Chapter2: Renal Tubular Peroxisomes Are Dispensable For Normal Kidney Function

For more info. contact tina.xiang@wecistanche.com

Discussion

A high abundance of peroxisomes in the proximal tubule together with the presence of renal abnormalities in patients with ZSD have strongly suggested an important role of peroxisomes in the kidney. Moreover, kidney-specific expression of several peroxisomal enzymes has supported the idea that renal peroxisomes may have biological functions partially distinct from those in other tissues(15). To test these hypotheses, we examined renal function in mice devoid of peroxisomes specifically in the renal tubule. Among different mouse models employed to study the role of peroxisomes, we selected mice allowing conditional genetic inactivation of Pex5 encoding peroxisome biogenesis factor PEX5 essential for peroxisome formation. This model was chosen because (i)Pex5-null mice exhibit multiple biochemical and functional abnormalities reminiscent of those observed in patients with ZSD(11) and (i) a large spectrum of mouse models with tissue-specific ablation of Pex5 has been analyzed(16), thereby allowing us to compare the functional consequences of a peroxisomal deficiency in the renal tubule with those observed in other tissues. As demonstrated by Baes et al., Pex5-null mice exhibited intrauterine growth retardation, malformation of the brain, and severe hypotonia and died within the first 72 hours of life (11). Biochemical analyses revealed several hallmarks of ZSD, including depletion of plasmalogens in the liver and brain and a strong increase in plasma levels of VLCFA. No cortical renal cysts were found in the kidney of Pex5-null mice at P0.5, but retardation in the intrauterine maturation of glomeruli was observed. The presence of calcium oxalate deposits was not investigated.

Click here to get to know cistanche herb benefits

In cKOm mice, ablation of peroxisomal biogenesis in the renal tubule did not cause any major pheno-type, except decreased KW and BW in infant cKOm mice and decreased KW/BW ratio in adult cKOm mice. In cKOf mice, incomplete excision of floxed Pex5 allele together with the much milder effect of Pex5 deletion on renal transcriptome and metabolome suggested that some residual peroxisomes may remain in the renal tubule. However, sex specificity of peroxisomal functions in the kidney cannot be ruled out as well. We did not find any significant morphological difference in the kidneys of cKOm mice that could explain the lower KW or KW/BW ratio. However, the tendency for reduced proximal cell width might be a possible cause of this difference. Importantly, the renal homeostatic function was maintained in cKOm mice despite marked changes in expression levels of transcripts encoding proteins involved in transepithelial transport of ions, water, amino acids, organic anions and cations, phosphate, urate, and in the megalin-mediated multili-gand endocytic pathway. Of note, a majority of downregulated transcripts encoded transporters expressed in the proximal tubule whereas most of the upregulated transcripts encoded transporters expressed in more distal parts of the nephron. For instance, increased expression of Nkcc2, Clc-Kb, Ncc, and βENaC transcripts suggested a compensatory upregulation in post-proximal sodium reabsorption. The integrated analysis of renal transcriptome and metabolome revealed profound changes in a variety of cellular pathways related to cellular metabolism, lipid membrane composition, and redox homeostasis. In kidneys of cKOm mice, about 8% of transcriptome and about 24% of detected metabolites exhibited different levels as compared with Ctrl mice. As expected, many of these transcripts and metabolites were related to peroxisomal function. A dramatic reduction was observed in the abundance of ether phospholipid plasmalogens that are synthesized in peroxisomes (~67% in cKOm mice and~50% in cKOf mice, unweighted sum of all detect-ed plasmalogen species). Plasmalogens are major cell membrane phospholipids with a growing number of attributed functions, including membrane integrity, cellular signaling, and antioxidant defense. In the kidney, plasmalogens represent approximately 20% of total phospholipids(17), thus suggesting that at least 10%of the phospholipid composition differed in kidneys of cKOm and cKOf mice from respective controls. The depletion of plasmalogens was potentially compensated by an increased abundance of structurally similar phosphatidylethanolamines (PEs) and by phosphatidylcholines(PCs), which exhibited a substantial enrichment in kidneys of both cKOm and cKOf mice. Of note, a similar compensatory increase of PEs and PCs was observed in the kidneys of patients with ZSD (17).

The metabolism of the proximal tubule mainly relies on fatty acid oxidation, which is characterized by a high generation of ROS produced both in mitochondria and peroxisomes. In cKO mice, pathways related to fatty acids degradation/biosynthesis were substantially upregulated, potentially suggesting compensatory enhancement of mitochondrial energy production from these substrates. Since one of the main functions of peroxisomes is detoxification of ROS, we hypothesized that ablation of peroxisome biogenesis would induce oxidative stress in the kidneys of cKO mice. However, molecular analyses did not reveal changes in the antioxidant capacity and biomarkers of oxidative stress. Challenge with HFD did not induce alterations in functional phenotype, except for a mild increase in urinary volume and urinary excretion of calcium and urate in cKOm mice. Integrated transcriptome and metabolome analysis allowed the identification of the glutathione pathway as a possible compensatory mechanism for the maintenance of the overall antioxidant capacity in renal proximal tubule cells lacking peroxisomes.

