Dietary Whey Protein Lowers the Risk for Metabolic Disease in Mice Fed a High-Fat Diet

Abstract

Consuming a high-fat (HF) diet produces excessive weight gain, adiposity, and metabolic complications associated with risk for developing type 2 diabetes and fatty liver disease. This study evaluated the influence of whey protein isolate (WPI) on systemic energy balance and metabolic changes in mice fed a HF diet. Female C57BL/6J mice received for 11 wk a HF diet, with or without 100 g WPI/L drinking water. Energy consumption and glucose and lipid metabolism were examined. WPI mice had lower rates of body weight gain and percent body fat and greater lean body mass, although energy consumption was unchanged. These results were consistent with WPI mice having higher basal metabolic rates, respiratory quotients, and hepatic mitochondrial respiration. Health implications for WPI were reflected in early biomarkers for fatty liver disease and type 2 diabetes. Livers from WPI mice had significantly fewer hepatic lipid droplet numbers and less deposition of nonpolar lipids. Furthermore, WPI improved glucose tolerance and insulin sensitivity. We conclude that in mice receiving a HF diet, consumption of WPI results in higher basal metabolic rates and altered metabolism of dietary lipids. Because WPI mice had less hepatosteatosis and insulin resistance, WPI dietary supplements may be effective in slowing the development of fatty liver disease and type 2 diabetes.

Introduction

The typical Western diet is high in fat content as well as refined cereals, sugars, and oils. Combined with an increasingly sedentary lifestyle and excessive energy intake, Western culture is suffering health threats in the form of increasing obesity and associated metabolic disorders, including various nonalcoholic fatty liver diseases (NAFLD),⁵ type 2 diabetes mellitus (T2DM), and cardiovascular disease. Childhood obesity associated with hepatic steatosis is of particular concern due to the lifelong risk for developing NAFLD, including nonalcoholic steatohepatitis, liver fibrosis and cirrhosis, as well as hepatocellular carcinoma (1, 2).

Mice, like humans, gain body weight and adiposity when consuming a high-fat (HF) diet. C57BL/6 mice are a useful model for examining changes that occur in body composition that predispose toward future metabolic disorders, including NAFLD and T2DM (3). In mice, we have shown the potential for pharmacological intervention in HF diet-induced gain in body weight and fat mass and the associated metabolic complications, utilizing over-the-counter analgesic drugs (4, 5) and the indolic antioxidant 4b,5,9b,10-tetrahydroindeno[1,2-b]indole (6). However, toxicological concerns regarding pharmacological intervention prompted us to examine potential dietary intervention strategies. Although lifestyle changes such as reducing energy and fat intake, and regular exercise are always advisable, dietary amino acid and protein supplements for intervention have shown promise in reducing weight gain and adiposity and associated metabolic complications for individuals at risk due to poor diet and nutrition and sedentary behavior (7). Studies in rodents suggest that whey protein isolate (WPI) may be beneficial in blood lipid and glucose lowering and in supporting muscle growth (8–13).

Materials and Methods

Chemicals.

All chemicals and reagents were from Sigma-Aldrich Chemical as the highest available grades.

Mice and treatment.

Female C57BL/6J mice were purchased from Jackson Laboratory. All experiments involving mice were conducted in accordance with the MIN standards for care and use of experimental animals and the University of Cincinnati Institutional Animal Care and Use Committee. Mouse groups were matched by initial body weight and maintained on a 12-h-light/-dark cycle. Mice consumed ad libitum a butterfat-supplemented HF diet with 40% energy derived from fat (Product D-03082706, Research Diets) (14). Mice were given either tap water or tap water supplemented with 100 g/L of Natural Pure WPI (Bioplex Nutrition). This protein source contains over 90% protein, 5% fiber, 4.5% salts, and no fat or sugar. Drinking water was used for WPI delivery, because it simplifies the monitoring of protein consumption independent from the HF diet. It is easily soluble in water and the normal route for human consumption. Body weight and feed and water consumption were measured twice weekly for the 11-wk duration of treatment.

