Curcumin, Cardiometabolic Health and Dementia

Current research indicates curcumin [diferuloylmethane; a polyphenolic compound isolated from the rhizomes of the dietary spice turmeric (Curcuma longa)] exerts a beneficial effect on health which may be partly attributable to its anti-oxidative and anti-inflammatory properties. The aim of this review is to examine potential mechanisms of the actions of curcumin in both animal and human studies. Curcumin modulates relevant molecular target pathways to improve glucose and lipid metabolism, suppress inflammation, stimulate antioxidant enzymes, facilitate insulin signalling and reduce gut permeability. Curcumin also inhibits Aβ and tau accumulation in animal models and enhances mitochondria and synaptic function. In conclusion, in high-dose animal studies and in vitro, curcumin exerts a potential beneficial effect on cardiometabolic disease. However, human studies are relatively unconvincing. More intervention studies should be conducted with the new curcumin formulation with improved oral bioavailability.


Introduction
Type 2 diabetes Mellitus (T2DM) is associated with impaired insulin signalling, leading to hyperglycaemia and micro and macrovascular diseases [1,2]. Insulin resistance is a major contributor to the pathogenesis of T2DM with secondary pancreatic failure [2,3]. The prevalence of diabetes mellitus will increase worldwide from 451 million people aged over 18 in 2017 to 693 million people in 2045 [4]. Diabetes is an enormous social, financial and health system burden across the world [4,5]. Lifestyle modification, including a healthy diet, can lower the risk of T2DM [6]. Dietary polyphenols have been a major research focus to reduce the risk of T2DM [7][8][9][10]. This review aims to present an update on the effect of curcumin (a polyphenol) on the prevention and treatment of T2DM and cardiovascular disease (CVD) in animal studies and human studies.

Curcumin Safety
The Allowable Daily Intake (ADI) value of curcumin is 3 mg/kg body weight [18]. Healthy subjects consumed curcumins ranging from 0.5 to 12 g/day for 72 h in order to assess the safety of curcumin. Up to 12 g/day of curcumin consumption for 72 h was deemed safe. About 30% subjects showed diarrhoea and headache, which were not dose-related [19]. Subjects who consumed curcumins at a dose ranging from 0.45 to 3.6 g/day for 1-4 months experienced nausea and diarrhoea. The serum concentrations of alkaline phosphatase and lactate dehydrogenase were elevated in 3-4 out of 15 patients [20].
A review article investigating the pharmacokinetic interactions of curcumin with conventional drugs (including cardiovascular drugs, antidepressants, anticoagulants, antibiotics, chemotherapeutic agents, and antihistamines) showed that curcumin can alter maximum serum concentrations (Cmax) and area under the curve (AUC) when used with those drugs. Curcumin can inhibit cytochrome P450 monooxygenases (drug-metabolising enzymes) and P-glycoprotein (an efflux pump from the ATP-binding cassette (ABC) super family which pumps various xenobiotics (e.g., drugs) out of cells.
Only one clinical trial has demonstrated a significant interaction between curcumin and drugs [21].

Curcumin Bioavailability
Detectable concentrations of curcumin and its metabolites in both blood and urine were observed with curcumin ingestion of ≥3.6 g/day in several studies [20,[22][23][24][25]. The mean plasma concentration in patients with advanced colorectal cancer refractory to standard chemotherapies who consumed 3.6 g/day of curcumin for up to 4 months was 4.3 ng/mL (i.e., about 0.01 µM/L). The mean plasma concentrations of curcumin glucuronide and curcumin sulphate were 5.8 and 3.3 ng/mL, respectively, 1 h after administration [20]. The plasma concentrations (mean ± SD) of curcumin for patients with high-risk or pre-malignant lesions who took 4 g, 6 g and 8 g daily for 3 months were 0.19, 0.20, and 0.60 µg/mL, respectively [22]. Healthy subjects (n = 6) ingested 10 g of curcumin extract after hydrolysis of conjugates. The Cmax (mean ± SE) of curcumin conjugates detected as glucuronide and sulphate were 3.2 ± 0.56 µg/mL. These values were 1000 times higher than those of other study subjects with lower doses of curcumin (mentioned above [20,22]). The time to reach maximum plasma concentrations (Tmax) was 4.33 ± 3.2 h [25].
