Ablation of the p16INK4a tumour suppressor reverses ageing phenotypes of klotho mice

The p16INK4a tumour suppressor has an established role in the implementation of cellular senescence in stem/progenitor cells, which is thought to contribute to organismal ageing. However, since p16INK4a knockout mice die prematurely from cancer, whether p16INK4a reduces longevity remains unclear. Here we show that, in mutant mice homozygous for a hypomorphic allele of the α-klotho ageing-suppressor gene (klkl/kl), accelerated ageing phenotypes are rescued by p16INK4a ablation. Surprisingly, this is due to the restoration of α-klotho expression in klkl/kl mice and does not occur when p16INK4a is ablated in α-klotho knockout mice (kl−/−), suggesting that p16INK4a is an upstream regulator of α-klotho expression. Indeed, p16INK4a represses α-klotho promoter activity by blocking the functions of E2Fs. These results, together with the observation that the expression levels of p16INK4a are inversely correlated with those of α-klotho throughout ageing, indicate that p16INK4a plays a previously unrecognized role in downregulating α-klotho expression during ageing.

T issue repair and regeneration are essential for longevity in complex animals, and often depend on the proliferative activity of stem or progenitor cells 1 . In many tissues, the proliferative activity of such cells declines with age, contributing to many ageing-associated pathologies [2][3][4] . In mammals, the p16 INK4a tumour-suppressor gene elicits irreversible cell-cycle arrest known as cellular senescence [5][6][7][8][9][10] , and its expression increases with age in many tissues [11][12][13] , along with the accumulation of dysfunctional senescent stem/progenitor cells [14][15][16] . However, recent studies using middle-aged mice lacking p16 INK4a (p16 À / À mice) revealed that the ageingassociated induction of p16 INK4a expression reduces the proliferative and regenerative capacities of certain stem/ progenitor cells during the ageing process [14][15][16] . These findings have led to speculation that the induction of p16 INK4a expression and the consequent cellular senescence are causally implicated in ageing-associated declines in stem/progenitor cell functions, thereby reducing longevity. However, since p16 À / À mice die of cancer long before they reach the age at which most normal mice start to die 17 , it remains unclear whether p16 INK4a truly limits longevity in mammals. One approach to circumvent this problem would be the use of short-lived mouse strains with acceleratedageing phenotypes. However, attempts towards extending the maximum lifespan of accelerated-ageing mouse strains by p16 INK4a ablation have so far been unsuccessful 18,19 , raising the question of whether p16 INK4a truly limits longevity in mammals.
Mutant mice homozygous for a severely downregulated hypomorphic allele of the a-klotho gene (referred to as kl kl/kl or klotho mice) manifest multiple age-related disorders that are also observed in humans, including infertility, growth retardation, osteoporosis, pulmonary emphysema, skin atrophy, ectopic calcification and shortened lifespan 20 . Furthermore, the levels of a-klotho expression decline with age in both humans and mice 21 , and overexpression of a-klotho extends the maximum lifespan in mice 22 , suggesting that a-klotho acts as an ageing-suppressor gene in mammals 23 . The a-klotho gene encodes a single-pass transmembrane protein that is predominantly expressed in the kidney 20 , and to a lesser extent in the brain 24 . Two forms of the a-klotho protein exist: a membrane-bound form and a secreted form 25 , and each has different functions 22,24,26,27 . As increases in senescent progenitor cells and decreases in stem cell numbers were observed in several tissues in kl kl/kl mice 28,29 , we wondered whether p16 INK4a contributes to the accelerated-ageing phenotypes in kl kl/kl mice, by eliciting cellular senescence in certain stem/progenitor cells.