Assays and procedures.

Blood glucose concentration in 8-h feed-deprived mice, glucose tolerance, and plasma insulin were determined as described (15). Insulin resistance was estimated by HOMA-IR (16). In vivo oxygen consumption and CO₂ release were determined under fed conditions to avoid any shift in metabolism toward fat utilization that may have been attributable to feed withdrawal rather than to the treatment condition (6). Body composition, including body fat, lean, and water, was assessed in live, unanesthetized mice by NMR (EchoMRI; EchoMedical Systems) (17). With this method, the physical characteristics of the lean component of the analysis is identical to lean chicken breast muscle, and therefore, lean represents primarily muscle mass. However, lean cannot be meaningfully expressed as a percentage of body weight when there are large differences in body fat. Therefore, we calculated lean as [(BW – body fat)/body water]. This index of muscle mass has extremely low inter-experimental variability.

Mice were killed by CO₂ asphyxiation. A 10% whole liver homogenate was prepared and one-half was used to prepare liver mitochondria and determine rates of oxygen consumption as described (18, 19). State 4 (ADP-limited) respiration was measured after addition of 3 mmol/L malate + 3 mmol/L glutamate; state 3 respiration was then determined following the addition of 0.25 mmol/L ADP. The respiratory control ratio was calculated as the ratio of state 3:state 4 respiration.

Aliquots of liver were also used to measure lipid content by 2 different procedures. First, total nonpolar and polar lipids were extracted using a modified Folch procedure (20) and then estimated gravimetrically. Briefly, a portion of the liver was extracted (2 h, dark, 4°C) with 2 vol chloroform and 1 vol methanol. Two vol (1 mmol/L) MgCl₂ was added, mixed, and centrifuged at 1500 × g for 15 min. The lower chloroform phase was treated with anhydrous Na₂SO₄ and taken to dryness under argon and the residue containing total liver lipids was dissolved in minimal vol diethyl ether. Acetone (10 vol) was added and the mixture stirred at 4°C to precipitate phospholipids. After centrifugation, the supernatant was removed and both supernatant (containing nonpolar lipids) and pellet (containing phospholipids) were separately dried under argon and weighed. Second, lipid was evaluated histologically. A 100-mg piece of liver was placed in 10% formalin, dehydrated through graded alcohols, embedded in paraffin, sectioned (5 μm), and mounted on a slide. The size and number of lipid droplets per unit area were calculated from several randomly selected fields per mouse by using a stage micrometer.

Statistics.

Values are presented as means ± SEM. Significance of the differences between group sample mean values was determined by Student’s t test. A P < 0.05 was considered significant. Comparison of the rates of body weight gain was evaluated using Friedman repeated-measures ANOVA on ranks with Tukey pairwise multiple comparison. Statistics were performed using SPSS software.

Results

WPI effect on body composition and energy homeostasis.

Body weight gains for HF and WPI mice were 0.39 and 0.23%/d, respectively (Fig. 1). This represents a 42% (P < 0.001) lower rate of body weight gain for the WPI mice. WPI mice also had 32% lower body fat and 7.4% greater lean body mass (primarily muscle) than HF mice (Table 1). WPI mice drank more water, resulting in 78% greater levels of protein consumption (Table 1). Consumption of fat plus carbohydrate in HF and WPI mice was 2.14 ± 0.13 kJ/(d⋅g body weight) and 1.79 ± 0.11 kJ/(d⋅g body weight), respectively (P = 0.059), such that WPI did not alter total energy consumption. This finding is consistent with the Protein Leverage Hypothesis, whereby total energy consumption is largely controlled by the levels of protein consumed (21). Furthermore, WPI did not affect calcium intake (P = 0.09).

FIGURE 1.

