Coordinated regulation of hepatic FoxO1, PGC-1α and SREBP-1c facilitates insulin action and resistance

Cellular Signalling 43 (2018) 62–70

Contents lists available at ScienceDirect

Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig

Coordinated regulation of hepatic FoxO1, PGC-1α and SREBP-1c facilitates insulin action and resistance

T

Mini P. Sajana, Mackenzie C. Leea, Fabienne Foufelleb, Joshua Sajana, Courtney Clelanda, ⁎ Robert V. Faresea, a

Medical and Research Services, James A. Haley Veterans Administration Hospital, Department of Internal Medicine, University of South Florida College of Medicine, Tampa, FL 33704, USA b Institut National de la Sante et de la Recherche Medicale (INSERM) UMRS 1138, Sorbonne Universites, UMPC Universite Paris 06, Sorbonne Paris Cite, Universite Paris Descartes, Universite Paris Diderot, Centre de Recherche Biomedicales des Cordeliers, Paris, France

A R T I C L E I N F O

A B S T R A C T

Keywords: Insulin Liver Gluconeogenesis Lipogenesis Akt Atypical PKC SREBP-1c FoxO1 PGC-1α PKC-ι

Type 2 diabetes is characterized by insulin resistance, hyperinsulinemia and hepatic overproduction of glucose and lipids. Insulin increases lipogenic enzyme expression by activating Akt and aPKC which activate SREBP-1c; this pathway is hyperactivated in insulin-resistant states. Insulin suppresses gluconeogenic enzyme expression by Akt-dependent phosphorylation/inactivation of FoxO1 and PGC-1α; this pathway is impaired in insulin-resistant states by aPKC excess, which displaces Akt from scaffolding-protein WD40/ProF, where Akt phosphorylates/ inhibits FoxO1. But how PGC-1α and FoxO1 are coordinated in insulin action and resistance is uncertain. Here, in normal mice, we found, along with Akt and aPKC, insulin increased PGC-1α association with WD40/ProF by an aPKC-dependent mechanism. However, in insulin-resistant high-fat-fed mice, like FoxO1, PGC-1α phosphorylation was impaired by aPKC-mediated displacement of Akt from WD40/ProF, as aPKC inhibition diminished its association with WD40/ProF, and simultaneously restored Akt association with WD40/ProF and phosphorylation/inhibition of both PGC-1α and FoxO1. Moreover, in high-fat-fed mice, in addition to activity, PGC-1α expression was increased, not only by FoxO1 activation, but also, as found in human hepatocytes, by a mechanism requiring aPKC and SREBP-1c, which also increased expression and activity of PKC-ι. In high-fat-fed mice, inhibition of hepatic aPKC, not only restored Akt association with WD40/ProF and FoxO1/PGC-1α phosphorylation, but also diminished expression of SREBP-1c, PGC-1α, PKC-ι and gluconeogenic and lipogenic enzymes, and corrected glucose intolerance and hyperlipidemia. Conclusion: Insulin suppression of gluconeogenic enzyme expression is facilitated by coordinated inactivation of FoxO1 and PGC-1α by WD40/ProF-associated Akt; but this coordination also increases vulnerability to aPKC hyperactivity, which is abetted by SREBP1c-induced increases in PGC-1α and PKC-ι.

1. Introduction Insulin maintains fasting blood glucose levels by suppressing hepatic gluconeogenesis. Impairments in this suppression provoke increases in hepatic glucose output, hyperinsulinemia and increases in SREBP-1c-dependent lipogenesis that promotes development of obesity, metabolic syndrome features and type 2 diabetes mellitus (T2DM). Hepatic gluconeogenesis is excessive in diet-induced obesity (DIO) in

adolescent children [1] and readily progresses to T2DM [2]. Increases in hepatic gluconeogenesis in insulin-resistant DIO/T2DM are provoked by impaired action or activation of Akt, which suppresses gluconeogenic enzyme expression by inactivating FoxO1 and PGC-1α [3–6], which, when active, facilitate transactivation of gluconeogenic genes, phophoenolpyruvate carboxykinase (PEPCK) and glucose-6phosphatase (G6Pase) by hepatic nuclear factor-4 (HNF4). Hepatic lipogenesis is also upregulated in insulin-resistant DIO/

Abbreviations: aPKC, atypical protein kinase C; ACC, acetyl-CoA carboxylase; ACPD, 2-acetyl-cyclopentane-1,3-dione; AMPK, AMP-activated protein kinase; Akt, protein kinase B; FAS, fatty acid synthase; FoxO1, forkhead homeobox class O1 protein; G6Pase, glucose-6-phosphatase; HFF, high-fat-fed; HNF4, hepatic nuclear factor-4; IRS-1, insulin receptor substrate-1; IRS-2, insulin receptor substrate-2; ICAPP, 1H-imidazole-4-carboxamide, 5-amino-1-[2,3-dihydroxy-4-[(phosphono-oxy)methyl]cyclopentyl-[1R-(1a,2b,3b,4a)]; LND, lean/nondiabetic; NCF, normal-chow-fed; PEPCK, phosphoenolpyruvate carboxykinase; PIP3, phosphatidylinositol-3,4,5-trisphosphate; PI3K, phosphatidylinositol 3-kinase; PKC, protein kinase C; PGC-1α, peroxisome proliferator-activated receptor-gamma (PPARα) coactivator-1α; aPKC, atypical protein kinase C; SREBP-1c, sterol receptor element binding protein-1c; T2DM, type 2 diabetes mellitus; WD40/ProF, [tryptophan-aspartate-x-x]-repeat, propeller-like, FYVE-containing protein ⁎ Corresponding author at: Research Service-151, James. A. Haley Veterans Medical Center, 13000 Bruce B. Downs Blvd., Tampa, FL 33612, USA. E-mail address: [email protected] (R.V. Farese). https://doi.org/10.1016/j.cellsig.2017.12.005 Received 19 October 2017; Accepted 17 December 2017 Available online 18 December 2017 0898-6568/ Published by Elsevier Inc.

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

hepatocytes of normal and T2D humans, treated with inhibitors of aPKC and Akt, and with adenoviruses that express dominant-negative or constitutively-active forms of aPKC and SREBP-1c.

