Amelioration of Diabetes by Protein S

Diabetes is a fast-growing global public health problem that is associated with increased morbidity and mortality (1). The prevalence of diabetes continues to increase, with an estimated rise in the worldwide population with diabetes from 285 million in 2010 to >400 million by 2030 (2). Diabetes is the fourth leading cause of death; the risk of death for people with diabetes is twice that of people without diabetes, and life expectancy is 5–10 years shorter among middle-aged patients with diabetes (2). Of the two major forms of diabetes, type 1 may be caused by genetic, environmental, or autoimmune factors, leading to selective apoptosis or destruction of insulin-producing pancreatic islet β-cells (3,4). Type 2 diabetes is associated with insulin resistance and abnormal insulin secretion, with evidence suggesting that reduced β-cell mass is linked to dysfunctional insulin secretion (3,5). Apoptosis of β-cells caused by glucotoxicity, lipotoxicity, advanced glycation end products, inflammatory cytokines, and intracellular deposition of islet amyloid polypeptide are believed to be the cause of the reduced number of islet β-cells in type 2 diabetes (5). Therefore, apoptosis of β-cells is a mechanistic pathway common to both type 1 and type 2 diabetes.

Protein S (PS) is a vitamin K–dependent glycoprotein that acts as an anticoagulant factor by enhancing the inhibitory activity of activated protein C on blood coagulation (6). PS may also directly stimulate the inhibition of the tissue factor pathway (7). In addition, PS regulates the inflammatory response and apoptosis pathways through Tyro3, Axl, and Mer (TAM) tyrosine kinase receptors (8). PS inhibits the expression of inflammatory cytokines from a variety of cells and is protective against lipopolysaccharide-induced acute lung injury (9). It supports the neuroprotective action of activated protein C (10) and circulates in plasma as both free and in complex with C4b-binding protein (C4BP), which is an inhibitor of the classic complement pathway (11). Localization of C4BP to cell membrane by PS may inhibit inflammation induced by complement activation (11). In addition, PS can suppress inflammatory and immune responses by enhancing the clearance of apoptotic cells by macrophages through binding to negatively charged phospholipids exposed on apoptotic cells (12). Furthermore, PS can directly inhibit cell apoptosis by activating the Akt signaling pathway (13). On the basis of the antiapoptotic activity of PS, we hypothesized that PS protects against diabetes by inhibiting apoptosis of pancreatic β-cells and the development of diabetic nephropathy.

Blood samples were obtained from 32 patients with type 1 (n = 6) and type 2 (n = 26) diabetes and controlled glycemia and 37 healthy volunteers for determination of PS and C4BP levels (Table 1). None of the patients had diabetic nephropathy or neuropathy. All patients and control subjects provided informed consent, and the protocol was approved by the Ethics Committee for Clinical Investigation of Mie University (Approval No. 2194).

Homozygous human PS (hPS) transgenic (TG) mice on a C57BL/6 background have been previously characterized (14). Briefly, the full-length hPS cDNA was cloned into a pCAG plasmid containing the CAG-promoter (cytomegalovirus enhancer + chicken + β-actin promoter) and rabbit β-globin polyadenylation sites. The plasmid was digested, purified, and microinjected into fertilized eggs from C57BL/6J mice, and TG founders were screened by Southern blotting. Most hPS TG organs express hPS, and its mean plasma concentration is 85 ± 3 μg/mL (14). Wild-type (WT) littermates were used as controls. Pancreatic islet mass is not different between WT and hPS mice. Male WT mice (8–10 weeks old) weighing 19–22 g and db/db male mice (6 weeks old) weighing 31–33 g were from Nihon SLC (Hamamatsu, Japan). All animals were maintained in a specific pathogen-free environment and subjected to a 12-h light/dark cycle in the animal house of Mie University. The research followed the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines for animal investigation. The Committee on Animal Investigation of Mie University approved the protocols (Approval No. 24-50), and animal procedures were performed in accordance with the institutional guidelines of Mie University. Mice were randomized, and researchers who measured parameters were blinded to treatment groups.