Modified curcumin chemical structures (analogues or derivatives of curcumin) have been developed with enhanced stability, solubility, bioavailability and biological effects. Rapid absorption (2 h and 57 min) and slow elimination (3 h and 39 min half-life) were shown [69]. Bioavailability was 60% when 32 mg/kg of curcumin analogue (EF-24) was orally administrated to mice [69].

Effects of Curcumin on Cardiometabolic Health
A summary of curcumin human intervention studies is shown in Table 1.

Anti-Oxidative Effects
Oxidative stress is characterised by an imbalance between reactive oxygen species (ROS) generation and anti-oxidative defence [72]. Hyperglycemia promotes autooxidation of glucose, glycation of protein and enhanced polyol pathways leading to the increased ROS [73]. Continuous oxidative stress can cause chronic inflammation which may result in chronic diseases such as T2DM, CVD and Alzheimer's disease (AD) [74][75][76][77].
The anti-oxidant effect of curcumin as a free radical scavenger appear to be related to its phenolic O-H and C-H groups [78].

Human Studies
There are fewer than 30 human studies, and most are small and uncontrolled, so the data is relatively unconvincing. However, it is noted that volunteer characteristics and experimental designs were rarely the same in different studies.
Thirty-eight healthy middle-aged subjects (40-60 years old) who consumed a low dose of lipidated curcumin (80 mg/day) for 4 weeks showed some favourable effects on cardiometabolic health, but no comparison was made with placebo, so no firm conclusions can be drawn [79]. Yang et al. 2015 [80] conducted a small open uncontrolled study to see if curcumin intake in subjects with T2DM can lower urinary microalbuminuria excretion. Urinary microalbuminuria was significantly decreased by 70% (n = 14; p < 0.05) by 500 mg/day of curcumin for 15 days. Moreover, decreased levels of malondialdehyde (MDA) and increased levels of nuclear factor erythroid 2-related factor 2 (Nrf2), NAD(P)H: quinone oxidoreductase (NQO1), superoxide dismutase (SOD) were observed. The levels of lipopolysaccharides (LPS) and caspase 3 decreased. The levels of IκBα and gut barrier function increased in a before and after comparison. There were no significant differences in fasting blood glucose, insulin, C-peptide, triglyceride (TG), total cholesterol (TC), HDL-C, low-density lipoprotein (LDL-C), alanine transaminase (ALT), aspartate transaminase (AST), creatinine and urea nitrogen (BUN) following 15-day curcumin ingestion compared with the baseline [80].
Curcumin (0-10 µm/L) reduced haemoglobin glycosylation and lipid peroxidation in erythrocytes under oxidative stress induced by high glucose concentrations [83], and inhibited aldose reductase which in turn leads to decreased intracellular sorbitol accumulation [84]. Elevated aldose reductase activities cause increased sorbitol production from glucose, which may increase diabetes complications [85,86]. It is noted that recent studies indicate aldose reductase inhibitors alone appear to be not effective for prevention of diabetic complications [96,97].
In alloxan-induced diabetic rats, a glucose-lowering of curcumin (1 g/kg body weight or 0.08 g/kg body weight) was noted, which led to reduced oxidative stress (decreased TBARS) through the prevention of glucose influx into the polyol pathway, as well as a decrease in sorbitol dehydrogenase (SDH-converts sorbitol to fructose) [82].
Curcumin may reduce diabetes complications related to vascular inflammation. Hyperglycaemic conditions induce the secretion of proinflammatory cytokines via epigenetic changes, which are mediated by the opposing actions of histone deacetylases (HDACs) and histone acetylases (HATs) [132,133]. Curcumin treatment of human monocytic (THP-1) cells under high-glucose conditions increased HDAC2, decreased HATs activity and expressions of p300 and acetylated CBP/p300 gene expression, consequently leading to decreased NF-κB transcription activity and proinflammatory cytokine secretion (IL-6, TNF-α) [132].
Oxidative stress and inflammation are contributors to cardiometabolic disease including insulin resistance, T2DM, CVD and AD. The potent anti-oxidative and anti-inflammatory effects of curcumin could beneficially influence glucose and lipid homeostasis and endothelial function.