T2DM, despite the fact that insulin normally increases lipogenesis by activating sterol receptor binding protein-1c (SREBP-1c), and insulin repression of PEPCK/G6Pase is impaired in T2D human liver [7,8]. This paradox in well-established T2DM partly reflects, in human liver [7,8]: (a) aPKC hyperactivity and activation of SREBP-1c; (b) persistent mTOR activation and SREBP-1c activation and action, despite reduced Akt activity; (c) aPKC-dependent impairment of insulin activation of insulin receptor substrate-1 (IRS-1), phosphatidylinositol 3-kinase (PI3K) and Akt; (d) impaired Akt-dependent FoxO1 phosphorylation/inhibition and subsequent increases in PEPCK/G6Pase; and (e) conserved insulin activation of IRS-2, IRS-2-dependent PI3K and aPKC. Combined activities of aPKC and mTOR1 in early and late stages of insulin resistance, and, increased activities of IRS-1/PI3K and Akt in early phases of DIO [9,10], may account for sustained activation of SREBP-1c and transcription of multiple lipogenic enzymes, including, fatty acid synthase (FAS) and acetyl-CoA carboxylase (ACC) in insulin-resistant states. Although PGC-1α-dependent increases in PEPCK/G6Pase are needed to maintain blood glucose levels during fasting, excessive activity and expression of hepatic PGC-1α contribute to development of fasting hyperglycemia and insulin resistance in DIO/T2DM [11,12]. Indeed, hepatic PGC-1α expression/abundance is increased in obese and T2D humans [8]. In this regard, PGC-1α activity is diminished by insulin/Akt-dependent phosphorylation/inactivation of both PGC-1α [4] and FoxO1 [3], which, when active, binds to and increases PGC-1α activity [3] and expression of 70 kDa human/liver-specific form of PGC-1α (hL-PGC-1α) [6]. In addition to FoxO1-dependent hL-PGC-1α expression, we presently found that mouse 91 kDa PGC-1α and human 70 kDa hL-PGC-1α expression is increased by an additional mechanism involving increases in aPKC and SREBP-1c. But if and how these signaling factors are integrated is enigmatic. Also germane to these issues, in early DIO that develops in mice fed a diet containing 40% of calories derived from fat over 3 months [13] and in 5–6 month-old ob/ob mice excessively consuming amounts of carbohydrate-rich mouse chow [14], insulin resistance, at the outset, when hepatic Akt activation is still intact, involves: increases in hepatic ceramide that directly activate aPKC; hyper-recruitment of aPKC to scaffolding-protein WD40/ProF; decreased Akt association with the WD40/ProF platform, to which both Akt and aPKC are recruited by insulin in adipocytes [15–17] and liver [13,14]; selective impairment of Akt-dependent FoxO1 phosphorylation, [but not glycogen synthase kinase-3β (GSK3β) and mTOR, phosphorylation, which are independent of WD40/ProF]; increased expression of PEPCK/G6Pase; hyperglycemia; hyperinsulinemia; insulin-dependent increases in hepatic activities of Akt and aPKC; and increases in aPKC + Akt-dependent activation of SREBP-1c and expression of lipogenic enzymes. In support of the importance of aPKC is the fact that all of these hepatic aberrations are reversed by treatment of HFF and ob/ob mice with liver-selective aPKC inhibitors [13,14]. Importantly, obese and T2D humans have the same hepatic signaling aberrations that are present in HFF and ob/ob mice, including, increases in aPKC and decreases in Akt association with WD40/ProF, but, additionally, there are decreases in IRS-1 levels and activities of IRS-1/PI3K and Akt [7,8]; similarly, impaired Akt activation is seen later in HFF mice [18]. Moreover, in human T2D liver, compounding the problem of increased aPKC activity, the expression and levels of the primate-specific aPKC isoform, PKC-ι, are increased by an auto-catalytic, feed-forward, negative feedback mechanism [7,8]; however, this aberration, as well as aberrations in IRS-1 and Akt, are reversed by aPKC inhibition [7,8]. Here, to elucidate mechanisms that: (a) coordinate activities of hepatic FoxO1 and PGC-1α, and alter levels of PGC-1α, during insulin regulation and DIO/T2DM-mediated dysregulation of gluconeogenic enzyme expression; and (b) contribute to hepatic aPKC hyperactivity via SREBP-1c-dependent increases in levels and activity of primatespecific PKC-ι; we studied livers of normal chow-fed (NCF) mice and HFF mice treated without or with a liver-selective aPKC inhibitor, and

2. Methods 2.1. Mouse studies C57Bl/6/SV129 4–5-month-old male and female mice (sex did not appreciably alter results) from a colony maintained in the Tampa VA/ USF Vivarium, were studied over 3 months while consuming diets (Harlan Industries, Madison WI, USA) supplying 10% or 60% of calories from fat. Mice were injected subcutaneously without or with aPKC inhibitor, 2-acetyl-cyclopentane-1,3-dione (ACPD), 10 mg/kg/day, which inhibits liver, but not muscle or adipose tissue, aPKC, without altering hepatic Akt or AMPK activities [13,14,19]. During the 10th week, glucose tolerance was measured after an overnight fast by intraperitoneal injection of 2 mg glucose per kg body weight and measurement of blood glucose, as described [13,14]. In week 12, 15 min before killing, mice were treated without or with insulin (1 U/kg) given intraperitoneally. All experimental procedures involving animals were approved by the Institutional Animal Care and Use Committees of the University of South Florida College of Medicine, and the James A. Haley Veterans Administration Medical Center Research and Development Committee, Tampa, Fl. 2.2. Human hepatocyte studies Cryo-preserved hepatocytes, harvested by Zen-Bio (NC, USA) from human transplant donors, 40–75 years of age, rejected for reason of immune incompatibility, were incubated as described [7,8]. These studies were approved by the Human Experimentation Internal Review Board (IRB) of the University of South Florida College of Medicine, as well as IRBs at organ-origination sites. 2.3. Lysate preparations and protein kinase assays As described [7,8,13,14,19], livers or hepatocytes were homogenized and analyzed for protein kinase activities of aPKC and Akt2, or, in Western analyses, aPKC activity/activation was assessed by immunoblotting for phospho-thr-555/560 [20], and Akt activity was assessed by immunoblotting for phospho-ser-473-Akt. 2.4. Antibodies Western analyses and immunoprecipitations were conducted as described [7,8,13,14,19] using the following antisera/antibodies: antiphospho-ser-473-Akt, anti-glyceraldehyde-phosphate dehydrogenase (GAPDH), anti-WD40/ProF, and anti-aPKC (Santa Cruz Biotechnologies, Santa Cruz, CA, USA); anti-phospho-thr-560/555-PKC-ζ/λ/ι (Invitrogen, Carlsbad, CA, USA); anti-p-ser-256-FoxO1, anti-FoxO1 (Abnova, Walnut, CA, USA); anti-phospho-ser-9-GSK3β, anti-GSK3β, antiphospho-ser-2248-mTOR, anti-mTOR, and mouse anti-Akt Mab (Cell Signaling Technologies, Danvers, MA, USA); anti-phospho-thr-571PGC-1α; (R&D Systems, Minneapolis, MN, USA); and anti-PGC-1α (Genetex, Irvine, CA, USA), a C-terminally-directed antiserum that measures both human liver-specific 77 kDa and mouse liver 91 kDa PGC-1α isoforms. 2.5. mRNA measurements As described [7,8,13,14,19], mouse tissues and human hepatocytes were added to Trizol reagent (Invitrogen) and RNA was extracted and purified with RNA-Easy Mini-Kit and RNAase-free DNAase Set (Qiagen, Valencia, CA, USA), and quantified by real-time reverse transcriptasepolymerase chain reaction (RT-PCR), using TaqMan reverse 63

