Nephron-Specific Deletion of Circadian Clock Gene Bmal1 Alters

BASIC RESEARCH www.jasn.org

Nephron-Speciﬁc Deletion of Circadian Clock Gene Bmal1 Alters the Plasma and Renal Metabolome and Impairs Drug Disposition

† ‡ Svetlana Nikolaeva,* Camille Ansermet,* Gabriel Centeno,* Sylvain Pradervand, | †† Vincent Bize,* David Mordasini,*§ Hugues Henry,¶ Robert Koesters,** Marc Maillard, †† ‡‡ Olivier Bonny,* Natsuko Tokonami,* and Dmitri Firsov*

*Department of Pharmacology and Toxicology and ‡Genomic Technologies Facility, University of Lausanne, Lausanne, Switzerland; †Institute of Evolutionary Physiology and Biochemistry, St. Petersburg, Russia; §Department of Nephrology, Hypertension and Clinical Pharmacology, Inselspital, Bern, Switzerland; |Department of Clinical Research, University of Bern, Bern, Switzerland; ¶Service of Biomedicine and ††Service of Nephrology, Department of Medicine, Lausanne University Hospital (CHUV), Lausanne, Switzerland; **Department of Nephrology, Tenon Hospital, Université Pierre et Marie Curie, Paris, France; and ‡‡Labeled Research Team (ERL) 8228–U1138 équipe 3, Centre de Recherche des Cordeliers, Paris, France

ABSTRACT The circadian clock controls a wide variety of metabolic and homeostatic processes in a number of tissues, including the kidney. However, the role of the renal circadian clocks remains largely unknown. To address this BASIC RESEARCH question, we performed a combined functional, transcriptomic, and metabolomic analysis in mice with inducible conditional knockout (cKO) of BMAL1, which is critically involved in the circadian clock system, in renal tubular cells (Bmal1lox/lox/Pax8-rtTA/LC1 mice). Induction of cKO in adult mice did not produce obvious abnormalities in renal sodium, potassium, or water handling. Deep sequencing of the renal transcriptome revealed significant changes in the expression of genes related to metabolic pathways and organic anion transport in cKO mice compared with control littermates. Furthermore, kidneys from cKO mice exhibited a significant decrease in the NAD+-to-NADH ratio, which reflects the oxidative phosphorylation-to-glycolysis ratio and/or the status of mitochondrial function. Metabolome profiling showed significant changes in plasma levels of amino acids, biogenic amines, acylcarnitines, and lipids. In-depth analysis of two selected pathways revealed a significant increase in plasma urea level corre- lating with increased renal Arginase II activity, hyperargininemia, and increased kidney arginine content as well as a significant increase in plasma creatinine concentration and a reduced capacity of the kidney to secrete anionic drugs (furosemide) paralleled by an approximate 80% decrease in the expression level of organic anion transporter 3 (SLC22a8). Collectively, these results indicate that the renal circadian clocks control a variety of metabolic/homeostatic processes at the intrarenal and systemic levels and are involved in drug disposition.

J Am Soc Nephrol 27: 2997–3004, 2016. doi: 10.1681/ASN.2015091055

The circadian timing system is an essential physi- Received September 23, 2015. Accepted January 10, 2016. ologic regulatory mechanism that provides cells, S.N., C.A., G.C., and S.P. contributed equally to this work. tissues, organs, and finally, the whole organism with Published online ahead of print. Publication date available at an important functional advantage of anticipation www.jasn.org. of circadian changes in the environment that are imposed by the Earth’s rotation. The circadian tim- Correspondence: Dr. Natsuko Tokonami, The National Institute for Health and Medical Research (INSERM)/Université Pierre et Marie ing system is organized in a hierarchical manner, in Curie (UPMC) Paris 6/National Center for Scientific Research (CNRS), that the central oscillator located in the suprachias- Centre de Recherche des Cordeliers Génomique, Physiologie et matic nucleus of hypothalamus coordinates subsid- Physiopathologie Rénales, Equipe 3 U1138, Labeled Research Team (ERL) 8228, 15 rue de l’Ecole de Médecine, 75270 Paris Cedex, France, iary oscillators located in peripheral tissues. On the or Dr. Dmitri Firsov, Department of Pharmacology and Toxicology, molecular level, both the central and peripheral University of Lausanne, 27 rue du Bugnon, 1005 Lausanne, Switzer- oscillators share a common molecular clock land.Email:[email protected] or dmitri.fi[email protected] mechanism on the basis of cell–autonomous and Copyright © 2016 by the American Society of Nephrology

