In Vivo Pancreatic β-Cell–Specific Expression of Antiaging Gene Klotho: A Novel Approach for Preserving β-Cells in Type 2 Diabetes

Diabetes affects ∼150 million people worldwide, and this figure is expected to double in the next 20 years (1). About 90–95% of all North American cases of diabetes are type 2 diabetes mellitus (T2DM) (1). Physiologically, pancreatic β-cells constantly synthesize insulin, which is stored within vacuoles and released once triggered by an elevation in blood glucose level. Insulin is the principal hormone that regulates uptake of glucose from the blood into most cells, including skeletal muscle cells and adipocytes. Insulin also is the major signal that promotes the conversion of glucose to glycogen for internal storage in liver and skeletal muscle cells. For many years, T2DM was recognized only owing to insulin resistance, but now, there exists a common agreement that T2DM is a complex pathophysiologic spectrum that includes insulin resistance and β-cell failure. Significant β-cell failure is now believed to take place at an early stage in disease progression; that is, β-cell function declines sharply before and after the diagnosis of T2DM (2). In the UK Prospective Diabetes Study, for example, the secretory capacity of β-cells was reduced by 50% at the time fasting hyperglycemia was diagnosed (3). Generally, the compensatory ability of the β-cell with respect to an increase in insulin resistance keeps blood glucose at the near-normal level through proportionate enhancements of β-cell function (4). No hyperglycemia exists without β-cell dysfunction (5). Maintaining recommended targets of blood glucose control is difficult for many patients with T2DM because of the progressive loss of β-cell function; thus, one of the goals in the treatment of T2DM is to preserve functional β-cells in pancreatic islets.

The mouse Klotho (also called α Klotho) gene contains five exons and encodes a single-pass transmembrane protein with 1,014 amino acids predominantly expressed in the kidney and the brain choroid plexus (6). The majority of amino acids in the Klotho peptide resides in the amino-terminal extracellular domain, which is followed by a 21-amino acid transmembrane domain and an 11-amino acid short intracellular carboxyl terminus (6). Two forms of Klotho exist—the full-length Klotho (130 kDa) and the short-form Klotho (65 kDa)—that can be generated by alternative RNA splicing or proteolytic cleavage (6,7). Overexpression of Klotho extends life span in mice, whereas mutation of the Klotho gene causes multiple premature aging phenotypes and shortened life span (6,8). Klotho has been reported to function as a cofactor for activation of fibroblast growth factor (FGF) receptor 1c by FGF23 in the regulation of calcium, phosphate, and vitamin D metabolism in the kidneys (9). Klotho^−/− mutant mice exhibit pancreatic islet atrophy, decreases in insulin content and mRNA levels in pancreatic islets, and decreases in serum insulin levels (10). Most recently, we reported that Klotho mRNA and proteins are expressed in mouse pancreatic islets and that silencing of Klotho impaired glucose-stimulated insulin secretion in MIN6 β-cells (11). However, whether Klotho expression is altered in pancreatic β-cells in T2DM is not known and whether it protects β-cell function in T2DM has never been investigated.

The db/db mouse was originally derived from an autosomal recessive mutation in the db gene, which encodes for leptin receptors. This model resembles key features of human T2DM, including peripheral insulin resistance and progressive deterioration of pancreatic β-cells (12). Our preliminary study showed that the Klotho level in pancreatic islets is decreased significantly in patients with T2DM and in db/db mice, an animal model of T2DM. The objective of the current study was to investigate whether β-cell–specific expression of Klotho protects β-cell function and attenuates the development of diabetes in db/db mice.

Pancreatic insulinoma MIN6 β-cells were provided by J. Miyazaki (School of Medicine, Kumamoto University, Kumamoto, Japan) and D.F. Steiner (University of Chicago, Chicago, IL) (13). MIN6 cells were cultured and maintained in DMEM containing 25 mmol/L glucose, 10% FBS, 1% penicillin/streptomycin, 2 mmol/L glutamine, and 100 μmol/L β-mercaptoethanol. MIN6 β-cells of <20 passages were used in this experiment. The 3T3-L1 preadipocytes and mouse renal inner medullary collecting duct (mIMCD3) cells were cultured in these media without β-mercaptoethanol.

The use of human pancreas was approved by the Institutional Review Board at the University of Oklahoma Health Sciences Center. Human pancreata from normal donors (age 37–50, both sexes) and T2DM donors (age 42–49, both sexes) were obtained from the National Disease Research Interchange, National Resource Center (Philadelphia, PA).