To evaluate susceptibility to diabetes induction, 200 μL (40 mg/kg body weight) streptozotocin (STZ) (Sigma, St. Louis, MO) was intraperitoneally administered for 5 consecutive days to hPS TG (hPS TG/STZ) and WT littermates (WT/STZ); control mice received 200 μL saline solution (hPS/SAL, WT/SAL) intraperitoneally for 5 consecutive days.

Fasting blood glucose levels were measured after STZ (hPS TG/STZ, WT/STZ) or saline (hPS/SAL, WT/SAL) injection once a week for 4 weeks. On the third week after STZ or saline injection, an intraperitoneal glucose tolerance (IPGT) test was performed after 16 h of fasting by intraperitoneal injection of glucose (1 g/kg mouse body weight), and tail vein blood was sampled on days 0, 7, 14, 21, and 28 for glycemia measurement. An insulin sensitivity test was performed in nonfasting mice by intraperitoneal injection of insulin (1 unit/kg mouse body weight), and an insulin secretion test was carried out after 16 h of fasting by intraperitoneal injection of 3 g/kg mouse body weight of glucose. Tail vein blood was sampled for measuring glucose or insulin level. All groups were killed 4 weeks after STZ or saline injection, and pancreatic tissues were excised for histochemistry.

hPS TG mice and WT littermates underwent unilateral nephrectomy, and after 4 weeks of recovery, WT/STZ and hPS TG/STZ mice received injections of STZ on 5 consecutive days, and then an average of five additional injections of STZ were given to hPS TG mice to induce diabetes of severity equal to that in WT mice (15). Control (hPS/SAL, WT/SAL) mice received intraperitoneally similar amounts of saline. Mice were killed 8 weeks after the last STZ or saline injection. In separate experiments, WT mice underwent unilateral nephrectomy, received intraperitoneally injected STZ or saline for 5 consecutive days after 4 weeks of recovery, and were then treated with either hPS or saline subcutaneously through osmotic minipumps for 4 weeks before they were killed (15).

After mice were killed by pentobarbital overdose, pancreas and kidneys were dissected, dehydrated, embedded in paraffin, and cut into 3-μm-thick sections, and the pancreas was prepared for hematoxylin-eosin staining and immunostaining. The kidneys were prepared for periodic acid Schiff (PAS) and Masson trichrome staining. An investigator blinded to the treatment group calculated the areas of pancreatic islets stained with hematoxylin-eosin and glomeruli (>30 per mouse) stained positive for PAS or trichrome by using an Olympus BX50 microscope with a plan objective and equipped with an Olympus DP70 digital camera (Tokyo, Japan) and WinROOF image processing software (Mitani Corp., Fukui, Japan).

Immunostaining of insulin, glucagon, and F4/80 was performed at Biopathology Institute Corporation by using specific antibodies from Dako Corporation (Carpinteria, CA) and Novus Biologicals (Littleton, CO) while following standard methods. An investigator blinded to the experimental group measured the area of immunoreactivity for insulin or glucagon in all visible islets from four pancreatic sections (40–50 fields per mouse) as described above.

Glucose in blood was measured by glucose oxidase method and insulin by a kit from ALPCO (Salem, NH). Thrombin-antithrombin complex (Cedarlane Laboratories, Burlington, Ontario, Canada) and total plasminogen activator inhibitor-1 (PAI-1) (E1/PAI-1; R&D Systems, Minneapolis, MN) were measured with enzyme immunosorbent assay (EIA) kits and total collagen and hydroxyproline as previously described (15). Complement C4BP was assayed with an EIA kit from Assaypro (St. Charles, MO) and total hPS as previously described (14). To assess free hPS, a 96-well microplate was coated with C4BP (ATGen Corp., Sampyeong-dong, Korea), and after appropriate washing and blocking, biotin-labeled polyclonal rabbit anti-hPS antibody (DakoCytomation, Glostrup, Denmark) was added. Urinary liver-type fatty acid–binding protein (L-FABP), a marker of diabetic nephropathy severity (16), was measured with an EIA kit (R&D Systems), tissue factor by EIA using primary and biotinylated antibodies (Santa Cruz Biotechnology, Dallas, TX), and tissue plasminogen activator activity by chromogenic substrate (S-2288; Carbiochem, Nottingham, U.K.). HOMA of insulin resistance (HOMA-IR) was determined as follows (17):