J Am Soc Nephrol 27: 2997–3004, 2016 ISSN : 1046-6673/2710-2997 2997 BASIC RESEARCH www.jasn.org self–sustained transcriptional/translational feedback RESULTS loops. This core clock mechanism controls circadian expression of so–called output genes that, in turn, im- Validation of the Bmal1lox/lox/Pax8-rtTA/LC1 pose cell–specific functional rhythms (reviewed in refs. (Conditional Knockout) Knockout Model 1and2). The conditional inactivation of the Arntl gene encoding Previous studies have shown that the circadian timing BMAL1 was induced by 2-week treatment with doxycycline system plays a major role in renal function. In mice, whole- (DOX; 2 mg/ml in drinking water) of 8-week-old Bmal1lox/lox/ body inactivation of different elements of the molecular clock Pax8-rtTA/LC1 mice (hereafter referred to as conditional leads to abnormal circadian patterns of urinary sodium and knockout [cKO] mice). In parallel, the same DOX treatment potassium excretion, loss of the circadian rhythmicity of was provided to their littermate controls (Bmal1lox/lox mice; plasma aldosterone levels, and significant changes in arterial hereafter referred to as control mice). The quantitative PCR BP (reviewed in refs. 3–5). In humans, growing evidence (qPCR) analysis revealed that Bmal1 mRNA expression was suggests a possible link between dysregulation of renal circa- significantly reduced in the kidneys of cKO mice (2 months dian rhythms and development of hypertension and accelera- after the end of DOX treatment) (Supplemental Figure 1). ted progression of CKD. The importance of the circadian clock Immunohistochemical staining for BMAL1 and CRE expres- system in the kidney has been recently shown on the molecular sion was performed on day 5 of the 2-week DOX treatment level. Zuber et al.6 and Nikolaeva et al.7 have shown that hun- period or 2 months after the end of DOX treatment. As shown dreds of transcripts in the microdissected distal convoluted in Supplemental Figure 2, at day 5 of DOX treatment, the CRE tubule/connecting tubule and the cortical collecting duct was ubiquitously expressed along the renal tubule but not in exhibit significant circadian oscillations in their expression glomeruli or blood vessels in cKO mice. The BMAL1 protein levels and that the whole-body inactivation of circadian tran- was ubiquitously expressed in all kidney cells in control mice, scriptional factor Clock leads to dramatic changes in the tran- but in cKO mice, all tubular cells became negative for BMAL1 scriptomes of the distal convoluted tubule/connecting tubule staining at day 5 of DOX treatment (Supplemental Figure 3). and the cortical collecting duct. Gumz and colleagues8–10 have The specific tubular inactivation of BMAL1 was maintained in shownthatthecircadianclockproteinPER1exertsdiverse cKO mice 2 months after the end of DOX treatment (Supple- effects on the renal handling of sodium. Zhang et al.11 have mental Figure 3). shown that, among 12 tested mouse tissues, the kidney ex- To assess the effect of BMAL1 deficiency on circadian hibits the second greatest number of circadian transcripts (the mechanisms per se, we performed qPCR analysis of expression greatest number was found in the liver). However, the role of levels of several genes involved in the core clock (Cry1, Per2, the renal circadian clocks in the generation of functional and and Nr1d1)andDbp, an output gene that is directly controlled molecular rhythms in the kidney as well as the development of by the circadian clock. As shown in Supplemental Figure 1, the kidney disease remains largely unknown. To our knowledge, BMAL1 deficiency results in a significant attenuation in cir- the only study that addressed this question is a recent paper by cadian oscillations of mRNA expression of Cry1, Per2, Nr1d1, Tokonami et al.,12 which showed that ablation of the circadian and Dbp. The expression level of Bmal2, a paralog of Bmal1 clock in renin–secreting granular cells results in several that is thought to be able to rescue the BMAL1 deficiency in abnormalities, including increased GFR, decreased plasma al- some tissues,13 was increased in kidneys of cKO mice, but the dosterone levels, and low BP. Here, we studied the role of absolute levels of Bmal2 expression remained significantly circadian clocks in the nephron by using conditional and lower than those of other tested core clock components (Sup- inducible inactivation of BMAL1 protein, an indispensable plemental Figure 1). Collectively, these results suggested that element of the molecular clock (Bmal1lox/lox/Pax8-rtTA/LC1 the activity of the circadian clock in renal tubular cells is sig- mice). nificantly impaired. The renal tubular cells are involved in a wide variety of PAX8 is a paired–box transcriptional factor crucial to the essential functions, including maintenance of homeostasis, organogenesis and development of the kidney, the thyroid disposition of xenobiotics, and synthesis of molecules that are gland, and the Müllerian system. Traykova-Brauch et al.14 released into the bloodstream (e.g., bicarbonate and arginine). used a partial Pax8 promoter to target the expression of a Because many of these processes exhibit circadian rhythmicity, tetracycline-dependent transactivator to the renal tubular the aim of this study was to identify the molecular pathways/ cells. The tissue specificity of the promoter-driven expression mechanisms that are directly controlled in the nephron by the was assessed in the kidney, liver, heart, lung, brain, spleen, tubular circadian clock. Using an approach combining func- thyroid, and colon. Among the tissues tested, the Pax8–driven tional, metabolomics, and deep sequencing analyses, we show CRE expression was found all along the renal tubule and in a that the tubular circadian clock is not a prerequisite for the subset of periportal hepatocytes. Hence, we tested the BMAL1 control of the GFR or the water, sodium, and potassium bal- expression in the liver. As shown in Supplemental Figure 4, ance. However, we show that the tubular circadian clock is indeed, some proportion of liver cells located in the vicinity of deeply involved in the control of both intrarenal and systemic large blood vessels appeared negative for BMAL staining in metabolisms as well as drug disposition. DOX–treated cKO mice.