Cumulative percent increases in body weight (BW) in HF and WPI mice. Initial body weights for HF and WPI mice were 23.4 ± 0.9 g and 24 ± 1.1 g, respectively. Values are means ± SEM, n = 4.

TABLE 1.

Body composition and energy derived from feed and fluid in HF and WPI mice¹

	HF	WPI
Percent body fat,2(g fat/g body weight)⋅100	24.3 ± 2.1	16.7 ± 1.3*
Lean mass,3(g body weight – g body fat)/g body water	1.20 ± 0.01	1.30 ± 0.03*
Feed consumed,3g/(d⋅g body weight)	0.129 ± 0.003	0.105 ± 0.002*
Energy from feed,3kJ/(d⋅g body weight)	2.52 ± 0.09	2.04 ± 0.12*
Water consumed,3mL/(d⋅g body weight)	0.12 ± 0.02	0.23 ± 0.01*
Energy from water,3,4kJ/(d⋅g body weight)	0	0.44 ± 0.01*
Energy from feed + water,3kJ/(d⋅g body weight)	2.52 ± 0.09	2.47 ± 0.03
Dietary fat energy,3kJ/(d⋅g body weight) (% energy)	1.00 ± 0.09 (39)	0.81 ± 0.08 (33)
Dietary protein energy,3kJ/(d⋅g body weight) (% energy)	0.39 ± 0.03 (15)	0.69 ± 0.05* (28)
Dietary carbohydrate energy,34kJ/(d⋅g body weight) (% energy)	1.14 ± 0.09 (45)	0.98 ± 0.08 (39)
Dietary calcium intake,3mg/(d⋅g body weight)	1.81 ± 0.07	1.60 ± 0.08
O2 consumed, kJ/(d⋅g body weight)	1.13 ± 0.08	1.35 ± 0.05*
Energy efficiency,3O2 consumed/energy consumed in feed + water	0.45 ± 0.01	0.55 ± 0.02*
Feeding efficiency,3mg body weight gain/kJ consumed	5.14 ± 0.12	3.65 ± 0.07*

¹Values are means ± SEM, = 11 weekly estimates unless otherwise noted. *Different from HF, P < 0.05.

²n = 4 mice.

³Calculation based on 418.7 kJ/22.2 g WPI, the value supplied by the manufacturer.

⁴Calculated by subtracting the fat and protein energy from total energy intake.

WPI mice had higher rates of oxygen consumption and lower respiratory quotients, indicating that the mice consuming the WPI diet utilized lower amounts of dietary fat for energy metabolism (Table 2). The differences in basal metabolic rates were reflected in energy efficiency and feeding efficiency (Table 1). The higher basal metabolic rate in HF mice, which gained weight at a rate greater than WPI mice, helps explain the ability of WPI to lower the rate of weight gain. The higher basal metabolic rate in WPI mice could be due to the greater rate of liver mitochondrial respiration (Table 2). However, respiratory coupling in WPI mice was higher than in mice fed the HF diet alone. Because WPI improves mitochondrial energy coupling in HF mice, uncoupling does not contribute to the elevated basal metabolic rate.

TABLE 2.

Systemic and mitochondrial respiration in HF and WPI mice¹

	HF	WPI
Systemic O2 consumption, mL/(h⋅g body weight)	2480 ± 80	2970 ± 80*
Systemic CO2 release, mL/(h⋅g body weight)	1790 ± 80	2460 ± 50*
Systemic respiratory quotient, CO2/O2	0.72 ± 0.01	0.83 ± 0.01*
State 3 respiration,2nmol/(min⋅mg protein)	169 ± 9	229 ± 11*
State 4 respiration,2nmol/(min⋅mg protein)	55 ± 4	59 ± 4
Mitochondrial respiratory control ratio, state 3:state 4	3.1 ± 0.2	3.9 ± 0.3*

¹Values are means ± SEM, = 4. *Different from HF, P < 0.05.