The mouse pancreatic β-cell line MIN6 (provided by J. Miyazaki, Osaka University), L6 rat skeletal myoblasts (provided by H. Ashida, Kobe University), HepG2 cells (RIKEN Cell Bank, Ibaraki, Japan), and murine 3T3-L1 and rat RAW264.7 cells (American Type Culture Collection) were cultured in DMEM (Sigma-Aldrich) containing 10% volume for volume heat-inactivated FCS. Differentiation to adipocytes was induced by treating with 0.5 mmol/L 3-isobutyl-1-methylxanthine, 4 mg/mL dexamethasone, 1 mg/mL insulin, and 10% FCS.

Cells were incubated in DMEM (low glucose) + 1% BSA for 6 h and placed in DMEM with high glucose (1 g/L) before adding plasma-derived hPS (20 μg/mL) (Enzyme Research Laboratories, South Bend, IN; purity >95%), and 30 min after hPS treatment, insulin (200 nmol/L) was added and cultured for 4 h. Glucose content of culture supernatant was measured by Glucose Colorimetric/Fluorometric Assay Kit (BioVision).

Pancreatic tissues were excised from killed WT and hPS TG mice after 4 weeks of intraperitoneal STZ or saline, cut into 1- to 2-mm pieces, and incubated for 30 min at 37°C in 1 mg/mL collagenase. After centrifugation and resuspension, the cells were placed on a discontinuous Percoll gradient and centrifuged, and the islet cell layer was collected, washed, and dispersed with trypsin/EDTA solution before use in assays (18).

Apoptosis in histological samples of pancreas islets was assayed by TUNEL method with a commercially available kit (Chemicon International, Temecula, CA). The number of TUNEL-positive cells within the islets was counted in six pancreatic sections (40–50 fields) per mouse by a blinded investigator as described previously.

Apoptosis of primary islet β-cells and MIN6 β-cell lines was analyzed by flow cytometry (BD Biosciences, Oxford, U.K.) after staining with fluorescein isothiocyanate-annexin V (BD Pharmingen) and propidium iodide. Apoptosis of MIN6 β-cell lines was assessed by immunofluorescence microscopy after similar staining. The mRNA expression of baculoviral inhibitor of apoptosis repeat-containing (BIRC) 3/inhibitor of apoptosis (IAP), B-cell lymphoma 2 (Bcl-2) family of regulator proteins, and apoptotic protease activating factor 1 (APAF1) were also measured.

The effect of exogenous hPS on insulin sensitivity was evaluated in db/db mice. Fasting db/db mice received subcutaneous hPS 2 mg/kg (n = 5) or saline (n = 5), and insulin sensitivity test and HOMA-IR measurement were performed at 0, 1, 2, 4, and 6 h. To study the in vivo effects of hPS on STZ-induced β-cell apoptosis, exogenous hPS 2 mg/kg (n = 6) or saline (n = 6) was subcutaneously administered for 5 days to WT mice 1 h before STZ injection and 9 days after the last STZ injection before mice were killed. Islets were isolated, incubated with trypsin-EDTA solution at 37°C, and gently dispersed, and apoptosis was assessed by flow cytometry.

RAW264.7 cells were pretreated with or without hPS (20 μg/mL) for 30 min and then treated with or without high glucose for 24 h. Expression of the M1 marker inducible nitric oxide synthase and the M2 marker arginase-1 was analyzed by RT-PCR.

MIN6 cells were pretreated with anti-Tyro3, anti-Axl, anti-Mer, or isotype IgG for 30 min before adding 20 μg/mL hPS and 2 mmol/L STZ and cultured for 24 h, and apoptosis was evaluated. All antibodies were goat IgGs from R&D Systems.