2998 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 2997–3004, 2016 www.jasn.org BASIC RESEARCH

Table 1. The 24-hour urine and plasma chemistry in control and cKO mice Blood collection was performed at Zeit- Measured Parameters Control cKO P Value geber time 4 (ZT4) and ZT16 (ZT0 is the Body wt, g 33.0763.28 (20) 31.8962.26 (21) NS time of lights on, and ZT12 is the time of Kidney weight, g 0.1960.02 (20) 0.1560.02 (21) ,0.001 lights off); 24-hour or hourly urine was Kidney weight/body wt, % 0.5860.06 (20) 0.4660.06 (21) ,0.001 collected from freely moving mice housed Urine in metabolic cages with free access to food Volume/body wt, ml/g 0.04960.019 (12) 0.05460.028 (12) NS and water. As shown in Table 1, the only 6 6 Osmolality, mOsm/kg H2O2734755 (12) 2636 575 (12) NS difference found in 24-hour urine samples pH 6.5360.27 (12) 6.1360.15 (12) ,0.001 was a significant reduction in urinary pH in + UV 3 Na /g body wt, mmol/g 6.363.25 (12) 5.661.67 (12) NS cKO mice. Because the whole-body knock- 3 + m 6 6 UV K /g body wt, mol/g 21.6 6.46 (12) 22.2 7.1 (12) NS out of Clock results in significantly different 3 2+ m 6 6 UV Ca /g body wt, mol/g 0.088 0.061 (12) 0.114 0.057 (12) NS circadian kinetics of urinary water, sodium, UV 3 Mg2+/g body wt, mmol/g 1.4660.54 (12) 1.6960.83 (12) NS 32 and potassium excretion, we also analyzed UV 3 PO4 /g body wt, mmol/g 3.8261.66 (12) 3.9662.12 (12) NS UV 3 creatinine/g body wt, mmol/g 0.3160.12 (12) 0.3160.13 (12) NS the hourly urine excretion pattern in con- fl UV 3 urea/g body wt, mmol/g 94.0641.6 (12) 97.0642.6 (12) NS trol and cKO mice. Toexclude the in uence UV 3 NOx/g body wt, mmol/g 0.10760.030 (12) 0.11860.046 (12) NS oflightonurineexcretoryrhythms,the UV 3 glucose/g body wt, mmol/g 0.1960.11 (12) 0.1560.08 (12) NS hourly urine collection was also performed UV 3 total protein/g body wt, mg/g 0.5060.13 (12) 0.3560.20 (12) NS on animals placed in constant darkness UV 3 ammonium/g body wt, mmol/g 0.60460.178 (11) 1.28860.407 (12) NS 30 hours before urine collection (dark/ UV 3 TA/g body wt, mmol/g 1.75760.325 (12) 2.34360.298 (12) NS dark conditions). As shown in Supplemen- Plasma tal Figures 5 and 6, the circadian patterns of 6 6 Osmolality ZT4, mOsm/kg H2O 306.8 6.9 (6) 304.8 4.2 (6) NS urinary excretion of water, sodium, and 6 6 Osmolality ZT16, mOsm/kg H2O 304.3 6.2 (6) 311.3 3.6 (6) 0.04 potassium were not different between con- Na+ ZT4, mM 153.162.52 (6) 153.861.44 (6) NS + trol and cKO mice in both light/dark and Na ZT16, mM 147.362.23 (6) 146.863.46 (6) NS K+ ZT4, mM 3.9360.63 (6) 3.7860.44 (6) NS dark/dark conditions, respectively. Basic K+ ZT16, mM 3.3560.33 (6) 3.2560.29 (6) NS analysis of plasma samples revealed signif- Ca2+ ZT4, mM 2.1860.06 (6) 2.2060.05 (6) NS icant increases in creatinine (ZT4 and Ca2+ ZT16, mM 2.0460.06 (6) 2.0260.07 (6) NS ZT16), urea (ZT4 and ZT16), and magne- Mg2+ ZT4, mM 1.1260.09 (6) 1.1460.09 (6) NS sium (ZT16) levels as well as osmolality Mg2+ ZT16, mM 1.1260.04 (6) 1.2360.07 (6) ,0.01 (ZT16) in cKO mice (Table 1) (t test). Dif- Creatinine ZT4, mM2162.1 (6) 28.562.5 (6) ,0.001 ferences between ZT4 and ZT16 for plasma Creatinine ZT16, mM 18.661.8 (5) 24.464.2 (5) 0.02 parameters were analyzed by two-way 32 6 6 PO4 ZT4, mM 2.15 0.26 (6) 2.22 0.34 (6) NS ANOVA. As shown is Supplemental Table 32 6 6 PO4 ZT16, mM 2.24 0.31 (6) 2.45 0.35 (6) NS 1, ANOVA revealed a significant effect of Urea ZT4, mM 9.2361.89 (6) 11.4060.62 (6) 0.02 time on plasma osmolality and plasma Urea ZT16, mM 10.7761.57 (6) 12.9761.08 (6) 0.02 Aldosterone ZT4, pg/ml 503.36116.1 (6) 435.06173.2 (6) NS concentration of sodium, potassium, cal- Aldosterone ZT16, pg/ml 312.96131.4 (6) 216.7688.2 (6) NS cium, creatinine, urea, and aldosterone. GFR This effect of time was similar between GFR ZT6, ml/min 264.0636.52 (7) 251.1643.03 (7) NS control and cKO mice. Importantly, there + Values are means6SEMs (t test). H2O, water; UV, 24-hour urine volume; Na , sodium concentration in the 24- was no difference in the GFR between con- hour urine; K+, potassium concentration in the 24-hour urine; Ca2+, calcium concentration in the 24-hour urine; trol and cKO mice as measured by inulin 2+ 32 Mag , magnesium concentration in the 24-hour urine; PO4 , phosphate concentration in the 24-hour urine; NOx, nitrogen oxides concentration in the 24-hour urine; TA, titratable acids concentration in the 24-hour urine. clearance(Table1).ThecKOmice showed a modest but statistically significant Basic Characteristics of cKO Mice decrease in systolic BP (false discovery rate ,0.001) but not in To avoid the potential side effects related to DOX toxicity, all diastolic BP; the inactive–phase BP dipping was not different additional analyses were performed 2 months after the end of between control and cKO mice (Supplemental Figure 7, Sup- DOX treatment. For all experiments, mice were adapted to a plemental Table 2). The 24-hour activity, heart rate, and pulse 12-hour light/12-hour dark cycle for 2 weeks. pressure were not different between genotypes (data not The cKO mice were overtly normal; their body weight was shown). not different from the control mice, but their kidney weight-to- body weight ratio was significantly decreased (Table 1). The Deep Sequencing Kidney Transcriptome Profiling and basic morphologic analysis did not reveal any obvious renal Plasma Metabolome Analysis of Control and cKO Mice abnormalities in cKO mice (tested parameters: gross renal The increased plasma creatinine and urea levels along with morphology, interstitial fibrosis, glomerular sclerosis, inflam- normal GFR suggested a tubular impairment in cKO mice. mation, and vascular lesions; data not shown). Because the circadian clock has been shown to control a variety