²Mitochondrial oxygen consumption.

The amino acid profiles for the diets and the calculated consumption of individual amino acids for each treatment group had distinct differences (Table 3). The largest differences were in consumption of the basic amino acids arginine and lysine, as well as the hydroxyl-containing amino acids serine and threonine. Aliphatic branched-chain amino acids (BCAA), as a group, were consumed at twice the level in the WPI mice [Ile + Leu + Val = 8.0 ± 0.5 mg/(d⋅g body weight)] than in the HF mice [Ile + Leu + Val = 3.9 ± 0.2 mg/(d⋅g body weight)].

TABLE 3.

Diet composition and dietary intake of amino acids by HF and WPI mice¹

Amino acid	Dietary amino acid, g/100 g		Amino acid consumption, mg/(d⋅g body weight)
HF	WPI	HF	WPI
Ala	0.48	4.7	0.6 ± 0.03	1.6 ± 0.11*
Arg	0.56	2.4	0.7 ± 0.04	1.1 ± 0.07*
Asp	1.1	9.7	1.5 ± 0.09	3.4 ± 0.22*
Cys + (Cys)2	0.32	1.2	0.4 ± 0.02	0.6 ± 0.04*
Glu	3.6	16.4	4.6 ± 0.03	7.4 ± 0.49*
Gly	0.25	1.8	0.3 ± 0.02	0.7 ± 0.05*
His	0.43	1.6	0.6 ± 0.03	0.8 ± 0.05*
Ile	0.71	6.6	0.9 ± 0.05	2.2 ± 0.14*
Leu	1.5	10.2	1.9 ± 0.11	3.8 ± 0.25*
Lys	1.2	8.2	1.6 ± 0.09	3.1 ± 0.20*
Met	0.48	1.8	0.6 ± 0.03	0.9 ± 0.06*
Phe	0.79	2.8	1.0 ± 0.06	1.5 ± 0.10*
Pro	1.7	5.4	2.1 ± 0.12	2.9 ± 0.19*
Ser	0.93	4.8	1.2 ± 0.07	2.1 ± 0.14*
Thr	0.67	6.2	0.9 ± 0.05	2.1 ± 0.14*
Trp	0.20	1.5	0.3 ± 0.02	0.5 ± 0.03*
Tyr	0.85	2.5	1.1 ± 0.06	1.5 ± 0.10*
Val	0.87	5.0	1.1 ± 0.06	2.0 ± 0.13*
Total	16.6	92.7	21.4 ± 1.2	38.3 ± 2.5*

¹Values are mean ± SEM, = 11 weekly estimates. *Different from HF, P < 0.05.

WPI mitigates biomarkers for metabolic diseases.

Livers from HF mice had an abundance of nonpolar lipid in the form of numbers of lipid droplets, size of lipid droplets, and total quantity of nonpolar lipids (Table 4). Livers from WPI mice contained about one-half the HF diet-associated number of lipid droplets and the tissue content of nonpolar lipids, mainly TG. By comparison, WPI did not alter the amount of phospholipids, which are primarily associated with cell membranes. Although WPI did not alter the blood glucose levels in feed-deprived mice, WPI mice had lower levels of blood glucose following a glucose challenge, as indicated by the lower area-under-the-curve value (Fig. 2). These changes in glucose tolerance can be explained by changes in insulin secretion. The plasma insulin concentrations in feed-deprived WPI mice were only 29% of the high levels in HF mice, whereas after a glucose challenge, concentrations did not differ between the groups (Table 4). HOMA-IR values reflected the differences in plasma insulin in feed-deprived mice. Mice receiving WPI had HOMA-IR values that were one-third the value observed with HF alone, indicating greater insulin sensitivity in the WPI group (Table 4).

TABLE 4.