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

phosphorylation of Akt substrates, GSK3β and mTOR, was partially increased in the resting/basal state and fully increased by acute insulin treatment, but unaltered by ACPD treatment.

transcription reagent and SYBR Green kit (Applied Biosystems, Carlsbad, CA, USA) with mouse or human nucleotide primers, and horse radish peroxidase transferase as an internal recovery standard. Note, for human liver-specific PGC-1α, we measured mRNA encoding the shortened liver-specific human isoform [in which exon 1 and 2 of the longer human isoform are replaced by a silent exon contained in the first intron of the longer isoform [6] by using primers specific for exon 1 of the shortened 77 kDa isoform; for mouse PGC-1α mRNA, we used primers in exon 1 of the full-length 91 kDa isoform. For other primers, see [7,8,13,14,19].

3.1.2. Insulin signaling to PGC-1α and FoxO1 in livers of NCF and HFF mice As seen in Fig. 1, despite increased resting/basal and normal insulinstimulated hepatic Akt activity in HFF mice, Akt-dependent PGC-1α phosphorylation, like that of FoxO1, was diminished in the resting/ basal state and poorly responsive to insulin. Inhibition of hepatic aPKC by ACPD treatment restored PGC-1α and FoxO1 phosphorylation.

2.6. Serum metabolites 3.1.3. Recruitment of aPKC, Akt, and PGC-1α to WD40/ProF, and phosphorylation of FoxO1 and PGC-1α in livers of NCF and HFF mice In livers of NCF mice, insulin acutely increased the association of aPKC (Fig. 2a), Akt (Fig. 2b) and PGC-1α (Fig. 1c), but not FoxO1 (Fig. 2d), with immunoprecipitable WD40/ProF. In HFF mice, the association of both aPKC (Fig. 2a) and PGC-1α (Fig. 2c) with WD40/ProF was increased both in the resting/basal state and further in response to acute insulin treatment. However, despite increases in total cellular Akt activity (Fig. 1b) and increases in phosphorylation of Akt substrates, GSK3β (Fig. 1c) and mTOR (Fig. 1f), Akt association with WD40/ProF was diminished in the resting/basal state, and, moreover, poorly responsive to insulin (Fig. 2b) in HFF mice. Further, along with decreased Akt association with WD40/ProF (Fig. 2b), the phosphorylation of FoxO1 (Fig. 1d) and PGC-1α (Fig. 1e) was diminished in HFF mice.

Glucose, triglycerides and total cholesterol were measured as described [13,14]. 2.7. Statistical evaluations Data are expressed as mean ± SEM, and P values were determined by one-way ANOVA and least-significant multiple comparison methods. 3. Results 3.1. Studies in mouse liver 3.1.1. Insulin signaling to aPKC and Akt in livers of NCF and HFF mice As seen in Fig. 1, resting/basal hepatic aPKC activity in HFF mice was increased near-maximally, as acute insulin treatment had little additional effect. Resting/basal Akt activity was also modestly elevated in HFF mice, and increased normally following acute insulin treatment. Administration of liver-selective [13,14,19] aPKC inhibitor, 2-acetylcyclopentane-1,3-dione (ACPD) to HFF mice reduced aPKC activity without altering Akt activity. Along with Akt in HFF mice,

b

p-aPKC

c

p-Akt Akt

*** *** *** ***

3

***

2 1

(7) (7) (15) - - - - + - + +

0

Insulin ACPD HF Diet

(8) (7) (16) + + + - - + - + +

d

4

p-Ser-473-Akt

p-thr-555-aPKC

***

**

p-Ser-9-GSK3•

GAPDH

aPKC 4

***

3 2

***

1 0

Insulin ACPD HF Diet

(7) (7) (15) (8) (7) (16) - - + + + - - + - - + - + + - + +

e

*

2 1.5

**

*

*

*

1 0.5 0

Insulin ACPD HF Diet

(7) -

(7)(15) - - + + +

(8) (7) (16) + + + - - + - + +

p-ser-571-PGC-1• PGC-1• Ratio

p-ser-256-FoxO1

GAPDH

*** **

**

*

**

1 0.5 0

Insulin ACPD HF Diet

2

*** *** ***

1 0

(7) (7) (15) (8) (7) (16) - - + + + - - + - - + - + + - + +

p-mTOR GAPDH

2 1.5

3

Insulin ACPD HF Diet

f

p-PGC-1• Total PGC-1•

p-FoxO1

p-GSK3•

(7) -

(7)(15) - - + + +

(8) (7) (16) + + + - - + - + +

p-Ser-2448-mTOR

a

3.1.4. Effects of inhibition of aPKC on recruitment of aPKC and Akt to WD40/ProF, and on phosphorylation of FoxO1 and PGC-1α in livers of HFF mice Treatment of HFF mice with aPKC inhibitor, ACPD, reduced aPKC association with WD40/ProF (Fig. 2a), and, importantly, this reduction was accompanied by increases in resting/basal and insulin-stimulated