Conditional deletion of Pex5 in the fetal liver(12), in pancreatic β cells(18), and in different cell types of the central nervous system (reviewed in ref.16) causes variable but mostly serious organ-specific functional deteriorations. This was not the case in the kidney. In this study, we identified several intrinsic renal mechanisms that potentially allow coping with the lack of peroxisomes. As the kidney is the organ that exhibits a high blood perfusion rate, another possibility lies in the extrarenal origin of at least a part of metabolites that the kidney does not produce in the absence of peroxisomes but that are required for normal renal function. Collectively, our data suggest that renal tubular peroxisomes are dispensable for normal renal function and that pathophysiological changes in kidneys of patients with ZSD are of extrarenal origin.

Methods

Animals. The procedures used to generate the Pex5a/la/Pax8-ntTA/LC-1 Cre mice were described previously (13). The 3 mouse lines used in this study are inbred strains, bred on the genetic background of the C57BL/6J mouse (The Jackson Laboratory). Before all experiments, mice were adapted to a 12-hour light/12-hour dark cycle. All experiments were performed at circadian time ZT3(ZT0is the time when the light is switched on, and ZT12 is the time when the light is switched off).

Pex5bxlo/Pax8-rtTA/LC1 triple transgenic male or female mice (referred to as cKOm and cKOf) and Pex5l/a mice (referred to as Ctrl or Ctrl) were used. Pex5 recombination in newborns was performed by addition of 0.2% DOX and 2% sucrose in the drinking water of pregnant female mice for 2 weeks, starting from their 14th gestation day. For Pex5 recombination in adult mice, DOX and sucrose were added to the drinking water of 8-week-old mice. The renal structure and function were evaluated 4 weeks after the end of the DOX treatment in4-week-old infant mice or 14-week-old adult mice. After their weaning at 3 weeks, mice were fed with a standard control diet containing 4.2% of fat(Sniff low-fat control diet) or when mentioned, for 4 weeks with a matching HFD containing 20.7% of fat (mostly LCFAs).

Metabolic cages and blood and urine chemistry. Mice were housed in individual metabolic cages(Tecniplast). Urine collection was performed after a 3-day adaptation period. Urine and blood chemistry were analyzed as previously described (19).Plasma and urine composition was measured in the Laboratoire Central de Chimie Clinique, Centre Hospitalier Universitaire Vaudoise University Hospital (Lausanne, Switzerland).

Electron microscopy. Kidney samples were cut into pieces around 1 mm³and fixed in glutaraldehyde solution (EMS)2.5% in phosphate buffer(PB,0.1 M pH7.4)(MilliporeSigma) for 2 hours at room temperature(RT). Then, they were rinsed 3 times, 5 minutes each, in PB buffer and then postfixed by a fresh mixture of osmium tetroxide 1%(EMS)with 1.5% of potassium ferrocyanide (MilliporeSigma)in PB for 2 hours at RT. The samples were then washed 3 times in distilled water and dehydrated in acetone solution (MilliporeSigma)at graded concentrations(30% 90 minutes;70%90 minutes;100%120 minutes;100%240 minutes). This was followed by infiltration in Epon resin(MilliporeSigma) at graded concentrations(Epon/acetone [1/3; v/v] 4 hours;Epon/acetone [3/1;v/v] 4 hours, Epon/acetone [1/1;v/v] 8 hours; Epon/acetone [1/1; v/v] 24 hours)and finally polymerization for 48 hours at 60°C in an oven. Ultrathin sections of 50 nm were cut on a Leica Ultracut (Leica Mikrosysteme GmbH) and picked up on a copper slot grid,2×1 mm(EMS), coated with a polystyrene film (MilliporeSigma). Sections were poststained with uranyl acetate (MilliporeSigma)2% in H, O for 10 minutes, rinsed several times with H, O followed by Reynolds lead citrate for 10 minutes, and rinsed several times with H, O.

Stereology. For stereology analysis,3 kidney cortex pieces per mouse were analyzed, 15 micrographs per sample with a pixel size of 4.08 nm, using a systematic uniform random sampling with a transmission electron microscope (Philips CM100, Thermo Fisher Scientific) at an acceleration voltage of 80 kV with a TVIPS TemCam-F416 digital camera (TVIPS GmbH).

