Abstract
Background Autosomal dominant polycystic kidney disease (ADPKD) is caused by mutations in the PKD1 and PKD2 genes, and often progresses to kidney failure. ADPKD progression is not uniform among patients, suggesting that factors secondary to the PKD1/2 gene mutation could regulate the rate of disease progression. Here we tested the effect of circadian clock disruption on ADPKD progression. Circadian rhythms are regulated by cell-autonomous circadian clocks composed of clock proteins. BMAL1 is a core constituent of the circadian clock.
Methods To disrupt the circadian clock, we deleted Bmal1 gene in the renal collecting ducts of the Pkd1RC/RC (RC/RC) mouse model of ADPKD (RC/RC;Bmal1f/f;Pkhd1cre, called DKO mice), and in Pkd1 knockout mouse inner medullary collecting duct cells (Pkd1Bmal1KO mIMCD3 cells). Only male mice were used.
Results Human nephrectomy ADPKD kidneys and Pkd1KO mIMCD3 cells showed reduced Bmal1 gene expression compared to normal controls. When compared to RC/RC kidneys, DKO kidneys showed significantly altered clock gene expression, increased cyst growth, cell proliferation, apoptosis and fibrosis. DKO kidneys also showed increased lipogenesis and cholesterol synthesis-related gene expression, and increased tissue triglyceride levels compared to RC/RC kidneys. Similarly, in vitro, Pkd1Bmal1KO cells showed altered clock genes, increased lipogenesis and cholesterol synthesis-related genes, and reduced fatty-acid oxidation-related gene expression compared to Pkd1KO cells. The Pkd1Bmal1KO cells showed increased cell proliferation compared to Pkd1KO cells, which was rescued by pharmacological inhibition of lipogenesis.
Conclusion Renal collecting duct specific Bmal1 gene deletion disrupts the circadian clock and triggers accelerated ADPKD progression by altering lipid metabolism-related gene expression.
Key points
Lack of BMAL1, a circadian clock protein in renal collecting ducts disrupted the clock and increased cyst growth and fibrosis in an ADPKD mouse model.
BMAL1 gene deletion increased cell proliferation by increasing lipogenesis in kidney cells.
Thus, circadian clock disruption could be a risk factor for accelerated disease progression in patients with ADPKD.
Introduction
Autosomal dominant polycystic kidney disease (ADPKD) is the most common inherited kidney disease which is caused by mutations in the PKD1 and PKD2 genes. Progressive cyst growth in ADPKD kidneys is accompanied by inflammation and fibrosis, which often leads to loss of renal function and kidney failure. However, disease progression to kidney failure varies significantly among ADPKD patients, even within families with the same PKD1 gene mutation 1. Based on the Mayo-Irazabal classification for age-adjusted assessment of ADPKD progression, the predicted incidence of kidney failure at 10 years ranges widely between 2.2 to 77.4% in older adults 2. Thus, factors secondary to the PKD1/2 gene mutation could regulate the rate of ADPKD progression to kidney failure. Here we tested if disruption of the circadian clock is a trigger for accelerated ADPKD progression.
Circadian rhythms are intrinsic, cyclical ∼24-hour oscillations in behavior and physiology that coordinate biological processes with the time of day. In mammals, circadian rhythms are regulated by cell-autonomous circadian clocks, which at the molecular level are comprised of clock proteins. The central clock, located in the suprachiasmatic nucleus in the hypothalamus, synchronizes peripheral clocks based on external timing cues. The canonical circadian clock involves heterodimerization of circadian proteins (transcription factors), CLOCK and BMAL1 (activators), which activate the gene expression of Per1,2,3 and Cry1,2 (repressors). PER and CRY proteins accumulate over time, repressing CLOCK and BMAL1 action, thereby inhibiting their own transcription. RORα/β and REV-ERBα/β function in an ancillary loop to mediate opposing action on BMAL1 transcription, thus driving rhythmic expression of BMAL1. These feedback loops result in ∼24-hour oscillations of clock gene expression, and rhythmic expression of ∼40% of protein-coding genes in mammals 3.
In mouse kidneys, ∼13% of all genes are expressed in a circadian manner 4, 5. Moreover, kidney functions including blood pressure rhythms, renal blood flow, glomerular filtration rate (GFR), and excretion of electrolytes such as sodium and potassium display circadian influence 6-9. Chronic disruption of the circadian clock (chronodisruption) is a well-known risk factor for accelerated ageing, memory loss, metabolic defects, obesity, cancer, cardiovascular defects, and metabolic syndrome in humans 10, 11. The role of chronodisruption in ADPKD pathogenesis is currently unclear.