2 1.5

*** * ***

1 0.5 0

Insulin ACPD HF Diet

(7) (7) (15) (8) (7) (16) - - + + + - - + - - + - + + - + +

Fig. 1. Effects of treatment of high-fat-fed (HFF) mice with liver-selective aPKC inhibitor, ACPD, on resting/basal and insulin-stimulated activities of hepatic aPKC (a) and Akt (b), and phosphorylation of Akt substrates, glycogen synthase kinase-3β (GSK3β) (c), FoxO1 (d), PGC-1α (e) and mTOR (f). Mice were fed normal chow or HF diets for 12 weeks, during which, HFF mice were treated subcutaneously ± aPKC inhibitor, ACPD (10 mg/kg/day). Mice were intraperitoneally injected ± insulin (1 U/kg) 15 min before killing. Relative values are mean ± SEM of (N) mice. Asterisks: *, P < 0.05; **, P < 0.01; ***, P < 0.001 (ANOVA).

64

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

b 3

***

***

Akt Associated With WD40/ProF Imm-ppt )

aPKC Associated With WD40/ProF Imm-ppt

a ***

***

2

1

0

(7) (7)

(15)

(8) (7) (16)

3

** 2

**

* *

1

0

(7) (7)

(15)

(8) (7) (16)

Insulin ACPD

-

-

+

+ -

+ -

+ +

Insulin ACPD

-

-

+

+ -

+ -

+ +

HF Diet

-

+

+

-

+

+

HF Diet

-

+

+

-

+

+

c ***

3

WD40/ProF ***

***

*** 2

***

***

1

0

(7) (7)

(15)

FoxO1 FoxO1 Associated With WD40/ProF Imm-ppt

PGC-1● Associated With WD40/ProF Imm-ppt

d

(8) (7) (16)

3

2

1

0

(7) (7)

(15)

(8) (7) (16)

Insulin ACPD

-

-

+

+ -

+ -

+ +

Insulin ACPD

-

-

+

+ -

+ -

+ +

HF Diet

-

+

+

-

+

+

HF Diet

-

+

+

-

+

+

Fig. 2. Effects of insulin, high-fat (HF) diet and treatment of HF-fed mice with liver-selective inhibitor, ACPD, on association of aPKC (panel a), Akt (panel b), PGC-1α (panel c) and FoxO1 (panel d) with immunoprecipitable WD40/ProF. Mice were treated as in Fig. 1, liver lysates were immunoprecipitated with anti-WD40/ProF antiserum, and immunoprecipitates were analyzed for indicated proteins (WD40/ProF serves as a loading control). Relative values are mean ± SEM of (n) mice. Asterisks: *, P < 0.05; **, P < 0.01; ***, P < 0.001 (ANOVA).

3.1.7. Improvements in glucose and lipid homeostasis following ACPD treatment of HFF mice Along with hepatic signaling alterations in HFF mice, glucose tolerance was impaired (Fig. 4a), and resting, ad lib-fed, serum glucose levels were elevated and poorly responsive to insulin (not shown). Further, serum levels of triglycerides and cholesterol were elevated (Fig. 4b), and the weight of perigonadal plus retroperitoneal fat (Fig. 4c) was increased in HFF mice, and, here too, these clinical abnormalities were corrected or largely improved by ACPD treatment (Fig 4a, b, c).

association of Akt with WD40/ProF (Fig. 2b), and, moreover, by restored phosphorylation of both PGC-1α (Fig. 1e) and FoxO1 (Fig. 1d). 3.1.5. Alterations in PGC-1α activity, expression and levels in livers of HFF mice Decreases in Akt-dependent phosphorylation, and thus activation, of FoxO1 in livers of HFF mice would be expected to increase PGC-1α activity [3,4]. In addition to this mechanism for increases in PGC-1α activity, we found that hepatic PGC-1α mRNA (Fig. 3a) and protein levels (Fig. 3b) were increased in HFF mice, and these increases were reversed by ACPD-induced inhibition of hepatic aPKC (Fig 3a,b). In this regard, unlike the dependence of expression of the shortened hL-PGC1α on Akt and FoxO1 [5,6], the expression of the longer murine form of PGC-1α is little influenced by Akt and FoxO1 [3,6], but other mechanisms (see hepatocyte studies below) may be operative.

3.2. Studies in human hepatocytes 3.2.1. aPKC-dependent increases in PGC-1α expression in isolated human hepatocytes Normally, by reducing FoxO1 activity, insulin-induced increases in hepatic Akt activity would be expected to diminish expression of 70 kDa hL-PGC-1α isoform in human hepatocytes [6]. Conversely, in insulin-resistant humans, increases in hL-PGC-1α expression may reflect FoxO1 activation owing to aPKC-induced decreases in Akt-dependent FoxO1 phosphorylation. On the other hand, FoxO1-dependent increases in hepatic PGC-1α expression are not evident in the mouse [3], which uses the longer 91 kDa PGC-1α isoform [6]. Thus, increases in PGC-1α expression presently seen in HFF mice are presumably FoxO1-independent. Relevant to this issue, we reported that 24-hour insulin treatment of hepatocytes of lean/non-diabetic (LND) humans provoked moderate increases, rather than decreases (as would be

3.1.6. Alterations in expression of gluconeogenic and lipogenic in livers of HFF mice Along with decreases in FoxO1 and PGC-1α phosphorylation (i.e., activation), and increases in PGC-1α expression, the expression of PEPCK and G6Pase, was increased in livers of HFF mice (Fig. 3c). Further, with increases in hepatic activities of aPKC and Akt, and, with increases in PGC-1α activity and expression, the expression of SREBP1c and FAS, was increased in HFF mice (Fig. 3d). Importantly, abnormalities in expression of both gluconeogenic and lipogenic enzymes were corrected by treatment of HFF mice with liver-selective aPKC inhibitor, ACPD (Fig 3c, d). 65

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

a

Fig. 3. Effects of high-fat (HF) diet and treatment of HF-fed mice with liver-selective aPKC inhibitor, ACPD, on levels of PGC-1α mRNA (a), PGC-1α protein (b), gluconeogenic enzyme expression, i.e., PEPCK and G6Pase mRNAs (c) and lipogenic enzyme expression, i.e., SREBP-1c and FAS mRNAs (d). Mice were treated as in Fig. 1, and liver samples were analyzed for indicated proteins and mRNAs. Relative values are mean ± SEM of (N) mice. Asterisks: *, P < 0.05; **, P < 0.01; ***, P < 0.001 (ANOVA).

