Promoter methylation confers kidney-specific expression of the Klotho gene

Abstract

The aging suppressor geneKlotho is predominantly expressed in the kidney irrespective of species. Because Klotho protein is an essential component of an endocrine axis that regulates renal phosphate handling, the kidney-specific expression is biologically relevant; however, little is known about its underlying mechanisms. Here we provide in vitro and in vivo evidence indicating that promoter methylation restricts the expression of the Klotho gene in the kidney. Based on evolutionary conservation and histone methylation patterns, the region up to −1200 bp was defined as a major promoter element of the human Klotho gene. This region displayed promoter activity equally in Klotho-expressing and -nonexpressing cells in transient reporter assays, but the activity was reduced to ∼20% when the constructs were integrated into the chromatin in the latter. Both endogenous and transfected Klotho promoters were 30–40% methylated in Klotho-nonexpressing cells, but unmethylated in Klotho-expressing renal tubular cells. DNA demethylating agents increased Klotho expression 1.5- to 3.0-fold in nonexpressing cells and restored the activity of silenced reporter constructs. Finally, we demonstrated that a severe hypomorphic allele of Klotho had aberrant CpG methylation in kl/kl mice. These findings might be useful in therapeutic intervention for accelerated aging and several complications caused by Klotho down-regulation.—Azuma, M., Koyama, D., Kikuchi, J., Yoshizawa, H., Thasinas, D., Shiizaki, K., Kuro-o, M., Furukawa, Y., Kusano, E. Promoter methylation confers kidney-specific expression of the Klotho gene.

The Klotho gene was originally identified as a gene whose inactivation caused premature aging phenotypes, such as vascular calcification, neural degeneration, impaired hearing, skin and muscle atrophy, osteoporosis, pulmonary emphysema, and hypogonadism, and shortened life span in mice (1). Conversely, Klotho overexpression significantly slows down the aging process via conferring resistance to oxidative stress and extends the life span of mice (2, 3). In humans, it has been reported that allelic variance and single-nucleotide polymorphisms of the KLOTHO gene are correlated with longevity (4–6), metabolic activity of lipid and glucose (5), and the incidence of aging-related disorders such as osteoporosis, coronary artery disease, cognitive impairment, and hypertension in various populations (7, 8). Moreover, Klotho expression decreases with age in the brain and other organs in nonhuman primates and rodents (9, 10). These findings clearly indicate that Klotho is a bona fide aging suppressor.

The Klotho gene encodes a type I transmembrane protein with a short cytoplasmic domain, which is expressed predominantly in the kidney (1, 11, 12). Chronic kidney disease causes down-regulation of Klotho expression (12–14), which may underlie accelerated aging and serious complications, such as arteriosclerosis, extensive cardiovascular calcification, and hyperphosphatemia, in patients with chronic renal failure (14–17). The kidney-predominant expression is consistent with the recent finding that Klotho protein forms a binary complex with fibroblast growth factor (FGF) receptors on distal tubules and converts them into high-affinity receptors for FGF23, thereby acting as a principal mediator of the homeostasis of inorganic phosphate (18–20). Several investigations point to the presence of environmental cues and factors that affect renal function and Klotho expression concomitantly. Those include the amounts of dietary phosphate, vitamin D₃, ischemia, iron overload, oxidative stress, angiotensin II, statins, and Rho kinase inhibitors (3, 21–23). Although these findings may explain the biological relevance of kidney-specific expression of Klotho, little is known about its underlying mechanisms at molecular levels. Elucidation of the molecular mechanisms of Klotho expression would be useful in therapeutic intervention for accelerated aging in patients with chronic renal failure and under other pathological and even physiological conditions.

Promoter methylation is one of the fundamental mechanisms that render tissue-specific expression of genes in higher eukaryotes (24). In the mammalian genome, DNA methylation occurs almost exclusively at the 5-position of cytosine in CpG dinucleotides, which are contiguously clustered in the regions called CpG islands (25). Transcription regulatory units are embedded in CpG islands in up to 70% of mammalian genes, especially housekeeping and tissue-restricted genes (26, 27). In general, dense methylation in CpG islands results in transcriptional silencing of downstream genes via recruitment of transcriptional repressor complexes composed of methyl-CpG binding domain (MBD) proteins and histone deacetylases (HDACs) and/or Polycomb complexes (28, 29). Methylation-dependent regulation is observed in some tissue- and developmental stage-specific genes, such as neuron-specific genes (26, 27). Promoter methylation should be highly context-dependent, and its importance is underscored by the recent finding that induced pluripotent stem (iPS) cells harbor a residual DNA methylation signature of their somatic cells of origin, which influences the differentiation potential and propensity of individual iPS cells (30). In light of these observations, we hypothesized that promoter methylation restricts Klotho gene expression in the kidney and substantiated this hypothesis using the combination of system biology and biochemical approaches.