The procedures for plasmid construction and adeno-associated virus (AAV) packaging were described in our previous studies (14,15). Plasmid of pAAV2.1-mINSULIN-nLacZ with 1.13-Kb mouse preproinsulin gene II promoter was provided by X. Xiao (Eshelman School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, NC) (16). A plasmid of pEFmKLcFT with the full-length mouse Klotho cDNA and COOH-terminal FLAG-tag was provided by M. Kuro-O (The University of Texas Southwestern Medical Center, Dallas, TX). The full-length Klotho cDNA with the COOH-terminal Flag-tag (3.1 kb in total) was cloned into AAV serotype 2 (AAV2) (Stratagene, La Jolla, CA). The mouse insulin II promoter was cloned into AAV2 by replacing the original CMV promoter and intron. The insulin II promoter and GFP cDNA (700 base pairs [bp]) were cloned into the AAV2 vector as the control constructs. The constructs of pAAV-mKL and pAAV-GFP were then packaged with pHelper and pAAV-RC to produce recombinant AAVs by following the manufacturer’s instruction manual (Stratagene). Recombinant viruses were purified through a CsCl gradient as previously described (17). The titers of recombinant viral genome particles were determined on a Bio-Rad CFX96 Real-Time PCR Detection System with a pair of primers targeted to the insulin II promoter region (forward [F]: 5′-AAATGCTCAGCCAAGGACAA-3′; reverse [R]: 5′-GGACTTTGCTGTTTGACCCATT-3′) and with the method previously described (18,19). For the remainder of this article, these recombinant viruses will be referred as rAAV-GFP and rAAV-mKL, respectively.

Plasmid DNA, including pAAV-mKL, pAAV-GFP, and pAAV-CMV-mKL, were purified with the QIAGEN Plasmid Maxi Kit. MIN6 cells, 3T3-L1 preadipocytes, and mIMCD3 cells cultured in a six-well plate were transfected with various plasmid DNAs at a concentration of 0.072 μg/mL using Optifect reagent according to the manufacturer’s protocol followed by 48-h incubation in DMEM with 10% FBS at 37°C in a 5% CO₂ incubator. Phase contrast images and fluorescence images of cells transfected with pAAV-GFP for 48 h were collected at equal exposure conditions under Nikon Eclipse Ti microscopy (magnification ×100) with NIS-Elements BR 3.0 software (Nikon).

Cells (or mouse pancreas) were lysed in radioimmunoprecipitation assay buffer 48 h after transfection. The lysates were directly subjected to SDS-PAGE followed by Western blotting with antibody against Klotho (R&D Systems), Rac1, and p-Rac1 as we described previously (20,21). The blot was rinsed and reprobed with antibody against β-actin or α-tubulin for loading controls.

This study was carried out in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. This project was approved by the Institutional Animal Care and Use Committee at the University of Oklahoma Health Sciences Center.

Eight-week-old male BKS.Cg-+Lepr^db/+Lepr^db/OlaHsd (db/db) mice and BKS.Cg-Dock^7m+/+Lepr^db/OlaHsd (lean) mice were purchased from Harlan Laboratories (Indianapolis, IN). All mice were housed in cages at room temperature (25 ± 1°C) and provided with Purina laboratory chow (No. 5001) and tap water ad libitum throughout the experiment. Three groups of db/db mice and three groups of lean mice were used (7–8 mice/group). Body weight was monitored weekly. Blood glucose was measured weekly from tail vein blood using a ReliOn Ultima glucose reader (Solartek Products, Inc., Alameda, CA). The mice were fasted for 12 h before glucose measurement. PBS, rAAV-GFP, or rAAV-mKL were carefully injected intraperitoneally into the region of pancreas of lean and db/db mice (10 weeks) at the dose of 2.57 × 10⁹ of viral genome copies/g body weight in a total volume of 500 μL.

The glucose tolerance test (GTT) was performed during weeks 2, 4, and 6 after the treatments. The insulin sensitivity test (IST) was performed during weeks 3 and 5 after the treatments. Briefly, blood glucose levels were measured at 30, 60, 90, and 120 min after subcutaneous injections of d-glucose 1 g/kg (Fisher Scientific) or insulin 1.0 unit/kg (Sigma-Aldrich). The baseline glucose levels were determined after a 12-h fast.

At the end of week 6 of the treatments, five animals from each treatment group were killed with an overdose of sodium pentobarbital (100 mg/kg i.p.), and blood was collected in EDTA. The plasma samples were stored at −80°C. After blood collection, animals were perfused transcardially with heparinized saline. One-fifth of the pancreas was embedded in optimal cutting temperature compound (Tissue-Tek) on dry ice and stored at −80°C until use. The rest of the pancreas was fixed in 4% PBS-buffered paraformaldehyde for 24 h and then embedded in paraffin.

For preliminary experiments, 20-week-old male db/db and lean mice were purchased from Harlan Laboratories (Indianapolis, IN). Four to five animals from each stain were killed with an overdose of sodium pentobarbital (100 mg/kg i.p.). The animals were perfused transcardially with heparinized saline. One-fourth of the pancreas was used for Western blot analysis of Klotho protein expression. The rest of the pancreas was fixed in 4% PBS-buffered paraformaldehyde for 24 h and then embedded in paraffin for immunohistochemical (IHC) analysis of Klotho and insulin protein expression. A total of 20–25 islets were examined for each mouse.