b 3

*

***

***

1.5

2

PGC-1 Protein

PGC-1 mRNA

*

1 (15)

0

(14)

1 0.5

(31) 0

ACPD HF Diet

-

-

+

-

+

+

(15)

(14)

(31)

-

-

+

-

+

+

ACPD HF Diet

c ***

***

1

(15)

0

(14)

(31)

3

***

2 1

(15)

0 3

2

***

***

1.5 1 0.5

(15)

(14)

*

(14)

(31)

**

2

*

1

(15)

(31)

(14)

(31)

0

0

-

ACPD HF Diet

+

+ +

b Serum Triglycerides (mg/dL)

a 400 ***

350

***

300

***

250

***

200

150*** 100

LF

50

HF

HF + ACPD

0 0

30

60 Minutes

90

120

+

200

**

+ +

c

***

6

150 100 50 0

(15)

(14)

(31)

*** **

300

Serum Cholesterol (mg/dL)

Blood Glucose (mg/dL)

-

ACPD HF Diet

250 200 150 100 50 0

(15)

ACPD

-

High Fat Diet

-

(14)

(31)

-

+ +

Epididymal + Retroperitoneal Fat (grams)

G6Pase mRNA

SREBP-1c mRNA

2

FAS mRNA

PEPCK mRNA

d

5

***

***

4 3 2 1 0

ACPD

(15)

(14)

(31)

-

-

+

High Fat Diet -

+

Fig. 4. Effects of high-fat (HF) diet and treatment of high-fat-fed (HFF) mice with liver-selective aPKC inhibitor, ACPD, on glucose tolerance (a), and ad lib-fed serum lipids (b) and weights of combined perigonadal plus retroperitoneal fat depots (c). Mice were treated as described in Fig. 1. Glucose tolerance tests were conducted during week 10; LF, low fat. Values are mean ± SEM of (N) mice. Asterisks: *, P < 0.05; **, P < 0.01; ***, P < 0.001 (ANOVA).

have contributed to the increases in hepatic PGC-1α expression seen in HFF mice, we presently found, in incubations of hepatocytes of LND humans, that both insulin and metformin [which, like insulin, activated hepatic aPKC (Fig. 5a), but, unlike insulin, did not activate Akt

expected from Akt-dependent FoxO1 inactivation) in PGC-1α expression, and these increases, as well as the greater increases seen in hepatocytes of T2D humans, were blocked by aPKC inhibitor, ACPD [8]. And, further suggesting that excessive increases in aPKC activity may 66

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

b

a

2.5

p-ser-473-Akt

p-thr-555-PKC-

4

* **

3 2 1

*** ***

***

2

***

1.5 1 0.5

0

Ins Met ACPD -

+ -

+ -

+

0

+ - + + +

-

+ - + - -

- + - - + + + +

d 3 2 1 0

Insulin Metformin ACPD KI-PKC- DN-SREBP -

+- -+-+ - - + - -+++ - - - - - - - - -

-+- -+ - - +++ - - -

-+- -+ - - - - +++

3

SREBP mRNA

4

e PGC-1 mRNA

PKC- mRNA

c

Ins Met ACPD

2 1 0

Insulin Metformin ACPD KI-PKC- DN-SREBP -

+- + - - - -

- + - -+ +++ - - - - -

- +- - + - - +++ - - -

- +- -+ - - - - +++

2.5 2 1.5 1 0.5 0

Insulin Metformin ACPD KI-PKCDN-SREBP

-

+-+ - - - -

- +- - + +++ - - - - -

- +- - + - - +++ - - -

- +- -+ - - - - +++

Fig. 5. Stimulatory effects of insulin and metformin on expression of PKC-ι and PGC-1α are dependent on aPKC and SREBP-1c in human hepatocytes. Hepatocytes of lean/non-diabetic humans were incubated 48 h ± 1 μM ACPD or adenovirus or adenovirus expressing kinase-inactive (KI) PKC-ζ (25MOI) or dominant-negative (DN) SREBP-1c (200 MOI) (total MOI kept constant at 200 MOI by adding non-expressing adenovirus) ± 200 nM insulin ± 100 μM metformin. Mean ± SEM (N = 4). Asterisks: *, P < 0.05; **, P < 0.01; ***, P < 0.001 versus control (ANOVA).

3.2.3. Effects of expression of DN and CA SREBP-1c on expression of PGC1α, PKC-ι, PKC-ζ, SREBP-1c, FAS, and ACC, and on aPKC activity in human hepatocytes In hepatocytes of LND humans, insulin and metformin increase SREBP-1c expression by an aPKC-dependent mechanism [19]. Presently, both agents increased expression of PGC-1α and PKC-ι by a mechanism requiring both aPKC and SREBP-1c, as adenovirally-mediated expression of kinase-inactive (KI) aPKC and dominant-negative (DN) SREBP-1c blocked stimulatory effects of insulin and metformin on these expressions (Fig. 5). Moreover, constitutively-active (CA) SREBP1c provoked dose-related increases in expression of PGC-1α and PKC-ι (but not PKC-ζ), as well as SREBP-1c, FAS and ACC in basal conditions; and, oppositely, adenovirally-mediated expression of DN SREBP-1c provoked dose-related decreases in insulin-stimulated increases in these expressions (Fig. 7a). Further note that SREBP-1c-provoked increases in PKC-ι expression were accompanied by increases in aPKC activity (Fig. 7b).

(Fig. 5b)] provoked increases in hL-PGC-1α expression that were blocked by both aPKC inhibitors, kinase-inactive aPKC and ACPD, and by adenovirally-mediated expression of kinase-inactive aPKC (Fig. 5d). These findings with metformin suggested that aPKC increases hL-PGC1α expression by a mechanism that does not necessarily require Akt coactivation, but, on the other hand, as seen below, Akt, as well as aPKC, contributes importantly to increases in hL-PGC-1α expression provoked by insulin in human hepatocytes.