MATERIALS AND METHODS

Bioinformatics

We used the public databases ArrayExpress (http://www.ebi.ac.uk/arrayexpress/), VISTA web server (http://genome.lbl.gov/vista/index.shtml), University of California–Santa Cruz Genome Bioinformatics (http://genome.ucsc.edu/index.html), and the U.S. National Center for Biotechnology Information (NCBI) genome database (http://www.ncbi.nlm.nih.gov/) to obtain information about Klotho gene expression and promoter structure.

Cells and cell culture

We purchased normal human renal proximal tubular epithelial cells (RPTECs), primary human mesangial cells, and human embryonic kidney fibroblasts from Lonza (Walkersville, MD, USA), Cambrex BioScience (Walkersville, MD, USA), and the Applied Cell Biology Research Institute (Kirkland, WA, USA), respectively. RPTECs were cultured in renal epithelial basal medium (Lonza) supplemented with 0.5% FCS, epithelial growth factor, insulin, transferrin, hydrocortisone, epinephrine, and triiodothyronine. Mesangial cells were maintained as described previously (31). Fibroblasts were cultured in fibroblast culture medium containing 10% FCS and attachment factors (Cell Systems Corp., Kirkland, WA, USA). We obtained human renal proximal tubular cell line HK-2, human embryonic kidney cell line HEK293, human endometrial carcinoma cell line HeLa, human chronic myeloid leukemia cell line K562, and nontumorigenic breast epithelial cell line MCF10A from the American Type Culture Collection (Manassas, VA, USA) and maintained them according to the manufacturer's instructions. Normal T lymphocytes were isolated from the peripheral blood of healthy volunteers as described previously (32).

To prepare the supernatants, we cultured HK-2 cells and RPTECs in appropriate media excluding FCS for 24 h. Proteins in the supernatants were concentrated by the addition of trichloroacetic acid to a final concentration of 10%, followed by centrifugation at 15,000 rpm for 15 min. The resulting pellets were washed with ethanol, dried, and reconstructed in 0.02 vol of 1× SDS-PAGE sample buffer in PBS in preparation for immunoblotting.

RT-PCR

We detected the expression levels of both membrane and secreted forms of human Klotho mRNA by a single RT-PCR using the following primer pair: sense 5′-ACAGAGGTTACAGCATCAGGCG-3′ and antisense 5′-TTCCTCTTCTTGGTCACAACC-3′ (19). We also used another sense primer (5′-TAGATACCACTCTGTCTCAGTTTACCG-3′) to detect the membrane form exclusively. Amplification was done with 35 cycles of denaturation at 94°C for 1 min, annealing at 62.5°C for 1 min, and polymerization at 72°C for 2 min. We simultaneously amplified GAPDH mRNA to verify the equal loading and integrity of samples (primer sequences: sense 5′-CCACCCATGGCAAATTCCATGGCA-3′ and antisense 5′-TCTAGACGGCAGGTCAGGTCCACC-3′) (31).

We performed real-time quantitative RT-PCR using TaqMan Gene Expression Assays (Hs00183100 for Klotho and Hs01922876 for GAPDH) and TaqMan Fast Universal PCR Master Mix as primers and buffer, respectively, on the StepOnePlus Real-Time PCR System (Applied Biosystems, Warrington, UK). The data were quantified with the 2^−ΔΔCt method using GAPDH as a reference (33).

Immunoblotting

We used a rat monoclonal antibody against human Klotho (clone KM2076; provided by Kyowa Hakko Kirin Co., Tokyo, Japan), which was generated with a part of human Klotho protein (aa 55-261) as an immunogen and recognizes only a 130-kDa membrane form (34), and a rabbit polyclonal antibody against GAPDH (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for immunoblotting.

