7) and was expressed in 79%–99% of PTs from control mice but only 20%–40% of PTs from KO mice. The Tn5 insertion was significantly greater in the PTs from control male mice (mPT = 1.26, P = 2.60 × 10−14) (Fig. 6, E and F and Supplemental Table S8) but was not significantly different for the female mice or other PT clusters (PT and PT5) (Fig. 6F and Supplemental Table S7) even though RNA expression was significantly greater in control mice than KO mice (Fig. 6G).

DISCUSSION

The central dogma is that the HDAC deacetylation of histone lysine tails functions to silence gene transcription via epigenetic modification of chromatin structure. But HDACs can also modify gene and protein transcription via protein-protein interactions with transcription factors and deacetylation of lysines of nonhistone proteins (10). Thus, to estimate the effect of HDAC1 and HDAC2 on regulating chromatin structure and the transcriptome, we profiled kidney nuclei from control and KO mice.

Knockdown of kidney epithelial Hdac1/Hdac2 in adulthood did not significantly reduce the ability to identify cell specific markers in the kidney. In other words, unsupervised clustering led to the identification of all the expected epithelial, endothelial, stromal, and immune cells in the kidney as reported previously (2, 3, 5, 6, 18) regardless of Hdac1/Hdac2 expression. However, cluster PT5 was highly enriched in the KO mice regardless of the sex of the mouse. This cluster has very high expression of DNA topoisomerase II and is enriched in biological pathways such as the cell cycle and chromosomal segregation (15). Interestingly, at the chromatin level, only inactive phospholipase c-like protein 2 (Plcl2) was significantly different as compared to all other clusters. PLCL2 functions as a novel inositol 1,4,5-trisphophate-binding protein but it lacks phospholipase activity (24). Although it is expressed in the kidney, there is very little information about this gene except that Plcl2 mRNA is downregulated in clear cell renal cell carcinoma and papillary renal cell carcinoma (25). Although the chromatin structure of this gene was “open” compared with other clusters, transcription of Plcl2 was not detected in our snRNA-seq data set. Thus, this cluster of nuclei expressed predominantly in the KO mice had significantly different transcriptional changes but the chromatin level was unremarkable.

Previously, Ransick et al. (18) determined at the single cell, kidney transcriptomic level, sex-specific diversity in predominantly the proximal tubules of mice. Likewise, studies with human kidney samples also have provided evidence that human proximal tubules have sex-specific transcriptomes (26). The snRNA-seq and snATAC-seq of mouse kidneys used in the current study also show sex-specific transcriptomes and chromatin structure changes. Recent physiological studies report that in males, there is increased reabsorptive capacity of the proximal tubules compared with females, which rely more on the distal nephrons for reabsorption of salts and water (27–29). Thus, even though urine composition between males and females is similar, the transport mechanisms along the nephron are sex-specific and this can be detected down to the chromatin structure level.

Although our snRNA-seq and snATAC-seq data sets captured the same types of nuclei when we compared within a cluster between the control and KO mice some interesting patterns emerged. Knockdown of kidney epithelial Hdac1/Hdac2 in adulthood results in kidney injury, fluid-electrolyte disorders, and death within 1 mo (15). Similar findings were found in developing mice where knockdown of Hdac1/Hdac2 in either the ureteric bud (14), which will form the collecting duct system, or in nephron progenitor cells (30), causes abnormal kidney development and death within a month of postnatal life. Thus, kidney HDAC1/HDAC2 are essential for kidney development, maintenance of a healthy kidney in adulthood, and survival. But given this extreme phenotype, it may be somewhat surprising that the kidney transcriptome and kidney chromatin landscape had fewer significant changes than would be predicted. For example, in the CDPC there were less than 1% of transcripts and chromatin accessibility changes between the control and KO mice. This is similar to microarray data from the E13.5 kidneys of control and ureteric bud Hdac1/Hdac2 knockdown mice were ∼1.2% of transcripts were differentially expressed (14). We interpret this to mean that HDAC1/HDAC2 regulate select genes that are necessary for kidney health and that they do not function in a genome-wide way to regulate all transcriptional processing. Part of the extreme phenotype may also be from loss of HDAC1/HDAC2 deacetylation of nonhistone proteins but this will require further investigation.

