Lupin protein isolate improves insulin sensitivity and steatohepatitis in vivo and modulates the expression of the Fasn, Gys2, and Gsk3b genes

Abstract Although studies on lupin protein isolate (LPI) have indicated the presence of a preventive effect on insulin resistance (IR) and lipid disturbances, their influence on established pathological traits has received little attention. Here, we evaluated the in vivo effects of LPI on IR and steatohepatitis as well as its influence on genes involved in lipid and carbohydrate metabolism. We first induced IR and steatohepatitis in rats by maintaining them on a high‐fat diet for 5 weeks. Thereafter, we administered LPI to the rats daily for 3 weeks. LPI improved insulin sensitivity (AUC: 6,777 ± 232 vs. 4,971 ± 379, p < .05, pre‐ vs. post‐treatment values) and reduced glucose and triglyceride levels by one‐third. In addition, LPI‐treated rats exhibited attenuated steatohepatitis. At the molecular level, LPI treatment reduced liver Fasn gene expression substantially but increased Gys2 and Gsk3b levels. We concluded that the hypolipidemic and hypoglycemic activities of LPI may be caused by reduced liver lipogenesis and modulation of insulin sensitization mechanisms.

on encouraging lifestyle changes, including forming healthy eating habits and performing daily exercise, followed by treatment with pharmacological agents (Cornier et al., 2005;Devlin, 1992;McAuley et al., 2003). In particular, the use of legume seed proteins for the management and treatment of noncommunicable diseases (NCDs) is an area of current interest, with an increasing number of recent reports regarding this (Hosseinpour-Niazi et al., 2015;McMacken & Shah, 2017;Pihlanto et al., 2017;Rizkalla et al., 2002).
Legumes have adequate nutritional value for a healthy diet because of their protein, fiber, complex carbohydrates, and micronutrient content (Polak et al., 2015). Besides their nutritional properties, legumes contain a wide variety of health-promoting bioactive compounds (Rebello et al., 2014). Regarding the importance of legume seeds in diabetes management, clinical studies have supported its inclusion in the diet because of its purported insulin sensitization actions (Clark et al., 2018).
The legume lupin (genus Lupinus) belongs to the Fabaceae family and includes more than 300 species distributed worldwide. The domestic species Lupinus albus, L. angustifolius, L. luteus, and L. mutabilis are the most widely cultivated and most widely used lupins for animal and human consumption in various countries (Duranti et al., 2008;Kohajdová et al., 2011). Furthermore, lupin and soybean represent excellent protein sources for the human diet (Erbaş et al., 2005;Sirtori et al., 2004).
Studies on humans and laboratory animals have shown that adding soybean protein to the diet lowers total cholesterol, LDLcholesterol (LDL-c), and triglyceride serum levels as well as ameliorates IR (Anderson et al., 1995;Tachibana et al., 2014). Moreover, soybean protein supplementation has been associated with clinical improvements in metabolic syndrome and T2D. Reduced values of fasting plasma glucose, insulin, homeostatic model assessment of IR (HOMA-IR), total cholesterol, low-density lipoprotein cholesterol, diastolic blood pressure, and C-reactive protein have also been found in subjects after consuming soybean protein (Zhang et al., 2016). Similarly, other studies have shown beneficial effects of lupin protein on hyperlipidemia and hyperglycemia (Bouchoucha et al., 2016;Sewani-Rusike et al., 2015). In rats fed with lupin proteinsupplemented pasta, body weight gain and food intake were reduced (Capraro et al., 2014). In addition, the consumption of a lupin proteinbased beverage caused acute reductions in serum glucose levels in T2D patients (Dove et al., 2011). Fornasini et al. (2012) found that lupin protein did not reduce glycemia in normoglycemic individuals but induced a hypoglycemic effect in dysglycemic individuals.
Previous research in the nutraceutical field has revealed the potential use of lupin protein in regulating IR, lipid, and glucose metabolism. One mechanism related to the metabolic effects of insulin is the modulation of liver glycogen synthesis, mainly through the activation of the glycogen synthase enzyme. The increase in liver glycogen synthesis has been associated with improved glucose tolerance (Ros et al., 2010). Specifically, Gys2 and Gsk3b are directly involved in the glycogen synthesis pathway. In contrast, altered lipid metabolism has also been implicated in IR. Fatty acid synthase protein, encoded by Fasn, is an enzyme that participates in the synthesis of long-chain saturated fatty acids. Interestingly, it has been reported that a luteolin-enriched artichoke leaf extract reduced Fasn gene expression and triglyceride levels in vivo by modulating lipogenesis and fatty acid oxidation, contributing to the amelioration of liver steatosis (Kwon et al., 2018).
Thus, we hypothesized that lupin protein isolate (LPI) exerts a beneficial effect on IR and steatohepatitis in an in vivo model through the modulation of Fasn, Gys2, and Gsk3b gene expression. Here, we aimed to analyze the effects of LPI on IR and steatohepatitis in vivo.
We also studied how lupin protein influences the expression of genes involved in lipid and carbohydrate metabolism such as Gys2, Gsk3b, and Fasn in the livers of animals under a pathological IR state.