At the end of week 6 of the treatments, three animals from each treatment group were sacrificed. Mouse pancreatic islets were isolated using a modified protocol as described previously (22). Briefly, collagenase P was injected into the common bile duct of a mouse. The pancreas was then excised and digested at 37°C. The islets were first purified with premixed Histopaque gradient and then purified by handpicking the separated islets with low-retention pipette tips under a dissecting microscope. When viewed under the microscope, spherical and golden-brown particles (darker color) with diameters of 100–300 μm were considered islets. Five size-matched islets from each animal were cultured with RPMI1640 with 10% FBS in a 24-well plate overnight. The rest of the isolated islets were stored at −80°C until use.

After overnight incubation in 10% FBS RPMI1640, the isolated islets were starved with Krebs-Ringer bicarbonate (125 mmol/L NaCl, 4.74 mmol/L KCl, 1 mmol/L CaCl₂, 1.2 mmol/L KH₂PO₄, 1.2 mmol/L MgSO₄, 5 mmol/L NaHCO₃, and 25 mmol/L HEPES; pH 7.4) supplemented with 0.1% BSA and 2.8 mmol/L glucose for 1 h. Islets were washed with PBS and incubated with Krebs-Ringer bicarbonate buffer supplemented with 2.8 or 16.7 mmol/L glucose for 1 h. Medium was collected and stored at −80°C for measuring insulin levels.

Insulin levels in plasma or supernatants from glucose-treated islets were measured using an insulin enzyme immunosorbent assay kit according to the manufacturer’s instructions (ALPCO Diagnostics, Salem, NH).

A series of 5-μm–thick sections of paraffin-embedded pancreas were cut at 100-μm intervals on three levels. For IHC analysis, consecutive pancreatic cross sections were deparaffinized, rinsed in xylene, and rehydrated. After heat-induced antigen retrieval in a microwave, the cross sections were blocked first with 3% hydrogen peroxide. The sections were incubated with antibodies against GFP (ab290; Abcam), Flag-tag (ET-DY100; Aves Labs), Klotho (R&D Systems), insulin (sc-9168; Santa Cruz Biotechnology), Pdx-1 (AB3503; Millipore), 4-hydroxyonenal (4-HNE) (ab48506; Abcam), DnaJ (Hsp40) homolog C3 (DNAJC3) (ab70840; Abcam), proliferating cell nuclear antigen (PCNA) (ab2426; Abcam), or microtubule-associated protein 1 light chain 3 (LC3) (L1564-50A; United States Biological) overnight at 4°C and then with appropriated secondary antibodies conjugated with horseradish peroxidase at room temperature for 60 min. Stable diaminobenzidine (Invitrogen) was used as a substrate for peroxidase. Hematoxylin was used as counterstaining. The islets of Langerhans in the cross sections of pancreas for each mouse were located under a microscope (Nikon Eclipse Ti). Images of islets from consecutive cross sections for each animal were collected at equal exposure conditions and at the same magnification (40× objective lens). The staining for Klotho, insulin, Pdx-1, or LC3 was quantified using NIH ImageJ freeware as mean gray value/pixel. Briefly, the selection line was drawn along the islet of Langerhans after the original red, green, and blue image was converted to a gray scale image. A TUNEL assay on the cross sections of mouse pancreas was performed using TACS XL Blue Label In Situ Apoptosis Detection Kit (Trevgen, Gaithersburg, MD). The number of cells with positive insulin, 4-HNE, DNAJC3, TUNEL, or PCNA staining in the islet was counted by NIS-Elements BR 3.0 software. A total of 20–25 islets were examined for each animal.

Dihydroethidium (DHE) (D7008; Sigma-Aldrich) was used to measure superoxide levels in pancreatic islets as we described previously (15,20,21,23,24). Briefly, 6-μm–thick cross sections of frozen optimal cutting temperature compound-embedded pancreas were cut on a cryomicrotome and fixed with 4% paraformaldehyde for 10 min. DHE 2.5 μmol/L was added to the sections and incubated at 37°C for 20 min. Fluorescence images of ethidium-stained islets for each sample were collected at equal exposure conditions under Nikon Eclipse Ti microscopy (magnification ×400) with NIS-Elements BR 3.0 software. The mean fluorescence density of pancreatic islets (20–25 islets/mouse) was also analyzed with NIS-Elements BR 3.0 software.