BMAL1 is encoded by the Arntl (Aryl hydrocarbon receptor nuclear translocator-like protein-1) gene and is a core component of the molecular circadian clock. Together with other clock proteins, BMAL1 plays an essential role in cellular metabolic processes by regulating time-dependent utilization of nutrients and metabolites to support cell growth 12. Bmal1 is a master clock gene, and the only clock gene whose deletion alone leads to complete loss of circadian rhythms 13. Here we examined the effect of collecting duct-specific gene deletion of Bmal1 on disease progression in an ADPKD mouse model and in Pkd1 gene knockout cells.
Methods
Human tissue samples
De-identified human normal and APKD kidney tissues were “medical discard” nephrectomy samples from the Kansas PKD Research and Translational Center. IRB approval #5929, 8/25/2020. The time of tissue collection is unknown.
Bmal1 gene knockout in WT and Pkd1KO mIMCD3 cells using Crispr-Cas12a
Murine Bmal1 (Arntl gene, ID 11865) was disrupted in mIMCD3 cells using a microhomology end joining-based strategy and Cas12a 14. The TTTGGCGTATCTACCACAGGAACT sequence in exon 13 was targeted (PAM in bold). CRISPR sequence was, 5’/AltR1/rUrArArUrUrUrCrUrArCrUrCrUrUrGrUrArGrArUrGrCrGrUrArUrCrUrArCrCrArCrArGrGrA rArCrU/AltR2/-3’. PCR products that showed deletion on PAGE were sequenced after alkaline phosphatase treatment (ExoSAP, # 78200, Applied Biosystems, Foster City, CA). Bmal1KO clone and Pkd1Bmal1KO had frame-shift deletions of 10bp and 16bp respectively.
Mouse model
RC/RC (p.R3277C) mice 15 were bred with Bmal1f/f mice (B6.129S4(Cg)- Bmal1tm1Weit/J, carrying loxP sites flanking exon 8 of Bmal1 gene, Strain#007668, The Jackson Laboratory), and mice carrying Pkhd1cre 16. The resultant Pkd1RC/RC;Bmal1f/f;Pkhd1cre mice (double knockout DKO) were compared with Bmal1f/f (wild type-WT), Bmal1f/f;Pkhd1cre (Bmal1KO) and RC/RC littermates. Only male mice were used and sacrificed at 8 months of age, between 11:30 AM and 12:30 PM. Studies were approved by our Univerisity’s IACUC committee.
Recombination specificity in Bmal1KO and DKO mice
Bmal1 recombination was confirmed by PCR on genomic DNA of whole papilla from mouse kidneys 17, 18. PCR reaction products run on a 2% agarose gel yielded a 431-bp band for floxed Bmal1 allele and 572 bp band for recombination product.
Statistics
Values are expressed as mean±SEM for in vivo and mean±SD for in vitro studies. Data was analyzed by two-tailed unpaired T-test with Welch’s correction; or one-way ANOVA followed by Tukey’s multiple comparison test and two-way ANOVA followed by Bonferroni’s multiple comparison test using GraphPad Prism software (Version 9). P≤0.05 was considered significant.
Supplementary methods
We provide detailed methods for in vitro BrdU incorporation 19, 20, MTT assay 21, BUN 22 quantification of cyst 22, Western blot 23, 24, TUNEL staining 21, 25, immunohistochemistry/immunofluorescence (IHC/IF) staining 26 and quantitative real time PCR (QRTPCR) 27 in supplementary methods, and primer list for QRTPCR in Supplemental Table 1.
Results
1. Renal collecting duct specific Bmal1 gene deletion in ADPKD mouse kidneys
Examination of BMAL1 (ARNTL gene) expression in human nephrectomy ADPKD kidneys showed over 2-fold decrease in BMAL1 mRNA levels, compared to normal human kidneys (Fig 1A). Consistent with this, the kidney interactive transcriptomics database 28 of single cell RNA-sequencing showed reduced Bmal1 expression in most renal cells of human ADPKD kidneys compared to normal control kidneys (Fig 1B). To examine the effect of Pkd1 gene deletion on Bmal1, we examined its mRNA levels in control and Pkd1-/- mouse renal inner medullary collecting duct cells (mIMCD3) 29 and mouse proximal tubular cells 16, 30. Synchronized cells were collected every 4h, for 48h. The control Pkd1+/+ mIMCD3 cells and Pkd1+/- proximal tubule cells showed clear circadian oscillation of Bmal1 mRNA, with ∼20h interval between peaks, while Bmal1 mRNA levels and circadian oscillation were significantly dampened in the corresponding Pkd1-/- cells (Fig 1C, D).