In the present study, we explore the roles of p16 INK4a in accelerated-ageing phenotypes of klotho mice. We show that ablation of the p16 INK4a gene reverses various ageing phenotypes, including maximum lifespan, of kl kl/kl mice. Surprisingly, however, this is due to the restoration of a-klotho expression in kl kl/kl mice and does not occur when p16 INK4a is ablated in knockout mice lacking a-klotho (kl À / À ), indicating that p16 INK4a is an upstream regulator of a-klotho expression. Thus, although p16 INK4a has an established role in the implementation of cellular senescence in stem/progenitor cells [5][6][7][8][9][10] , which are likely to reduce longevity 13 , our results reveal that p16 INK4a has an additional function in promoting ageing phenotypes by downregulating a-klotho expression in mice. Our findings advance our understanding of the molecular mechanisms underlying the development and progression of ageing in mammals.

Results
Ablation of p16 INK4a reverses ageing phenotypes of klotho mice. To investigate whether p16 INK4a contributes to the accelerated-ageing phenotypes in kl kl/kl mice, we first generated kl kl/kl mice lacking p16 INK4a (kl kl/kl p16 À / À mice) by cross-breeding heterozygous klotho (kl kl/ þ ) mice with heterozygous p16 INK4a knockout (p16 þ / À ) mice, and tested whether the accelerated-ageing phenotypes of kl kl/kl mice can be reversed by p16 INK4a ablation. Various accelerated-ageing phenotypes of kl kl/kl mice, such as growth retardation, osteoporosis, pulmonary emphysema and severe atrophy of the intestinal wall and skin, were remarkably mitigated in the kl kl/kl p16 À / À mice as compared with those in the kl kl/kl p16 þ / þ (referred to as kl kl/kl ) or kl kl/kl p16 þ / À littermates with the same genetic background (98.375% C57BL/6J, 1.625% C3H/J) (Figs 1 and 2, and data not shown). Moreover, the maximum lifespan of the kl kl/kl p16 À / À mice was significantly extended as compared with those of the kl kl/kl or kl kl/kl p16 þ / À littermates, irrespective of gender (Fig. 1b, and data not shown). These results indicated that the accelerated-ageing phenotypes of the kl kl/kl mice depend strongly on the p16 INK4a status. The simplest explanation for these results would be that the a-klotho deficiency caused the ageing phenotypes by elevating p16 INK4a expression, and thus p16 INK4a ablation would mitigate the ageing phenotypes in kl kl/kl mice.
To test this hypothesis, we took advantage of using the p16-luc mice 12 , in which the levels of p16 INK4a expression can be monitored throughout the body using a bioluminescence imaging (BLI) technique. The p16-luc mice were crossed into the klotho genetic background and were subjected to BLI. Unexpectedly, however, we were unable to detect any substantial increase of p16 INK4a expression throughout the body in kl kl/kl mice as judged by BLI (Fig. 3). Because a-klotho is predominantly expressed in the kidney 20,24 (see also Supplementary Fig. 1 ( Supplementary Fig. 2). These results raise a question of how the p16 INK4a ablation reversed the accelerated-ageing phenotypes of kl kl/kl mice.