Glucose Homeostasis
In a randomised double-blind, placebo-controlled study, subjects with non-alcoholic fatty liver disease (NAFLD) (mean age 46.37 ± 11.57 years; mean BMI 31.35 ± 5.67 kg/m 2 ; n = 77) consumed 500 mg/day of an amorphous dispersion curcumin formulation (equivalent to 70 mg curcumin) for 8 weeks. Curcumin consumption significantly reduced glucose, glycated haemoglobin (HbA1c), TC, LDL, TG, liver fat content, BMI, aspartate aminotransferase (AST), alanine aminotransferase (ALT) compared with the placebo [134]. All the changes seen may be due to the loss of weight seen in the curcumin group (over 2 kg difference over 8 weeks), which may be caused by nausea and anorexia, which caused 3 dropouts as well. Ultrasound is not a reliable method amount of liver fat. Large changes in HbA1c were seen in both groups. In a randomised double-blind, placebo-controlled trial of 100 overweight/obese subjects with T2DM (average age: 54.72 ± 8.34 years; BMI ≥ 24.0; curcuminoids (300 mg/day; n = 50) supplementation for 12 weeks significantly reduced fasting glucose, HbA1c and homeostasis model assessment insulin resistance (HOMA-IR) with decreased levels of serum free fatty acids (FFAs) and TG, and increased lipoprotein lipase (LPL) activity compared with a placebo. However, levels of TC, LDL-C, HDL-C, Apo A-I or Apo B did not differ. Anthropometric variables such as body weight and waist and hip circumferences were also not altered on curcuminoids supplementation compared with a placebo [135].
In a randomised, double-blind, placebo-controlled trial of prediabetic subjects (n = 237), curcumin treatment (a total of 6 capsules/day-250 mg curcuminoid/capsule) significantly decreased HbA1c, fasting glucose and OGTT at 3, 6, and 9 months compared with a placebo (p < 0.01) and reduced the diagnosis of T2DM from 16.9% to 0%. At 3 months, curcumin treatment significantly increased HOMA-β (p < 0.01) compared with a placebo. At 9 months, C-peptide and insulin were significantly attenuated with curcumin treatment compared with a placebo (p < 0.05). Moreover, curcumin treatment significantly reduced HOMA-IR at 6 and 9 months and elevated adiponectin at 9 months compared with a placebo [136]. Curcumin treatment significantly decreased body weight by 6.2 kg at 9 months compared with a placebo which could account for all of the observations on glucose and HbA1c. AST, ALT, creatinine, bone mineral density and waist circumference were not altered with curcumin treatment compared with a placebo. This clinical study supports the preventive effect of curcumin on the development of T2DM in individuals with prediabetes [136].
Adiponectin acts as an insulin sensitiser to suppress hepatic gluconeogenesis via AMP-activated protein kinase (AMPK)-dependent and -independent pathways. It stimulates fatty acid oxidation predominantly in skeletal muscle to activate glucose uptake [137]. Increased adiponectin levels were observed with curcumin supplementation (1 g/day) for 6 weeks in subjects with metabolic syndrome compared with a curcumin-phospholipid complex group or a placebo group, whereas no differences in BMI, body weight, waist circumference, fasting blood glucose, fat (%) were observed compared with a curcumin-phospholipid complex group or a placebo group [138].
In a randomised double-blind, placebo-controlled design, subjects with T2DM (mean age 59 ± 10.6 years; n = 107) were treated with 250 mg of curcumin (3 times/day) for 6 months. Curcumin treatment increased adiponectin and decreased leptin levels compared with the placebo. Pulse wave velocity (PWV), HOMA-IR, TG, uric acid, abdominal obesity (visceral fat and total body fat) were significantly reduced with curcumin treatment compare with the placebo [139].
In a randomised crossover study, 11 healthy subjects supplemented with turmeric (2.8 g/day) for 4 weeks showed no changes in fasting plasma glucose, TC and TG compared with the control (water only) [140]. In a crossover design, 14 healthy subjects who consumed 6 g of encapsulated turmeric with a standard 75 g oral glucose tolerance test (OGTT) showed increased insulin responses with no difference in postprandial glucose responses compared with the reference OGTT with placebo capsules [141].

Potential Mechanisms of Curcumin
On the other hand, in vitro and high-dose animal studies supports the glucose lowering effects of curcumin. A high-fat diet (HFD) was given to male Wistar rats (7 weeks of age) for 6 weeks, and then diabetes was induced by streptozotocin injection (30 mg/kg body weight). Three different concentrations of curcumin (50, 150, or 250 mg/kg body weight) were used for 7 weeks. Curcumin (especially, 150 mg/kg body weight) significantly improved glucose and insulin tolerance compared with normal control rats [147]. In addition, in the in vitro study, L6 myotube cells were treated with different concentrations of curcumin (5, 10, 20, or 40 µm/L) in the presence of palmitate (0.25 mM/L). The 2-deoxy-[3H] D-glucose uptake was enhanced by curcumin in a dose dependent manner and glucose transporter 4 (GLUT4) protein expression increased on the cell membrane of L6 myotubes [147].