(A) Bmal1 mRNA relative to 18S mRNA in normal human kidneys (NHK) and ADPKD kidneys measured by QRTPCR. ***P<0.001 by T-test. (B) Bmal1 gene expression in human normal control and ADPKD kidneys using kidney interactive transcriptomics database of single cell RNA-seq data. (C) Time course showing fold change in Bmal1 mRNA relative to 18S mRNA in mouse inner medullary collecting duct cells (mIMCD3) (n=4), and (D) proximal tubule cells (n=3). ***P<0.001 by 2-way ANOVA for C and D. Cells were synchronized using dexamethasone (100nM, for 2h) and released into media with 5% FBS. After 12h, cell lysates were collected every 4 hours for 48h. (E) Agarose gel showing recombination PCR products of genomic DNA obtained from mouse kidney. The 431 bp band denotes the floxed Bmal1 allele and the 572 bp band represents the recombination product. A single 431 bp band indicates absence of recombination. (F) Immunostaining for BMAL1 (brown) in kidney tissue sections of WT, Bmal1KO, RC/RC and RC/RC;Bmal1KO (DKO) mice (male, 8 months old) (Scale bar = 100μm). (G) QRTPCR analysis of mRNA levels of clock genes relative to 18S mRNA in whole kidney tissue lysates. *=vs WT, # =vs Bmal1KO and $=vs RC/RC. *P<0.05, **P<0.01, ***P<0.001 by One-way ANOVA. n=5 in WT and Bmal1KO, and n=7 in RC/RC and DKO.
To determine the role of BMAL1 in ADPKD progression, we chose the Pkd1 gene hypomorphic Pkd1RC/RC (RC/RC) mouse model of ADPKD. The RC/RC mice showed no significant difference in renal Bmal1 levels at 4 or 8 months of age compared to wild type (WT) littermates (Supplemental Fig 1A,B,C,D,E). It is possible that total loss of Pkd1 gene expression is required for Bmal1 expression to decrease. Renal collecting duct specific Bmal1 gene knockout in WT mice- Bmal1f/f;Pkhd1cre (Bmal1KO) and RC/RC mice-RC/RC;Bmal1f/f;Pkhd1cre (DKO) mice were generated. To test the effect of Bmal1 gene deletion, the DKO mice were compared with WT (Bmal1f/f), RC/RC and Bmal1KO littermates. Only male mice were used. We detected recombination of exon-8 of the Bmal1 gene (disrupted allele) in the Bmal1KO and DKO kidneys, while recombination was not detected in the WT or RC/RC kidneys (Figure 1E). BMAL1 expression was detected in the nuclei of tubular epithelial cells in WT and RC/RC mouse kidneys (Fig 1F). In the Bmal1KO and DKO kidneys, BMAL1 was not detected in some tubular or cyst epithelial cells (Fig 1F). Most large cysts in DKO kidneys showed Dolichos biflorus agglutinin (DBA) staining, implying collecting ducts (Supplemental Fig 1F).
To determine the effect of Bmal1 gene deletion on the circadian clock, we measured clock gene expression. When compared to RC/RC kidneys, the DKO kidneys showed significantly increased mRNA levels of Clock, Cry1, RevErba and Tef (Fig 1G). RC/RC kidneys showed significant decreases in Per1, Dbp and Tef, and increases in Clock mRNA levels when compared to WT kidneys (Fig 1G). Clock, Cry1 and RevErba genes were upregulated in both Bmal1KO and DKO kidneys compared to WT kidneys and RC/RC kidneys respectively (Fig 1G).
2. Renal collecting duct specific Bmal1 gene deletion increased cyst growth in ADPKD mouse kidneys
At 8 months of age, the DKO kidneys were larger and more cystic (Fig 2A, 2B), with significantly higher cyst number (Fig 2C), cystic index (Fig 2D), kidney weight (Fig 2E) and kidney to body weight ratio (Fig 2F) compared to RC/RC kidneys. No significant changes were observed in the Bmal1KO kidneys compared to WT kidneys (Fig 2A,E,F). Morever, no change was observed in the BUN (blood urea nitrogen) levels in DKO mice when compared to RC/RC, WT and BMAL1KO mice (Supplementary Fig 1G). The results show that gene deletion of Bmal1 in collecting ducts accelerates renal cyst growth in the ADPKD male mouse kidney but has no overt effect on WT kidneys.