Min6 cells were transfected with 33 nmol BIRC3 small interfering RNA (siRNA) or scrambled siRNA cultured in the presence or absence of hPS for 30 min and then treated with or without 3 mmol/L STZ for 24 h. Apoptotic cells were assessed by flow cytometry.

Activation of Akt/PKB and nuclear factor-κB (NF-κB) in MIN6 cells was evaluated by Western blot while following standard methods using antibodies against p-Akt/PKB, Akt/PKB, p-inhibitor of κB (IκB), cleaved form of caspase-3, and β-actin from Cell Signaling (Danvers, MA) and antibody against human/mouse cIAP-2/HIAP-1 from R&D Systems. Activation of Akt/PKB and NF-κB in primary islet β-cells by hPS was assessed by flow cytometry in the presence or absence of anti-hPS.

Total RNA was extracted from pancreas tissues by TRIzol Reagent (Invitrogen, Carlsbad, CA) before reverse transcription using oligo(dT) primers and DNA amplification by PCR using a SuperScript preamplification system kit (Invitrogen). Applied Biosystems StepOne Real-Time PCR System, Taqman Master Mix, and SYBR Green were used for quantitative amplification. Supplementary Table 1 describes primer sequences. The data were analyzed by using 7500 software from Applied Biosystems, and gene expression was normalized by the GAPDH transcription level.

Data are expressed as the mean ± SEM unless otherwise specified. The statistical difference among several variables was calculated by ANOVA with post hoc analysis by Tukey test and between two variables by Mann-Whitney U test. Statistical analyses were done using StatView 4.1 software for Macintosh (Abacus Concepts, Berkeley, CA). Statistical significance was considered at P < 0.05.

There was no difference in BMI or age between patients with type 2 diabetes and healthy subjects (Table 1). The plasma concentration of free hPS was significantly decreased in patients with type 1 and type 2 diabetes, and that of C4BP was significantly increased in patients with type 2 diabetes compared with control subjects (Table 1). There was no significant difference by sex. A significant inverse correlation between free PS and hemoglobin A_1c (HbA_1c) and a significant proportional correlation between C4BP and HbA_1c were observed in patients with type 2 diabetes (Supplementary Table 2). Age was significantly correlated with free PS and C4BP in patients with type 2 diabetes. No significant correlation was found between parameters in patients with type 1 diabetes (Supplementary Table 3).

Diabetes was induced with STZ in hPS TG and WT mice, and diabetes severity was compared. The nonfasting blood glucose level was significantly decreased in the hPS TG/STZ group compared with WT/STZ group on days 14, 21, and 28 after the first STZ injection (Fig. 1A). The blood glucose level in WT/STZ mice was significantly increased from day 7 after STZ injection compared with WT/SAL, whereas the significant increase occurred from day 21 after STZ injection in the hPS TG/STZ group compared with the hPS/SAL group. The IPGT test disclosed significantly lower levels of glycemia in the hPS TG/STZ group than in the WT/STZ group at all time points after glucose injection; the levels of glycemia were significantly higher in the hPS TG/STZ and WT/STZ groups than in the saline groups (Fig. 1B). The blood glucose levels were also significantly increased in the WT/STZ group compared with the hPS TG/STZ group during the insulin sensitivity test at all time points after glucose administration (Fig. 1C), and the plasma insulin was significantly higher in the hPS TG/STZ group than in the WT/STZ after 30 min of intraperitoneal glucose injection (Fig. 1D).

Insulin sensitivity was significantly improved in db/db mice treated with exogenous hPS 15 min after insulin injection and remained better over time compared with control mice. HOMA-IR was improved in hPS-treated db/db compared with controls, although it was not at a significant level (Supplementary Fig. 1A and B).

Treatment of rat L6 skeletal muscle cells, HepG2 human liver cells, and 3T3-L1 mouse adipocytes with hPS increased their insulin sensitivity (Fig. 2A). This effect was specific to hPS because it could be inhibited by anti-hPS antibody (Fig. 2B).