Lipid accumulation and insulin sensitivity in HF and WPI mice¹

	HF	WPI
Lipid droplets, n/mm2 liver	864 ± 65	537 ± 93*
Lipid vacuole diameter, μm	15.8 ± 1.6	15.3 ± 1.7
% Droplet area,22 droplet/mm2 liver	18.9 ± 2.7	9.8 ± 3.1*
Nonpolar lipids, mg/g liver	137 ± 15	63 ± 7*
Phospholipids, mg/g liver	42 ± 3	41 ± 4
Plasma insulin (zero-time),3pmol/L	134 ± 7	39 ± 4*
Plasma insulin (20 min),4pmol/L	176 ± 9	196 ± 15
HOMA-IR	3.0 ± 0.2	1.0 ± 0.1*

¹Values are mean ± SEM, = 4. *Different from HF, P < 0.05.

²Calculated from droplet size and number.

³Used to calculate insulin resistance as HOMA-IR.

⁴After i.p. glucose administration.

FIGURE 2.

Glucose tolerance in HF and WPI mice. Blood glucose was determined after 8-h feed deprivation. Following i.p. glucose administration, blood glucose concentrations were measured at intervals. The insert shows the calculated area-under-the-curve (AUC) value for each line. Values are means ± SEM, n = 4.

Discussion

High-energy and HF diets, coupled with sedentary lifestyles, contribute substantially to early onset insulin resistance, an increase in blood lipids, and hepatosteatosis, which in turn are predictive biomarkers for T2DM and NAFLD. Approximately 1 in 3 patients with NAFLD have T2DM (22), and ∼3 in 4 patients with T2DM have some form of NAFLD (23). In this paper, we show that WPI in female mice reduced the risk for developing fatty liver disease and diabetes associated with consumption of a HF diet.

Although peripheral insulin resistance in skeletal muscle and adipose tissue occurs early, compensatory hyperinsulemia generally masks clinical symptoms of diabetes for years, until metabolic oxidative stress and inflammation cause reduction in pancreatic β-cell mass. Eventually, reduced insulin production and increased insulin resistance manifest in diabetic hyperglycemia. Glucose and lipid hyperemia accelerate the decline in β-cell mass and the severity of T2DM (24). The etiology of fatty liver involves the inability of insulin in obese patients to suppress hormone-sensitive lipase-mediated release of nonesterified FFA from adipose tissue (25) as well as insulin-stimulated overproduction of VLDL-1 (26). In HF mice, hepatic TG lipase is involved in the development of hepatosteatosis and increased adiposity, likely mediated by the increased production of LDL (27). Steatosis is a requisite first step in the development of various NAFLD, including nonalcoholic steatohepatitis and cirrhosis, which is itself a strong predictor for the development of hepatocellular carcinoma (28).

The importance of insulin resistance and fatty liver in the etiology of T2DM and NAFLD has prompted human epidemiological and rodent studies, especially using the C57BL/6 mouse model (3), to develop strategies for disease intervention. Several diverse pathways have been targeted for intervention and remediation. The tumor suppressor protein PTEN (phosphatase and tensin homolog) is a phosphoinositide and protein phosphatase that regulates the PI3K/AKT signaling pathway, with loss in activity producing insulin resistance, NAFLD, and hepatocellular carcinoma (29). In mice, hepatocyte growth factor administration blocks HF diet-induced fatty liver by activation of microsomal TG transfer protein and apo B (30). In humans and mice, PPARγ-mediated upregulation of adipose differentiation-related protein promotes hepatosteatosis (31). In mice, HF diet-induced hepatosteatosis is reduced by administration of recombinant IL-22 via a mechanism involving downregulating hepatic lipid synthesis (32) and by reducing hepatic inflammation (33).