(A) Images of H&E staining of kidney sections and (B) magnified images. Scale bar = 1 mm for A and 400µM for B. (C) Cyst number, (D) cystic index, (E) total kidney weight, and (F) two kidney to body-weight ratio. **P<0.01, ***P<0.001 by T-test for C and D, and One-way ANOVA for E and F.
3. Bmal1 gene deletion increased cell turnover and fibrosis in ADPKD mouse kidneys
To examine the effect of Bmal1 gene deletion on cell proliferation, kidney tissue sections were immunostained for Ki67 and quantified. DKO mouse kidneys showed significantly increased Ki67 stained nuclei compared to RC/RC kidneys (Fig 3A,B). Immunoblot of kidney tissue lysate showed increased cyclin D1 and reduced P53 protein levels in DKO kidneys compared to RC/RC kidneys (Fig 3C,D,E). In DKO kidneys, pERK/ERK, pCREB/CREB, and pS6/S6 ratios were significantly increased, suggesting higher activity compared to RC/RC kidneys (Fig 3C,F,G,H). No change was observed in c-Myc levels (Supplemental Fig 2A). No significant difference was observed between Bmal1KO and WT kidneys in Ki67 staining (Supplemental Fig 2B, C), or in cyclin D1, pERK/ERK or pCREB/CREB protein levels (Supplemental Fig 2D,E). TUNEL staining of kidney tissue sections (Fig 4A,B) and mRNA levels of renal injury markers, Kim1 and Ngal1 (Fig 4C,D) were significantly increased in DKO kidneys compared to RC/RC kidneys. The increased cell proliferation and apoptosis indicate high turnover of cells in the DKO kidneys.
(A) Immunostaining for Ki-67 in mouse kidney tissue sections (brown nuclear staining denotes Ki-67). Scale bar = 100 µm. (B) Quantitation of Ki-67-stained nuclei per kidney section. (C) Immunoblot on kidney tissue lysate and quantitation of band density for (D) Cyclin D1, (E) P53, (F) pERK/ERK ratio, (G) pCREB/CREB ratio, (H) pS6/S6 ratio. *P<0.05, **P<0.01, ***P<0.001 by T-test.
(A) TUNEL staining (green) (Scale bar 400µm), and immunostaining for αSMA (red), and Collagen1 (green) (Scale bar 100µm) in mouse kidney tissue sections. (B) Quantitation of TUNEL staining per kidney tissue section. (C) QRTPCR for Kim1, (D) Ngal1 (E) αSMA, (F) Collagen 1a1, (G) Collagen 3a1, and (H) Fibronectin in kidney tissue lysates. *P<0.05, **P<0.01, ***P<0.001 by T-test in B and One-way ANOVA in C, D, E, F, G and H.
DKO mouse kidneys also showed increased renal fibrosis. Immunostaining for the myofibroblast marker αSMA and the extracellular matrix (ECM) protein, collagen1 were increased in the DKO kidneys compared to RC/RC kidneys (Fig 4A). Moreover, mRNA levels of ECM proteins, collagen1a, collagen3a1 and fibronectin were significantly higher in the DKO kidneys compared to RC/RC kidneys (Fig 4E, F, G, H).
4. Bmal1 gene knockout altered the expression of circadian clock genes and gene targets of BMAL1 in mIMCD3 cells
To better understand the mechanism by which Bmal1 gene knockout in the renal collecting ducts accelerates disease progression in RC/RC male mice, we deleted Bmal1 gene in WT and Pkd1 gene knockout mIMCD3 cell lines using Crispr-Cas12a (Fig 5A, Supplemental Fig 3). WT, Bmal1KO, Pkd1KO, and Pkd1Bmal1KO cells were compared. Consistent with results in mouse kidneys, cell proliferation was significantly increased in Pkd1Bmal1KO cells compared to Pkd1KO cells, while Bmal1KO cells showed no significant difference compared to WT cells (Fig 5B). Examination of circadian clock gene expression showed significantly reduced Per1, Per2, RevErba and Rora, and increased Cry1 mRNA levels in Pkd1Bmal1KO cells when compared to Pkd1KO cells (Fig 5C). The Pkd1KO cells showed significant increase in Per1, Per2, RevErba, Rora and Hlf mRNA levels compared to WT cells (Fig 5C). Significant difference in Per2, Cry1 and RevErba, members of the repressor arm of the clock were common to both Bmal1KO and Pkd1Bmal1KO cells when compared to WT and Pkd1KO cells respectively (Fig 5C). Rorb and Dbp were not detected.