3.2.2. Phosphorylation of PGC-1α in hepatocytes of LND and T2D humans As the above-described results showed that insulin increases hLPGC-1α expression, the simultaneous phosphorylation/inhibition of PGC-1α [4] by Akt presumably overshadows any increases in PGC-1α levels that may occur during normal insulin action, as gluconeogenic enzyme expression diminishes. It was therefore interesting that increases in PGC-1α phosphorylation (Fig. 6a) during 24-hour insulin action in hepatocytes of LND humans were commensurate with, or slightly in excess of, increases in PGC-1α protein levels (Fig. 6b), thus maintaining a normal phospho-PGC-1α to total PGC-1α ratio during 24hour insulin action (Fig. 6c). In contrast, in hepatocytes of T2D humans, increases in PGC-1α levels (Fig. 6b) were accompanied by decreases in insulin-dependent PGC-1α phosphorylation (Fig. 6a), and the ratio of phospho-PGC-1α to total PGC-1α was markedly diminished (Fig. 6c), indicating activation, as well as increased levels of PGC-1α. Importantly, reversal of aPKC-induced increases in PGC-1α expression elicited by treatment of hepatocytes of T2D humans with aPKC inhibitor, ACPD, was accompanied by improvement in the ratio of phospho-PGC-1α to total PGC-1α in the resting/basal state, and to a greater degree with insulin treatment (Fig. 6c).

3.2.4. Effects of aPKC and Akt on expression of SREBP-1c, PGC-1α and PKC-ι in human hepatocytes In keeping with findings suggesting that transcriptional activity of SREBP-1c requires inputs from both Akt and aPKC [10], insulin- and T2D-induced increases in expression of SREBP-1c (Fig. 8c), PGC-1α (Fig. 8b) and PKC-ι (Fig. 8c) were inhibited by Akt inhibitor, Akti, and aPKC inhibitor, ICAPP. 4. Discussion Alterations in hepatic insulin signaling in mice consuming a 60%fat-cal diet were similar to those in mice consuming a 40%-fat-cal diet [13] in that, whereas Akt activity and phosphorylation of GSK3β and mTOR were intact, Akt-dependent FoxO1 phosphorylation was 67

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

c

b

1.5 1 0.5 0

*

*** ***

2 1.5

p-PGC-1 /PGC-1

***

***

2

PGC-1 Protein

p-ser-571-PGC-1

a

1 0.5

+

-

+

Insulin -

Non-diabetic T2DM

+

-

*** ***

1.5 1

*** **

0.5

0 Insulin - + - + - + - + ACPD - - + + - - + +

0

Insulin -

2

+

Non-diabetic

Non-diabetic T2DM

T2DM

Fig. 6. Effects of 24-h treatment of hepatocytes of lean non-diabetic and type 2 diabetic (T2DM) humans with 200 nM insulin, and, where indicated in panel c, aPKC inhibitor, 1 μM ACPD, on levels of phospho-PGC-1α (panel a), total PGC-1α (panel b), and ratios of phospho-PGC-1α to total PGC-1α (panel c). Relative bar values are mean ± SEM of 4 determinations. Asterisks: *, P < 0.05; **, P < 0.01; ***; P < 0.001 (ANOVA).

both early and late phases of insulin resistance, with excessive aPKC activation by diet-derived ceramide and resultant decreases in Akt presence on the WD40/ProF platform, the concomitant impairments in Akt-dependent FoxO1 and PGC-1α phosphorylation would be expected to act in an integrated, complementary fashion in provoking increases in activity and expression of PGC-1α that in turn increase expression of gluconeogenic and lipogenic enzymes. Thus, from the present findings, it may be surmised that the regulation of FoxO1 and PGC-1α phosphorylation is closely coordinated by their co-dependence on Akt association with WD40/ProF, and this coordination would be expected to facilitate their subsequent interactions

diminished. But here, we also found: (a) in HFF mice, that Akt-dependent PGC-1α phosphorylation, like FoxO1, was diminished and reflected aPKC-induced decreases in Akt association with WD40/ProF; (b) in normal mice, PGC-1α association with WD40/ProF was increased by insulin by an aPKC-dependent mechanism; and (c) increases in hepatic PGC-1α expression in HFF mice were dependent on aPKC, and, in human hepatocytes, dependent on both aPKC and SREBP-1c. Importantly, Akt-dependent PGC-1α, like FoxO1, phosphorylation was compromised at 3 months here and in a previous study of HFF mice [13], relatively early, i.e., before Akt activation was impaired, as well as in T2D human hepatocytes, when Akt activity declines [7…8]. Thus, in

1.5 1 0.5

*

0 0

10

50

2 1.5 1

2

*

mRNA

0

*

0.5

10

50

100

*

1.6

*

200

* *

1.5 1

0.5

0

CA SREBP-1c MINUS INSULIN

1

0

PKC mRNA

ACC mRNA

2 1.5

100 200

* *

DN SREBP-1c PLUS INSULIN

*

2

2.5

*

1.2

* 0.8

*

*

0.4

0 0

10

50

100 200

0

10

50

100

200

2.5

*

2 1.5 1

*

0.5

mRNA

1.5

1

PKC

3

2.5

0.5

*

3

FAS mRNA

b

3

*

2

PGC1

SREBP1c mRNA

2.5

p-thr-555/560-PKC-

a

0.5

Adenovirus MOI

0

0 0

10

50

100 200

Adenovirus MOI

0

10

50

100

200

Adenovirus MOI

0

10

50

100

200

- - - - - + + + + + Insulin Adenovirus MOI 0 10 50 100 200 0 10 50 100 200 CA SREBP-1c

DN SREBP-1c

Fig. 7. Effects of adenovirally-mediated expression of dominant negative (DN) SREBP-1c on insulin-stimulated levels, and effects of adenovirally-mediated expression of constitutively active (CA) SREBP-1c on basal levels, of indicated mRNAs (a) and aPKC activity (b) in human hepatocytes. Hepatocytes of lean/non-diabetic humans were incubated 48 h without and with 200 nM insulin and indicated MOI of DN or CA SREBP-1c (total MOI kept constant for all at 200). [Unchanged PKC-ζ mRNA as a control reference point.] Mean ± SEM (N = 4). Asterisk: *, P < 0.05 vs non-expressing adenovirus.

