Aqueous and methanol extracts of Vernonia amygdalina leaves exert their anti-obesity effects through the modulation of appetite-regulatory hormones

Abstract

Context: Aqueous and methanol extracts of Vernonia amygdalina Del. (Asteraceae) (AEVA and MEVA, respectively) leaves are reported to possess anti-obesity properties, exerted through unknown mechanisms.

Objective: This study investigated the effects of AEVA and MEVA on relevant hormones and enzymes in high-fat diet (HFD)-induced obese rats.

Materials and methods: Forty-two Wistar rats were placed into seven groups. The test groups received 100 mg/kg.bw AEVA (AEVA100), 500 mg/kg.bw AEVA (AEVA500), 50 mg/kg.bw MEVA (MEVA50) and 200 mg/kg.bw MEVA (MEVA200), respectively. The positive control (PC) group received 20 mg/kg.bw Orlistat, while the negative control (NeC) and normal control (NoC) groups received distilled water. The extracts were given orally daily for 12 weeks. Thereafter, the concentrations/activities of relevant hormones/enzymes in their sera were determined.

Results: Insulin concentrations (ng/ml) in the test groups ranged from 1.08 ± 0.01 (AEVA100) to 1.09 ± 0.01 (AEVA500). They were all similar (p > .05) to the NoC and PC controls. Leptin concentrations (pg/ml) in the test rats ranged from 0.02 ± 0.01 (AEVA500) to 0.03 ± 0.00 (MEVA50), and were all similar to the NoC group. The ghrelin concentrations of only the AEVA500 and MEVA200 groups were similar to those of the PC group (0.10 ± 0.01 pg/ml). AEVA100 and MEVA200 resulted in adiponectin concentrations (ng/ml) of the rats (0.27 ± 0.04 and 0.28 ± 0.04 respectively) that were similar to the PC group. The activities of lipoprotein lipase and the concentrations of intestinal amylase in the test rats were similar to values obtained for the control groups.

Conclusion: Appetite regulation may be the mechanism through which the weight-loss properties of AEVA and MEVA are expressed.

Keywords:

• Appetite regulation
• high-fat diet
• obesity
• phytochemicals
• Vernonia amygdalina

Introduction

Globally, between 1980 and 2013, the proportion of overweight/obese men and women [defined as a body mass index (BMI) of 25 kg/m² or greater] increased from 28.8% to 36.9% and 29.8% to 38.0%, respectively (Ng et al. 2014). In fact, obesity is reported to be the leading risk for deaths globally, accounting for the death of 2.8 million people annually (WHO 2012). Low and middle income countries (LMICs) are thought to be experiencing increases in the obesity epidemic that match, or in some cases surpass, those of some high income countries. A recent study in young-adult Nigerians reported overweight/obesity in 17% of the population (13% for males and 20.9% for females) (Ejike et al. 2015). A slightly earlier study in adult Nigerians (Akarolo-Anthony et al. 2014) reported significantly higher values [64% (74% for women and 57% for men)]. Clearly, obesity is currently a major public health challenge even in LMICs such as Nigeria, where thinness and wasting in children on the one hand, and infections and infestations on the other hand, compete for the often meager healthcare budgets.

In response to the alarming increase in the prevalence of obesity and the associated healthcare burden, there has been an increase in research activities geared towards developing phyto-pharmacological agents that are better than existing therapies for the disorder. There is currently a litany of plants, many of them found in the tropics, with reported anti-obesity potential. Their use appears to gain traction due to their production from ‘natural’ sources and purported safety (Yun 2010; Hasani-Ranjbar et al. 2013). Recently, our group reported that aqueous and methanol extracts of Vernonia amygdalina Del. (Asteraceae), a popularly consumed dark green leafy vegetable in Nigeria, was effective in causing significant weight loss (while sparing vital organs) and normalizing metabolic markers of obesity in a murine model fed a high-fat diet (HFD) (Egedigwe et al. 2016). Subsequent to that report, the possible mechanism(s) of action of the extracts were investigated. Here, the results of that investigation are reported.

Materials and methods

The materials used for this study and the study design were published recently (Egedigwe et al. 2016) and will, therefore, be described only briefly here.

Preparation of the VA extract

Fresh mature leaves of V. amygdalina were harvested (voucher specimen, FHI 28786-VA) cleaned and shade-dried to a constant weight before milling to fine powder. The constituents of the milled leaves were extracted in distilled water in one case and absolute methanol in another case for 48 h with occasional shaking, and at room temperature. Thereafter, the mixtures were filtered and concentrated. The yields were 20% and 10.4% for the aqueous and methanol extracts, respectively. The extracts, designated as aqueous extract of VA (AEVA) and methanol extract of VA (MEVA), respectively, were then reconstituted in 2% DMSO in normal saline to get the stock concentrations used in the study.