A) Western blot shows WT and Bmal1KO clones in WT and Pkd1 gene knockout mIMCD3 cells. WT-Bmal1KO clone #4 and Pkd1Bmal1KO clone# 3 were used for further studies. (B) Cell proliferation assessed by BrdU incorporation assay. ***P<0.001 by One-way ANOVA. (C) QRTPCR analysis for mRNA levels of clock genes relative to 18S mRNA in cells. (D) mRNA levels of genes that regulate fatty acid oxidation, (E) lipogenesis, and (F) cholesterol synthesis relative to 18S mRNA. * = vs WT, # = vs Bmal1KO and $ = vs Pkd1KO. *P<0.05, **P<0.01, ***P<0.001 by One-way ANOVA. n=4 in C,D,E and F.
We also examined BMAL1’s prospective gene targets. BMAL1, a transcription factor binds to E-box elements on gene promoters 31. The Pkd1 gene promoter contains an E-box element 32, although it is unknown if Bmal1 is a transcriptional regulator of Pkd1. Contrary to our expectation, the Bmal1KO cells had significantly increased Pkd1 mRNA levels when compared to WT cells (Supplemental Fig 4A). We also examined gene expression of Ets1, Sp1, Rxra, Cebpa, Hnf4a, Fli1, Hnf1a, Hnf1b, Hnf4a, Pax8 and Tead4, based on a previous Chip-seq analysis 33 that identified these genes as targets of BMAL1 in WT mouse kidneys 9. Of the genes detected, Cebpa, Ets1 and Hnf1b were significantly reduced in DKO kidneys and Pkd1Bmal1KO cells when compared to RC/RC kidneys and Pkd1KO cells respectively (Supplemental Fig 4A,B).
5. Bmal1 gene knockout altered the lipid metabolism-related gene expression in Pkd1Bmal1KO mIMCD3 cells and DKO mouse kidneys
Reduced fatty-acid oxidation and increased fatty acid synthesis enhance cyst growth in ADPKD 34-36. Since Bmal1 is known to regulate fatty-acid oxidation, lipogenesis and cholesterol metabolism 37-41, we examined genes regulating these processes. In Pkd1Bmal1KO cells, mRNA levels of fatty-acid oxidation-related genes, Acox1, Acox2, Acadsb, Acadvl, Cpt1a, Cypt2, Fabp3, and Ppara showed significant decreases, while Pgc1a was increased compared to Pkd1KO cells (Fig 5D). Pkd1KO cells showed significantly reduced Acox2, Ppara and Pgc1a mRNA levels compared to WT cells (Fig 5D). Acox2, Acadsb, Acadvl and Cpt1a are common genes that were significantly decreased in both Bmal1KO and Pkd1Bmal1KO cells compared to WT and Pkd1KO cells respectively (Fig 5D).
Lipogenesis associated genes, Acaca, Acss2, Dagla, Daglb, Fads1, Fads3, Scd1, and Srebp1c were significantly increased in Pkd1Bmal1KO cells compared to Pkd1KO cells (Fig 5E). Acss2, Fads3 and Srebp1c are common genes that were upregulated in Pkd1Bmal1KO cells, but downregulated in the Bmal1KO cells compared to WT and Pkd1KO cells respectively (Fig 5E). Compared to WT cells, Pkd1KO cells showed significantly increased Srebp1c, and reduced Dagla mRNA levels (Fig 5E). Genes related to cholesterol synthesis including Hmgcs1, Hmgcr, Mvk and Mvd were significantly increased in Pkd1Bmal1KO cells when compared to Pkd1KO cells (Fig 5F). When compared to WT cells, the Pkd1KO cells showed significantly reduced levels of Hmgcs1, Hmgcsr, Mvk, Pmvk, Fdps and Mvd (Fig 5F). Other fatty acid oxidation and lipogenesis-related genes that remained unchanged are shown in Supplemental Fig 5A,B.