To explain the effects of the p16 INK4a ablation on the acceleratedageing phenotypes of kl kl/kl mice, we next took a closer look at the biochemical characteristics of kl kl/kl mice. In kl kl/kl mice, the level of 1a-hydroxylase gene expression is reportedly increased in the kidneys, resulting in an elevated serum level of 1,25-dihydroxyvitamin D, the active metabolite of vitamin D that regulates calcium and phosphate homeostasis 30 . Since these changes are known to be associated with the accelerated-ageing phenotypes in kl kl/kl mice 31 , we examined the levels of these biochemical hallmarks in kl kl/kl mice. Notably, the levels of these hallmarks were substantially reduced in kl kl/kl p16 À / À mice, as compared with those in kl kl/kl mice (Fig. 4a,b). Moreover, the aberrant activation of calpain-1 and the ectopic calcification in kidneys, signs of the abnormal calcium homeostasis observed in kl kl/kl mice 20,27,32 , were absent in kl kl/kl p16 À / À mice (Figs 2 and 4c), implying that the a-klotho function might be somewhat restored in the kl kl/kl p16 À / À mice. Since the kl kl/kl mice are not a complete null, but have a severe hypomorphic mutation for a-klotho expression, the entire a-klotho-coding sequence is intact in kl kl/kl mice 20 . Thus, we next wondered whether p16 INK4a ablation could restore the levels of a-klotho expression in kl kl/kl mice. Indeed, the levels of both a-klotho mRNA and protein were substantially increased in the kidneys of kl kl/kl p16 À / À mice, as compared with those in kl kl/kl mice, albeit to lesser extents as compared with those in wt mice (Fig. 4a,c). Notably, a-klotho expression was observed only in the renal distal convoluted tubules in kl kl/kl p16 À / À mice (Fig. 5a), which are the major sources of a-klotho expression in wt mice 20 . Thus, it appears that p16 INK4a ablation restores the normal a-klotho expression pattern in kl kl/kl mice. Importantly, in stark contrast to the kl kl/kl mice, the p16 INK4a ablation failed to reverse the accelerated-ageing phenotype in mice lacking the a-klotho gene (a-klotho knockout mice (kl À / À ; ref. 30; Supplementary Fig. 3). These results indicate that p16 INK4a ablation mitigates the acceleratedageing phenotypes of kl kl/kl mice, by restoring a-klotho expression. The histograms indicate the quantitative analysis of X-ray transparency of femur (WT (n ¼ 3), p16 À / À (n ¼ 3), kl kl/kl (n ¼ 6) and p16 À / À kl kl/kl (n ¼ 3)), the mean linear intercept (Lm) in lung tissue (WT (n ¼ 9), p16 À / À (n ¼ 9), kl kl/kl (n ¼ 7) and p16 À / À kl kl/kl (n ¼ 9)), intestinal villi length (WT (n ¼ 14), p16 À / À (n ¼ 7), kl kl/kl (n ¼ 11) and p16 À / À kl kl/kl (n ¼ 3)), epidermal and subcutaneous fat layer thickness (WT (n ¼ 6), p16 À / À (n ¼ 6), kl kl/kl (n ¼ 8) and p16 À / À kl kl/kl (n ¼ 3)) and the percentages of calcified areas in kidneys (WT (n ¼ 3), p16 À / À (n ¼ 3), kl kl/kl (n ¼ 3) and p16 À / À kl kl/kl (n ¼ 3)). For graphs of X-ray transparency of femur and Lm in lung tissues, data were analysed by Mann-Whitney U-test and are displayed as mean ± s.e.m. For graphs of intestinal villi length, epidermal and subcutaneous fat layer thickness and the percentages of calcified areas in kidneys, data were analysed by Student's t-test and are displayed as mean±s.e.m. For all graphs: *Po0.05, **Po0.01.