J Am Soc Nephrol 27: 2997–3004, 2016 Circadian Clocks in the Kidney 2999 BASIC RESEARCH www.jasn.org of physiologic processes, we performed unbiased kidney transcriptome and plasma metabolome analyses in control and cKO mice to identify metabolic/homeostatic pathways and/or transporter systems that are controlled by the renal circadian clock. Transcriptome profiling of RNAs extracted from the whole kidney (ZT4 and ZT16) was performed by deep sequencing, and the partial metabolome profiling (180 plasma metabolites from five substance classes; i.e.,hexose, amino acids, biogenic amines, acylcarnitines, and lipids; BIOCRATES Life Sciences AG, Innsbruck, Austria) was performed on plasma samples collected at ZT4 and ZT16. For each sample, from 34 to 50 million sequencing reads were aligned to the mouse genome (Supplemental Table 3). As shown in Figure 1A, 721 and 765 transcripts exhibited differential expression levels between control and cKO mice at ZT4 and ZT16, respectively (false discovery rate ,0.05) (Supple- mental Table 4). Among them, 552 transcripts were commonly changed at ZT4 and ZT16 (Figure 1A, Supplemental Table 4). Gene ontology analysis with these 552 genes followed by summarization in REVIGO15 revealed enrichment of processes related to cellular metabolism (carboxylic acid metabolism cluster shown in the blue box in Figure 1B, Sup- plemental Table 5) and organic anion transport (organic anion transport cluster shown in the pink box in Figure 1B, Supplemental Table 6). Differential expression of six selected genes, namely nicotinamide phosphoribosyltrans- ferase, peroxisome proliferator–activated receptor-d, mitochondrially encoded reduced nicotinamide adenine di- nucleotide (NADH) dehydrogenase 1, carnitine transporter Slc22a5, monocarboxylate transporter 1 (or Slc16a1), and Figure 1. Analysis of genes differentially expressed in kidneys of cyclin–dependent kinase inhibitor 1a (or p21), was validated Control and cKO mice revealed enrichment of processes related by qPCR (Supplemental Figure 8). Analysis of transcripts to the cellular metabolism and organic anion transport. (A) Heat encoding proteins located in mitochondria showed that maps of normalized expression values of genes significantly af- transcripts encoded by both mitochondrial and nuclear ge- fected in kidneys of cKO mice at ZT4, ZT16, or both time points (false , nomes are significantly over-represented among the most discovery rate 5%). Expression values were mean centered, vari- ance normalized, and subjected to hierarchical clustering (complete downregulated genes in cKO mice at both ZT4 and ZT16 , – – linkage) using Pearson correlation as the similarity metric. (B) Gene (P 0.001; two tailed Mann Whitney U test) (Supplemen- ontology analysis showing significantly enriched biologic processes tal Figure 9). Because the circadian clock has been shown to (Bonferroni correction) among 552 genes differentially expressed in control mitochondrial biogenesis,16 we performed a quan- kidneys of control and cKO mice at ZT4 and ZT16. Significant bi- titative analysis of mitochondrial genomic DNA in kidneys ologic processes were summarized and classified with REVIGO. of control and cKO mice. As shown in Figure 2A, the mito- Ctrl, control. chondrial DNA content was not different between control and cKO mice. The NAD+-to-NADH ratio, a marker reflect- Comparison of metabolome and transcriptome data al- ing the oxidative phosphorylation-to-glycolysis ratio, was lowed us to identify several potential mechanisms through significantly decreased in the kidney of cKO mice but not which tubular circadian clocks participate in the control of in the liver (Figure 2B). blood metabolome. Mechanisms involved in the control of Metabolome analysis revealed significant differences in the plasma urea/arginine and creatinine levels were selected for plasma levels of amino acids (arginine [ZT16], glutamate validation and more in–depth analysis. [ZT4], and methionine [ZT 16]), biogenic amines (asymmet- ric dimethylarginine [ZT16] and creatinine [ZT4 and ZT16]; Urea and Arginine thereby confirming results of blood chemistry analysis) (Table Urea is the end product of ammonia detoxification in the liver, a 1), carnosine (ZT16), taurine (ZT16), carnitine (ZT4 and process that depends on the activity of Arginase I, an enzyme ZT16), and several acylcarnitines of different species of phos- that converts L-arginine formed in the liver urea cycle into urea phatidylcholine, lysophosphatidylcholine, and sphingomye- and L-ornithine. The kidney expresses the second arginase iso- line (Supplemental Table 6). form, namely ARGII, but its activity is significantly lower than