Transcriptomic analysis. All sequencing data are publicly available through NIH Gene Expression Omnibus (accession number GSE179202). RNA-Seq libraries were prepared using 200 ng of total kidney RNA as described in Nikolaeva et al. (19). Illumina TruSeq SR Cluster Kit v4 reagents were used. Sequencing data were processed using Mus musculus. GRCm38.86 gene annotation, Statistical analysis was performed in R (version 3.4.0). Transcripts with low counts were filtered out according to the rule of 1 count per million(CPM) in at least 1 sample. Library sizes were scaled using TMM normalization and log-transformed into CPM using voom (20). Principal component analysis showed a small variability between replicate samples. Differential expression between cKO and Ctrl mice was computed into a 1 F test using limma (21). P values were adjusted as described (22) into FDR values to consider for multiple comparisons, using results of all transcripts in males and females or results of selected transporters or peroxisomal related transcripts in males only. Transcripts with an FDR<5% were considered significant. Comparisons of expression levels were done using the mean FC of CPM in cKO mice as compared with Ctrl mice. For a graphical representation of individual values in box-and-whisker plots, individual values were normalized between 0 and 1. A weighted GSEA was performed using the unfiltered transcript list ranked by signed P-value (product of P-value and sign of the FC). The analyses were performed using the gseKEGG function from the R package cluster profile(version 3.16.1)(23). KEGG pathway collections were restricted to gene sets with minimum and maximum sizes of 10 and 500, respectively. The enrichment scores were normalized by gene set size, and their statistical significance was assessed by permutation tests (n = 1000).

Metabolomic analysis. Metabolites from frozen kidney samples were identified by ultrahigh-performance liquid chromatography-tandem mass spectroscopy and quantified using AUC (Metabolon Inc). Raw values were log-transformed and normalized in terms of raw area counts. Two-way ANOVA contrast tests were used to identify biochemicals that differed significantly between cKO and Ctrl mice of both sexes. P values were adjusted as described (22) into FDR values to consider for multiple comparisons, and metabolites with FDR<5% were considered significantly modulated. For Log2FC calculation, missing values, if any, were replaced with the minimum observed value in mice of the same sex, for each compound. For a graphical representation of individual values in box-and-whisker plots, normalized values were divided by the median value obtained in mice of the same sex, for each compound.

Joint transcriptome-metabolome analysis. An integrated metabolic pathway analysis combining results obtained from gene expression and metabolomics studies was conducted using the Joint Pathway Analysis module on the MetaboAnalyst 5.0 website(24). Compound name of metabolites and official gene symbol of transcripts significantly more or less abundant (FDR<0.05), along with their respective FCs in cKOm mice as compared with Ctrl mice, were inputs. The integration was performed using KEGG metabolic pathways with default settings(hypergeometric test for overrepresentation analysis, degree centrality for pathway topology analysis, and tight integration by combining queries). Pathways with FDR<0.1 were considered significantly affected in cKOm mice.

Stainings. Stainings of global kidney structure, calcium oxalate deposits, or neutral lipid deposition were performed on transverse kidney slices that were 5 μm thick. Except for neutral lipid staining， mice were anesthetized, and their left kidney was perfused through the abdominal aorta with a 4% paraformaldehyde solution before tissue harvesting. For analysis of global kidney structure, kidney slices embedded in paraffin were deparaffinized, rehydrated, stained with H&E, dehydrated, and mounted in ROTI Histokitt(Carl Roth GmbH). Neutral lipids were stained in kidney slices embedded in OCT compound (Tissue-Tek) using an Oil Red O solution. Calcium oxalate deposits were stained according to the method described by Pizzolato (25). Briefly, kidney slices embedded in paraffin were deparaffinized and rehydrated, then submerged in a 1:1 solution of silver nitrate 5% w/v and hydrogen peroxide 30%and placed under a 60 W light bulb for 30 minutes. Slides were thoroughly washed in distilled water and counterstained with Nuclear Fast Red (MilliporeSigma), then dehydrated before mounting in ROTI Histokitt. Longitudinal kidney slices from a mouse treated with a calcium oxalate nephropathy-induced-ing diet(1.5% calcium+1.5% hydroxyproline mixed in standard chow for 3 weeks)served as a positive control of calcium oxalate kidney deposits. Results were analyzed using a Zeiss AxioScan.Z1 Slide Scanner at 100× or 200× original magnification.

ELISA and Trolox assay.4-HNE were measured using Lipid Peroxidation(4-HNE)Assay Kit from Abcam (ab238538). Total and nonenzymatic antioxidant capacities were assessed by a Trolox assay kit from Abcam (ab65329).

Statistics. All data are expressed as mean ± SEM. Statistical tests and threshold for significance are described in figure legends or appropriate Methods sections. Statistical analysis was performed using R packages or GraphPad Prism software version 8.2.1. All t-tests were 2 tailed.

Study approval. All experiments with animals were performed by the Swiss guidelines for animal care, which conform to the National Institutes of Health animal care guidelines, and approved by the Swiss cantonal(Canton de Vaud) and federal veterinary authorities (authorization 31827 to DF)(Direction genérale de P'agriculture, de la viticulture et des affairs vetérinaires, Epalinges, Switzerland).

Prior publication: A part of this work has been submitted as an abstract to the 2021 Annual Meeting of the American Society of Nephrology (November 4, 2021.https://www.asn-online.org/education/kidney-week/2021/program-abstract.aspx?controlId=3605774).