As seen in the histological images, Hdac1/Hdac2 knockdown mice present with proximal tubule damage such as a reduction in the brush border. Moreover, the male mice develop interstitial fibrosis within 2 wk of Hdac1/Hdac2 knockdown. In agreement with this, GO analyses of genes significantly upregulated in the KO were enriched in the TGFβ receptor signaling pathway. Following kidney injury, TGFβ is strongly induced by damaged proximal tubules and if sustained may mediate the progression to chronic kidney disease (31). One of the most highly induced genes in the proximal tubules of KO mice was Abi3bp. Abi3bp encodes a newly identified extracellular matrix protein that may also function to promote cellular senescence (23). Single nucleotide polymorphisms in ABI3BP are associated with neuropsychological distress including stress-related hypertension (32, 33) and early-onset preeclampsia (34). Thus to date, there is an association between ABI3BP and cardiovascular disease so it is intriguing to find this gene upregulated in a kidney that is damaged. Only 1%–2% of proximal tubule cells from control mice expressed this gene, which had minimal Tn5 insertion (e.g., closed chromatin), but Hdac1/Hdac2 knockdown resulted in up to 40% of proximal tubules expressing this gene, regardless of sex. This suggests that HDAC1/HDAC2 may normally suppress Abi3bp at least in part by keeping the Abi3bp promoter inaccessible. Future studies are needed to determine if Abi3bp is causing kidney damage or is expressed in response to kidney damage and whether it represents a novel extracellular matrix protein leading to the development of interstitial fibrosis.

One of the most highly expressed genes in the proximal tubules of the control mice was Acsm2. ACSM2 is a member of a large family of enzymes that generate acyl-CoA from free fatty acids although the specific function of ACSM2 is unclear (35). GO analyses of genes upregulated in control mice were enriched in metabolic pathways compared with KO mice. ACSM2 is highly expressed in the proximal tubules of the adult kidney (36) with very low expression during development (35). Moreover, Acsm2 is downregulated in mouse models of kidney disease, and protein abundance of ACSM2A (the human ortholog) is less in diabetic nephropathy patients (35). Our KO mice, which have kidney disease, also have reduced Ascm2 expression in the proximal tubule consistent with these publishes studies. Interestingly, study by Watanabe et al. (35) determined that the Acsm2 gene is likely under epigenetic regulation as in the adult mouse kidney the Acsm2 gene locus has significant peaks for H3K27 acetylation, H3K4me3, and RNA polymerase II that are all known markers of active transcription (37). Although there is evidence that HDAC1 may deacetylate H3K27 (38), we would predict that our Hdac1/Hdac2 KO mice would have increased in H3K27 acetylation and greater Acsm2 expression. Taken together, our study demonstrates that HDAC1/HDAC2 may promote proximal tubule Acsm2 expression that is independent of histone lysine acetylation but still involves changes to chromatin structure at the Acsm2 locus.

Fluid-electrolyte disorders are frequent, sometime serious, adverse events experienced by patients on HDAC inhibitors to treat cancers and neurological disorders (15, 39–41). The kidney specific Hdac1/Hdac2 KO mouse has significant fluid-electrolyte disorders, namely, hypernatremia and hyperchloremia (15). Greater plasma sodium and chloride are predominantly from a loss of water from the body (42, 43). Normally, plasma osmolality is tightly regulated via thirst sensors in the body that when there is an increase in osmolality the pituitary releases arginine vasopressin that acts in the CDPC to increase AQP2 expression in the apical membrane to drive water reabsorption and concentrate the urine (43). Hdac1/Hdac2 knockdown resulted in a closing of chromatin structure along the Aqp2 gene, and this correlated with a significant decreased in Aqp2 transcript abundance. Likewise, Aqp1 is expressed in the proximal tubule and thin limbs of Henle and contributes to a bulk ∼60% reabsorption of fluid from the ultrafiltrate was also closed and had significantly less transcript abundance. These chromatin and transcriptional changes are consistent with Hdac1/Hdac2 knockdown mice having hypernatremia and hyperchloremia. However, not all aquaporins were affected. Aqp4 is constitutively expressed in the CDPC and its chromatin structure was not significantly different between the mice and transcript abundance was also similar. Thus, Hdac1/Hdac2 expression does not affect all aquaporins in the kidney, again highlighting that these enzymes do not globally affect transcription.