Compared with the WT/SAL, hPS/SAL, and hPS TG/STZ groups, the total area of pancreatic islets (Supplementary Fig. 2A and B) and the area stained positive for insulin were significantly reduced, whereas the area with apoptotic cells was significantly increased in the WT/STZ group at the time of kill on day 28 after STZ injection (Fig. 3A and B). As the insulin stained area decreased, the area of glucagon staining was significantly increased in the WT/STZ group compared with the WT/SAL group but remained unchanged in the hPS TG/STZ compared with the saline-treated group (Fig. 3A).

Infiltration of macrophages was significantly decreased in pancreatic islets from hPS TG/STZ mice compared with WT/STZ and WT/SAL mice; there was no difference between the hPS/SAL and hPS TG/STZ groups (Supplementary Fig. 3A and B). In vitro, hPS promoted differentiation of M2-type macrophages (Supplementary Fig. 3C).

C57BL/6 WT mice were treated subcutaneously with exogenous hPS or saline during and after STZ injection, and islet β-cell apoptosis was assessed. Mice treated with hPS had significantly lower β-cell apoptosis than saline-treated controls (Supplementary Fig. 4A–C).

Apoptosis of MIN6 cells induced by STZ was significantly suppressed when the cells were pretreated with hPS compared with control cells, and cleavage of caspase-3 was significantly inhibited by hPS (Fig. 4A). There was increased phosphorylation of Akt/PKB and IκB in MIN6 cells treated with hPS compared with cells treated with vehicle (Fig. 4B). Akt/PKB and IκB were also significantly phosphorylated by hPS in primary β-cells (Supplementary Fig. 5A and B). MIN6 cells expressed all three TAM PS receptors (Supplementary Fig. 6), but only Mer mediated the inhibitory activity of hPS on apoptosis (Supplementary Fig. 7A and B).

We determined changes in the expression of all members of the mouse BIRC (IAP) family in MIN6 cells treated with either STZ or hPS or both reagents. The expression of BIRC1b and BIRC3 mRNA were significantly increased in cells pretreated with hPS (hPS/SAL) compared with cells pretreated with saline (SAL/SAL) before additional saline. BIRC3 expression was significantly increased, and BIRC1b tended to be high (P = 0.05) in cells pretreated with hPS (hPS TG/STZ) compared with cells pretreated with saline (SAL/STZ) before addition of STZ (Supplementary Fig. 8). The significantly increased expression of BIRC3 mRNA in hPS-treated MIN6 cells (hPS/SAL, hPS TG/STZ) compared with controls was confirmed by Western blot (Fig. 4C). Knockdown of BIRC3 with specific siRNA abolished the hPS-mediated inhibition of MIN6 cell apoptosis (Supplementary Fig. 9A–C). BIRC3 mRNA expression was also significantly increased in primary islet β-cells from hPS TG/STZ mice compared with cells isolated from WT/STZ mice (Fig. 4D). hPS increased the expression of antiapoptotic Bcl-2 but blocked the expression of proapoptotic Bax in MIN6 cells (Supplementary Fig. 10).

We assessed whether hPS can improve diabetic nephropathy independently of its protective activity on β-cell apoptosis. WT C57BL/6 mice were unilaterally nephrectomized, and after complete recovery from surgery, they were made diabetic with STZ or kept as controls with saline. After 4 weeks, mice from the STZ/p-hPS and STZ/p-SAL groups, both with equal diabetes severity, and mice from the SAL/p-hPS and SAL/p-SAL groups, with no diabetes, were treated with hPS or saline through subcutaneous osmotic minipumps for a second 4-week period (Fig. 5A). Both STZ/p-SAL and STZ/p-hPS mice became diabetic, with significant weight loss and increased blood glucose levels compared with the saline-treated groups. There was a significant difference in body weight but not in blood glucose between the STZ/p-SAL and STZ/p-hPS groups (Fig. 5B). The renal tissue content of collagen and hydroxyproline, plasma creatinine, and L-FABP and the renal areas positive for PAS staining and apoptotic cells were increased in the STZ/p-SAL group compared with control and STZ/p-hPS groups (Fig. 5C–E). Tumor growth factor-β1 (TGF-β1) mRNA expression was significantly elevated, whereas that of podocin was significantly reduced in the STZ/p-SAL group compared with the control and STZ/p-hPS groups (Fig. 5F). There were no differences in the level of thrombin-antithrombin complex, a marker of coagulation activation (Supplementary Fig. 11).