Promoter isolation and reporter assays

We isolated a 5′-upstream region of the human Klotho gene (−1261 to +88) by BglI digestion of the bacterial artificial chromosome clone RP11-270H22 (BACPAC Resources, Children's Hospital Oakland Research Institute, Oakland, CA, USA) and inserted the fragment into pGL4.14 firefly luciferase vector (Promega, Madison, WI, USA) to construct a reporter plasmid. HK-2 and HeLa cells were transfected with the reporter plasmid along with pGL4.73 Renilla luciferase vector (Promega), which served as a positive control to determine transfection efficiency. After 48 h, we harvested small fractions of cells to measure firefly and Renilla luciferase activities discriminatorily using the Dual-Luciferase Reporter Assay System (Promega) (32). The remainder of the cells were continuously cultured in the presence of hygromycin B to establish stable transformants harboring the reporter constructs.

Methylation-specific PCR and bisulfite sequencing

We analyzed the methylation status of 5′-flanking regions of the human Klotho gene using methylation-specific PCR (MSP) and bisulfite sequencing. In brief, DNA was isolated from cells with phenol/chloroform extraction, denatured with NaOH, and incubated at 50°C for 16 h in a buffer containing 3.1 M sodium bisulfite and 0.5 mM hydroquinone (35). A proximal GC-rich region of the Klotho gene was amplified in the presence of the outer primer set [sense 5′-GAGTGGGAGAAAAGTGAGAG-3′ (−476 to −457 of the ATG start coden) and antisense 5′-CACCTATTTCTCCCAACTCCC-3′ (−278 to −258)] with 40 cycles of denaturation at 94°C for 1 min, annealing at 45°C for 1 min, and polymerization at 72°C for 2 min, followed by amplification with 30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and polymerization at 72°C for 2 min in the presence of the inner primers to discriminate CpG methylation [sense, 5′-TTTTTTTAGCGGCGCGTTTCGTTAGGGTTC-3′ (−450 to −418) and antisense, 5′-CGCGCGTCCTCTAAAAACAACCCTAAAACG-3′ (−359 to −331)] and demethylation [sense 5′-TTTTTTTAGTGGTGTGTTTTGTTAGGGTTT-3′ (−450 to −418) and antisense 5′-CACACATCCTCTAAAAACAACCCTAAAACA-3′ (−359 to −331)]. As a control, we performed MSP for p16/Ink4a promoter using a MethylDetector Kit (Active Motif, Carlsbad, CA, USA). The PCR products were subcloned into pCRII vector (Invitrogen, Carlsbad, CA, USA) for nucleotide sequencing. Methylation-specific primers were designed using the MethPrimer program (http://www.urogene.org/methprimer/index1.html).

Methylation analyses of Klotho-deficient (kl/kl) mice

Genomic DNA was isolated from kidney and liver of kl/kl mice and their littermates with normal Klotho alleles and treated with sodium bisulfite as described above. We performed pyrosequencing to analyze the methylation status of 3 different regions of the Klotho gene: transgene (−8777 to −8613), intermediate segments (−4251 to −4072), and proximal region (−336 to −193), using the PCR primers and conditions described in Supplemental Table S1. The PCR products were subjected to DNA sequencing with or without subcloning into pCRII vector to obtain single DNA clones.

RESULTS

The Klotho gene is preferentially expressed in renal tubular cells irrespective of species