Preparation of high-fat diet (HFD)

Basic feed materials were used to prepare the basal diet and the HFD, following standard protocol. The HFD was designed such that 35% of the total energy in the diet came from fats. The compositions of both diets are shown in Table 1.

Table 1. Basal and HFDs composition.

Component	High fat diet (g/kg)	Basal diet (g/kg)
Maize	389	667
Groundnut cake	134	89
Egg yolk powder	58	0
Crayfish	22	29
Vitamin/mineral	20	11
Bone meal	20	11
Non-nutritive cellulose	4	2
Palm kernel oil	70	0
Palm oil	70	0
Corn starch	215	192

Animal treatment and feeding

Forty two male Wistar rats weighing 110–140 g were purchased from the University of Nigeria, Nsukka, and acclimatized to the animal house for 2 weeks. Thereafter, they were randomized into seven groups of six rats each (Table 2). The rats were housed in standard cages, had access to feed (test and control diets, as per the respective groups) and water ad libitum, and were maintained in the animal house under humid tropical conditions. The experiment lasted for 12 weeks during which the extracts, Orlistat or distilled water were administered daily to the appropriate groups of rats orally.

Table 2. Protocol for grouping the animals and treatments administered.

Group	Normal control	AEVA 100	AEVA 500	MEVA 50	MEVA 200	Negative control	Positive control
Diet	Basal diet	HFD	HFD	HFD	HFD	HFD	HFD
Treatment	Distilled water	100 mg/kg bw AEVA	500 mg/kg bw AEVA	50 mg/kg bw MEVA	200 mg/kg bw MEVA	Distilled water	20 mg/kg bw Orlistat

Treatments were given per os. Orlistat was purchased from a reputable commercial pharmaceutical vendor. AEVA, MEVA and orlistat were each dissolved in 2% DMSO. AEVA, MEVA and HFD represent Aqueous extract of V. amygdalina, Methanol extract of V. amygdalina and HFD, respectively.

At the end of the experiments, the rats were fasted, and then euthanized humanely. The rats were thereafter bled exhaustively by cardiac puncture. The sera from the blood were separated by centrifugation after clotting and were thereafter placed in labeled tubes to be used for hormonal and enzyme assays.

Hormonal and enzyme assays

Serum insulin, leptin, unacylated ghrelin and serum adiponectin concentrations were determined by the enzymatic immunoassay methods as described in the EIA kits procured from Bertin Pharma, Montigny-le-Bretonneux, France. Serum lipoprotein lipase, α-amylase and intestinal α-amylase activities were assayed for using test kits supplied by Agappe Diagnostics GmbH, Cham, Switzerland. For all determinations and assays, the manufacturer’s instructions were followed.

Statistical analyses

Descriptive statistical tests were carried out on the data generated. The results are represented as means ± standard deviations within the respective groups. To test for significant differences between the groups, the One-Way ANOVA test (and post hoc multiple comparisons) was employed with the significant threshold fixed at p < .05. The IBM-SPSS version 20.0 software (IBM Corp. Atlanta, GA) was used for all data analyses. The results are presented in figures.

Results

The serum concentrations of insulin in all the test groups were significantly (p < .01) lower than those of the negative control (NeC) group. The extracts resulted in all the test rats having insulin concentrations that were statistically similar (p > .05) (hereafter referred to simply as similar) to the non-obese rats (Figure 1). When compared to the positive control (PC) group, only the MEVA200 group had mean insulin values that were significantly lower (p < .05), the other test groups were similar to it.

Figure 1. Serum concentrations of insulin in HFD fed rats treated with different concentrations of AEVA and MEVA. [** and *** indicate significant differences at p < .01 and <.001, respectively; comparisons are made to the NeC group. Data for the test groups are statistically similar (p > .05) to the NoC group and the PC group (except for MEVA 200 versus PC)].

Serum leptin concentrations of the test rats were all significantly (p < .05) lower than that of the NeC group. Though the said values were all significantly (p < .05) higher than the PC group, they were in fact similar to those of the normal control (NoC) group (Figure 2). AEVA500 and MEVA200 resulted in serum concentrations of ghrelin that were similar to those of the PC group and significantly (p < .05) higher than those of non-obese rats. All the test groups nonetheless had serum ghrelin concentrations that were significantly (p < .05) higher than the NeC group (Figure 3). Treatment with AEVA100 and MEVA200 (but not the other extracts/concentrations) resulted in a significantly (p < .05) lowered serum adiponectin concentrations of the rats relative to the normal and NeC rats. The values were similar to those of the PC group (Figure 4).

Figure 2. Serum concentrations of leptin in HFD fed rats treated with different concentrations of AEVA and MEVA. [*** indicates significant difference at p < .001; comparisons are made to the NeC group. Data for the test groups were all statistically similar (p > .05) to the NoC group but significantly (p < .01) higher than the PC group].