p16 INK4a downregulates a-klotho expression in wt mice. The obvious next question is whether p16 INK4a downregulates a-klotho expression in wt mice. Note that there is an inverse correlation between the levels of p16 INK4a expression and a-klotho expression during the ageing process in kidneys (Fig. 5b). However, because p16 À / À mice die prematurely from cancer 17 (see also Fig. 1b), we cannot examine whether p16 INK4a ablation ameliorates the ageing-associated decline of a-klotho expression in mice harbouring wt a-klotho. To circumvent this problem, we employed the mouse model of chemically induced kidney injury. It was previously reported that patients with chronic renal failure develop multiple age-related disorders resembling those of kl kl/kl mice, with a marked reduction of a-klotho expression in kidneys 33,34 . Moreover, treatment with cisplatin, a chemotherapeutic agent that causes severe adverse actions with nephrotoxicity, is known to provoke a significant reduction of a-klotho expression in kidneys, accompanied by the accumulation of DNA damage 35 . Since persistent DNA damage induces p16 INK4a expression in many different cell types 12 , we analysed the effect of p16 INK4a expression on the levels of a-klotho expression in the cisplatin-induced kidney injury model. Indeed, the cisplatin treatment resulted in a marked reduction of aklotho expression in the renal distal convoluted tubules of wt mice, coinciding with the accumulation of gH2AX foci, a sign of the DNA damage response, and the induction of p16 INK4a expression ( Supplementary Fig. 4). Notably, however, the cisplatin-induced reduction of a-klotho expression was substantially attenuated in the p16 À / À mice ( Supplementary Fig. 4), although this level of a-klotho restoration was insufficient to block the cisplatin-induced nephrotoxicity in this experimental condition (Supplementary    p16 INK4a downregulates the a-klotho promoter in murine cells. To further verify this notion, we next sought evidence that p16 INK4a downregulates a-klotho promoter activity. As a-klotho expression is rather limited in the renal distal convoluted tubules 20,24 , we were unable to find any established murine cell lines expressing substantial levels of a-klotho ( Supplementary  Fig. 6). Therefore, primary mouse renal tubular epithelial cells (mRTECs) were prepared from the kidneys of wt mice, and were used for a promoter-reporter analysis. The luciferase activity of a reporter plasmid containing the region 1,035 nucleotides upstream of the mouse a-klotho translation start site was reduced by the ectopic expression of p16 INK4a in a dosedependent manner in early-passage primary mRTECs (Fig. 6a), indicating that p16 INK4a indeed downregulates a-klotho expression at the promoter level. Although the a-klotho promoter sequences are not well conserved between human and mouse, both include potential binding sites for E2F transcription factors, which are critical downstream mediators of the p16 INK4aretinoblastoma tumour suppressor pathway 6 , at the same position from the a-klotho translation start sites (Figs 6a and 7b). Notably, newborn mice lacking both E2F1 and E2F3a, a subset of the activator E2Fs (referred to as E2F1 À / À E2F3a À / À mice), reportedly exhibited normal weight and appearance; however, by their third week of life the proliferative index of most tissues was significantly reduced, and 90% of the mice became severely runted and died within 2 months 36 . Furthermore, white adipose tissues were absent and lung alveolar branching was severely reduced in E2F1 À / À E2F3a À / À mice 36 . Since these phenotypes are reminiscent of the kl kl/kl phenotypes 20 , we analysed whether E2F1 and/or E2F3 activate the a-klotho promoter activity. Indeed, the ectopic expression of either E2F1 or E2F3 increased the activity of the a-klotho promoter in cultured primary mRTECs (Fig. 6b). However, this was not the case when the E2F-binding element was disrupted by a nucleotide substitution in the reporter plasmid (Fig. 6b, -674 E2F-Mut). Moreover, increasing amounts of E2F3 blocked the trans-repressing activity of co-transfected p16 INK4a in early-passage primary mRTECs ( Supplementary  Fig. 7). Note that endogenous E2F1 and E2F3 were found to bind to the a-klotho promoter, as judged by a chromatin immunoprecipitation (ChIP) analysis using cultured primary mRTECs or kidney tissues prepared from wt mice (Fig. 6c). These results, in conjunction with the observation that the levels of endogenous aklotho mRNA and protein expression were substantially reduced in the kidneys of E2F1 À / À E2F3a À / À mice (Fig. 6d), strongly suggest that p16 INK4a downregulates a-klotho expression, at least partly by blocking the function of activator E2Fs in wt mice.