Total RNA was purified from isolated mouse pancreatic islets using TRIzol Reagent followed by QIAGEN RNeasy Mini Kit. RNA (2 μg) was reverse transcribed using SuperScript III Reverse Transcriptase with random hexamer in the presence of 10 μL deoxyribonucleotide for 1 h at 50°C. The resulting cDNAs were used as templates for real-time PCR, with oligonucleotides primers used to amplify the mRNAs of insulin I (F: 5′-CCTGTTGGTGCACTTCCTAC-3′; R: 5′-TGCAGTAGTTCTCCAGCTGG-3′; size: 317 bp), insulin II (F: 5′-AGCCCTAAGTGATCCGCTACAA-3′; R: 5′-CATGTTGAAACAATAACCTGGAAGA-3′; size: 178 bp), Pdx-1 (F: 5′-CCACCCCAGTTTACAAGCTC-3′; R: 5′-ACGGGTCCTCTTGTTTTCCT-3′; size: 315 bp), DNAJC3 (F: 5′-AAGCCCGTGGAAGCCATTAG-3′; R: 5′-GGTCATTTTCATTGTGCTCCTGAG-3′; size: 160 bp), PCNA (F: 5′-TAAAGAAGAGGAGGCGGTAA-3′; R: 5′-TAAGTGTCCCATGTCAGCAA-3′; size: 175 bp), LC3 (F: 5′-CGAGCGCTACAAGGGTGAG-3′; R: 5′-CCGGATGATCTTGACCAAC-3′; size: 100 bp), and β-actin (F: 5′-AGGTCATCACTATTGGCAACGA-3′; R: 5′-CACTTCATGATGGAATTGAATGTAGTT-3′; size: 118 bp) (25–29). Real-time PCR was performed on a Bio-Rad CFX96 Real-Time PCR Detection System. PCRs were cycled 40 times under the following conditions: 95°C for 5 s and 58°C for 5 s. Homogeneity of PCR products from each reaction was confirmed by melt curve analysis and 1.5% agarose gel analysis.

Data from collected human and mouse Klotho and insulin in pancreases were analyzed using the unpaired t test. Blood glucose and body weight were analyzed by a repeated-measures one-way ANOVA. The remaining data were analyzed by one-way ANOVA. The Newman-Keuls procedure was used to reveal differences between groups. P < 0.05 was considered to be statistically significant.

We performed IHC staining of Klotho and insulin in human pancreas. Both Klotho and insulin staining were significantly decreased in pancreatic islets of patients with T2DM versus those of healthy donors (Fig. 1A and B). We further quantified Klotho protein expression levels in human pancreas using Western blot. Klotho protein expression was significantly decreased in pancreas in patients with T2DM (Fig. 1C).

We next assessed Klotho expression in pancreata in a mouse model of T2DM (db/db mice). The IHC analysis showed a significant decrease in Klotho protein expression in pancreatic islets in db/db mice (Fig. 1D and E). The insulin storage was also significantly decreased in pancreatic islets in db/db mice compared with lean mice (Fig. 1D and E). We previously reported that Klotho protein is expressed in mouse pancreatic islets of Langerhans with an apparent molecular weight of 65 kDa (11). Western blot analysis indicated that Klotho protein expression levels were significantly decreased in pancreas of db/db mice (20 weeks) versus age-matched lean mice (Fig. 1F). These results indicate that downregulation of Klotho protein expression is associated with a decrease in insulin storage in pancreatic β-cells in patients with T2DM and in db/db mice.

To test the β-cell specificity of the mouse insulin II promoter, we transfected MIN6 β-cells, 3T3-L1 preadipocytes, and mIMCD3 cells with 0.072 μg/mL pAAV-GFP, pAAV-mKL, or pAAV-CMV-mKL DNAs for 48 h. In pAAV-GFP and pAAV-mKL, the original CMV promoter was replaced by the mouse insulin II promoter. pAAV-GFP (GFP protein) was specifically expressed in MIN6 cells but not in 3T3-L1 or mIMCD3 cells (Supplementary Fig. 1A), indicating that the insulin II promoter is β-cell specific. The constructed insulin II promoter and the conventional CMV promoter are equally potent in driving mKL expression in MIN6 β-cells (Supplementary Fig. 1B).

We carefully injected rAAV-GFP or rAAV-mKL into the region of pancreas intraperitoneally in lean and db/db mice. These db/db mice were developing severe hyperglycemia at the age of 10 weeks (Fig. 2A). Of note, rAAV-mKL significantly attenuated hyperglycemia and dampened the development of overt diabetes in db/db mice within 2 weeks compared with the PBS- and rAAV-GFP–treated control groups (Fig. 2A). The antihyperglycemic effects of rAAV-mKL were sustained for 6 weeks (length of the study), although it did not eventually prevent the rise of blood glucose levels (Fig. 2A). The rAAV-mKL did not alter blood glucose levels significantly in lean mice (Fig. 2A).

To gain insight into the mechanism of the Klotho action, we performed a GTT at weeks 2, 4, and 6 and an IST at weeks 3 and 5 following Klotho gene delivery. The db/db mice showed overt glucose intolerance versus the lean mice (Fig. 2B–D). Treatments with rAAV-mKL markedly improved glucose tolerance in db/db mice, and glucose tolerance was not altered by rAAV-mKL in lean mice (Fig. 2B–D). The db/db mice developed severe insulin resistance; however, β-cell–specific expression of Klotho did not affect insulin sensitivity in either db/db or lean mice (Fig. 2E and F). These data suggest that the treatments with rAAV-mKL improved β-cell function but did not affect insulin sensitivity in peripheral tissues in diabetic mice.

The db/db-PBS group exhibited slightly higher levels of plasma insulin compared with the lean-PBS group (Fig. 2G and H). rAAV-mKL further increased plasma insulin levels significantly in db/db mice but not in lean mice (Fig. 2G and H). These results suggest that rAAV-mKLs increase insulin release in response to hyperglycemia in db/db mice.