Figure 3. Serum concentrations of ghrelin in HFD fed rats treated with different concentrations of AEVA and MEVA. [*** indicates significant difference at p < .001; comparisons are made to the NeC group. Data for AEVA 500 and MEVA 200 were significantly (p < .05) higher than the NoC group while the other two were similar. AEVA 100 and MEVA 50 were significantly (p < .01) lower than the PC group].

Figure 4. Serum concentrations of adiponectin in HFD fed rats treated with different concentrations of AEVA and MEVA. [* indicates significant difference at p < .05; comparisons are made to the NeC group. Data for the test groups were all significantly (p < .05) lower than the NoC group. AEVA 500 and MEVA 50 were significantly (p < .05) higher than the PC group (while the other two were similar)].

The activities of lipoprotein lipase in the sera of the test rats (except MEVA50) were all similar to values obtained for all the control groups. The MEVA50 group had significantly (p < .05) higher activity compared to the controls (Figure 5). The serum amylase concentrations of the test rats were significantly (p < .05) lower than that of the NeC group, but similar to those of the normal and PC groups (Figure 6). Conversely, the intestinal amylase concentrations of the test rats were all similar to the negative and PC groups (Figure 7). Non-obese rats had significantly (p < .05) higher intestinal amylase concentrations compared to the obese groups (treated or untreated).

Figure 5. Serum lipoprotein lipase activity in HFD fed rats treated with different concentrations of AEVA and MEVA. [* indicates significant difference at p < .05; comparisons are made to the NeC group. Data for the test groups (except MEVA 50) were all statistically similar (p > .05) to both the NoC and the PC groups].

Figure 6. Serum concentrations of α-amylase in HFD fed rats treated with different concentrations of AEVA and MEVA. [*** indicates significant difference at p < .001; comparisons are made to the NeC group. Data for the test groups were statistically similar (p > .05) to both the NoC and the PC groups].

Figure 7. Serum concentrations of intestinal amylase in HFD fed rats treated with different concentrations of AEVA and MEVA. [Data for the test groups were statistically similar (p > .05) to the NeC and PC groups; but significantly (p < .01) lower than the NoC group].

Discussion

The weight loss reported in rats given extracts of VA is believed to be due to the rich milieu of phytochemicals present in the leaves of the plant (Egedigwe et al. 2016). Such phytochemicals include (but are not limited to) saponins, sesquiterpenes, lactones and flavonoids (Ijeh & Ejike 2011). The goal of the present study was to identify the possible mechanism(s) of action of both MEVA and AEVA. The strategy was to investigate the role of some hormones involved in appetite regulation and energy metabolism, and enzymes involved in the absorption of energy molecules from the gastrointestinal tract. We opted for HFD-induced obesity in rats as our model, knowing that it has substantial similarity to what is observed in the pathophysiology of human obesity (Buettner et al. 2007).

The extracts lowered the insulin concentrations of the test rats such that they were similar to the PC and the non-obese rats. Hyperinsulinemia often precedes the development of obesity. Ordinarily, insulin is secreted post-prandially, and functions as a negative feedback signal which reduces feed intake (and ultimately the development of obesity) (Clegg et al. 2005). In diet-induced obese rats, however, loss of sensitivity to insulin and leptin cues results in hyperphagia and the attendant failure to reduce their energy intake and positively regulate their body weight. The rats remain hyperphagic despite early spikes in the concentration of insulin and leptin (Levin et al. 2004). This explains the observation that insulin concentrations are high in such rats, yet they remain obese. It is therefore interesting to observe that both MEVA and AEVA lowered the concentration of insulin in the test rats, plausibly by restoring the rats’ sensitivity to insulin and thereby relaxing the feedback loop that ensures an increased production and release of insulin in the insulin-resistant state.

Leptin was reduced significantly and was similar to the PC. Leptin is synthesized by adipose tissue, and it acts on the hypothalamus, resulting in reduced food intake and increased energy expenditure. In well fed individuals, circulating leptin concentrations is proportional to the quantity of adipose tissue present in the individual (Webber 2003). Leptin is produced subsequent to post-prandial insulin secretion and its concentration in serum decreases as insulin decreases during fasting (French & Castiglione 2002). Leptin has been shown in many rodent studies to function as a feedback signal that inhibits feed intake when secreted, and in doing so it regulates body weight gain (Klok et al. 2006). A reduced sensitivity to the anorectic effects of leptin is a known method through which obesity is induced in rats (Levin & Dunn-Meynell 2002; Levin et al. 2004). It is therefore interesting to find that the extracts at the tested concentrations significantly reduced the circulating concentrations of leptin to values that were statistically similar to that of the non-obese rats. The very high concentrations of leptin found in the NeC are likely due to increased production of the hormone in response to a reduced sensitivity to leptin as mentioned earlier. The extracts may have restored leptin sensitivity to the test rats; a fall-out of which would be the reduced feed intake in the test groups reported earlier (Egedigwe et al. 2016).