C57BL/KsJ-db/db mice and their age-matched lean non-diabetic db/+ mice were fed a diet with curcumin or without curcumin (0.02%, wt/wt) for 6 weeks. In db/db mice, curcumin decreased glucose and HbA1c, improved HOMA-IR and glucose tolerance as assessed by intraperitoneal glucose tolerance test (IPGTT) and increased insulin levels. Curcumin did not alter glucose tolerance and insulin levels in db/+ mice. Curcumin increased hepatic glucokinase (GK) activity and suppressed the elevation of hepatic gluconeogenic enzyme activities, glucose-6-phosphatase (G6Pase) and phosphoenolpyruvate carboxykinase (PEPCK) in db/db mice. Curcumin increased glycogen storage in the liver in the db/db mice, while curcumin did not alter hepatic GK, P6Pase, PEPCK and glycogen content in non-diabetic db/+mice. These findings indicate that curcumin can lower glucose levels in db/db mice [93].
AMPK plays a key role in glucose and fatty acid metabolism [148]. Its roles in activation of catabolic processes aimed at ATP production (e.g., glucose uptake, glycolysis, fatty acid uptake, and β-oxidation) and inhibition of anabolic pathways (e.g., synthesis of glycogen, proteins, fatty acids, and cholesterol) have been well documented [149].
In high-fat diet-induced obese and leptin-deficient ob/ob male C57BL/6J mice, the improved glucose control following curcumin treatment (3% dietary curcumin) was associated with anti-inflammatory effects of curcumin by decreasing macrophage infiltration into adipocytes, and by increasing adiponectin production in white adipocytes, by decreasing ER stress (elevated ER stress in adipocytes and hepatocytes is related to obesity) and perigonadal fat expression of Sirtuin 1 (SIRT1), heat shock proteins (HSP70), HSP90 and Foxo1 and decreasing NF-κB activity in liver [156]. SIRT1 is a regulator of adiponectin transcription through the activation of Foxo1 and Foxo1 and C/EBPalpha interaction in adipose tissue [157]. SIRT1 plays a role in glucose homeostasis, inducing hepatic glucose production through the deacetylation of PGC-1α and by repressing peroxisome proliferator-activated receptor gamma (PPAR γ) [158,159].
A further meta-analysis of RCTs also showed no effect of curcumin on lipid profiles [162]. According to a 2017 position paper from an International Lipid Expert Panel, the lipid lowering effect of curcumin in human intervention studies is inconsistent, but several recent interventions report favourable effects on lipid profiles [134,143,146,163,164].
In a randomised double-blind, placebo-controlled parallel study of 117 subjects with metabolic syndrome (aged 25-75 years), supplementation of 1000 mg/day of curcuminoids and piperine (bioperine ® Sami Labs LTD, Bangalore, Karnataka, India) (100:1 ratio combination) for 8 weeks significantly lowered LDL-C, non-HDL-C, TC, TG and lipoprotein(a) [Lp(a)] and increased HDL-C, compared with the placebo. The lipid changes were significant after adjustment for baseline of BMI and lipids. No change in small dense LDL (sdLDL) was observed [163].
In a randomised placebo-controlled study of 87 subjects with non-alcoholic fatty liver disease, 1000 mg/day of curcumin plus dietary and lifestyle intervention for 8 weeks significantly decreased TC, non-HDL-C, LDL-C, TG and uric acid compared with placebo. No differences in HbA1c, fasting glucose, insulin, HOMA-IR, HOMA-β, quantitative insulin sensitivity check index (QUICKI) were seen compared with placebo [146].
In a randomised, double-blind, placebo-controlled trial in 70 subjects with T2DM, nano-curcumin (80 mg/day) for 3 months significantly reduced HbA1C, fasting blood glucose, TG, and BMI before and after the treatment [57].
In a randomised, double-blind, placebo-controlled, parallel study investigating the effect of curcumin (400 mg/day as Longvida ® ) on cognition, mood and biomarkers in 60 elderly subjects (mean age: 68.5 years) for 4 weeks significantly lowered TC and LDL-C with the significant improvement memory and mood compared with the placebo [68].