There are some limitations to this study. First, the single nucleus analyses were completed on a single mouse for each group and thus were not biologically replicated at the level of the animal, but rather were replicated at the level of the nucleus. However, the mice used in the seq experiments matched the phenotype of the larger cohort (previously published in Ref. 15) used for transcript validation. Second, we only determined the snRNA-seq and snATAC-seq at one time point, between 7:00 AM and 8:00 AM when the mice are starting to go to sleep. Circadian rhythmicity in transcript abundance especially in the kidney is well appreciated (44), and there is emerging evidence of circadian variation in chromatin structure (45). Thus, the relationship between chromatin structure and transcript abundance over a 24-h period requires further investigation and may explain why there are some genes that appear closed at the chromatin level that are highly abundant at the transcript level (i.e., time lag between these processes).

In conclusion, during development and in the adult kidney, appropriate HDAC1/HDAC2 activity is necessary for kidney health. These HDACs function in the regulation of gene expression in part through modifying chromatin structure. However, changes in HDAC1/HDAC2 expression does not result in a complete loss of transcriptional regulation but rather results in changing chromatin structure at specific gene loci and/or alter transcript expression of specific genes. These HDACs appear to be suppressors of certain genes (e.g., Ab13bp) but activators of other genes (e.g., Aqp2, Ascm2); thus, identifying these gene targets will be important for determining the benefit of HDAC inhibitor therapies to treat the spectrum of diseases from cancer (46) to kidney (11) and cardiovascular diseases (47).

DATA AVAILABILITY

The data that support this study are available as follows: snRNA-seq libraries, National Center for Biotechnology Information Gene Expression Ombnibus (GEO) under Accession no. GSE148354; fastQ files of the snATAC-seq libraries at GEO under Accession no. GSE181557. All code used in the study analyses is deposited at https://github.com/hyndmank/snATAC_Hyndman.

SUPPLEMENTAL DATA

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.15141645.v1.

Supplemental Table S2: https://doi.org/10.6084/m9.figshare.15141660.v1.

Supplemental Table S3: https://doi.org/10.6084/m9.figshare.15141663.v1.

Supplemental Table S4: https://doi.org/10.6084/m9.figshare.15141669.v1.

Supplemental Table S5: https://doi.org/10.6084/m9.figshare.15141672.v1.

Supplemental Table S6: https://doi.org/10.6084/m9.figshare.16811488.v1.

Supplemental Table S7: https://doi.org/10.6084/m9.figshare.15141678.v1.

Supplemental Table S8: https://doi.org/10.6084/m9.figshare.15141687.v1.

Supplemental Figs. S1–S3: https://doi.org/10.6084/m9.figshare.15173679.v1.

GRANTS

Research reported in this publication was supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Numbers R01DK126664 and R01DK128001 (to K.A.H.). The BERD is funded by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR003096, and the CFCC is funded by NIH P30 AR048311 and NIH P30 AI27667.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

K.A.H. conceived and designed research; K.A.H. performed experiments; K.A.H. and D.K.C. analyzed data; K.A.H. and D.K.C. interpreted results of experiments; K.A.H. prepared figures; K.A.H. drafted manuscript; K.A.H. and D.K.C. edited and revised manuscript; K.A.H. and D.K.C. approved final version of manuscript.