Examination of lipogenesis-related genes in our mouse study showed significantly increased mRNA levels of Acaca, Dagla, Fads1, Scd1, Srebp1c, Mag1 and Lpcat in DKO mouse kidneys compared to RC/RC kidneys (Fig 6A-G). Furthermore, tissue triglyceride levels showed a 2-fold increase in DKO mouse kidneys when compared to RC/RC mouse kidneys (Figure 6H). To determine if elevated lipogenesis contributes to increased cell proliferation, we inhibited lipogenesis in Pkd1Bmal1KO cells. Firsocostat (GS-0976) is a pharmacological inhibitor of acetyl-CoA carboxylase (ACACA & ACACB), a lipogenesis regulating enzyme 42. We found ∼2 fold increase in Acaca gene in both DKO mouse kidneys and in Pkd1Bmal1KO cells when compared to RC/RC mouse and Pkd1KO cells respectively (Fig 5E & 6A). Firsocostat (10nM) treatment for 24h significantly reduced cell proliferation in Pkd1Bmal1KO cells, indicated by significant reduction in BrdU incorporation compared to vehicle treated cells, while Pkd1KO cells remained unaffected (Fig 6I). Upon extended Firsocostat treatment for 48h, cell viability indicated by MTT assay was significantly reduced in both Pkd1KO and Pkd1Bmal1KO cells (Figure 6J). The results suggest that the increased cell proliferation in Pkd1Bmal1KO cells can be rescued by inhibition of lipogenesis.
The mRNA levels of genes that regulate lipogenesis, (A) Acaca, (B) Dagla, (C) Fads1, (D) Scd1, (E) Srebp1c, (F) Magl, (G) Lpcat, relative to 18S. (H) Mouse kidney tissue triglyceride levels. (I) Pkd1KO and Pkd1Bmal1KO IMCD3 cells were treated with Firsocostat (Fir), (10nM) or vehicle for 24h. Cell proliferation (% BrDU/ DAPI) shown. (J) MTT assay for cell viability of Pkd1KO and Pkd1Bmal1KO IMCD3 cells treated with Firsocostat (10nM) or vehicle for 48h. (K) Cholesterol metabolism related genes Mvk, (L) Pmvk, and (M) Fdps. *P<0.05, **P<0.01, ***P<0.001 by One-way ANOVA.
Cholesterol metabolism-related Mvk, Pmvk and Fdps6 genes were also significantly increased in DKO kidneys compared to RC/RC kidneys (Fig 6K,L,M). Fatty-acid oxidation-related genes were not significantly different between DKO and RC/RC kidneys, although RC/RC kidneys showed significantly reduced expression of Acox1, Acadsb, Ppara, Acat1, Acat2 and Acads compared to WT kidneys (Supplemental Fig 6 A-N). Unchanged lipogenesis and cholesterol synthesis-related genes in mouse kidneys are shown in Supplemental Fig 7A-J
Discussion
Our studies show that disruption of the circadian clock stimulates accelerated cyst growth in ADPKD. We found reduced Bmal1 gene expression in human nephrectomy ADPKD kidneys, and in Pkd1 gene knockout mIMCD3 and proximal tubule cells, but not in Pkd1 gene hypomorphic RC/RC mouse kidneys. Renal collecting duct specific gene deletion of Bmal1 in RC/RC male mice increased renal cell proliferation, apoptosis, fibrosis, and cyst growth. Bmal1 gene deletion in RC/RC mouse kidneys significantly altered the expression of other clock genes, reduced the expression of BMAL1 target genes and increased lipogenesis. In vitro, proliferation of Pkd1Bmal1KO mIMCD3 cells was higher than Pkd1KO mIMCD3 cells, which could be rescued by inhibition of lipogenesis.
Our finding that Bmal1 gene knockout in the renal collecting ducts increased cyst growth in ADPKD mice is significant because it shows for the first time that chronodisruption accelerates ADPKD progression. Chronodisruption is disease inducing, and known to directly drive metabolic dysfunction, cause type-2 diabetes, obesity, various cancers and neurodegenerative conditions, and adversely affect fetal development 43-47. Since chronodisruption by alterations in behavior or the environment are systemic effects, we used renal collecting duct specific Bmal1 gene deletion as a model for genetic disruption of the molecular clock so as to focus on the kidney. In our study, Bmal1 gene knockout altered the expression of clock genes including Cry1, Per1, Per2, and RevErba that are in the negative arm of the clock, and are known to be regulated by Bmal1.