In a randomised double-blind, placebo-controlled crossover trial, 30 subjects aged 18-65 years who were not taking lipid-lowering agent, as well as who had any conditions including BMI ≥ 30 kg/m 2 or 2 risk factors (except for T2DM) for coronary heart disease (CHD) or ≥ 2 risk factors (except for T2DM) for CHD and 130 mg/dL < LDL-C < 160 mg/dL, were supplemented with curcuminoids (1 g/day) for 30 days. Curcuminoids supplementation decreased TG levels with no differences, LDL-C, HDL-C and CRP. Anthropometric variables such as body weight, BMI, waist circumference, arm circumference, and fat percentage also were not altered with curcuminoids supplementation compared with the placebo [100].
In a randomised double-blind, controlled trial, subjects with acute coronary syndrome, curcumin ingestion (45-180 mg/day) for 1 year showed no changes in lipids [145]. In a randomised, double-blind study of 36 elderly subjects (mean age: 73.4 ± 8.8 years), either 4 g/day or 1 g/day of curcumin supplementation did not significantly alter TG, LDL-C and HDL-C over 1 month or 6 months compared with a control [160]. In a double-blind, randomised, placebo-controlled, 2 × 2 factorial trial, 70 hypercholesterolemia subjects (mean fasting TC: 6.57 ± 0.13 mM/L) were randomised to either curcumin (200 mg/day: Meriva ® Indena) plus phytosterols (2 g/day; n = 17) or curcumin (200 mg/day; n = 18) or phytosterols (2 g/day; n = 17) or placebo (n = 18) for 4 weeks. Phytosterol group and curcumin plus phytosterol group showed significant reductions in TC, LDL-C and TC: HDL-C compared with the baseline, whereas the curcumin group did not significantly alter TC and LDL-C. HDL-C and TG were not altered in any group [161].
In the human hepatoma cell line (HepG2), a potential hypocholesterolemic effect of curcumin (0.02% wt/wt) was observed with elevated gene expressions of the LDL-receptor. Curcumin (80 mg/kg) has been shown to downregulate fatty acid synthase (FAS-related to increased plasma lipid levels) and increase skeletal muscle LPL [93,167].
Statins are lipid-lowering medications known as HMG-CoA reductase inhibitors. Curcumin (0.02% wt/wt) has been shown to decrease HMG-CoA reductase activity leading to reduced plasma and hepatic cholesterol levels [93,175].
In animals fed a high-fat diet, curcumin treatment (0.05 g/100 g diet) has been shown to decrease hepatic acyl-CoA: cholesterol acyltransferase (ACAT) which may decrease intestinal cholesterol uptake and transport in the intestine [93,172].
Curcumin treatment (0.5% dietary curcumin or 0.1% wt/wt) also has been shown to increase hepatic cholesterol 7 α-hydroxylase (CYP7A) which is a rate-limiting enzyme responsible for the biosynthesis of bile acid from cholesterol [171,176,177].
Curcumin treatment (100 or 400 mg/kg body weight) has been shown to decrease serum TG and fetuin-A (α2-Heremans-Schmid glycoprotein) levels in liver of rats fed a high-fat diet [174]. Fetuin-A produced in the liver is associated with insulin resistance and fatty liver [183,184].
In a randomised double-blind, placebo-controlled trial of subjects with T2DM, curcuminoids (1000 mg/day combined with piperine 10 mg/day) for 12 weeks significantly decreased body weight, BMI, TC and Lp(a) and increased HDL-C compared with a placebo. However, curcumin did not alter TG and LDL-C compared with the placebo [45].
In a randomised parallel trial, overweight subjects with metabolic syndrome (mean BMI 25-29.9 kg/m2; mean age 39.1 ± 16.8 years; n = 44) who adhered to a diet plus lifestyle intervention for 30 days (adherence rate > 80% and a weight loss < 2%) underwent either curcumin treatment (800 mg twice per day; n = 22) plus lifestyle intervention or phosphatidylserine plus lifestyle intervention for 4 weeks. Significant reductions in body weight, body fat, waistline and BMI were observed with curcumin treatment compared with the phosphatidylserine group [46].
On the other hand, in a randomised, double-blind, placebo-controlled, crossover study, obese subjects (mean BMI: 33.95 ± 3.81) who were treated with curcuminoids (1 g/day) for 30 days showed no differences in BMI and weight compared with the placebo [186].