3000 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 2997–3004, 2016 www.jasn.org BASIC RESEARCH

Figure 2. Altered renal metabolism in cKO mice. (A) Quantifi- cation of mitochondrial DNA (mtDNA) in kidneys of control and cKO mice (ZT8). Relative amounts of mtDNA and nuclear DNA were quantitated by qPCR of mitochondrially encoded NADH dehydrogenase1(mt-Nd1) and nuclear Ppia (cyclophilin) genes16 (mean6SD; n=6; t test). (B) NAD+-to-NADH ratios in the kidney (ZT12) and the liver (ZT12) of control and cKO mice (mean6SD; n=6; t test). the activity of Arginase I in the liver. Hence, it is generally accepted that, in the normal physiologic state, Arginase II (ARGII) does not influence plasma urea levels.17 Most of the Figure 3. Altered arginine production/degradation in cKO mice. (A) Western blot analysis of ARGII protein expression in the circulating L-arginine is synthesized in the kidney proximal proximal straight tubules (PSTs) microdissected from kidneys of tubule from L-citrulline that is absorbed from the small in- control and cKO mice (ZT8). Three mice per genotype were used testine. for this analysis. (B) Arginase activity in the kidney (ZT0) and the Analysis of transcriptomes revealed that ARGII transcript liver (ZT0) of control and cKO mice (mean6SD; n=3; t test). (C) fi levels are signi cantly upregulated in kidneys of cKO mice L-arginine (L-Arg) levels in kidney tissue of control and cKO mice (Supplemental Figure 10A, Supplemental Table 5). Western (ZT8; mean6SD; n=6; t test). (D) nitric oxide synthase–specific blotting performed on microdissected renal tubules (Figure L-citrulline (L-Citr) formation in kidney tissue of control and cKO 3A, Supplemental Figure 10B) and immunohistochemical mice (ZT8; mean6SD; n=6; t test). prot, Protein. staining (Supplemental Figure 10C) showed that ARGII protein expression is dramatically increased specifically in the proximal straight tubule. As shown in Figure 3B, the enzymatic OAT3 expression is reduced in cKO mice at both ZT4 and arginase activity was significantly increased in kidneys of cKO ZT16 (Supplemental Table 4). These results were confirmed mice and reached approximately 25% of arginase activity in the by qPCR (Figure 4A). Western blotting showed an approxi- liver; the liver arginase activity was not different between con- mately 80% reduction in OAT3 protein expression in cKO trol and cKO mice. In parallel, arginine levels in kidney tissue of mice (Figure 4B) (P=0.004). Because OAT3 is involved in cKO mice were significantly increased (Figure 3C), thereby the basolateral transport of a variety of organic acids, includ- providing a potential link with hyperargininemia in the cKO ing several clinically important drugs, we tested the possibility mice shown in the metabolomic profiling (Supplemental Table that the tubular circadian clocks control pharmacologic prop- 6). Interestingly, the activity of the nitric oxide synthase, the erties of furosemide, a diuretic that is actively secreted in the second major arginine–consuming enzyme in the kidney,18 was kidneybyOAT3.AsshowninFigure4C,thecKOmice also significantly increased in cKO mice (Figure 3D). exhibited rightward shift in the natriuretic dose-response curve for furosemide (control mice: half maximal inhibitory Creatinine and Furosemide concentration (IC50)=0.11760.048 mg furosemide per 1 g Creatinine, a weak organic acid, is eliminated in the urine by body wt; cKO mice: half maximal inhibitory concentration glomerular filtration and in part, tubular secretion. Vallon (IC50)=0.26360.094 mg furosemide per 1 g body wt; et al.19 have shown that organic anion transporter 3 (OAT3 mean6SD; n=5; P=0.01; t test). In parallel, cKO mice showed [Slc22a8]) significantly contributes to the creatinine secretion lower urinary furosemide excretion after a single 0.3-mg/g in the mouse kidney. Analysis of transcriptome revealed that body wt furosemide bolus (Figure 4D).

J Am Soc Nephrol 27: 2997–3004, 2016 Circadian Clocks in the Kidney 3001 BASIC RESEARCH www.jasn.org