ACKNOWLEDGMENTS

The authors thank the technical help of Luciano Mendoza, Jack Colson, and McKenzi King with mouse husbandry and genotyping, Dr. Michael Crowley Director of the UAB Genomics Core Facility for sequencing services, and Dr. Shanrun Liu, the Single Cell Manager in the UAB Comprehensive Flow Cytometry Core (CFCC). A statistical consult was provided by Dr. David Redden as part of the UAB Center for Clinical and Translational Science Biostatistics, Epidemiology, and Research Design (BERD). Graphical abstract image created with BioRender and published with permission.

REFERENCES

  • 1.Park J, Liu CL, Kim J, Susztak K. Understanding the kidney one cell at a time. Kidney Int 96: 862–870, 2019. doi: 10.1016/j.kint.2019.03.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Muto Y, Wilson PC, Ledru N, Wu H, Dimke H, Waikar SS, Humphreys BD. Single cell transcriptional and chromatin accessibility profiling redefine cellular heterogeneity in the adult human kidney. Nat Commun 12: 2190, 2021. doi: 10.1038/s41467-021-22368-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Lake BB, Chen S, Hoshi M, Plongthongkum N, Salamon D, Knoten A, Vijayan A, Venkatesh R, Kim EH, Gao D, Gaut J, Zhang K, Jain S. A single-nucleus RNA-sequencing pipeline to decipher the molecular anatomy and pathophysiology of human kidneys. Nat Commun 10: 2832, 2019. doi: 10.1038/s41467-019-10861-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Malone AF, Humphreys BD. Single-cell transcriptomics and solid organ transplantation. Transplantation 103: 1776–1782, 2019. doi: 10.1097/TP.0000000000002725. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Park J, Shrestha R, Qiu C, Kondo A, Huang S, Werth M, Li M, Barasch J, Suszták K. Single-cell transcriptomics of the mouse kidney reveals potential cellular targets of kidney disease. Science 360: 758–763, 2018. doi: 10.1126/science.aar2131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Wilson PC, Wu H, Kirita Y, Uchimura K, Ledru N, Rennke HG, Welling PA, Waikar SS, Humphreys BD. The single-cell transcriptomic landscape of early human diabetic nephropathy. Proc Natl Acad Sci USA 116: 19619–19625, 2019. doi: 10.1073/pnas.1908706116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Cramer P. Organization and regulation of gene transcription. Nature 573: 45–54, 2019. doi: 10.1038/s41586-019-1517-4. [DOI] [PubMed] [Google Scholar]
  • 8.O’Brien J, Hayder H, Zayed Y, Peng C. Overview of MicroRNA biogenesis, mechanisms of actions, and circulation. Front Endocrinol (Lausanne) 9: 402, 2018. doi: 10.3389/fendo.2018.00402. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Gil N, Ulitsky I. Regulation of gene expression by cis-acting long non-coding RNAs. Nat Rev Genet 21: 102–117, 2020. doi: 10.1038/s41576-019-0184-5. [DOI] [PubMed] [Google Scholar]
  • 10.Drazic A, Myklebust LM, Ree R, Arnesen T. The world of protein acetylation. Biochim Biophys Acta 1864: 1372–1401, 2016. doi: 10.1016/j.bbapap.2016.06.007. [DOI] [PubMed] [Google Scholar]
  • 11.Hyndman KA. Histone deacetylases in kidney physiology and acute kidney injury. Semin Nephrol 40: 138–147, 2020. doi: 10.1016/j.semnephrol.2020.01.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Hyndman KA, Knepper MA. Dynamic regulation of lysine acetylation: the balance between acetyltransferase and deacetylase activities. Am J Physiol Renal Physiol 313: F842–F846, 2017. doi: 10.1152/ajprenal.00313.