Fasting urine glucose levels in db/db mice were significantly higher than those of lean mice. rAAV-mKL significantly decreased urine glucose levels in db/db mice at weeks 2 and 4 after gene delivery (Fig. 2I).

The control db/db mice had much greater body weights than the control lean mice (Supplementary Fig. 2A). β-Cell–specific delivery of mKL did not affect the body weights significantly in either lean or db/db mice (Supplementary Fig. 2A). The rAAV-mKL slightly but significantly decreased food intake in db/db mice (normalized to body weight) at week 5 after gene delivery (Supplementary Fig. 2B–D). The control db/db-PBS group had greater water intake and urine output than the lean mice (Supplementary Fig. 2E–J). The rAAV-mKL attenuated both water intake and urine output in diabetic mice at weeks 3 and 5 after gene delivery (Supplementary Fig. 2E–J).

Different serotypes of rAAV with single- or double-stranded DNA have been used in pancreatic islet gene transfer with various efficiencies through different routes (16,30,31). Given that the insert genes ranged from 0.7 to 3.1 kilobases, AAV2 vector with single-stranded DNA was used in the gene transfer. A FLAG-tag sequence was inserted at the 3′ end of the mouse Klotho gene. At 6 weeks after gene delivery, we first examined GFP and FLAG-tag protein expression in cross sections of paraffin-embedded pancreatic islets, livers, and kidneys through IHC analysis. The rAAV-GFP drove GFP expression in pancreatic islets of lean and db/db mice, whereas GFP was not detectable in the livers and kidneys of animals injected with rAAV-GFP (Supplementary Fig. 3A and B). In addition, rAAV-mKL drove FLAG-tag expression in islets of animals treated with rAAV-mKL (Supplementary Fig. 3C). Thus, the intraperitoneal delivery of rAAV coupled with mouse insulin II promoter led to islet-specific gene transfer in mice.

We next studied Klotho expression in pancreatic islets in diabetic mice. Klotho staining in pancreatic islets of Langerhans in control db/db mice was significantly decreased compared with that of control lean mice (Fig. 3A and B). The treatments with rAAV-mKL increased Klotho staining in pancreatic islets of both lean and db/db mice (Fig. 3A and B). Western blot analysis also showed that the treatments with rAAV-mKL increased Klotho protein expression in pancreas of lean and db/db mice (Supplementary Fig. 3D and E).

To investigate whether the β-cell–specific expression of mKL exerts beneficial effects on pancreatic islets of Langerhans, we performed insulin staining in pancreatic cross sections. Insulin staining in pancreatic islets was significantly decreased in db/db mice compared with that of lean mice (Fig. 3C and D). Of note, β-cell–specific expression of mKL increased insulin staining in pancreatic islets of db/db mice by 1.26-fold (Fig. 3D). In addition, the number of insulin-positive cells in pancreatic islets of Langerhans in db/db mice was significantly less than that of lean mice (Fig. 3E). In contrast, β-cell–specific expression of mKL partially restored the number of insulin-positive cells in pancreatic islets in db/db mice (by 55%) (Fig. 3E).

We further tested the glucose-stimulated insulin secretion in isolated mouse pancreatic islets (ex vivo). Insulin secretion was lower at both 2.8 and 16.7 mmol/L glucose in islets isolated from db/db mice compared with that of lean mice (Fig. 3F). β-Cell–specific expression of mKL promoted insulin secretion in response to 16.7 mmol/L glucose but not 2.8 mmol/L glucose in islets of lean and db/db mice (Fig. 3F). These results reveal, for the first time, that the pancreatic islets of db/db mice lose compensatory ability in response to increased blood glucose levels and that β-cell–specific expression of mKL improves the impaired response of pancreatic β-cells to glucose challenge in db/db mice.

To study the mechanisms for the preservation of β-cell function by Klotho in db/db mice, we evaluated oxidative stress markers (4-HNE), intracellular superoxide (DHE staining), and insulin transcription factors (Pdx-1) in pancreatic islets. The number of 4-HNE–positive cells and the intracellular superoxide level were significantly increased in pancreatic islets of db/db mice (Fig. 4A–D), indicating oxidative damage. In vivo expression of mouse Klotho attenuated the oxidative stress levels in pancreatic islets of db/db mice (Fig. 4A–D). In addition, the Pdx-1 expression level (staining) was significantly lower in pancreatic islets of db/db animals, whereas β-cell–specific expression of mKL increased Pdx-1 in islets of db/db mice (Fig. 4E and F). Thus, the beneficial effects of Klotho on β-cells may involve suppression of oxidative stress and enhancement of Pdx-1 expression in diabetic mice.