68

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

*

6000

*

5000

b *

Akt2 Activity (cpm/immunoppt.)

aPKC Activity (cpm/immunoppt.)

a

*

4000 * 3000 2000 1000 0

*

*

6000 5000 *

*

*

*

4000 3000 2000 1000 0

Insulin - + - + - + - + - + - + Akti - - + + - - - - + + - ICAPP - - - - + + - - - - + +

Insulin - + - + - + - + - + - + Akti - - + + - - - - + + - ICAPP - - - - + + - - - - + +

Non-Diabetic

Non-Diabetic

T2D d

5

*

* *

2 1

e

*

*

* 2 1

*

*

2.5

0

0

Insulin - + - + - + - + - + - + Akti - - + + - - - - + + - ICAPP - - - - + + - - - - + + Non-Diabetic

*

*

3

4 3

*

PGC-1 mRNA (relative values)

* PKC- mRNA (relative values)

SREBP-1c mRNA (relative values)

c

T2D

*

2

*

*

1.5 1 0.5 0

Insulin - + - + - + - + - + - + Akti - - + + - - - - + + - ICAPP - - - - + + - - - - + +

Insulin - + - + - + - + - + - + Akti - - + + - - - - + + - ICAPP - - - - + + - - - - + +

Non-Diabetic

Non-Diabetic

T2D

T2D

T2D

Fig. 8. Effects of PKC-ι inhibitor, 100 nM ICAPP, and Akt inhibitor, 10 μM Akti, on insulin-stimulated and T2D-induced increases in expression of SREBP-1c (c), PGC-1α (d) and PKC-ι (e) mRNAs during 6-hour incubation of hepatocytes of lean/non-diabetic and T2D humans ± 100 nM insulin. Effectiveness and selectivity for inhibition of aPKC and Akt2 activities by ICAPP and Akti, respectively, are illustrated in panels a and b. Values are mean ± SE of 4 determinations. Asterisks: *, P < 0.05; **, P < 0.01; ***, P < 0.001 (ANOVA).

Coordinang effects of the WD40/ProF Plaorm in insulin ACTION and diet-induced insulin RESISTANCE p-Akt SREBP-1c

PIP3

FoxO1

p-Akt

p-PKCPKC-

Fig. 9. Coordinated regulation of FoxO1 and PGC-1α by WD40/ProF-associated Akt during insulin regulation of gluconeogenic enzyme expression and its dysregulation by aPKC excess abetted by SREBP-1c.

p-FoxO1 (inacve) p-PGC-1 (inacve)

PEPCK G6Pase

PGC-1

CERAMIDE, PA activity and expression of lipogenic enzymes (Fig. 9). Interestingly, in isolated human hepatocytes, we found that both insulin and metformin increase hL-PGC-1α expression by a mechanism dependent on aPKC, and for metformin, is independent of Akt, which,

and actions on hepatic enzymes, such as HNF-4, that increase gluconeogenic enzyme expression [3–6]. In addition, alterations in activity and expression of PGC-1α may impact transcriptional effects of PPARγ and liver X-receptor [6], which are involved in regulating SREBP-1c 69

Cellular Signalling 43 (2018) 62–70

M.P. Sajan et al.

Conflicts of interest

by inactivating FoxO1, would be expected to decrease hL-PGC-1α expression [5,6]. Accordingly, insulin effects on hL-PGC-1α expression were less than those elicited by metformin, perhaps reflecting that insulin activation of Akt and Akt phosphorylation/inhibition of FoxO1 may limit PGC-1α expression [5,6]. As metformin, insulin and AMPK activator, AICAR, were previously found to increase SREBP-1c mRNA in human hepatocytes by an aPKCdependent mechanism [19], it was interesting to find (a) along with increases in expression of SREBP-1c and lipogenic enzymes, hL-PGC-1α expression was increased by constitutively-active SREBP-1c, and (b) stimulatory effects of insulin and metformin on these expressions were blocked by dominant-negative SREBP-1c, kinase-inactive aPKC and aPKC inhibitor, ACPD. It was of further interest that insulin-stimulated expression of PGC-1α in human hepatocytes was dependent on both aPKC and Akt. This co-dependence of insulin-induced PGC-1α expression on aPKC and Akt may reflect a need for inputs from both aPKC and Akt during insulin activation of SRE by SREBP-1c [10]. On the other hand, metformin increased PGC-1α expression independently of Akt. Of further note, insulin-induced increases in expression of PKC-ι and aPKC activity were inhibited by dominant-negative SREBP-1c, and constitutively-active SREBP-1c provoked increases in. PKC-ι expression and aPKC activity in human hepatocytes. This interdependency between aPKC and SREBP-1c may explain how insulin increases expression of primate-specific PKC-ι by a feed-forward, positive-feedback, aPKC-dependent mechanism [7]. In keeping with this idea, we found that insulin activated a PKC-ι promoter/reporter complex by an aPKC and SREBP-1c-dependent mechanism (unpublished). Note that this cycle can become “vicious” in insulin-resistant states: thus, persistent increases in hepatic aPKC would impair both Akt actions on the WD40/ProF platform and Akt activation by IRS-1/PI3K [7,8], and resultant increases in expression of gluconeogenic and lipogenic enzymes would provoke further insulin + lipid-induced increases in aPKC and SREBP-1c activity, and so on. To summarize, our findings suggest that co-compartmentalization of Akt-dependent FoxO1 and PGC-1α phosphorylation on the WD40/ProF platform serves to functionally coordinate the activations and actions these hepatic signaling factors. Thus, co-inactivation of FoxO1 and PGC-1α owing to Akt-dependent phosphorylation would facilitate insulin suppression of gluconeogenic enzyme expression, but, on the other hand, the ability of aPKC excess in insulin-resistant states to coactivate FoxO1 and PGC-1α and increase gluconeogenic enzyme expression would be similarly enhanced. Compounding these aberrations in insulin-resistant states, increases in expression of PGC-1α and, in human hepatocytes, PKC-ι, are dependent on aPKC and SREBP-1c.and activation of aPKC and SREBP-1c are co-dependent. Thus, in addition to increasing lipogenic enzyme expression, SREBP-1c may amplify mechanisms that increase gluconeogenic enzyme expression, and, in humans, SREBP-1c may reciprocate with aPKC to induce a vicious cycle.