In a separate experiment, unilaterally nephrectomized hPS TG and WT mice were made diabetic by using a higher dose of STZ in hPS TG mice than that used in WT mice so that both WT and hPS would have similar blood glucose levels during the entire duration of diabetes (Fig. 6A). As planned, the blood glucose levels were not significantly different between the WT/STZ and hPS TG/STZ groups (Fig. 6B). In contrast, despite the similar levels of hyperglycemia, plasma creatinine, L-FABP, and renal hydroxyproline content, the glomerular mesangial expansion and collagen deposition were significantly decreased in hPS TG/STZ mice compared with WT/STZ mice (Fig. 6B and C). There was no significant difference in signal for F4/80-positive macrophages in the kidneys among groups (data not shown). There were no differences in markers of coagulation (tissue factor) and fibrinolysis (PAI-1, tissue plasminogen activator) (Supplementary Fig. 12).

This study showed that hPS attenuates experimental diabetes by inhibiting apoptosis of pancreatic islet β-cell and the development of diabetic nephropathy.

hPS is a 69-kDa glycoprotein with anticoagulant, anti-inflammatory, and antiapoptotic properties that is expressed by a variety of cells (8). It inhibits inflammation by decreasing leukocyte infiltration and the release of cytokines (e.g., interleukin-6, tumor necrosis factor-α [TNF-α]) and chemokines (MCP-1) by stimulating apoptotic cell clearance through macrophage-mediated phagocytosis or by blocking cell apoptosis through TAM receptors and the Akt/PKB pathway (9,11–13). PS suppression of apoptosis may prevent autoimmune responses, but it may also be deleterious in some conditions, such as cancer or hepatitis (14,19). In agreement with its antiapoptotic property, we show that hPS protects pancreatic islet β-cells from apoptosis and attenuates STZ-induced diabetes. hPS TG mice have significantly less hyperglycemia and more insulinemia and require high doses of STZ to develop diabetes severity similar to their WT counterpart. STZ induced significantly less apoptosis in hPS-pretreated MIN6 β-cell lines and in islet β-cells from mice overexpressing hPS and from mice treated systemically with exogenous hPS compared with controls, as demonstrated by pathological, immunostaining, and flow cytometry studies. Overall, these findings show for the first time in our knowledge that hPS attenuates diabetes induced by STZ-mediated apoptosis of pancreatic islet β-cells.

Apoptosis of β-cells plays a critical role in the pathogenesis of type 1 diabetes, but its significance in type 2 diabetes remains unclear (20). Insulin resistance, hyperinsulinemia, and β-cell hyperplasia are typical features of preclinical type 2 diabetes (21). However, diabetes becomes clinically overt when there is relative insulin deficiency, which is believed to result from apoptosis-associated decreased β-cell mass (22). Islet β-cell apoptosis can be triggered by death receptor signaling, by imbalance between pro- and antiapoptotic proteins, and by proapoptotic effectors activated during endoplasmic reticulum stress, with the three pathways converging to activate caspase-3 to execute apoptosis (23). Critical pathways that protect β-cells from apoptosis in diabetes are Akt/PKB and NF-κB, IAP proteins, and some proteins from the Bcl-2 family (24–28). Previous studies demonstrated that hPS inhibits apoptosis by activating TAM receptor/Akt/PKB signaling, but whether hPS affects expression of IAP or Bcl-2 proteins is unknown. In the current study, we show that hPS inhibits caspase-3 activation, elicits increased phosphorylation of both Akt/PKB and IκB, enhances the expression of the antiapoptotic BIRC3 and Bcl-2 proteins, and reduces proapoptotic Bax in β-cells, and this antiapoptotic activity of hPS is abolished by Mer receptor antibody. These results are consistent with the concept that hPS ameliorates diabetes by inhibiting β-cell apoptosis through a TAM receptor–mediated mechanism, leading to increased activation of Akt/PKB and NF-κB signaling and upregulation of IAP and Bcl-2 proteins (25).