Finally, to further support our murine data and to extend the analysis to human physiology, we tested whether p16 INK4a downregulates a-klotho mRNA expression in human cells. Similar to murine cells, primary human RTECs (hRTECs), but not other human cell lines, express substantial levels of a-klotho ( Supplementary Fig. 8). We thus used primary hRTECs in the following experiments. We found that the levels of a-klotho mRNA expression declined when cultured primary hRTECs were rendered senescent by serial passage, accompanied by the induction of p16 INK4a mRNA expression (Fig. 7a, left). However, the short interfering RNA (siRNA)-mediated depletion of p16 INK4a substantially increased the levels of a-klotho mRNA expression in latepassage hRTECs, coinciding with the increased expression of cdc6, an established E2F target gene (Fig. 7a, middle). Conversely, the ectopic expression of p16 INK4a reduced the levels of a-klotho mRNA expression in early-passage hRTECs (Fig. 7a, right). Together, these results suggest that p16 INK4a downregulates aklotho expression, by blocking the function of the activator E2Fs in human cells, as well as in mouse cells. Indeed, the ectopic expression of either E2F1 or E2F3 substantially increased the transcriptional activity of the human a-klotho promoter in cultured hRTECs (Fig. 7b, À 1,028 wt and À 473 wt). These effects were blunted when the putative E2F-binding site within the human a-klotho promoter was mutated or deleted in the reporter plasmid (Fig. 7b, -473 E2F-Mut and-360 wt). Furthermore, the ChIP analysis revealed that endogenous E2F1 and E2F3 bind to the human a-klotho promoter in cultured hRTECs (Fig. 7c). Interestingly, the G to A single-nucleotide polymorphism (SNP), in the putative E2F-binding site of the human a-klotho promoter, reportedly impaired the DNA-protein interaction and is associated with the reduction of bone mineral density (BMD) in aged postmenopausal women 37 . Indeed, the G to A substitution in the E2F-binding site of the reporter plasmid greatly reduced the response to E2F overexpression (Fig. 7b,-473 E2F-SNP). These results, in conjunction with previous observations that there is the potential inverse correlation between the levels of renal p16 INK4a expression and a-klotho expression in elderly people 21,38-40 , suggest that p16 INK4a is likely to have the potential to downregulate a-klotho expression by blocking the transcriptional activity of E2Fs in human kidneys.

Discussion
The ageing process is multifactorial, with genetic background and environmental stress as two critical components 4,41 . The mutation of the a-klotho gene causes multiple premature ageing phenotypes, including a shortened lifespan in mice 20 , and some SNPs in the human a-klotho gene are associated with reduced lifespans 37,42,43 . Moreover, the levels of plasma a-klotho decrease with increasing age and are associated with longevity in humans 21,40 , indicating that the a-klotho gene is an important antiageing gene in both mouse and human. However, it remained unclear how the a-klotho gene could be linked to environmental stress. Here we show that the p16 INK4a tumour-suppressor gene, a stress sensor known to induce cellular senescence 5-12 , downregulates a-klotho gene expression in both mouse and human renal tubular epithelial cells. Ablation of the p16 INK4a gene mitigates various accelerated-ageing phenotypes of kl kl/kl mice, including shortened maximum lifespan, by partially restoring a-klotho expression (Figs 1,2 and 4). Furthermore, cell culture studies reveal that p16 INK4a represses a-klotho gene expression at the promoter level by blocking the function of activator E2F, most likely through activation of retinoblastoma protein in mouse and human cells (Figs 6-8). These results, together with previous epidemiological studies 38,39 , suggest that this previously unrecognized function of p16 INK4a is likely to play a role in humans as well as in mice. However, p16 INK4a has an established role in the implementation of cellular senescence in stem/progenitor cells 5-10 , thereby causing dysfunctional tissue regeneration and repair 13 , which are likely to reduce longevity. Indeed, a series of studies using middleaged p16 À / À mice revealed that the ageing-associated induction of p16 INK4a expression reduces the proliferative and regenerative capacities of certain progenitor cells during the ageing process [14][15][16] , further illustrating the importance of the p16 INK4a -cellular senescence pathway in the development of ageing phenotypes 44 . Nevertheless, our present study revealed that p16 INK4a plays another role in promoting ageing phenotypes, through the downregulation of the expression of the a-klotho ageing suppressor (see model in Fig. 8). This previously unrecognized pathway, linking p16 INK4a to a-klotho expression, enhances our understanding of the molecular mechanisms underlying the development and progression of ageing phenotypes in mammals and opens up new possibilities for their control.