We found that collecting duct-specific gene deletion of Bmal1 dramatically increased cyst growth, cell proliferation and fibrosis in the male RC/RC mouse model of ADPKD. In prior studies, Bmal1 gene knockout reduced locomotor activity, and induced tendon calcification, sarcopenia, hypersensitivity to insulin-shock, fragmented sleep, and rapid ageing in mice 48, 49. In the kidneys, nephron-specific Bmal1 gene deletion caused Na+ retention in response to a K+- restricted diet 50, lowered blood pressure 50-52 and altered the urinary metabolome in mice 48. However, the effect of altering Bmal1 gene expression on kidney disease progression is context dependent, as shown in experimental mouse models 53. In a mouse model of AKI, shRNA mediated Bmal1 gene knockout alleviated AKI, while adenoviral overexpression of Bmal1 accelerated renal injury 54. Similarly in unilateral ureteral obstruction model of CKD, global gene knockout of Bmal1 reduced fibrosis 55 or had no effect on fibrosis or disease progression 56. In contrast to the above mentioned studies, and supporting our results, renal tubule specific Bmal1 gene knockout worsened hyperglycemia, and accelerated renal hypertrophy in mice with streptozotocin-induced type-1 diabetes, compared to similarly treated WT mice 57. Moreover, global or proximal tubule specific gene knockout of Bmal1 worsened disease progression in the adenine induced model of CKD in mice 58,59. We found that in the presence of Pkd1 gene mutation, Bmal1 gene knockout in renal collecting ducts increased cell proliferation indicated by increase in Ki-67 immunostaining, and pERK/ERK1/2, pCREB/CREB, pS6/S6 and CyclinD1 levels, all factors known to increase cell proliferation in ADPKD kidneys 25. Our data is consistent with previous studies where Bmal1 gene deletion or inactivation increases the proliferation rate of cancer cells 60, 61.
BMAL1 binds to E-box elements on promoters of rhythmically and non-rhythmically expressed genes 31. Over 7 million E-box elements exist in the mouse genome, ∼ 0.7% of which are bound by CLOCK-BMAL1 or NPAS2-BMAL1 dimers in mouse peripheral tissues 33, 62. Based on this published data, we examined 12 prospective gene targets of BMAL1 and identified Ets1, Hnf1b and Cebpa to be significantly reduced in Bmal1 deficient Pkd1KO cells and RC/RC mouse kidneys. C/EBPα plays an essential role in embryonic development and supresses cell proliferation by stabilizing p21 and inhibiting cdk2, cdk4 and E2F 63. Inhibiting C/EBPα contributes to tumor progression in liver, pancreatic, breast and lung cancers 63. Similarly, ETS1 and HNF1 transcription factors have oncogenic or tumor supressor roles in a context-dependent manner in human cancers 64, 65. HNF1β is known to regulate Pkd2 and Pkhd1 gene expression, while and ETS1 regulates Pkd1 gene expression 32, 66, 67.
Fatty-acid metabolism is impaired in ADPKD and is thought to enhance renal cyst growth 34-36. A previous study showed reduced fatty-acid oxidation-related gene expression in ADPKD mouse kidneys, and treatment with fenofibrate, a PPARα agonist inhibited cyst growth in these mice 34. Similarly, pups born from female mice fed with a high fat diet had higher cystic index compared to those fed with low fat diet 35. In vitro studies have also shown that Pkd1 gene mutation leads to defects in fatty-acid oxidation 35, 36. Consistent with these studies, we found fatty-acid oxidation-related genes to be significantly reduced in Pkd1KO cells compared to WT cells, and in RC/RC mouse kidneys compared to WT mouse kidneys. BMAL1 is known to regulate fatty-acid oxidation 37-41. Bignon et al 41 showed that renal tubule specific Bmal1 gene deletion in WT mice significantly reduced mRNA and protein expression of ACADS, ACADM, ACADL, and ACADVL which catalyze the rate-limiting step of β-oxidation, suggesting disrupted fatty-acid oxidation. However in our study, renal collecting duct specific Bmal1 gene deletion in the Bmal1KO mice or DKO mice did not alter fatty-acid oxidation-related genes when compared to WT or RC/RC mouse kidneys. The reason for this disparity is presently unclear. Our results from the mIMCD3 cells were different, with 4 of the 14 fatty acid oxidation-related genes showing significant decreases in the Bmal1KO and Pkd1Bmal1KO cells compared to WT and Pkd1KO cells respectively.