Type 2 diabetes is characterized by impaired insulin secretion and increased insulin resistance, leading to relative hypoinsulinemia, decreased uptake of blood glucose in muscles and adipose tissue, and enhanced glucose output from the liver (3). The Akt/PKB pathway is critical for insulin sensitivity and for the maintenance of normal glucose homeostasis. Apart from enhancing secretion of insulin from β-cells, activation of the Akt/PKB pathway promotes glucogen synthesis by inactivating glycogen synthase kinase-3, glucose uptake by increasing the expression or translocation of glucose transporters, and glycolysis by activating 6-phosphofructo-2-kinase (29–31). Through these mechanisms, Akt/PKB can improve insulin sensitivity in peripheral tissues. We show that systemic administration of hPS improves sensitivity to insulin in db/db mice with reduced HOMA-IR values and that pretreatment with hPS improves insulin sensitivity in adipocytes, skeletal muscle, and liver cell lines, supporting the role of hPS in the amelioration of insulin resistance. Possibly, hPS improves insulin sensitivity by activating the Akt/PKB pathway (32).

Renal deposition of matrix protein with thickened glomerular basement membrane, mesangial expansion, and nodular sclerosis are characteristic findings of diabetic nephropathy (33). Growth factors and chemokines, including TGF-β1, stimulate matrix protein release from myofibroblasts following apoptosis of renal cells of glomeruli (33). In the current study, the protective effect of hPS was evaluated by using two models of diabetic nephropathy. In one model, WT mice were treated with exogenous hPS or saline after becoming fully diabetic, and in the second model, diabetes of equal severity was induced in both WT and hPS TG mice by administering higher STZ doses to hPS TG than to WT mice. The development of nephropathy was then compared over the same time period in the two groups. The results showed significantly less glomerulosclerosis, hydroxyproline content, and TGF-β1 expression in the kidneys and less plasma creatinine and L-FABP in diabetic WT mice treated with exogenous hPS than in those treated with saline and in diabetic hPS TG mice than in diabetic WT mice with equal severity and duration of diabetes, suggesting that hPS attenuates diabetic nephropathy independently of its protective effects on islet β-cells. Although the mechanism by which hPS attenuates diabetic nephropathy was not clearly defined by these results, the observation of a decreased number of apoptotic cells and improved podocin expression in diabetic mice treated with exogenous hPS compared with controls suggests that inhibition of glomerular cell apoptosis is the probable mechanism.

We found significantly reduced circulating levels of free PS in patients with type 1 and type 2 diabetes and a significant inverse correlation of free PS with HbA_1c in patients with type 2 diabetes. The significant inverse correlation between the circulating levels of free PS and HbA_1c suggests that lower systemic availability of PS is clinically relevant in patients with diabetes.

We believe that these novel findings combining observations of supplementing PS levels by either administration of exogenous hPS or by overexpressing PS are consistent with our original hypothesis that hPS ameliorates diabetes and its renal complications.

Funding. This research was supported in part by a Kakenhi Grant-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (No. 24591128 and No. 15K09170).

The funders had no role in study design, data analysis, decision to publish, or preparation of the manuscript.

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

Author Contributions. T.Y. contributed to the preparation of the diabetes mouse models and writing of the manuscript. Y.Y. and Y.T. contributed to the study concept and design and data analysis. C.N.D.-G. and K.N. contributed to the preparation of the diabetes mouse models. M.T., J.A.H., and Z.R. contributed to the measurement of several parameters in mouse model samples. P.G-B., R.M.-M., and I.M. contributed to the preparation of the models of diabetic nephropathy. T.K. contributed to the data interpretation and critical revision of the manuscript. J.M. and I.C. contributed data interpretation and writing and critical revision of the manuscript. E.C.G. contributed to the study concept and design, data analysis, and writing and critical revision of the manuscript. E.C.G. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.