Stearoyl-coenzyme A desaturase 1 (SCD-1) is a rate enzyme in the synthesis of unsaturated fatty acids [190]. Upregulation of SCD-1 may be associated with obesity, insulin resistance and atherosclerosis [191]. Curcumin (80 mg/kg) treatment for 12 weeks in HFD-induced obese mice has been shown to downregulate the expression of SCD-1 in brown adipose tissue and white adipose tissue [167].
Norepinephrine binding to beta 3 adrenoreceptors (β3AR) present in white adipose tissues are known to exert a key role in the browning of white adipose tissues [202]. Curcumin treatment (50 or 100 mg/kg) in C57BL/6J mice elevated plasma norepinephrine levels and upregulated β3AR gene expression in inguinal white adipose tissues and induced browning with the decreased body weight and fat accumulation compared with control mice [192].
Gut microbial dysbiosis appears to increase gut permeability, leading to an increased inflammatory response [218]. A positive association between a high-fructose diet and/or a high-fat diet and increased gut permeability has been reported [217][218][219][220].
In vitro, LPS increases IL-1 β which can activate p38 MAP kinases and subsequently myosin light chain kinases (MLCK). The phosphorylation of myosin light chains can increase gut permeability. Curcumin treatment decreased LPS-induced release of IL-1 β from intestinal epithelial cells and intestinal macrophages. Curcumin (5 µm/L) also suppressed p38 MAPK activation by IL-1 β and myosin light chain kinase in intestinal epithelial cells [221].
Mitogen-activated protein kinase phosphatase 1 (MKP-1) exerts an essential role in dephosphorylating MAPK and inactivating ERK, JNK and p38 in response to stress [226,227]. Curcumin (100 mg/kg) decreased the impairment of intestinal mucosa barrier by methotrexate in rats through the activation of MKP-1 and suppression of p38 and NF-kB [222]. Therefore, curcumin appears to reduce gut permeability induced by external dietary factors (e.g., a western diet) or exogenous injury by altering signal pathways, consequently leading to the prevention of a chronic inflammatory state [228].
In a randomised, controlled, double-blind parallel study of 59 healthy subjects, curcumin supplementation (200 mg/day) for 8 weeks improved endothelial function as assessed by flow-mediated dilation (FMD) compared with placebo [237]. Sixty-seven subjects with T2DM who ingested NCB-02 (300 mg/day of curcumin) for 8 weeks showed significantly enhanced endothelial function (measured using digital plethysmography) with lower levels of MDA, ET-1, IL-6 and TNF-α compared with the baseline, but statistical comparison was not made with placebo. These beneficial effects of curcumin (NCB-02) on endothelial function through anti-inflammatory and antioxidant actions are comparable to those of atorvastatin (10 mg/day). However, in comparison with the baseline, NCB-02 did not significantly alter fasting glucose, HbA1c, TC, LDL-C HDL-C and TG, while atorvastatin significantly lowered TC, LDL-C and TG [108].
In a double-blind, parallel, randomised study of healthy middle-aged and older adults (45-74 years) curcumin supplementation (2000 mg/day Longvida ® ) for 12 weeks improved resistance artery endothelial function with enhanced forearm blood flow to brachial artery infusion of acetylcholine (FBFACh) when infused with vitamin C following curcumin compared with baseline (but not placebo in any variable). Curcumin also increased brachial artery FMD. Curcumin reversed the decrease in FBFACh from the nitric oxide synthase inhibitor, NG monomethyl-L-arginine. Curcumin did not alter the levels of adiponectin, leptin, insulin, HOMA-IR, oxidised LDL-C, total antioxidant status, GPx, IL-6, TNF-α, cortisol, epinephrine, norepinephrine and endothelin-1 (ET-1) compared with baseline or placebo [144].
Curcumin treatment (30 and 300 mg/kg) in streptozotocin-induced diabetic rats reduced vascular superoxide anion (O2 −) production and inhibited vascular protein kinase C (PKC-bII) resulting in improved endothelial function [232]. Vascular endothelial cell damage resulting from oxidative stress can be protected by curcumin treatment via autophagy activation. In human umbilical vein endothelial cells (HUVECs) under the condition of oxidative stress induced by hydrogen peroxide H 2 O 2 , curcumin treatment (1, 5 and 10 µm/L) increased microtubule-associated protein 1 light chain 3-II (LC3-II-an autophagosomal marker) in a dose-dependent manner.