There are no conflicts of interest amongst the authors. Contributions R.V.F. conceived, designed and provided overall direction of the studies, analyzed, and interpreted data, and wrote the paper. M.P.S. helped in planning, and designing of the studies, conducted and/or supervised M.C.L., J.S. and C.C. in the performance of experiments and assays, assembled data, and assisted in analysis and interpretation of data. F.F. provided adenoviruses encoding constitutively-active and dominant-negative SREBP-1c. References [1] A.L. Sunehag, G. Toffolo, M. Campioni, D.M. Bier, M.W. Haymond, Effects of dietary macronutrient intake on insulin sensitivity and secretion and glucose and lipid metabolism in healthy obese adolescents, J. Clin. Endocrinol. Metab. 90 (2006) 4496–4502. [2] S.T. Chung, D.S. Hsia, S.K. Chacko, L.M. Rodriguez, M.W. Haymond, Increased gluconeogenesis in youth with newly diagnosed type 2 diabetes, Diabetologia 58 (2015) 596–603. [3] P. Puigserver, J. Rhee, J. Donovan, C.J. Walkey, J.C. Yoon, Y. Kitamura, J. Altomonte, H. Dong, D. Accili, B.M. Spiegelman, Insulin-regulated gluconeogenesis through FOXO1-PGC-1α interaction, Nature 423 (2003) 550–555. [4] X. Li, B. Monks, Q. Ge, M.J. Birnbaum, Akt/PKB regulates hepatic metabolism by directly inhibiting PGC-1a transcription activator, Nature 447 (2007) 1012–1016. [5] H. Daitoku, K. Yamagata, H. Matsuzaki, M. Hatta, A. Fukamizu, Regulation of PGC1 promoter activity by protein kinase B and the forkhead transcription factor FKHR, Diabetes 52 (2003) 642–649. [6] T.K. Felder, S.M. Soyal, H. Oberkofler, P. Hahne, S. Auer, R. Weiss, G. Gadermaier, K. Miller, F. Krempler, H. Esterbauer, W. Patsch, Characterization of novel peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) isoform in human liver, J. Biol. Chem. 286 (2011) 42923–42938. [7] M.P. Sajan, R.V. Farese, Insulin Signalling in hepatocytes of type 2 diabetic humans. Excessive expression and activity of PKC-ι and dependent processes and reversal by PKC-ι inhibitors, Diabetologia 55 (2012) 1446–1457. [8] M.P. Sajan, R.A. Ivey III, R.V. Farese, BMI-related progression of atypical PKC-dependent aberrations in insulin signaling through IRS-1, Akt, FoxO1 and PGC-1α in livers of obese and type 2 diabetic humans, Metabolism 64 (2015) 1454–1465. [9] S. Li, M.S. Brown, J.L. Goldstein, Bifurcation of insulin signaling pathway in rat liver: mTORC1 required for stimulation of lipogenesis, but not inhibition of gluceoneogenesis, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 3441–3446. [10] Y. Wang, J. Viscarra, S.-J. Kim, H.S. Sul, Transcriptional regulation of hepatic lipogenesis, Nat. Rev. Mol. Cell Biol. 16 (2015) 678–689. [11] C.T. De Souza, E.P. Araujo, P.O. Prada, M.J. Saad, A.C. Boschero, L.A. Velloso, Short-term inhibition of peroxisome proliferator-activated receptor-γ coactivator1α expression reverses diet-induced diabetes mellitus and hepatic steatosis, Diabetologia 48 (2005) 1860–1871. [12] B.N. Finck, D.P. Kelly, PGC-1 coactivators: inducible regulators of energy metabolism in health and disease, J. Clin. Invest. 116 (2006) 615–622. [13] M.P. Sajan, M.E. Acevedo-Duncan, M.L. Standaert, R.A. Ivey III, M. Lee, R.V. Farese, Akt-dependent phosphorylation of hepatic FoxO1 is compartmentalized on a WD40/Propeller/FYVE scaffold and is selectively inhibited by atypical PKC in early phases of diet-induced obesity. A mechanism for impairing gluconeogenic but not lipogenic enzyme expression, Diabetes 63 (2014) 2690–2701. [14] M.P. Sajan, R.A. Ivey III, M.C. Lee, R.V. Farese, Hepatic insulin resistance in ob/ob mice involves increases in ceramide, atypical PKC activity and selective impairment of Akt-dependent FoxO1 phosphorylation, J. Lipid Res. 56 (2015) 70–80. [15] T. Fritzius, G. Burkard, E. Haas, J. Heinrich, M. Schweneker, M. Bosse, S. Zimmerman, A.D. Frey, A. Caelers, A.S. Bachmann, K. Moelling, A WD-FYVE protein binds to the kinases Akt and PKCζ/λ, Biochem. J. 399 (2006) 9–20. [16] T. Fritzius, K. Moelling, Akt and Foxo1-interacting WD-repeat-FYVE protein promotes adipogenesis, EMBO J. 27 (2008) 1399–1410. [17] T. Fritzius, A.D. Frey, M. Schweneker, D. Mayer, K. Moelling, WD-repeat-propellerFYVE protein, ProF, binds VAMP2 and protein kinase Cζ, FEBS J. 274 (1552–1566) (2007) 2007. [18] L. He, A. Sabet, S. Djedjos, et al., Metformin and insulin suppress hepatic gluconeogenesis through phosphorylation of CREB binding protein, Cell 15 (2009) 635–646. [19] M.P. Sajan, R.A. Ivey III, R.V. Farese, Metformin action in human hepatocytes. Coactivation of atypical protein kinase C alters 5′-AMP-activated protein kinase effects on lipogenic and gluconeogenic enzyme expression, Diabetologia 56 (2013) 2507–3010. [20] R.A. Ivey III, M.P. Sajan, R.V. Farese, Pseudosubstrate arginine residues are required for auto-inhibition and are targeted by phosphatidylinositol-3.4.5-(PO4)3 during aPKC activation, J. Biol. Chem. 289 (2014) 25021–25030.

Acknowledgements Dr. Robert V. Farese is the guarantor of this work and, as such, had full access to all the data in the study and takes full responsibility for the integrity of the data and the accuracy of the data analysis. This work does not represent the views of the Department of Veteran Affairs or the United States government.

Funding Supported by funds from the Department of Veterans Affairs Merit Review Program to R.V. Farese, and the National Institutes of Health Grants DK 065969-09 to R.V. Farese.

70