| IR induction and experimental procedure
The study consisted of a 5-week IR induction period followed by a 3-week experimental phase. In the induction period, a group of ten healthy rats (180-220 g body weight) were provided with free access to a high-fat, high-cholesterol diet (HFD) to induce IR, whereas a control group (Ctrl) of five rats were fed a standard chow diet (Purina LabDiet ® 5001). After 5 weeks, we randomly assigned the induced IR rats to an IR-LPI (lupin-treated) or IR (saline placebo-treated) experimental group (five rats per group). During the 3-week experimental phase, animals in the IR-LPI and IR groups were maintained on the HFD and given daily doses of LPI (2 g/kg BW of LPI dissolved in 5 ml saline) or sterile saline (5 ml), respectively. The Ctrl group was maintained on standard chow diet.
We performed an insulin tolerance test (ITT) at the start and end of the induction period as well as at the end of the experimental phase. We also determined the blood biochemical parameters at the beginning and end of the experimental phase. At the end of the treatment period, we excised and collected the livers for histological analysis and quantification of Fasn, Gys2, and Gsk3b gene expression ( Figure 1).

| Insulin tolerance test
The IR status was verified in the animals by performing ITT at the beginning and end of the 5-week IR induction period as well as at the end of the treatment period. ITT was done by administering intraperitoneal insulin injections (0.5 IU/kg BW, Humulin R ® , Eli Lilly and Company) after a 12 hr fasting period. Blood glucose concentrations were measured using a glucometer (One Touch Ultra ® , Johnson & Johnson) at 0, 30, 60, and 90 min after insulin injection. Finally, we calculated the area under the curve (AUC) using the glucose values and determined the IR status using the Ctrl group as a reference.

| Blood sample collection
The animals were fasted for 12 hr and anesthetized via intraperitoneal administration of Zoletil ® 50 (80 mg/kg BW; tiletaminezolazepam; Virbac). Blood samples were withdrawn from the retro-orbital plexus. Serum was collected from the centrifuged blood samples (1,300 x g for 10 min, 4°C) and stored at −70°C until analysis.

| Determination of biochemical parameters
Serum glucose, triglyceride, total cholesterol, LDL-c, HDL-c, and alanine aminotransferase levels (ALT) were determined using a semiautomatic spectrophotometer (BTS-350) and commercial reagents Fasn; 95°C for 10 min and 45 cycles at 95°C for 10 s, 61°C for 10 s, and 72°C for 6 s for Rps18. In the negative controls, sterile water was used instead of the cDNA. All amplification reactions were performed in triplicates using a 2.0 LightCycler ® instrument (Roche).
We normalized the target gene Ct values with the Rps18 Ct values and calculated the relative Fasn, Gys2, and Gsk3b gene expressions using the 2 −ΔΔCt method (Livak & Schmittgen, 2001). Single product amplification for each gene was confirmed via melting curve analysis.

| Liver histological assessment
At the end of the 3-week experimental period, the animals were anesthetized, and their livers were excised by laparotomy. Next, we fixed tissue fragments in 4% paraformaldehyde (1× phosphatebuffered saline) and embedded them in paraffin. A certified pathologist evaluated the 4μm thick paraffin-embedded liver tissue sections separately stained with hematoxylin and eosin (H&E) and Masson's trichrome. We also semi-quantitatively determined liver glycogen changes by staining tissue sections with periodic acid-Schiff (PAS)-diastase. In this staining procedure, the tissue sections were treated with diastase before application of the PAS stain.

| Statistical analysis
Results are presented as the mean ± standard error of the mean. We established which biochemical parameters significantly differed between pre-and post-treatment, using the dependent t-test. The significance of the inter-group effects in Fasn, Gys2, and Gsk3b gene expression and AUC was established using ANOVA and Bonferroni post hoc test.
We used IBM SPSS Statistics software for Windows (version 20.0; NY, USA) for data analysis and considered p < .05 as significant.