granular cells results in modified circadian rhythm of urinary sodium excretion, polyuria, increased GFR, and low BP.12 Ac- cumulating evidence from the work by Titze et al.21 suggests that total body sodium content fluctuates independently of in- take or body weight, thereby pointing to a possible role of extracellular sodium storage/release dynamics. Also, there is strong evidence that the circadian rhythms of urinary potassium excretion are determined, at least in part, by the net potassium fluxes between intracellular and extracellular compartments.22 Collectively, these results suggest that circadian clocks in many different organs/cell types may participate in the generation/ maintenance of circadian rhythms of urinary sodium, potassium, and water excretion, but the role of the intrinsic renal tubular clocks in these processes remains unclear. However, we found that the tubular circadian clock is deeply involved in several other essential physiologic functions, including control of both the intrarenal and systemic metabolisms, elimination of xenobiotics/drugs, and maintenance of homeostasis of compounds secreted into the blood by the kidney (nota bene,itremainspossiblethatsomeofthe Figure 4. Lower OAT3 mRNA and protein expression correlates observed changes result from circadian clock–independent with impaired natriuretic response to furosemide and decreased activity of BMAL1). The critical role of the circadian clock urinary excretion of furosemide in cKO mice. (A) qPCR analysis of system in the control of metabolism has been shown in a Oat3 (Slc22a8) mRNA expression in kidneys of control (white number of tissues23; however, the novelty of our study is circles) and cKO (gray circles) mice (mean6SD; n=6). (B) Western that the suppression of the tubular circadian clock greatly blotting with anti-OAT3 or anti-glyceraldehyde 3-phosphate dehydrogenase antibodies on protein extracts prepared from kid- affects the plasma metabolome in a way that, in turn, may neys of control and cKO mice (n=5 mice in each group) at ZT16. affect metabolic and homeostatic processes at the systemic , (C) Rightward shift in the natriuretic response to furosemide in level. Our partial metabolome analysis covered 5% of the cKO mice (mean6SD; n=5). The half maximal inhibitory con- total plasma metabolome estimates (180 versus approxi- centration (IC50) values were calculated using the GraphPad mately 4000 metabolites, respectively), but even this re- PRISM 6 software (GraphPad Software, La Jolla, CA) at ZT16. (D) stricted approach allowed us to identify .50 metabolites Lower urinary excretion of furosemide in cKO mice. Furosemide that are differentially represented in plasma of cKO mice. was dosed in the urine collected for 10 minutes followed by the Comparison of transcriptome and metabolome data allows m 6 0.3- g/g body wt furosemide bolus (mean SEM; n=8; t test). us to establish tubular mechanisms that are disturbed in cKO Body wt, body weight; Ctrl, control; iv, intravenous; UNaV, so- mice. For instance, carnitine deficiency in cKO mice correlates dium excretion rate; v, vehicle. with significantly decreased expression of the Scl22a5 carnitine transporter involved in the carnitine reabsorption in the DISCUSSION proximal tubule. Of note, the loss of function of Slc22a5 is a well known cause of primary carnitine deficiency.24 Two of An intriguing result of this study is that the renal phenotype of these mechanisms, namely those that are involved in the renal cKO mice differs significantly from that of Clock null mice.7 control of plasma creatinine and urea levels, were selected for The major differences are in the normal circadian dynamics of in-depth analysis. These two mechanisms were chosen, urinary water, sodium, and potassium excretion, the normal because (1) plasma creatinine and urea concentrations are plasma aldosterone levels, and the normal GFR in cKO mice, commonly used as surrogate biomarkers of glomerular whereas these functions were significantly impaired in Clock (dys)function in clinical settings and (2) creatinine secretion null mice. This difference suggests that either the tubular cir- in the kidney occurs through the same transporter systems as cadian clocks are not involved in the generation of urinary those used by many clinically important drugs (see below). excretory rhythms or the inactivation of tubular circadian We found that the increase in plasma urea levels correlates clocks could be fully compensated for by other mechanisms. with tubular dysfunction and does not correlate with glomer- These findings raise the question about the tissue/cellular or- ular dysfunction. We show that cKO mice exhibit a parallel igin of renal excretory rhythms. Okamura and colleagues20 increase in the renal ARGII activity and renal and plasma showed that mice lacking Cry1 and Cry2 exhibit hyperaldoster- arginine levels. Arginine, a semiessential amino acid, is onism and blunted circadian oscillations in plasma aldosterone synthesized in the proximal convoluted tubule and released into levels caused by increased aldosterone synthesis in adrenal the bloodstream, where it can be, in part, recovered by the glands. We have recently shown that the deletion of Arntl in renal proximal straight tubule and converted to urea and L-ornithine

3002 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 2997–3004, 2016 www.jasn.org BASIC RESEARCH