Curcumin treatment (2.5, 5, 10 and 20 µm/L) of HUVECs and lymphocytes (Jurkat cells) exposed to either high glucose or advanced glycation end products (AGEs) restored transmembrane potential. Curcumin treatment decreased membrane fluidity in AGE-exposed Jurkat cells or glucose-exposed Jurkat cells. In addition, curcumin treatment inhibited MCP-1 release from Jurkat cells and HUVECs exposed to AGEs or glucose showing an anti-inflammatory action [238].
Curcumin has shown cardioprotective effects [246,247]. Curcumin treatment (150 mg/kg or 100 µm/L) of diabetic rat hearts decreased oxidative DNA and protein damage by decreasing levels of endothelial nitric oxide synthase (eNOS) and inducible nitric oxide synthase (iNOS) via reduced NF-kB and AP-11 in both diabetic rat hearts and in microvascular endothelial cells induced by high glucose. [246]. A vasoconstrictor, ET-1 levels increased with curcumin treatment in diabetic rat hearts and microvascular endothelial cells but ET-1 levels in the kidneys and the retina decreased, indicating that curcumin acts differently on organs [246].

Neurodegenerative Diseases
Diabetes mellitus enhances the risk of dementia. The impairments in glucose metabolism, insulin signalling, insulin sensitivity and lipid metabolism, as well as increases in inflammation and oxidative stress in central nerve and peripheral system, contribute to the risk of Alzheimer's disease (AD) [249][250][251][252].
The etiology of AD includes accumulation of fibrillar amyloid-β (Aβ) peptides (amyloid plaques), decreased Aβ degradation enzymes, Aβ oligomer-promoted synaptic dysfunction, neurotoxic mediators from glial cells, apoE4 (lipid transport protein)-enhanced Aβ deposits, impaired mitochondrial function, mis-localised tau protein from axons to neuronal soma and dendrites, increased neurofibrillary tangles (NFTs-comprised of hyperphosphorylated tau) formation, self-assembled α-synuclein, vascular abnormalities, impaired supply of nutrients, impaired metabolic by-product removals, as well as activated glial cells [252]. Loss of neurons in certain brain areas such as pyramidal cells in lamina II of the entorhinal cortex and in the CA1 region of the hippocampus are responsible for early AD pathology [253,254].
Moreover, the accumulation of advanced glycation end products (AGEs) and the receptor for advanced glycation end products (RAGE) are associated with the T2DM, CVD, degenerative disease and ageing [255][256][257][258]. The increased AGE binding to microglia, neurons and vascular endothelia cells stimulates inflammatory actions and Aβ influx, which in turn leads to neuronal disfunction, cognitive decline and brain damage [259,260].
Aβ oligomers in cultured hippocampal neurons lead to phosphorylation of tau and inhibition of IRS-1 (Ser616) through the activation of c-Jun N-terminal kinase signalling, while curcumin treatment of triple transgenic -AD mice on a high-fat diet showed reductions in phosphorylated JNK, IRS-1, and tau in their brain [261].
Islet amyloid polypeptide (IAPP) or amylin is a hormone consisting of 37 amino acid residues which is co-released with insulin from pancreatic β cells [262]. The accumulation of amyloid as a consequence of the IAPP misfolding is associated with T2DM, AD and Parkinson's disease [263,264]. Curcumin inhibited the self-assembly of IAPP [265,266].
In the hippocampal CA1 area of the brain of AD mice (PS1dE9 double transgenic mice model), curcumin upregulated the expressions of GLUT1 and GLUT3 indicating an improvement in cerebral glucose uptake. Curcumin also stimulated insulin-like growth factor (IGF)-1R, IRS-2, phosphoinositide 3-kinase (PI3K), p-PI3K, Akt and p-Akt protein, and suppressed IR and IRS-1 which implicates the improvement in insulin signalling pathways [267]. In addition, curcumin enhanced spatial memory and learning as assessed by the water maze behaviour test [268]. Aβ-derived diffusible ligands (ADDLs) are known to dysfunction insulin signalling [269][270][271] and act as neurotoxins [272] in AD.
Curcumin treatment showed decreases in Aβ40, Aβ42, ADDLs and γ-secretase [presenilin (PS2)] expression, as well as increases in Aβ degradation enzymes such as insulin-degrading enzymes and neprilysin in the hippocampus CA1 region. The changes in these molecules are related with improved behaviour functions [268].