Bioluminescence imaging. For the detection of luciferase expression, mice were anaesthetized, injected intraperitoneally with D-luciferin sodium salt (75 mg kg À g ) 5 min before beginning photon recording. Mice were placed in the light-tight chamber and a grey-scale image of the mice was first recorded with dimmed light followed by acquisition of luminescence image using a cooled CCD (chargedcoupled device) camera (PIXIS 1024B; Princeton Instruments) 12,46 . The signal-tonoise ratio was increased by 2 Â 2 binning and 5-min exposure. For colocalization of the luminescent photon emission on the animal body, grey scale and pseudocolour images were merged by using IMAGE-PRO PLUS (Media Cybernetics).
Histology and immunofluorescence analysis. Samples were fixed in 10% formalin for a 24 h or longer, progressively dehydrated through gradients of alcohol and embedded in paraffin. Samples were then sectioned on a microtome (5-mm thick), deparaffinized in xylene, rehydrated and then stained with haematoxylin and eosin (HE). For immunofluorescence, the relevant Alexa Fluor 488 goat antimouse or 546 goat anti-rabbit antibodies (1:1,000, Molecular probes) were used for detection of primary antibodies. Fluorescence images were observed and photographed using an immunofluorescence microscope (Carl Zeiss). The primary antibodies used for mouse samples were as follows: Klotho (1:100, TransGenic Inc., KO603), E-cadherin (1:100, Cell Signaling no. 3195), g-H2AX (1:100, Millipore, 05-636). E-cadherin was used as a marker of the renal distal tubes [47][48][49] . Calcium deposition was visualized with the von Kossa staining. Paraffin-embedded sections were deparaffinized and rehydrated. Fixed sections were incubated with 5% silver nitrate during exposure to light for 60 min, and washed with distilled water. Excess silver was washed out with 5% sodium thiosulfate for 2 min. The sections were then stained with Kernechtrot dye. Epidermal and subcutaneous fat-layer thickness and intestinal villi length were determined using 10-74 random measurements along the length of skin and small intestine from at least three mice per age group and genotype. For skin sections, the skin was cut parallel to the spine and sections were cut perpendicular to the skin surface. For villi sections, intestinal tracts were flushed with PBS and rolled up in a compact circle using longitudinally oriented jejunal sections for analysis; 5-mM sections were used for HE staining, and the Image J software was used for length measurements 50 . The mean linear intercept (Lm) in the lung tissue was calculated using light microscopy and Image J software. An overlay consisting of horizontal and vertical, parallel lines was placed over the photographed image of each region. All intercepts with alveolar septal tissue were counted. The total length of all the lines together divided by the number of intercepts gives the Lm for the region studied. The overall mean of the Lm for each of the three regions studied for each tissue block was used as the Lm for the corresponding tissue block 51 .
Reverse transcription and quantitative real-time PCR. Total RNA was extracted from mouse tissues using TRIzol reagent (Life Technologies). Reverse transcription and quantitative real-time PCR (RT-qPCR) was performed using the SYBER Premix EX Taq system (TAKARA) and a Prism 7900HT (ABI) 52,53 . Amplified signals were confirmed to be single bands with gel electrophoresis and were normalized to the levels of glyceraldehyde 3-phosphate dehydrogenase. The data were analysed using the SDS2.1 software (ABI) 46 . The PCR primer sequences used are shown in the Supplementary Table 1.
Measurement of serum phosphate, calcium and BUN. The concentration of serum phosphate, calcium and blood urea nitrogen (BUN) was measured with a phosphate assay kit (Serotec UPi-L; Serotec Co. Ltd., Japan), a calcium assay kit (Metalloassay Ca-CPZIII; AKJ Global Technology, Japan) and a BUN assay kit (Iatoro LQ UN rate (A) II; LSI medience Co. Ltd., Japan), respectively, according to the manufacturers' instruction.