by ARGII or oxidized to nitric oxide and L-citrulline by nitric Metabolic Cages oxide synthases in tubular and nontubular renal cells.18 Inter- Mice were housed in individual metabolic cages (Tecniplast). Urine estingly, the activity of the nitric oxide synthase was also sig- collection was performed after a 3-day adaptation period. Urine and nificantly upregulated in kidneys of cKO mice. Collectively, blood chemistry was analyzed as previously described.7 Hourly urine these results suggest that the circadian clock in tubular cells collection was performed as previously described.7 controls both the intrarenal arginine metabolism and systemic arginine and urea levels. Blood and Urine Chemistry Similarly, we found that the increase in plasma creatinine Partial plasma metabolome profiling (180 plasma metabolites from levels has tubular origin and results from the decreased five substance classes; i.e., hexose, amino acids, biogenic amines, acylcar- expression levels of OAT3. This finding has a second important nitines, and lipids) was performed by BIOCRATES Life Sciences AG. facet that supports the link between the circadian clock and the Plasma aldosterone levels were measured by radioimmunoassay pharmacokinetics of drugs. The OAT3transporter is involved in (DPC). tubular secretion of a variety of clinically important drugs, including diuretics (furosemide and bendroflumethiazide), Measurements Performed on Kidney and/or Liver antiherpetics (acyclovir), antiretrovirals (tenofovir), antineo- Tissue Extracts plastics (methotrexate), antibiotics (benzylpenicillin), and The NAD+ and NADH levels in kidney and liver tissues were deter- many others (reviewed in ref. 25). Here, we show that the re- mined by the PicoProbe Colometric Kit from BioVision, Inc. (San duced expression of OAT3 in cKO mice correlates with a less Francisco, CA). Kidney tissue spermine levels were determined potent natriuretic effect and a lower rate of urinary elimina- by Ansynth Service B.V. Kidney tissue L-arginine levels were deter- tion of furosemide, a diuretic that is not filtered in glomer- mined by a kit from MyBioSource. Arginase activity in the kidney and uli and must be secreted into tubular lumen to inhibit the liver tissues was determined as arginase–specific urea formation 14 sodium-potassium-chloride cotransporter 2. To our knowl- from C-L-arginine as previously described.29,30 NOS activity was edge,thisisthefirst demonstration that the renal circadian determined as nitric oxide synthase–specific L-citrulline formation 14 clocks are involved in drug pharmacokinetic. Hence, together from C-L-arginine as described in Supplemental Material. with the well documented effects of the circadian clock on drug metabolism in the liver,26 drug absorption in the intes- BP tine,27 and drug distribution,28 our study shows that the cir- The BP was measured in conscious unrestrained mice using telemetry cadian timing system interferes with drug pharmacokinetics (DSI System). at very different levels and highlights the attention that must be given to chronopharmacology and chronotherapy. GFR The GFR was determined according to the method described by Qi et al.31 CONCISE METHODS Gene Analysis by RNA Sequencing Animals Gene analysis was by RNA sequencing (Supplemental Material). The procedures used to generate the characterization of Bmal1lox/lox, 14 Pax8-rtTA, and LC-1 Cre mice were described previously. The Natriuretic Effect of Furosemide and Furosemide three mouse lines used in this study are inbred strains bred on the Dosage genetic background of the C57BL/6J mouse. The animals were These experiments were performed according to the protocol maintained ad libitum on the standard laboratory chow diet (KLIBA established by Vallon et al.32 NAFAG Diet 3800). Before all experiments, mice were adapted to a 12-hour light/dark cycle. All experiments with animals were performed in accordance with the Swiss Guidelines for Animal Care, which conform to the National Institutes of Health Animal Care ACKNOWLEDGMENTS Guidelines. This work was supported by Swiss National Science Foundation Antibodies Research Grant 31003A-149440 (to D.F.). Anti–CRE recombinase antibody was from Novagen. Anti-BMAL1 Selected data in this manuscript were previously presented at the antibody was described elsewhere.12 Anti-ARGII antibody was from November 11–16, 2014 (Philadelphia, PA) and November 3–8, 2015 Santa Cruz Biotechnology (Santa Cruz, CA). Antiactin antibody was (San Diego, CA) Annual Meetings of the American Society of from Sigma-Aldrich (St. Louis, MO). Anti-OAT3 (SLC22a8) anti- Nephrology. body was from Abcam, Inc. (Cambridge, MA). Anti-OCTN2 (SLC22a5) antibody was from GeneTex (Irvine, CA). Immunohisto- chemistry and Western blotting protocols were identical to those de- DISCLOSURES scribed previously.12 None.

J Am Soc Nephrol 27: 2997–3004, 2016 Circadian Clocks in the Kidney 3003 BASIC RESEARCH www.jasn.org