| Gene expression
The effects of LPI on Gys2 and Gsk3b gene expression are shown in

| Histological analysis
In the histological analysis of the liver sections, we found that our experimental model reproduced histopathological changes compatible with NASH in the induced IR groups ( Figure 5). All IR animals not treated with LPI (IR group, Figure 5b) presented grade III steatohepatitis with necroinflammatory activity, based on the histopathological criteria stipulated by Kleiner et al. (2005). Interestingly, we observed decreased lobular inflammation in the IR-LPI group (Figure 5c).
Notably, only two out of five LPI-treated animals had grade III steatohepatitis, whereas the other three experienced amelioration of the disease (exhibiting grade I or II steatohepatitis).
Expression of the glycogen-related Gys2 gene led us to test for   However, PAS positivity was located intercellularly between hepatocytes ( Figure 6b). In this case, PAS staining remained even after diastase treatment (Figure 6d).

| D ISCUSS I ON
It is widely accepted that IR is associated with metabolic disorders such as hyperglycemia and hypertriglyceridemia and is essential in the development of T2D. Thus, improving a patient's insulin sensitivity is crucial for preventing the progression of IR to NCD. This progression is associated with hyperglycemia and lipid alterations, such as increased LDL-c and triglyceride levels as well as reduced HDL-c concentrations (Marotta et al., 2010).
To evaluate the effects of nutraceuticals and functional foods on metabolic pathways related to IR, in vivo and in vitro models were employed. In this study, we used a high-fat, high-cholesterol diet to induce IR. As expected, we observed that the induced IR animals exhibited higher glucose, triglyceride, total cholesterol, and LDL-c serum levels. The observed variations in the IR rats before the treatment could be explained by individual metabolic responses to in vivo IR induction.
After publication of this study, several groups have attempted to discover the mechanisms behind the regulation of glucose homeostasis by Cγ.
Recent reports have described the effects of lupin extracts on IR. In this regard, the biological effects of other legumes on IR have also been previously explored. Chickpea seed flour was found to prevent development of an IR model induced by a HFD (Yang et al., 2007).
Recently, Terruzzi et al. (2018) reported that lupin flour reduced HOMA-IR values in C57BL/6 mice that were fed an IR-inducing experimental diet.
However, its applicability to experimental research has been questioned because of its weak validation for small samples (Wallace et al., 2004). ITT is a procedure used to determine insulin sensitivity in in vivo studies. This is an acceptable method for IR evaluation since it has been compared and validated against the hyperinsulinemic-euglycemic clamp, which is the gold standard (Akinmokun et al., 1992;Bonora et al., 1989;Gelding et al., 1994;Hirst et al., 1993).
In this study, we used ITT to evaluate the effect of LPI on IR.
We found a significant improvement in insulin sensitivity after LPI administration, indicating that lupin proteins exert an insulinsensitizing effect.