While such genetic and pharmaceutical approaches target specific pathways involved in metabolic diseases, intervention by dietary modification can target multiple pathways and offer less toxicological concerns. High-protein diets increase satiety and protein has low energy efficiency (metabolizable energy). The approximate relative energy efficiencies for starch, protein, and lipid are 0.842, 0.520, and 0.883, respectively (34). These values can explain the differences in energy efficiency reported in this study and could explain the effect of dietary protein on body weight. However, not all protein sources are similar in preventing or reversing metabolic disease. In overweight men and women, legume lupin kernel protein does not benefit body weight or composition, or lipid or glucose profiles in overweight men and women (35). In rats fed a HF diet, whey protein is better than red kangaroo meat for reducing the increase in body weight and insulin resistance (8). In aging humans and muscle builders, whey protein is better than soy protein in reducing sarcopenia and increasing muscle protein synthesis (9). Both whey and soy proteins support protein synthesis in skeletal muscle of male rats following exercise, possibly mediated by the stimulation via phosphorylation of the mammalian target of rapamycin (36). In sedentary young men, whey protein is better than casein in supporting postprandial protein synthesis (37).

The benefits from whey protein have resulted in a number of human and animal studies designed to elucidate the active components of whey and their mechanisms of action. In rats, whey protein supplements reduced liver lipid content but only when lipids were increased with a high-cholesterol diet (10). Obese human individuals lost less lean muscle during energy restriction when their whey protein supplement contained calcium and BCAA, especially leucine (38). Dietary supplementation by leucine alone was able to stimulate protein synthesis in skeletal muscle of neonatal pigs consuming a low-protein diet (39). In obese mice, weight loss under conditions of energy restriction was promoted by high calcium and WPI (11). In mice receiving a HF diet, high calcium plus whey protein inhibited the accumulation of body fat that is associated with changes in adipocyte gene expression, specifically by upregulating the leptin signaling pathway, and increasing the adrenergic signaling pathway via the β3 adrenergic receptor (12). A series of rodent and human epidemiological studies showed that high doses of dietary calcium inhibit the synthesis of bioactive vitamin D, which promotes Ca entry into adipocytes. Inhibiting adipocyte calcium levels decreased lipogenesis and promote lipolysis (7). Although this mechanism has the potential to contribute to leanness and weight management, it is not involved in this study, because dietary calcium was the same in the treatment groups.

BCAA levels are high in whey protein. In humans, whey protein decrease postprandial blood glucose levels, with the glycemic index-lowering peptide fraction high in aliphatic BCAA (13). BCAA can alter patterns of gene expression by activating the mammalian target of rapamycin pathway, thus promoting fatty acid oxidation and protecting pancreatic β-cells from AMP activated kinase-mediated apoptosis (40).

The differences between HF and WPI mice in systemic and mitochondrial oxygen consumption could be due to NO. In mice, a HF diet increases production of NO, leading to nitrosylation of mitochondrial proteins and heme binding to cytochrome oxidase, thus inhibiting and partially uncoupling respiration (41). The increase in NO could result from the release of proinflammatory cytokines (TNFα, IL-1β, and IFNγ) generated by the HF diet, promoting transcriptional activation of inducible NO synthase (42). Whey protein and whey-derived protein supplements reduce hepatic oxidative stress and systemic inflammation (43).

In mice and humans, HF diets contribute to the development of insulin resistance and hepatosteatosis, biomarkers and major risk factors for T2DM and NAFLD. In this study, WPI supplementation in mice reduced the severity of several biomarkers, including gain in body weight and adiposity, insulin resistance, and fatty liver. The protective effect of whey protein was consistent with higher basal metabolic rates and mitochondrial oxygen consumption and lower metabolic utilization of dietary lipid, leading to an overall lower feeding efficiency. Because the diets utilized in this study were not isonitrogenous, it is possible that supplementation with any protein would have been effective. Certainly, the active component(s) of whey responsible for these results have yet to be identified. Nevertheless, whey protein may have therapeutic potential to reduce the incidence of diabetes and fatty liver diseases, especially in at-risk individuals who consume excess energy and fat and lead a sedentary lifestyle.