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hyndman KA, Kasztan M, Mendoza LD, Monteiro-Pai S. Dynamic changes in histone deacetylases following kidney ischemia-reperfusion injury are critical for promoting proximal tubule proliferation. Am J Physiol Renal Physiol 316: F875–F888, 2019. doi: 10.1152/ajprenal.00499.2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Chen S, Yao X, Li Y, Saifudeen Z, Bachvarov D, El-Dahr SS. Histone deacetylase 1 and 2 regulate Wnt and p53 pathways in the ureteric bud epithelium. Development 142: 1180–1192, 2015. doi: 10.1242/dev.113506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Hyndman KA, Speed JS, Mendoza LD, Allan JM, Colson J, Sedaka R, Jin C, Jung HJ, El-Dahr S, Pollock DM, Pollock JS. Fluid-electrolyte homeostasis requires histone deacetylase function. JCI Insight 5: e137792, 2020. doi: 10.1172/jci.insight.137792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Buenrostro JD, Giresi PG, Zaba LC, Chang HY, Greenleaf WJ. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat Methods 10: 1213–1218, 2013. doi: 10.1038/nmeth.2688. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Korsunsky I, Millard N, Fan J, Slowikowski K, Zhang F, Wei K, Baglaenko Y, Brenner M, Loh PR, Raychaudhuri S. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat Methods 16: 1289–1296, 2019. doi: 10.1038/s41592-019-0619-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Ransick A, Lindström NO, Liu J, Zhu Q, Guo JJ, Alvarado GF, Kim AD, Black HG, Kim J, McMahon AP. Single-cell profiling reveals sex, lineage, and regional diversity in the mouse kidney. Dev Cell 51: 399–413.e7, 2019. doi: 10.1016/j.devcel.2019.10.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Hao Y, Hao S, Andersen-Nissen E, Mauck WM 3rd, Zheng S, Butler A, Lee MJ, Wilk AJ, Darby C, Zager M, Hoffman P, Stoeckius M, Papalexi E, Mimitou EP, Jain J, Srivastava A, Stuart T, Fleming LM, Yeung B, Rogers AJ, McElrath JM, Blish CA, Gottardo R, Smibert P, Satija R. Integrated analysis of multimodal single-cell data. Cell 184: 3573–3587.e29, 2021. doi: 10.1016/j.cell.2021.04.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4: 44–57, 2009. doi: 10.1038/nprot.2008.211. [DOI] [PubMed] [Google Scholar]
  • 21.Mendoza LD, Hyndman KA. The contribution of collecting duct NOS1 to the concentrating mechanisms in male and female mice. Am J Physiol Renal Physiol 317: F547–F559, 2019. doi: 10.1152/ajprenal.00180.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Wang X, Spandidos A, Wang H, Seed B. PrimerBank: a PCR primer database for quantitative gene expression analysis, 2012 update. Nucleic Acids Res 40: D1144–D1149, 2012. doi: 10.1093/nar/gkr1013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Delfín DA, DeAguero JL, McKown EN. The extracellular matrix protein ABI3BP in cardiovascular health and disease. Front Cardiovasc Med 6: 23, 2019. doi: 10.3389/fcvm.2019.00023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Murakami A, Matsuda M, Harada Y, Hirata M. Phospholipase C-related, but catalytically inactive protein (PRIP) up-regulates osteoclast differentiation via calcium-calcineurin-NFATc1 signaling. J Biol Chem 292: 7994–8006, 2017. doi: 10.1074/jbc.M117.784777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Dmitriev AA, Rudenko EE, Kudryavtseva AV, Krasnov GS, Gordiyuk VV, Melnikova NV, Stakhovsky EO, Kononenko OA, Pavlova LS, Kondratieva TT, Alekseev BY, Braga EA, Senchenko VN, Kashuba VI. Epigenetic alterations of chromosome 3 revealed by NotI-microarrays in clear cell renal cell carcinoma. Biomed Res Int 2014: 735292, 2014. doi: 10.1155/2014/735292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Huang L, Liao J, He J, Pan S, Zhang H, Yang X, Cheng J, Chen Y, Mo Z. Single-cell profiling reveals sex diversity in human renal proximal tubules. Gene 752: 144790, 2020. doi: 10.1016/j.gene.2020.144790. [DOI] [PubMed] [Google Scholar]