First, we performed expression profiling of the Klotho gene using the data collected from previous literature (1, 11, 12) and the public database ArrayExpress. As illustrated in Fig. 1A, Klotho mRNA is strongly expressed in the kidney in all three species for which data are available, whereas Klotho expression is also detectable in the placenta, prostate, and intestines in humans; in the brain, placenta, and reproductive organs in mice; and in the brain, lung, and ovary in rats to a lesser strength than in the kidney. In human and monkey brains, the strength of Klotho expression varies regionally: very strong in the hippocampus, choroid plexus and caudate nucleus, and weak or absent in the cerebellum and hypothalamus. These data confirm the notion that the kidney is the major site of Klotho production irrespective of species. Next, we investigated which cell types express Klotho in the human kidney. To this end, we subjected HEK293 embryonic kidney cells, HK-2 proximal tubular cells, RPTECs, mesangial cells, embryonic kidney fibroblasts, and 4 types of nonrenal epithelial and mesenchymal cells (HeLa cells, MCF10A cells, K562 cells, and T lymphocytes) to semiquantitative RT-PCR (Fig. 1B and Supplemental Fig. S1A) and real-time quantitative RT-PCR (Fig. 1C) for Klotho mRNA expression. RT-PCR analyses clearly indicated that Klotho was preferentially expressed in renal tubular cells but not in the other cell types examined, except for marginal expression in mesangial cells. This observation is compatible with the previous results of immunohistochemical staining of mouse and human kidneys, which yielded positive signals in some parts of the cortical renal tubules (1, 13). We then attempted to verify tubular cell-specific expression of Klotho at protein levels. Immunoblotting revealed that Klotho protein was abundantly expressed in HK-2 cells but not in HEK293 cells, mesangial cells, fibroblasts, and nonrenal epithelial and mesenchymal cells (Fig. 1D and Supplemental Fig. S1B). Unexpectedly, RPTECs expressed only a tiny amount of Klotho protein in immunoblot analysis of whole-cell lysates. We reasoned that this was due to the ability of RPTECs to secrete Klotho protein or shed it from the membrane (36, 37). To validate this assumption, we performed immunoblotting of culture supernatants of HK-2 cells and RPTECs. As shown in Fig. 1E, almost equal amounts of Klotho protein were detected in the supernatant and cell pellet fractions of RPTECs, whereas <10% of Klotho protein was secreted or shed in the supernatant of HK-2 cells. The mechanism underlying the difference in the ability of secreting or shedding Klotho protein between RPTECs and HK-2 cells is at present unknown.

Figure 1.

Predominant expression of the Klotho gene and its product in the kidney. A) Schematic view of tissue distribution of Klotho mRNA expression based on data from the literature (1, 13, 14) and the public database ArrayExpress (http://www.ebi.ac.uk/arrayexpress/). Red, yellow, and green indicate high, low, and no mRNA expression, respectively. Black indicates data not available. B) Total cellular RNA was isolated from the indicated cells and subjected to RT-PCR for the expression of Klotho and GAPDH (internal control). We used a primer pair that yields secreted and membranous forms of Klotho transcripts (354- and 308-bp bands, respectively) in a single PCR (19). The data shown are representative of 5 independent experiments. C) RNA samples prepared as described above were subjected to real-time quantitative RT-PCR using commercial probe sets. The data were quantified with the 2^−ΔΔCt method using simultaneously amplified GAPDH as a reference. The y axis indicates relative Klotho mRNA expression in each cell type, with the abundance in HK-2 cells set at 100 (%). The means ± sd (bars) of 5 independent experiments are shown. n.d., not detectable as defined by values of C_t > 30. D) Whole-cell lysates were prepared from the indicated cells and subjected to immunoblotting using anti-Klotho antibody KM2076 (34). The membrane was reblotted with anti-GAPDH antibody to serve as a loading control. E) Protein samples were prepared from the culture supernatants and simultaneously collected cell pellets as described in Materials and Methods and subjected to immunoblotting using anti-Klotho antibody. The data shown are representative of 3 independent experiments.

Klotho promoter is located in a CpG island and lacks ordinary regulatory elements