An AD brain shows an overexpression of PI3K/Akt/mTOR signalling, which is associated with insulin resistance and the pathology of Aβ and tau [273,274]. Curcumin decreased cognitive impairment as assessed by Morris water maze test, inhibited Aβ generation, and suppressed PI3K, phosphorylated Akt and mammalian target of rapamycin (mTOR-Akt regulating cell growth, proliferation, survival, angiogenesis, as well as autophagy [275]) leading to stimulated autophagy, in APP/PS1 double transgenic mice (an AD mice model) [276].
The modification of synapse structure and function is associated with the development of AD. The decreases in synapse-related proteins, including PSD95 and Shank1, contribute to the pathology of AD. Curcumin treatment improved the quantity and structure of the synapse by facilitating the PSD95 and Shank1 in the hippocampal CA1 regions of the APPswe/PS1dE9 double transgenic mice [277]. Curcumin treatment improved synaptic plasticity and neurogenesis resulting in improved memory function [278][279][280].
Brain-derived neurotrophic factor (BDNF) is a protein which stimulates neurogenesis, synaptic plasticity and memory in hippocampus and frontal cortex (FC) [281]. Decreased levels of BDNF are reported in obese and diabetic conditions [282][283][284][285]. Franco-Robles et al. 2014 [142] conducted both animal and human studies to investigate effects of curcumin on BDNF levels in the hippocampus and FC of diabetic db/db mice and in sera of obese subjects. Compared to baseline, curcumin treatment (50 mg/kg daily) for 8 weeks normalised BDNF levels in the hippocampus and FC of diabetic db/db mice. However, nondiabetic obese human subjects who consumed curcumin (500 and 750 mg/day) for 12 weeks did not have altered BDNF levels. Curcumin (500 mg) significantly decreased LDL-C in weeks 2 and 12. Curcumin (500 mg) in weeks 6 and 12 significantly decreased TBARS and oxLDL. Curcumin (500 mg and 750 mg) significantly lowered protein carbonyls levels in week 2, 6 and 12. Curcumin (500 mg and 750 mg) did not alter body weight, BMI, fat, glucose, TC, TG, HDL-C, VLDL and uric acid during the study period [142].
Membrane integrity is important to mainting the normal function of mitochondria and synapse in brain [286,287]. Curcumin facilitated the DHA biosynthesis in liver and DHA accumulation in the brain indicating that curcumin (1-10, 20, or 40 µm/L) can improve cell membrane integrity in the brain, consequently leading to the improvement of neurodegenerative disease by enhancing mitochondria and synaptic function [288,289].

Conclusions
A summary of possible mechanisms of curcumin on glucose homeostasis, lipid metabolism, oxidative stress, inflammation and endothelial function is shown in Figure 2.
In conclusion, based on high-dose animal and in vitro studies, curcumin appears to be a promising therapeutic agent to decrease the risk of T2DM, CVD and neurodegenerative disease by improving glucose homeostasis, lipid metabolism, endothelial function and insulin signalling, and by inhibiting Aβ aggregation. These favourable effects of curcumin could be attributable to potent anti-oxidant and anti-inflammatory actions.
To avoid misinterpretation of results from in vitro studies, it should be noted that cells used in most of the in vitro studies were exposed to very high level of curcumin, i.e., 10 to 100 times greater than the circulating dose measured in plasma after consumption of curcumin supplements or curcumin-rich meal.
There is variability in curcumin metabolism between humans and animals. Thus, the results of in vitro and animal studies cannot be directly related to human physiology. The formula to convert animal doses into human doses for the determination of the equivalents of animal doses in human doses is as follows: human effective dose (mg/kg body weight) = animal dose (mg/kg) X animal km/human km. The correction factor (km) is calculated by dividing the average body weight (kg) of species to its body surface area (m 2 ) [299]. Frequently, animal doses are well beyond normal human doses.
Curcumin is metabolised to curcumin glucuronide or curcumin sulphate by glucuronidase and sulfatase, respectively. The predominant curcumin glucuronides metabolites are terahydrocurcumin (THC) and hexahydrocurcumin (HHC). The minor curcumin metabolites are dihydroferulic acid and ferulic acid. These water-soluble metabolites are excreted through the urine [30]. Ninety-nine percent of plasma curcumin is glucuronide-conjugates. The maximum plasma concentrations of curcuminoid conjugates were observed within 1 h of oral administration in humans [25]. It is unclear whether curcumin metabolites are more or less bioactive than native curcumin [30,[300][301][302].
Author Contributions: All authors conceived of the manuscript structure and contributed to the writing and editing.
Funding: This research received no external funding.