F I G U R E 3
Relative mRNA expression levels of the Gys2 and Gsk3b genes. The bars represent the mean ± SEM. We performed ANOVA and Bonferroni post hoc tests. *p <.05, **p <.01. Ctrl, control group; IR, insulin resistance group; IR-LPI, insulin resistance group treated with LPI; LPI, lupin protein isolate F I G U R E 4 Relative mRNA expression levels of the Fasn gene. The values represent the mean ± SEM. Statistical analysis was performed via ANOVA and Bonferroni post hoc tests. ***p <.001. Ctrl, control group; IR, insulin resistance group; IR-LPI, insulin resistance group treated with LPI; LPI, lupin protein isolate In addition to IR amelioration, an attenuated postprandial hyperglycemia was reported in a clinical study wherein oral administration of lupin kernel flour or soybean proteins was found to exert an acute glucose-reducing effect evaluated by the oral glucose tolerance test (OGTT) in T2D patients. The authors concluded that these results were probably due to similarities in protein content between these legumes (Dove et al., 2011). Furthermore, another study has reported that rats fed with a high-fat diet supplemented with raw chickpea seeds exhibited reduced serum glucose levels as compared with a nonsupplemented group under measurement with OGTT and ITT (Yang et al., 2007). In accordance with these findings, our results showed that LPI treatment improved IR from experimental induction by a high-fat diet.
Insulin modulates liver glycogen synthesis by activating glycogen synthase enzyme. The Gys2 gene encodes liver glycogen synthase, an enzyme responsible for directing synthesis of glycogen, one of the primary sources of stored energy in the body. In contrast, Gsk3b encodes a serine-threonine kinase with negative regulatory activity on glycogen synthesis. Thus, an increase in liver glycogen synthesis is associated with improved glucose tolerance (Ros et al., 2010).
Since treatment with LPI has exhibited improvements in insulin sensitivity, we evaluated the gene expressions of Gys2 and Gsk3b.
Here, we found that LPI treatment increased Gys2 and Gsk3b gene expression in the liver. Similar results were also found in the gene expression and protein levels in diabetic mice treated with Bauhinia holophylla extract, a plant belonging to the Fabaceae family (Camaforte et al., 2019). Furthermore, our results showed that LPI induced a recovery of insulin sensitivity and a subsequent decrease in hyperglycemia, a finding that may have involved the activation of hepatic glycogenesis. Therefore, we decided to analyze the glycogen content in liver tissues from the experimental groups. Unexpectedly, we did not observe differences in the liver PAS positivity of the IR and IR-LPI groups. This may be attributed to the development of prominent fibrosis when inducing IR with a high-fat diet, hindering the histological evaluation and ability to differentiate glycogen from other extracellular matrix components. To better understand the role of LPI in glycogen metabolism, future studies should consider quantifying liver glycogen levels and assessing the protein levels of GSK3β and GYS2.
We cannot rule out higher glucose uptake by other insulinsensitive tissues such as skeletal muscle and adipose tissue as an explanation for our observations (Petersen et al., 2007;Tachibana et al., 2014;Terruzzi et al., 2011). Based on our data, we can hy-  Metabolic changes associated with IR also correlate with NAFLD and its progression to NASH (Ota et al., 2007). In the present study, we also evaluated the histological changes caused by a high-fat diet and the effect of LPI administration on liver tissue. Using the NASH staging scoring system (Takahashi & Fukusato, 2014), an amelioration of the degree of steatohepatitis in the IR-LPI group was observed. The relationship between IR and dyslipidemia is widely known (Franch-Nadal et al., 2018;Garg, 1996). We observed a decrease in Other studies have also shown that rats fed a western diet and treated with lupin proteins exhibited lower triglyceride and VLDL levels (Fontanari et al., 2012;Spielmann et al., 2007). In accordance with our results, Spielmann et al. (2007) also found lower mRNA levels of Fasn and Srebp-1c genes as well as lower liver triglyceride content in animals treated with lupins. In addition, it has been reported that HFDs are not only associated with NAFLD development but have also been found to induce IR (Liu et al., 2016). Therefore, the insulin-sensitizing effect observed in LPI-treated animals and reduced Fasn levels might be associated with the attenuation of liver damage.
Finally, our data contribute to the understanding of the biological effects exerted by lupin proteins in a state of IR. To the best of our knowledge, this is the first study to provide evidence that LPI could be useful in the treatment of established pathological states, such as IR and steatohepatitis. Moreover, our results add to the body of evidence indicating the presence of a beneficial effect of lupins on insulin sensitivity, as suggested by some clinical studies. Likewise, there is a need for further basic and clinical research aimed at clarifying whether lupin-based therapies are beneficial in different clinical conditions such as obesity, dyslipidemia, IR, and diabetes.
In conclusion, our results show that LPI exerts insulin-sensitizing and hypoglycemic effects on IR-induced rats. Moreover, administration of lupin proteins promoted a hypolipidemic effect by decreasing serum lipid levels, involving a marked reduction in the mRNA levels of the lipogenic Fasn gene in induced IR animals. Thus, we provide evidence that lupin proteins might be useful in restoring insulin sensitivity and attenuating the histopathological changes induced by IR.
Further molecular and metabolic characterizations of the liver and other insulin-sensitive tissues represent an exciting research area to better understand the effects of lupin proteins and their potential use in the management of metabolic disorders.
F I G U R E 6 Histological comparison of glycogen content in the IR and IR-LPI groups. Representative photomicrographs from the tissues stained with PAS and PAS-diastase are shown. In the IR group, loss of staining in the cytoplasmic glycogen deposits (arrows) of hepatocytes after diastase treatment was observed (a and c). In contrast, panels b and d show an intercellular PAS positivity in tissues from the IR-LPI group (arrows). IR, insulin resistance group; IR-LPI, insulin resistance group treated with LPI; LPI, lupin protein isolate (a) (b) (c) (d)