The fact that Klotho is commonly expressed in the kidney irrespective of species infers the involvement of evolutionarily conserved mechanisms. We therefore used the VISTA web server to align genomic sequences of human chromosome 13 region 32,480,500-32,490,500, which contains exon 1 of the Klotho gene and up to 8 kb of upstream sequences, with corresponding regions in rhesus monkey, dog, horse, mouse, rat, and chicken (Fig. 2A). Several peaks indicated >50% homology among humans, mice, and rats in this region: genomic positions at 32,488,573 (−6 of the ATG start codon of human Klotho), 32,488,321 (−258), 32,488,167 (−412), 32,488,113 (−466), 32,487,236 (−1347), 32,486,984 (−1595), 32,486,809 (−1770), 32,486,469 (−2110), 32,485,955 (−2624), 32,485,706 (−2873), 32,485,560 (−3019), and 32,482,132 (−6447) (indicated by arrowheads in Fig. 2A). These regions may act as evolutionarily conserved promoters and/or enhancers of the Klotho gene. Moreover, chromatin immunoprecipitation and direct sequencing analyses revealed that trimethylation of histone H3-lysine 4, a hallmark of active promoter (38), was detected in the sequence between 32,487,446 (−1133) and 32,488,516 (−63) according to the data deposited in University of California–Santa Cruz Genome Bioinformatics (Supplemental Fig. S2A). Taken together, these data suggest that the region up to −1200 bp of the ATG codon acts as a canonical promoter of the Klotho gene. As shown in Fig. 2B, this region is extremely GC-rich: G+C percentage is 65.9%, and the ratio of observed/expected CpG is 0.915 in the sequence between −848 and +88, which fulfills the requirements for a CpG island by definition (25). A database search confirmed that the promoter region and the entire exon 1 of the Klotho gene constitutes a CpG island (Supplemental Fig. S2B). These features are virtually identical to those of the monkey Klotho promoter characterized by King et al. (39). In addition, Klotho promoter lacks ordinary regulatory elements such as TATA and CAAT boxes and has 5 GC boxes, cognate binding sites of Sp-1 transcription factor (Fig. 2B). These are typical features of the promoters of housekeeping genes as well as tissue-specific and developmental genes (28).

Figure 2.

Structure and activity of Klotho promoter. A) Using the VISTA web browser, we aligned homologous sequences in Klotho promoter regions (up to 8 kb upstream of the entire exon 1 of the Klotho gene; exon 1 is marked in purple) in human, rhesus monkey, dog, horse, mouse, rat, and chicken. Peaks with >50% homology among human, mouse, and rat are indicated by arrowheads. The region harboring histone H3-lysine 4 trimethylation is shown by a bracket (see Supplemental Fig. S2A for details). B) Nucleotide sequence of the 5′ untranslated region of the Klotho gene with the ATG start codon as +1. CpG dinucleotides and exon 1 sequences are shown in boldface italics and italics, respectively. G-395 is indicated by an arrowhead. GC boxes, putative binding-sites for Sp-1, are underlined. Red and blue arrows correspond to outer and inner primers for MSP, respectively. This sequence has been deposited in the DDBJ/EMBL/GenBank database under accession number AB673445. C) We transfected pGL4.14 plasmid carrying Klotho promoter sequences between −1261 and +88 (pGL4-KL) or an empty vector (pGL4) into HK-2 and HeLa cells along with an SV40 promoter-driven Renilla luciferase vector and measured luciferase activities after 48 h. Klotho (KL) promoter activity (y axis) was calculated as firefly luciferase activity of cells transfected with pGL4-KL with the values of pGL4 being set at 1.0 after normalization of transfection efficiencies using Renilla luciferase activity. Data shown are the means ± sd of 5 independent experiments. *P < 0.05 vs. pGL4; Student's t test. D) We continued to culture HK-2 and HeLa cells after transfection of pGL4-KL and Renilla luciferase vectors in the presence of hygromycin B at 1000 μg/ml (IC100) and measured luciferase activities serially. Representative results of passages 10 and 42 are shown. Klotho promoter activity (y axis) was calculated as normalized luciferase activity with the values of transient transfection assays of corresponding cells being set at 1.0. Data shown are the means ± sd of 5 independent experiments. *P < 0.01 vs. p10; Student's t test.

We subcloned the region between −1261 and +88 into a pGL4 luciferase vector to perform reporter assays. In transient transfection assays, this region exhibited considerable promoter activity in both Klotho-expressing HK-2 and Klotho-nonexpressing HeLa cells (Fig. 2C). This may be explained by the fact that Sp-1 is equally expressed in HK-2 and HeLa cells (data not shown). Therefore, it is unlikely that the genetic mechanisms exemplified by Sp-1 binding to GC boxes contribute to renal tubule-specific expression of the Klotho gene. Notably, Klotho promoter activity remained high after long-term culture of HK-2 cells, whereas it was repressed to ∼20% in HeLa cells after 30–40 passages when transfected plasmids were integrated into the chromatin (Fig. 2D). From these results, it is obvious that epigenetic mechanisms play a pivotal role in renal tubule-specific expression of the Klotho gene.