  • 27.Hu R, McDonough AA, Layton AT. Sex differences in solute and water handling in the human kidney: modeling and functional implications. iScience 24: 102667, 2021. doi: 10.1016/j.isci.2021.102667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Hu R, McDonough AA, Layton AT. Sex differences in solute transport along the nephrons: effects of Na+ transport inhibition. Am J Physiol Renal Physiol 319: F487–F505, 2020. doi: 10.1152/ajprenal.00240.2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Veiras LC, Girardi ACC, Curry J, Pei L, Ralph DL, Tran A, Castelo-Branco RC, Pastor-Soler N, Arranz CT, Yu ASL, McDonough AA. Sexual dimorphic pattern of renal transporters and electrolyte homeostasis. J Am Soc Nephrol 28: 3504–3517, 2017. doi: 10.1681/ASN.2017030295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Liu H, Chen S, Yao X, Li Y, Chen CH, Liu J, Saifudeen Z, El-Dahr SS. Histone deacetylases 1 and 2 regulate the transcriptional programs of nephron progenitors and renal vesicles. Development 145: dev153619, 2018. doi: 10.1242/dev.153619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Gewin LS. Transforming growth factor-β in the acute kidney injury to chronic kidney disease transition. Nephron 143: 154–157, 2019. doi: 10.1159/000500093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Kimbrel NA, Garrett ME, Dennis MF; VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center Workgroup, Hauser MA, Ashley-Koch AE, Beckham JC. A genome-wide association study of suicide attempts and suicidal ideation in U.S. military veterans. Psychiatry Res 269: 64–69, 2018. doi: 10.1016/j.psychres.2018.07.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Perlis RH, Huang J, Purcell S, Fava M, Rush AJ, Sullivan PF, Hamilton SP, McMahon FJ, Schulze TG, Potash JB, Zandi PP, Willour VL, Penninx BW, Boomsma DI, Vogelzangs N, Middeldorp CM, Rietschel M, Nöthen M, Cichon S, Gurling H, Bass N, McQuillin A, Hamshere M; Wellcome Trust Case Control Consortium Bipolar Disorder Group, Craddock N, Sklar P, Smoller JW. Genome-wide association study of suicide attempts in mood disorder patients. Am J Psychiatry 167: 1499–1507, 2010. [Erratum in Am J Psychiatry 168: 100, 2011]. doi: 10.1176/appi.ajp.2010.10040541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Nevalainen J, Skarp S, Savolainen ER, Ryynänen M, Järvenpää J. Intrauterine growth restriction and placental gene expression in severe preeclampsia, comparing early-onset and late-onset forms. J Perinat Med 45: 869–877, 2017. doi: 10.1515/jpm-2016-0406. [DOI] [PubMed] [Google Scholar]
  • 35.Watanabe H, Paxton RL, Tolerico MR, Nagalakshmi VK, Tanaka S, Okusa MD, Goto S, Narita I, Watanabe S, Sequeira-Lopez MLS, Gomez RA. Expression of Acsm2, a kidney-specific gene, parallels the function and maturation of proximal tubular cells. Am J Physiol Renal Physiol 319: F603–F611, 2020. doi: 10.1152/ajprenal.00348.2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Hilgers KF, Nagaraj SK, Karginova EA, Kazakova IG, Chevalier RL, Carey RM, Pentz ES, Gomez RA. Molecular cloning of KS, a novel rat gene expressed exclusively in the kidney. Kidney Int 54: 1444–1454, 1998. doi: 10.1046/j.1523-1755.1998.00143.x. [DOI] [PubMed] [Google Scholar]