Cell culture. For primary mRTECs 47 , kidneys of 6-to 10-week-old male mice were taken and placed in cold PBS containing antibiotics (Penicillin and streptomycin). The medulla of the kidneys were removed and the cortex of the kidney was taken, minced and transferred to 2 ml serum-free DMEM containing 0.1% collagenase (Sigma). Minced kidney cortex tissue was incubated at 37°C with shaking for 30 min. The tissue suspension was mixed with 10 ml of DMEM containing 10% serum to inactivate collagenase. The remaining tissue mass was removed through the 70-mm strainer, and cells were pelleted using centrifugation. The cell number was counted and seeded on the poly-L-lysin-coated dishes (Corning, USA) for ChIP analysis or for transfection. For normal hRTECs, primary hRTECs were purchased from Lonza (Lonza, Switzerland) and were cultured according to the manufacturer's instruction.
In vivo ChIP analysis. Sixty milligrams of kidney tissue was chopped into 1-to 2-mm pieces using razor blades and were transferred into a new tube containing 1 ml PBS with proteinase inhibitor cocktail (Nacalai tesque). Crosslinking was performed in 1% of formaldehyde by rotating at room temperature for 15 min. This crosslinking reaction was stopped by adding fresh glycine to a final concentration of 0.125 M by continuous rotation at room temperature for 5 min. Tissue pieces were then washed and suspended in PBS-containing proteinase inhibitor cocktail, and were grinded by Dounce homogenizing seven times. The cell pellet was resuspended into Lysis buffer containing 5 mM PIPES pH8.0, 85 mM KCl, 0.5% NP40 and proteinase inhibitor cocktail, and lysed by Dounce homogenizing four times to aid in nuclei release 52 . After this procedure, ChIP was performed using the EZ-ChIP kit (Millipore). The immunoprecipitation of cross-linked chromatin was conducted with anti-mouse E2F1 (1:1,000, Santa Cruz, sc-193X), anti-mouse E2F3 (1:1,000, Santa Cruz, sc-878X) and rabbit IgG (1:1,000, Cell Signaling Technology, 2729) as a negative control. After immunoprecipitation, DNA was extracted using the QIAquick PCR purification kit (Qiagen), and an aliquot was amplified by qPCR using the following primers flanking the putative mouse E2F-binding site position at À 388 to À 396 bp of mouse a-klotho gene promoter: 5 0 -TGTTCTCTGAAA GATTCCCC-3 0 and 5 0 -TCCCTTTGCCTTCCTGGGAC-3 0 . The p16 INK4a has an established role in provoking cellular senescence, which is likely to cause stem cell ageing and thereby contributing to organismal ageing.
Here we show that, in addition, p16 INK4a also contributes to organismal ageing through blocking the expression of ageing suppressor, a-klotho.
Luciferase-reporter assays. The human and mouse a-klotho gene promoter sequence was amplified with PCR using genomic DNA extracted from the mouse tail or BAC clone (RP11-720E2) containing the entire human a-klotho gene as templates. Deletion mutants were prepared with standard PCR procedures. Promoter sequences containing point mutations were generated using the Quick Change Site-directed Mutagenesis kit (Agilent Technologies). The promoter fragments were inserted into PGL3 basic firefly luciferase reporter plasmid (Promega). All inserted DNAs were sequenced and verified. Transfection of reporter plasmids was performed using the X-treamGENE9 DNA transfection reagent (Roche) according to the manufacturer's instructions. The luciferase assays were performed using the Luciferase assay systems kit (Promega). Cytomegalovirus promoterrenilla luciferase plasmid or SV40 promoter-b-galactosidase plasmid was used as an internal control.
Cisplatin (cis-diamminedichloro-platinum II) treatment. Cisplatin was purchased from Wako pure chemical, Japan and dissolved in saline. Mice were intraperitoneally injected with Cisplatin solution (12 mg kg À g ) three times (every other day) for a week and killed. The kidneys were immediately taken from mice and used for analysis.