Promoter methylation restricts Klotho gene expression in renal tubular cells

Given that Klotho promoter constitutes a CpG island, we hypothesized that DNA methylation was largely responsible for the epigenetic mechanisms underlying renal tubule-specific expression of the Klotho gene. To verify this hypothesis, we determined the methylation status of the sequence between −450 and −331 (indicated by blue arrows in Fig. 2B) using MSP. As shown in Fig. 3A, this GC-rich region of Klotho promoter was completely demethylated in Klotho-expressing renal tubular cells (RPTECs and HK-2 cells), whereas methylation-specific bands were dominant (1.5- to 2.0-fold) in Klotho-nonexpressing cells (HEK293 cells and fibroblasts). The same pattern was detected by bisulfite sequencing, showing that 30–40% of CpG dinucleotides were methylated in Klotho-nonexpressing cells. Klotho promoter was not methylated in some DNA clones from HEK293 cells and fibroblasts (Fig. 3B). Instead, different parts of Klotho promoter (−903 to −484, −250 to −101, and −90 to +39) were methylated in these clones (data not shown), suggesting that partial methylation is sufficient for Klotho gene silencing.

Figure 3.

CpG methylation renders renal tubule-specific expression of the Klotho gene. A) We performed MSP using bisulfite-treated DNA from the indicated cells to evaluate the methylation status of Klotho and Ink4a promoter regions. H₂O means PCR performed without template DNA. M, DNA size marker. B) Methylation status of the Klotho gene between nt −450 and −331 was analyzed by bisulfite sequencing using DNA from the indicated cells. The nucleotide sequence is shown at the top. Methylated and unmethylated cytosines are indicated by solid and open squares, respectively. Each line shows the result of independent DNA clones isolated from each cell type. C) HEK293 cells, normal fibroblasts, and HeLa cells were treated in the absence (untreated) or presence of 1 μM azacytidine or 1 mM valproic acid for 48 h. We performed semiquantitative RT-PCR (top panel) and real-time quantitative RT-PCR (bottom panel) as described in the legend to Fig. 1. Bottom panel: data are means ± sd of 5 independent experiments. *P < 0.05 vs. untreated control; Student's t test). D) We performed methylation-specific PCR of transgenes. To specifically analyze the methylation status of transfected Klotho promoter, we first performed PCR with a pGL4 vector-specific primer (5′-CTAGCAAAATAGGCTGTCCCC-3′) and the outer antisense primer (−278 to −258), followed by second PCR with internal primer pairs corresponding to the region between −450 and −331. E) Passage 42 HK-2 and HeLa cells were cultured in the absence (Aza−) or presence (Aza+) of 1 μM azacytidine for 48 h and subjected to reporter assays. Klotho (KL) promoter activity (y axis) was calculated as normalized luciferase activity with the values of transient transfection assays of corresponding cells being set at 1.0. Data are means ± sd of 3 independent experiments. *P < 0.01; Student's t test.

To confirm that Klotho promoter methylation is indeed responsible for gene silencing, we investigated whether DNA demethylating agents could restore the expression of Klotho in nonexpressing cells. As anticipated, the DNA methyltransferase inhibitor azacytidine increased Klotho gene expression 1.5- to-3.0 fold in HEK293 cells, fibroblasts, and HeLa cells (Fig. 3C). It is well known that DNA methylation-induced gene silencing is primarily mediated through the recruitment of HDACs to methylated cytosine residues by the aid of MBDs (28, 29, 35). Therefore, we tested whether HDAC inhibitors were also effective for the restoration of Klotho gene expression. The data shown in Fig. 3C indicate that this is the case with valproic acid-treated HEK293 cells but not fibroblasts and HeLa cells. Another HDAC inhibitor romidepsin also induced the expression of the Klotho gene in HEK293 cells (data not shown).

In addition, we investigated whether promoter methylation caused silencing of the exogenous Klotho gene in HeLa cells after the integration of reporter plasmids into the chromatin. As shown in Fig. 3D, MSP revealed that transduced Klotho promoter was at least partly methylated after 30–40 passages of culture in HeLa cells, whereas methylation was not detectable in early-passage HeLa cells, in which the transfected Klotho gene was still active before chromosomal integration. It is of note that Klotho promoter activity was restored by treatment with azacytidine in long-term cultured HeLa cells (Fig. 3E), validating the pivotal role of methylation in gene silencing. In striking contrast, promoter methylation and gene silencing did not occur in HK-2 cells even after chromosomal integration of transfected reporter plasmids (Fig. 3D, E). This implies that renal tubular cells possess a mechanism that actively protects Klotho promoter from CpG methylation in a natural context.