  • 37.Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifications. Cell Res 21: 381–395, 2011. doi: 10.1038/cr.2011.22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Li F, Wu R, Cui X, Zha L, Yu L, Shi H, Xue B. Histone deacetylase 1 (HDAC1) negatively regulates thermogenic program in brown adipocytes via coordinated regulation of histone H3 lysine 27 (H3K27) deacetylation and methylation. J Biol Chem 291: 4523–4536, 2016. doi: 10.1074/jbc.M115.677930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Pili R, Salumbides B, Zhao M, Altiok S, Qian D, Zwiebel J, Carducci MA, Rudek MA. Phase I study of the histone deacetylase inhibitor entinostat in combination with 13-cis retinoic acid in patients with solid tumours. Br J Cancer 106: 77–84, 2012. doi: 10.1038/bjc.2011.527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Ree AH, Dueland S, Folkvord S, Hole KH, Seierstad T, Johansen M, Abrahamsen TW, Flatmark K. Vorinostat, a histone deacetylase inhibitor, combined with pelvic palliative radiotherapy for gastrointestinal carcinoma: the Pelvic Radiation and Vorinostat (PRAVO) phase 1 study. Lancet Oncol 11: 459–464, 2010. doi: 10.1016/S1470-2045(10)70058-9. [DOI] [PubMed] [Google Scholar]
  • 41.Ramalingam SS, Maitland ML, Frankel P, Argiris AE, Koczywas M, Gitlitz B, Thomas S, Espinoza-Delgado I, Vokes EE, Gandara DR, Belani CP. Carboplatin and paclitaxel in combination with either vorinostat or placebo for first-line therapy of advanced non-small-cell lung cancer. J Clin Oncol 28: 56–62, 2010. doi: 10.1200/JCO.2009.24.9094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Sonani B, Naganathan S, Al-Dhahir MA. Hypernatremia. In: StatPearls. Treasure Island (FL): StatPearls Publishing, 2021. [PubMed] [Google Scholar]
  • 43.Seay NW, Lehrich RW, Greenberg A. Diagnosis and management of disorders of body tonicity-hyponatremia and hypernatremia: core curriculum 2020. Am J Kidney Dis 75: 272–286, 2020. doi: 10.1053/j.ajkd.2019.07.014. [DOI] [PubMed] [Google Scholar]
  • 44.Firsov D, Bonny O. Circadian rhythms and the kidney. Nat Rev Nephrol 14: 626–635, 2018. doi: 10.1038/s41581-018-0048-9. [DOI] [PubMed] [Google Scholar]
  • 45.Weizman EN, Tannenbaum M, Tarrant AM, Hakim O, Levy O. Chromatin dynamics enable transcriptional rhythms in the cnidarian Nematostella vectensis. PLoS Genet 15: e1008397, 2019. doi: 10.1371/journal.pgen.1008397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Bondarev AD, Attwood MM, Jonsson J, Chubarev VN, Tarasov VV, Schiöth HB. Recent developments of HDAC inhibitors: emerging indications and novel molecules. Br J Clin Pharmacol 87: 4577–4597, 2021. doi: 10.1111/bcp.14889. [DOI] [PubMed] [Google Scholar]
  • 47.Li P, Ge J, Li H. Lysine acetyltransferases and lysine deacetylases as targets for cardiovascular disease. Nat Rev Cardiol 17: 96–115, 2020. doi: 10.1038/s41569-019-0235-9. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.15141645.v1.

Supplemental Table S2: https://doi.org/10.6084/m9.figshare.15141660.v1.

Supplemental Table S3: https://doi.org/10.6084/m9.figshare.15141663.v1.

Supplemental Table S4: https://doi.org/10.6084/m9.figshare.15141669.v1.

Supplemental Table S5: https://doi.org/10.6084/m9.figshare.15141672.v1.

Supplemental Table S6: https://doi.org/10.6084/m9.figshare.16811488.v1.

Supplemental Table S7: https://doi.org/10.6084/m9.figshare.15141678.v1.

Supplemental Table S8: https://doi.org/10.6084/m9.figshare.15141687.v1.

Supplemental Figs. S1–S3: https://doi.org/10.6084/m9.figshare.15173679.v1.

Data Availability Statement

The data that support this study are available as follows: snRNA-seq libraries, National Center for Biotechnology Information Gene Expression Ombnibus (GEO) under Accession no. GSE148354; fastQ files of the snATAC-seq libraries at GEO under Accession no. GSE181557. All code used in the study analyses is deposited at https://github.com/hyndmank/snATAC_Hyndman.

Articles from Physiological Genomics are provided here courtesy of American Physiological Society

RESOURCES