Because oxidative stress could damage cells, we further assessed the endoplasmic reticulum (ER) stress marker DNAJC3, autophagy marker LC3, cell proliferation, and cell apoptosis in pancreatic islets in diabetic mice. The number of DNAJC3-positive cells in the pancreatic islets was significantly increased in db/db mice versus lean mice. In contrast, β-cell–specific expression of mKL decreased the number of DNAJC3-positive cells in islets of db/db mice, suggesting that expression of Klotho suppresses ER stress in islets of db/db mice (Fig. 5A and B). LC3 staining was lower in islets of db/db mice compared with that of lean mice, suggesting that the autophagic activity was decreased in pancreatic islets of db/db mice. Of note, β-cell–specific expression of mKL restored autophagic activity in db/db mice (Fig. 5C and D).

On the other hand, the number of PCNA-positive cells in islets was increased in db/db mice compared with lean mice, and β-cell–specific expression of Klotho further increased the number of PCNA-positive cells in islets of db/db mice (Fig. 6A and B). Therefore, in vivo expression of Klotho further promoted cell proliferation in pancreatic islets of db/db mice. The number of apoptotic cells was increased in islets of db/db mice, which can be dampened by expression of Klotho (Fig. 6C and D). Thus, rAAV-mKL treatments attenuated apoptosis in pancreatic islets of db/db mice (Fig. 6C and D). Therefore, the preservation of β-cell function in db/db mice may be partially attributed to suppression of ER stress and apoptosis, restoration of autophagic activity, and enhancement of β-cell proliferation.

We also assessed whether Klotho affects the corresponding gene expressions in line with the changes in several proteins involved in the preservation of pancreatic islets. The db/db mice exhibited significantly lower mRNA expression levels of insulin I, insulin II, Pdx-1, and LC3, whereas the treatments with rAAV-mKL increased mRNA levels of these genes in islets of diabetic mice (Fig. 7A–D). In addition, db/db mice had higher mRNA expression levels of DNAJC3 and PCNA, whereas the rAAV-mKL treatments decreased DNAJC3 mRNA expression but further increased PCNA mRNA expression levels in islets (Fig. 7E and F). These data suggest that β-cell–specific expression of mKL preserves β-cells through regulating the gene expression of insulin I, insulin II, Pdx-1, PCNA, and LC3 mRNA.

Because the beneficial effects of Klotho on β-cells involved suppression of superoxide production and oxidative stress (Fig. 4A–D), we further investigated the underlying mechanism in MIN6 β-cells. High glucose levels increased the activity of NADPH oxidases (Fig. 8A), an important source of superoxide in MIN6 β-cells. Of note, overexpression of mKL abolished high glucose–induced activation of NADPH oxidases and superoxide production in MIN6 β-cells (Fig. 8A–C), suggesting that the NADPH oxidase is involved in the upregulation of superoxide generation. Furthermore, overexpression of mKL eliminated high glucose–induced activation of Rac1 (p-Rac1) without altering the total Rac1 level (Fig. 8D–F), a key regulator of the NADPH oxidase activity. These results suggest that the suppressor effect of mKL on NADPH oxidase activity is mediated by inhibition of Rac1 phosphorylation.

We next assessed the subcellular localization of the short-form Klotho (65 kDa) using confocal microscopy. Endogenous 65-kDa Klotho (red) was found in the cytosol and nucleus (Supplementary Fig. 4A). Our previous study indicated that the short-form Klotho is also expressed in the plasma membrane (11), although the current IHC method could not show its membrane localization. The transgene-expressed Klotho detected using FLAG-tag staining (green) was located in the cytosol and outside the nucleus (Supplementary Fig. 4B).

We further assessed the effects of recombinant short-form Klotho (65 kDa) and full-length Klotho (130 kDa) protein on β-cell function in the β-islets isolated from lean mice. The β-islets were treated with 65- and 130-kDa Klotho proteins. We found that only short-form Klotho promoted insulin secretion in β-islets (Supplementary Fig. 5). The full-length Klotho did not have obvious effects on insulin secretion in β-islets (Supplementary Fig. 5). Exogenous short-form Klotho could bind to the cell membrane. We demonstrated previously that short-form Klotho in the cell membrane enhances glucose-induced insulin secretion by upregulating membrane levels of transient receptor potential V2 (TRPV2), which increases glucose-induced calcium responses (11).

Pancreatic β-cells are essential to the regulation of glucose homeostasis. Substantial β-cell failure is now believed to occur at an early stage in the progression of T2DM (2). Thus, one of the goals in the treatment of T2DM is to preserve functional β-cells. Klotho, a recently discovered aging-suppressor gene, was believed to be expressed in kidneys (21). Our most recent study indicates that the Klotho gene and protein are expressed in pancreatic islets (11). Notably, Klotho protein expression in β-cells was decreased in both patients with T2DM and db/db mice, a mouse model of T2DM (Fig. 1A–F). The novel finding of the current study is that β-cell–specific expression of Klotho attenuated the development of diabetes and enhanced glucose tolerance in db/db mice. The beneficial effects of Klotho is likely due to the increases in number of insulin-positive β-cells, insulin storage levels in pancreatic islets, and glucose-stimulated insulin secretion from pancreatic islets, which led to the increased blood insulin levels in db/db mice.