Hypermethylation of the Klotho gene in kl/kl mice

Finally, we investigated whether promoter methylation is responsible for the loss of Klotho gene expression in the original kl/kl mice. In kl/kl mice, a >50-kb transgene was inserted 6472 bp upstream of the start codon of the Klotho gene (1); however, little is known about the mechanisms by which the transgene caused silencing of the downstream gene. We hypothesized that de novo methylation of the transgene, which might occur during development to maintain genomic stability, spread downstream to the promoter/enhancer elements of the Klotho gene to cause a severe hypomorph. To test this idea, we assessed the methylation states of a transgene and the region between the transgene and Klotho promoter in kl/kl mice using bisulfite sequencing. As shown in Fig. 4A, both transgene (−8777 to −8613) and intermediate GC-rich regions (−4251 to −4072) were heavily methylated in the kidney and liver of kl/kl mice. We then examined whether the promoter region is methylated in kl/kl mice. Because the region between −470 and +778 constitutes a CpG island in mice (http://www.ncbi.nlm.nih.gov/epigenomics/view/genome/16591/?term=KL), we investigated the methylation status of the 5′ -end of this region using bisulfite sequencing and determined that 4-7 and 1-6 of 8 CpG dinucleotides were methylated in DNA subclones from different kl/kl mice and wild-type littermates, respectively (Fig. 4A). Furthermore, we quantified the methylation levels at each CpG site using direct sequencing of PCR products for statistical analysis (Fig. 4B depicts a representative result). As shown in Table 1, kl/kl mice, especially in the liver, showed significantly higher methylation levels throughout the analyzed region than normal littermates (P<0.01 by 1-way repeated-measures ANOVA with the Bonferroni post hoc test). It is of note that methylation levels of 3 sites (−302, −300, and −265) were significantly lower in the kidney than in the liver of normal mice (P<0.05 by 1-way repeated-measures ANOVA with the Student-Newman-Keuls post hoc test), suggesting that this region plays a pivotal role in kidney-specific expression of the mouse Klotho gene.

Figure 4.

Methylation analyses of kl/kl mice. A) We isolated genomic DNA from the kidney and liver of kl/kl mice (Klotho) and their littermates with normal Klotho alleles (wild type) and performed bisulfite sequencing to analyze methylation status of the Klotho gene between −8777 and −8613 (transgene), −4251 and −4072 (intermediate segments), and −336 and −195 (promoter region). Nucleotide sequence is shown at the top of each column. ■, methylated cytosines; ☐, unmethylated cytosines. Each line shows the result of a single DNA clone from different mice (n=2−4). B) We directly sequenced PCR products of the region between −336 and −195 using sense primers. Methylation levels of each position were calculated from the quantified data of bisulfite sequencing according to the formula (height of C peak)/[(height of C peak)+(height of T peak)] and are shown as percentages in Table 1. The data shown are representative results of positions −318 and −316 of a kidney sample from Klotho mice.

Table 1.

CpG methylation levels of Klotho promoter

Mouse/tissue	CpG methylation level at nucleotide position (%)
	−318	−316	−305	−302	−300	−265	−211	−208
Klotho/kidney	74	60	75	78	61	20	0	0
Klotho/liver	76	67	73	68	57	71	40	35
Wild type/kidney	40	26	60	0	0	0	0	0
Wild type/liver	66	33	67	61	32	12	0	0

Methylation levels of each position were calculated from the quantified data of bisulfite sequencing according to the formula (height of C peak)/[(height of C peak) + (height of T peak)] and are shown as percentages. Data are mean values of 3 mice; sd values are <10% and were omitted.

DISCUSSION

CpG methylation has been considered a fundamental mechanism of tissue- and developmental stage-specific gene expression (24); however, only a few definite examples have been reported so far, including SERPINB5 (40), MCJ (41), and 14-3-3σ (42). In this study, we obtained in vitro and in vivo evidence indicating that promoter methylation confers kidney-predominant expression of the Klotho gene. In particular, it is noteworthy that the mechanism underlying the loss of Klotho expression in the original kl/kl mice was clarified for the first time. Recently, King et al. (39) reported that CpG methylation is implicated in age-related down-regulation of Klotho in brain white matter of the rhesus monkey. They also demonstrated that a relatively small increase (0.4% overall) and site-specific CpG methylation (4 of 36 sites in the sequence between −466 and −115) resulted in a significant decrease (∼20%) in Klotho expression. This result is fully compatible with our observation in the present study. Taken together, these findings underscore the importance of promoter methylation in spatiotemporal regulation of Klotho gene expression.