Unlike fatty-acid oxidation, our in vitro and in vivo results consistently showed increased lipogenesis and cholesterol metabolism-related genes in Bmal1 gene deleted cells and mouse kidneys. The mRNA levels of 7 out of the 14 lipogenesis-related genes tested in Pkd1Bmal1KO cells and in DKO kidneys showed significant increase compared to WT cells and RC/RC kidneys. Of the lipogenesis-related genes, Acaca, Fads1, Scd1 and Srebp1c were upregulated in both DKO mouse kidneys and in Pkd1Bmal1KO cells, suggesting regulation by Bmal1. Tissue triglyceride levels were also higher in DKO mouse kidneys compared to RC/RC mouse kidneys, supporting increased lipogenesis. Importantly, pharmachological inhibition of lipogenesis rescued the increased cell proliferation in Pkd1Bmal1KO cells. The Pkd1Bmal1KO cells and DKO kidneys showed significant increases in several cholesterol synthesis-related genes. Mvk was a frequently upregulated genes in the cells and kidneys. Renal collecting duct specific Bmal1 gene deletion in WT mice (Bmal1KO mice) did not alter any lipogenesis, cholesterol synthesis or fatty-acid oxidation-related genes when compared to WT mouse kidneys, suggesting that Bmal1 gene knockout in the ADPKD kidneys is important for these changes.
Although Bmal1 gene deletion altered the expression of multiple clock genes, BMAL1 regulated genes and lipid metabolism-related genes in the mIMCD3 cells and mouse kidneys, their changes were not always similar in the cells and kidneys. This could be because Bmal1 is deleted only in the collecting duct cells of the mouse kidneys, while the IMCD3 cells had complete gene deletion of Bmal1. Moreover, the mice were all sacrificed around 12 PM, while the mIMCD3 cells were collected 24h after synchronization in vitro.
While patients with ADPKD often develop non-dipping hypertension, polyuria, polydipsia, and nocturnal polyuria 68-70, it is not known if chronodisruption is a modifier of ADPKD progression. Our results show that Bmal1 gene knockout in the renal collecting ducts disrupt the circadian clock and expression of key elements of cell proliferation, fibrosis and lipid metabolism, thereby accelerating cyst growth in ADPKD. It is possible, therefore, that chronodisruption in individuals with ADPKD might be a cause of, or contribute to more rapid disease progression. Environmental or behavioral factors such as mistimed feeding, shift work or jet lag are known to induce chronodisruption 71. Shift work for instance, reduces the estimated GFR in humans and is a risk factor for CKD 72-74, cardiovascular disease 75, multiple cancers 76 and metabolic disorders 77. It is unclear if environmental or behavioral factors have any effect on ADPKD progression. Future studies could examine if poor quality or duration of sleep or mistimed eating habits act as chronodisruptors and accelerate ADPKD progression, and if methods to re-set the clock using pharmacological approaches, or behavioral modifications can slow progression of ADPKD.
Disclosure Statement
The authors have declared that no conflict of interest exists.
Funding
This study was supported by, Department of Defence Discovery grant # W81XWH2310011, National Institutes of Health grants # R56DK128962-01 and # R01DK135308-01 to RR.
Author contributions
RR conceptualized and designed studies, analyzed the results, and wrote the paper. AJ performed experiments, analyzed results and wrote parts of the paper. CJW, VR, MMS and NSP performed some studies, analyzed results and edited the paper. MLG conceptualized some studies and reviewed the paper. All authors read, edited and approved the paper.
Data sharing statements
No proteomics, transcriptomics or GWAS data was generated in this paper. All in vivo studies using mouse models were done, and tissues analyzed at the University of Kansas Medical Center. All data that was generated are provided in the figures in the main text or in Supplementary Material. The datasets used to examine Bmal1 expression levels in human samples are openly available on the kidney interactive transcriptomics database (https://humphreyslab.com/SingleCell/) and permission was obtained to use it. If any added information is required to re-examine or analyze the data shown in this paper, they will be readily available from the lead author. All requests for data will be processed based on institutional policies for noncommercial research purposes. Data sharing could require a data transfer agreement as determined by the University of Kansas Medical Center’s legal department.
Acknowledgments
We thank Dr. Peter Harris and Dr. Katharina Hopp for providing Pkd1RC/RC mice; Dr. Stefan Somlo and Dr. Yiqiang Cai (George M. O’Brien Kidney Center, Yale University, New Haven, CT) for Pkd1KO and control proximal tubular and inner medullary collecting ducts cells; Dr. Peter Igarashi for the Pkhd1cre mice and Dr. Darren Wallace for human kidney tissues (NIDDK PKD-Research Resource Consortium and the Kansas PKD Research and Translational Core Center-U54 DK126126).
Footnotes
Figure 6H, 6I, 6J added. Parts of discussion updated