The depleted Klotho protein expression was associated with decreased insulin storage in pancreatic islets (Fig. 1A, B, D, and E) and the impaired glucose-stimulated insulin release in pancreatic islets of db/db mice (Fig. 3F). β-Cell–specific expression of Klotho improved or preserved β-cell function (Figs. 2 and 3). Of note, a new finding is that β-cell function may be regulated by Klotho. These findings are supported by those of our previous cell culture study that silencing of the Klotho gene impaired glucose-stimulated insulin release and that overexpression of Klotho promoted glucose-stimulated insulin secretion in MIN6 β-cells (11). Klotho enhances glucose-induced insulin secretion by regulating plasma membrane levels of TRPV2 and intracellular levels of calcium (11). Klotho-deficient mice exhibit hypoinsulinemia and pancreatic islet atrophy with diminished insulin protein and mRNA levels (10). The current study further demonstrated that β-cell–specific expression of Klotho increased Pdx-1, insulin I, and insulin II mRNA levels and their corresponding proteins levels in pancreatic islets of db/db mice. Pdx-1 is the major regulator of glucose-stimulated insulin gene transcription. Specific point mutations in Pdx-1 are associated with maturity-onset diabetes of the young 4 and late-onset T2DM, as characterized by a decline in β-cell function (32). The present study suggests that the promoting effects of Klotho on insulin synthesis may be partially attributed to the increased Pdx-1 expression. Further study is required to elucidate the mechanism for the regulation of Pdx-1 gene expression by Klotho in pancreatic islets.

β-Cell–specific expression of Klotho did not alter insulin sensitivity. The insulin II promoter was specific in driving gene expression in pancreatic β-cells because GFP was exclusively detected in pancreatic islets but was not detectable in peripheral tissues (liver and kidneys) in mice treated with rAAV-GFP (Supplementary Fig. 3A). rAAV-mKL was expressed in pancreatic islets as indicated by expression of FLAG-tag (Supplementary Fig. 3C). The insulin II promoter was also potent in driving Klotho gene expression in β-islets (Supplementary Fig. 3D and E). This promoter is as potent as the CMV promoter (Supplementary Fig. 1B).

We further explored the potential mechanisms by which β-cell–specific expression of Klotho protected against β-cell failure in pancreatic islets of db/db mice. rAAV-mKL decreased reactive oxygen species (ROS) and oxidative damage as measured by DHE and 4-HNE, respectively, in pancreatic islets of db/db mice. Oxidative stress induced by ROS is critically involved in the impairment of β-cell function during the development of diabetes (33). Because of their low antioxidant capacity, β-cells are extremely susceptible to oxidative stress (34). Hyperglycemia and hyperlipidemia cause oxidative damage to proteins, lipids, and DNA in β-cells as the result of a combination of increased free radical production and an impaired ability of cells to detoxify the radicals and repair damaged molecules (33). By covalently modifying membrane-associated proteins, the membrane lipid peroxidation product 4-HNE may play particularly sinister roles in the metabolic syndrome and associated disease processes (35).

Superoxide is the major source of ROS that causes oxidative damage, so we further explored the mechanism of Klotho-induced reduction of superoxide production under hyperglycemic conditions. Because a study of the mechanistic link of Klotho and superoxide generation may be compromised in the in vivo animal experiment, we investigated how Klotho attenuates high glucose–induced superoxide production in MIN6 β-cells. The data suggest that Klotho decreases high glucose–induced upregulation of NADPH oxidase activity and superoxide production likely through suppressing phosphorylation of Rac1 (Fig. 8), a key regulator of NADPH oxidase. Thus, this result reveals a previously unidentified role of Klotho in the regulation of Rac1 and NADPH oxidase activity in β-cells. Of note, Klotho attenuates high glucose–induced upregulation of NADPH oxidase activity and superoxide production without alteration of glucose levels, suggesting that Klotho has a direct protective effect in β-cells. On the other hand, in the animal study, the beneficial effects of overexpression of Klotho in β-cells may also be partially attributed to the euglycemic effect of Klotho due to increased insulin secretion. Collectively, these data indicate that β-cell–specific expression of Klotho may preserve β-cells partially by suppressing superoxide production and oxidative stress in pancreatic β-islets in T2DM.

Oxidative stress could impair ER function, leading to ER stress (36,37). As a secretory cell that synthesizes and releases a large amount of insulin, the β-cell is expected to be susceptible to alterations in ER homeostasis, which can result in the accumulation of unfolded, misfolded, and/or aggregated proteins (a phenomenon known as ER stress) (38). Eukaryotic cells respond to ER stress by activating the unfolded protein response (UPR), a process that allows cells to adapt to and attempt to relieve ER stress conditions (38). Hyperactivation of the UPR is indispensible for ER homeostasis and may be involved in β-cell dysfunction and death during the progression of T2DM. The level of ER chaperone protein DNAJC3 is elevated in pancreatic islets of db/db mice and human T2DM (39). β-Cell–specific expression of mKL attenuated diabetes-induced increases in DNAJC3 expression in pancreatic islets (Fig. 5). Thus, this result reveals that Klotho attenuated ER stress, which may contribute to the preservation of β-cells and protection against diabetes by β-cell–specific expression of Klotho.