Renal tubular cells appear to possess a mechanism that actively protects Klotho promoter from CpG methylation. According to comprehensive analyses using cDNA microarrays (http://www.ebi.ac.uk/arrayexpress/), Klotho mRNA is also expressed in the brain, mammary glands, uterine cervix, stomach, colon, skeletal muscle, and skin less abundantly than in the kidney. As in renal tubular cells, Klotho promoter may be protected from CpG methylation in these tissues. Indeed, recent investigations revealed that Klotho promoter is unmethylated in the normal uterine cervix, gastric mucosa, colon, and mammary glands and becomes aberrantly methylated during malignant transformation of these tissues (43–46). Moreover, Sun et al. (47) have demonstrated that uremic toxins induce hypermethylation of CpG islands of the Klotho gene via increasing the expression of DNA methyltransferases, proposing a novel mechanism of Klotho down-regulation in uremic patients. These observations underscore the general importance of DNA methylation-dependent Klotho gene regulation in the maintenance of tissue homeostasis.

Recently, it was found that the active DNA demethylation machinery is present in mammals (48, 49). For example, 5-methylcytosine is hydroxylated by the 10-11 translocation (TET) family of enzymes to 5-hydroxymethylcytosine, which may be ultimately converted to unmethylated cytosine through further oxidation, formylation, and carboxylation (50). This system is particularly active in neurons and embryonic stem cells to maintain cellular identity, such as plasticity and self-renewal (51, 52). In this regard, the recent observation that 5-hydroxymethylcytosine is detected at moderate levels in the kidney is very intriguing (53, 54). The TET-mediated pathway is shown to act at early stages of development and, more importantly, may constitute a mechanism directing the spatiotemporal specificity of DNA demethylation. After fertilization, paternal pronuclei undergo extensive loss of 5-methylcytosine due to TET3 activation (54, 55), whereas maternal pronuclei are protected from demethylation by the DNA-binding protein PGC7/Stella (56). A similar mechanism may be involved in the protection of Klotho promoter from DNA methylation in renal tubular cells during nephrogenesis.

On the other hand, some studies have delineated the regulation of Klotho gene expression by upstream enhancer elements. Zhang et al. (57) reported that thiazolidinediones increase Klotho mRNA expression in HEK293 and other renal cell lines through the binding of peroxisome proliferator-activated receptor-γ (PPAR-γ) to a noncanonical PPAR-responsive element at −3698/−3686 of the Klotho gene. Recently, Moreno et al. (58) demonstrated that Klotho expression is down-regulated by an NF-κB-dependent mechanism during inflammation. Proinflammatory cytokines promote the binding of RelA to NF-κB consensus sequences at −2255/−2058 of the Klotho gene. This site corresponds to one of the evolutionarily conserved promoters/enhancers of the Klotho gene in our present analysis (the peak at 32,486,469 in Fig. 2A). Their findings provide a novel mechanistic view of inflammation-induced acute kidney injury in terms of the kidney-protective function of Klotho (22, 59).

According to our present findings and previous investigations by others (57, 58, 60), it seems feasible to modulate Klotho expression for therapeutic purposes. We demonstrated that epigenetic drugs, such as DNA demethylating agents and HDAC inhibitors, induced Klotho expression in Klotho-nonexpressing cells and restored the promoter activity in HeLa cells in which the activity of the transfected reporter was silenced after culture. HDAC inhibitors are also effective for preventing Klotho down-regulation caused by uremic toxins and proinflammatory cytokines, which is accompanied by hyperacetylation of Klotho promoter (47, 58). Zhang et al. (57) showed that PPAR-γ agonists increased Klotho expression in the murine kidney, whereas PPAR-γ antagonists attenuated Klotho expression in vitro and in vivo. These findings would be useful in therapeutic intervention for accelerated aging and several complications caused by Klotho down-regulation in patients with chronic renal failure and other pathological conditions.