REFERENCES 17. Iyer RK, Yoo PK, Kern RM, Rozengurt N, Tsoa R, O’Brien WE, Yu H, Grody WW, Cederbaum SD: Mouse model for human arginase deficiency. Mol Cell Biol 22: 4491–4498, 2002 1. Bollinger T, Schibler U: Circadian rhythms - from genes to physiology 18. Levillain O: Expression and function of arginine-producing and con- and disease. Swiss Med Wkly 144: w13984, 2014 suming-enzymes in the kidney. Amino Acids 42: 1237–1252, 2012 2. Peek CB, Ramsey KM, Levine DC, Marcheva B, Perelis M, Bass J: 19. Vallon V, Eraly SA, Rao SR, Gerasimova M, Rose M, Nagle M, Anzai N, Circadian regulation of cellular physiology. Methods Enzymol 552: Smith T, Sharma K, Nigam SK, Rieg T: A role for the organic anion 165–184, 2015 transporter OAT3 in renal creatinine secretion in mice. Am J Physiol 3. Firsov D, Bonny O: Circadian regulation of renal function. Kidney Int 78: Renal Physiol 302: F1293–F1299, 2012 640–645, 2010 20. Doi M, Takahashi Y, Komatsu R, Yamazaki F, Yamada H, Haraguchi S, 4. Bonny O, Firsov D: Circadian regulation of renal function and Emoto N, Okuno Y, Tsujimoto G, Kanematsu A, Ogawa O, Todo T, potential role in hypertension. Curr Opin Nephrol Hypertens 22: Tsutsui K, van der Horst GT, Okamura H: Salt-sensitive hypertension in 439–444, 2013 circadian clock-deficient Cry-null mice involves dysregulated adrenal 5. Richards J, Diaz AN, Gumz ML: Clock genes in hypertension: Novel Hsd3b6. Nat Med 16: 67–74, 2010 insights from rodent models. Blood Press Monit 19: 249–254, 2014 21. Titze J, Rakova N, Kopp C, Dahlmann A, Jantsch J, Luft FC: Balancing 6. Zuber AM, Centeno G, Pradervand S, Nikolaeva S, Maquelin L, wobbles in the body sodium [published online ahead of print Sep- Cardinaux L, Bonny O, Firsov D: Molecular clock is involved in pre- tember 25, 2015]. Nephrol Dial Transplant doi:gfv343 dictive circadian adjustment of renal function. Proc Natl Acad Sci U S A 22. Moore Ede MC, Brennan MF, Ball MR: Circadian variation of inter- 106: 16523–16528, 2009 compartmental potassium fluxes in man. J Appl Physiol 38: 163–170, 7. Nikolaeva S, Pradervand S, Centeno G, Zavadova V, Tokonami N, 1975 Maillard M, Bonny O, Firsov D: The circadian clock modulates renal 23. Peek CB, Affinati AH, Ramsey KM, Kuo HY, Yu W, Sena LA, Ilkayeva O, sodium handling. JAmSocNephrol23: 1019–1026, 2012 Marcheva B, Kobayashi Y, Omura C, Levine DC, Bacsik DJ, Gius D, 8. Gumz ML, Stow LR, Lynch IJ, Greenlee MM, Rudin A, Cain BD, Weaver Newgard CB, Goetzman E, Chandel NS, Denu JM, Mrksich M, Bass J: DR, Wingo CS: The circadian clock protein Period 1 regulates expres- Circadian clock NAD+ cycle drives mitochondrial oxidative metabolism sion of the renal epithelial sodium channel in mice. JClinInvest119: in mice. Science 342: 1243417, 2013 2423–2434, 2009 24. Stanley CA: Carnitine deficiency disorders in children. Ann N Y Acad Sci 9. Richards J, Ko B, All S, Cheng KY, Hoover RS, Gumz ML: A role for the 1033: 42–51, 2004 circadian clock protein Per1 in the regulation of the NaCl co-transporter 25. Burckhardt G: Drug transport by organic anion transporters (OATs). (NCC) and the with-no-lysine kinase (WNK) cascade in mouse distal Pharmacol Ther 136: 106–130, 2012 convoluted tubule cells. JBiolChem289: 11791–11806, 2014 26. Gachon F, Olela FF, Schaad O, Descombes P, Schibler U: The circadian 10. Richards J, Cheng KY, All S, Skopis G, Jeffers L, Lynch IJ, Wingo CS, PAR-domain basic leucine zipper transcription factors DBP, TEF, and Gumz ML: A role for the circadian clock protein Per1 in the regulation of HLF modulate basal and inducible xenobiotic detoxification. Cell aldosterone levels and renal Na+ retention. Am J Physiol Renal Physiol Metab 4: 25–36, 2006 305: F1697–F1704, 2013 27. Iwasaki M, Koyanagi S, Suzuki N, Katamune C, Matsunaga N, Watanabe 11. Zhang R, Lahens NF, Ballance HI, Hughes ME, Hogenesch JB: A cir- N, Takahashi M, Izumi T, Ohdo S: Circadian modulation in the intestinal cadian gene expression atlas in mammals: Implications for biology and absorption of P-glycoprotein substrates in monkeys. Mol Pharmacol – medicine. Proc Natl Acad Sci U S A 111: 16219 16224, 2014 88: 29–37, 2015 12. Tokonami N, Mordasini D, Pradervand S, Centeno G, Jouffe C, Maillard 28. Baraldo M: The influence of circadian rhythms on the kinetics of drugs in M, Bonny O, Gachon F, Gomez RA, Sequeira-Lopez ML, Firsov D: Local humans. Expert Opin Drug Metab Toxicol 4: 175–192, 2008 fl renal circadian clocks control uid-electrolyte homeostasis and BP. 29. Morris SM Jr., Gao T, Cooper TK, Kepka-Lenhart D, Awad AS: Arginase- – JAmSocNephrol25: 1430 1439, 2014 2 mediates diabetic renal injury. Diabetes 60: 3015–3022, 2011 13. Shi S, Hida A, McGuinness OP, Wasserman DH, Yamazaki S, Johnson 30. Kepka-Lenhart D, Ash DE, Morris SM: Determination of mammalian CH: Circadian clock gene Bmal1 is not essential; functional re- arginase activity. Methods Enzymol 440: 221–230, 2008 – placement with its paralog, Bmal2. Curr Biol 20: 316 321, 2010 31. Qi Z, Whitt I, Mehta A, Jin J, Zhao M, Harris RC, Fogo AB, Breyer MD: 14. Traykova-Brauch M, Schönig K, Greiner O, Miloud T, Jauch A, Bode M, Serial determination of glomerular filtration rate in conscious mice us- Felsher DW, Glick AB, Kwiatkowski DJ, Bujard H, Horst J, von Knebel ing FITC-inulin clearance. Am J Physiol Renal Physiol 286: F590–F596, fi Doeberitz M, Niggli FK, Kriz W, Gröne HJ, Koesters R: An ef cient and 2004 versatile system for acute and chronic modulation of renal tubular 32. Vallon V, Rieg T, Ahn SY, Wu W, Eraly SA, Nigam SK: Overlapping in – function in transgenic mice. Nat Med 14: 979 984, 2008 vitro and in vivo specificities of the organic anion transporters OAT1 š 15. Supek F, Bo njak M, Skunca N, Smuc T: REVIGO summarizes and vi- and OAT3 for loop and thiazide diuretics. Am J Physiol Renal Physiol sualizes long lists of gene ontology terms. PLoS One 6: e21800, 2011 294: F867–F873, 2008 16. Woldt E, Sebti Y, Solt LA, Duhem C, Lancel S, Eeckhoute J, Hesselink MK, Paquet C, Delhaye S, Shin Y, Kamenecka TM, Schaart G, Lefebvre P, Nevière R, Burris TP, Schrauwen P, Staels B, Duez H: Rev-erb-a modulates skeletal muscle oxidative capacity by regulating mito- This article contains supplemental material online at http://jasn.asnjournals. chondrial biogenesis and autophagy. Nat Med 19: 1039–1046, 2013 org/lookup/suppl/doi:10.1681/ASN.2015091055/-/DCSupplemental.

3004 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 2997–3004, 2016