The current study demonstrates that β-cell–specific expression of Klotho enhances cell proliferation and decreases cell apoptosis in pancreatic islets of db/db mice (Fig. 6A–D). These results suggest for the first time that β-cell–specific expression of Klotho may preserve β-cells partially through attenuating apoptosis and promoting cell proliferation in pancreatic islets of db/db mice. This beneficial effect may be partially mediated by Klotho-induced suppression of superoxide production and oxidative stress, which are known to cause cell apoptosis and impair cell proliferation (40). Regulation of β-cell mass is dynamic and tightly matched to meet the body’s demand for insulin (41). The rates of β-cell apoptosis or necrosis and β-cell proliferation or neogenesis equilibrate at a frequency of 0.5% under steady-state conditions (42). It was reported that β-cell apoptosis contributes to the reduction of β-cell mass in patients with T2DM (43,44). It has been shown that Klotho and FGF23 together promote cell proliferation in vitro (45,46). The current diabetic model revealed that the frequency of proliferative and apoptotic events per islet is relatively low (1–2%). We believe that the accumulated effects of proliferation and apoptosis in β-cell mass may take time.

An unexpected finding was that the expression of LC3, a marker of autophagy, decreased in pancreatic islets of db/db mice, whereas β-cell–specific expression of mKL reversed the downregulation of LC3 expression in islets of db/db mice (Figs. 5D and 7D). Autophagy is a physiologically preserved process that maintains homeostatic functions such as protein degradation and organelle turnover (47). A major member of this family is LC3, which is associated with the autophagosome from its formation to its maturation into autolysosome and serves as a bona fide marker for autophagy (48). Loss-of-function experiments (Atg7^f/f:RIP-Cre mice) have demonstrated that autophagy in β-cells is critical in the preservation of pancreatic β-cell function (49). Accumulation of p62, a substrate for autophagy, in β-cells of db/db mice has been observed (50). Autophagy deficiency may be involved in lipotoxicity-induced β-cell failure in T2DM (2,48). Therefore, the current study indicates that the restoration of autophagic activity by expression of Klotho may contribute to the preservation of β-cells in pancreatic islets of diabetic mice. The regulation of autophagy by Klotho may be partially attributed to suppression of superoxide production and oxidative stress, which are known to disrupt autophagy (40).

The major type of endogenous Klotho in pancreatic β-cells is ∼65 kDa based on Western blot analysis (Figs. 1C, 1F, and 8D). Of note, overexpression of Klotho produced two types of Klotho, short form (65 kDa) and full length (130 kDa) (Fig. 8D, Supplementary Fig. 3D). However, the 65-kDa Klotho seems to be the functional protein because treatment with 65-kDa, not 130-kDa, Klotho protein promoted glucose-induced insulin secretion (Supplementary Fig. 5). Therefore, the beneficial effects of in vivo β-cell–specific expression of Klotho in T2DM may be mediated by the 65-kDa Klotho. Of note, Klotho promotes high glucose-stimulated insulin secretion in β-cells but does not affect insulin secretion at low glucose levels (Fig. 3F, Supplementary Fig. 5). This finding is consistent with our previous observation in MIN6 β-cells (11).

In summary, β-cell–specific expression of Klotho preserved β-cell function and protected against the development of T2DM in db/db mice. Particularly, β-cell–specific expression of Klotho attenuated hyperglycemia and improved glucose tolerance in diabetic mice. This protection was associated with significant increases in number of β-cells, Pdx-1 levels, and insulin storage levels in pancreatic islets; the glucose-stimulated insulin secretion from pancreatic islets; and blood insulin levels. β-Cell–specific expression of Klotho preserved β-cell function likely by suppressing oxidative stress, ER stress, and apoptosis; increasing cell proliferation; and normalizing autophagy in pancreatic islets of db/db mice. Therefore, in vivo expression of Klotho in pancreatic β-cells may offer a new and effective therapeutic strategy for β-cell dysfunction in T2DM. These promising findings warrant further mechanistic investigation into the role of Klotho in regulating β-cell function. β-Cell–specific expression of Klotho attenuates but does not prevent the development of T2DM; therefore, simultaneous management of hyperglycemia and insulin resistance is also important for the protection of β-cells in T2DM.

Funding. This work was supported by NIHhttp://dx.doi.org/10.13039/100000002 grants R01-DK-093403, HL-105302, HL-102074, and HL-118558. This publication was made possible by NIHhttp://dx.doi.org/10.13039/100000002 grant 9P20-GM-104934-06 from the Centers of Biomedical Research Excellence program of the National Institute of General Medical Sciences.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.L. contributed to the experiments, data analysis, and writing of the manuscript. Z.S. contributed to the study concept and hypotheses, study design, and writing of the manuscript. Z.S. is the guarantor of this work and, as such, takes responsibility for the integrity of the data and the accuracy of the data analysis.