Serum ghrelin concentrations were higher in the test rats than in the NeC and non-obese rats; but similar to the PC group. Ghrelin is produced principally by the stomach. It stimulates the growth hormone secretagogue receptor (GHS-R) on neuropeptide Y (NPY)/agouti-related peptide (AGRP)-producing neurons located in the arcuate nucleus (Drazen & Woods, 2003; Hagemann et al. 2003). This means, it produces its profound orexigenic, adipogenic and somatotropic properties, in humans and rodents, ultimately increasing food intake and body weight. Circulating ghrelin levels increase during fasting and decrease after a meal. Ghrelin is known to function in opposition to leptin (that is, while leptin is satiety-inducing, ghrelin is appetite-stimulatory) (Yildiz et al. 2004). The results show a significantly higher ghrelin concentration in the test rats relative to the negative and NoC groups. As leptin has been shown to be an upstream regulator of ghrelin in rodents (Ueno et al. 2005), it appears that the higher circulating ghrelin levels in the test rats may be as a result of the low leptin concentrations in the said rats. Ghrelin antagonism is known to decrease the increased appetite that ordinarily occurs with decreased feeding (Bays 2004). The above fact is, however, without prejudice to the expected spike in the ghrelin concentration of all the rats as a result of the fact that they were fasted prior to sacrifice.

Serum adiponectin concentrations were significantly lowered by the extracts and the values were similar to that of the Orlistat-treated control. Adiponectin is produced largely in adipocytes. One of its major physiological effects is that it decreases lipogenesis and gluconeogenesis in the liver, thereby resulting in decreased blood glucose and free fatty acid concentrations. The serum adiponectin concentration is a significant correlate of insulin sensitivity (Klöting et al. 2010) as insulin (and insulin-like growth factor, IGF-1) increases its synthesis in white adipose tissue. Unlike leptin, adiponectin levels are lower in obese individuals. In fact, there are reports in the literature of a negative correlation between circulating adiponectin and percentage body fat and central fat distribution – markers of adiposity (Faraj et al. 2003; Bouassida et al. 2010). The lowered concentration of insulin in the test rats may have caused the lowering of circulating adiponectin in the test rats. It is important to note, however, that though one would expect higher concentrations of adiponectin in the serum of the test rats, given that they had lost some weight, they still had values that were lower than the NeC group. This only makes sense when viewed in the context of the understanding that adiponectin synthesis and secretion are decreased in cases of lowered leptin concentrations (Faraj et al. 2003), as seen in this study. In fact, all the rats in which obesity was induced had adiponectin concentrations that were significantly lower than that of the non-obese groups. Interestingly, the AEVA500 and MEVA50 groups had concentrations that were significantly higher than those of the PC group.

As the reduction in the availability of dietary energy is desirable in the management of obesity, the inhibition of some digestive enzymes is thought to be beneficial. We, therefore, investigated the effects of MEVA and AEVA in inhibiting some enzymes involved in lipid and carbohydrate absorption and uptake, namely lipoprotein lipase, serum α-amylase and intestinal amylase. Lipases are enzymes responsible for the digestion of lipids, including triacylglycerol and phospholipids. Lipoprotein lipase is the rate-limiting enzyme in the sequence of reactions involved in the uptake and storage of lipoprotein triacylglycerol by the adipocytes. The pancreatic lipase is reported to be responsible for the hydrolysis of 50–70% of dietary lipids (Shi & Burn 2004). It is, therefore, not surprising that recent approaches to the treatment of obesity have targeted the inhibition of lipases (Birari & Bhutani 2007) as a way of preventing the absorption of excess calories. It is, therefore, surprising that despite the weight loss observed in the test rats, the lipoprotein lipase activity in the serum of the test rats were all statistically similar to that of all the control groups. This finding suggests that the inhibition of lipoprotein lipase may not be a mechanism of action of the extracts.

Inhibiting carbohydrate-hydrolyzing enzymes such as the amylases reduces the absorption of glucose. This is particularly important when viewed in the context of the reduced insulin sensitivity that is the hallmark of diet-induced obesity and the need to reduce the availability of precursors for the synthesis and accumulation of triacylglycerol by the adipose tissue. Therefore, amylases offer a veritable target for obesity treatment. In fact, the usefulness of plant extracts in inhibiting amylases, and thus modulating disorders of energy metabolism has been reported (Barrett & Udani 2011). The data on amylases are slightly discordant as the intestinal amylase activities of the test rats were similar to that of all the control groups, while the serum α-amylase activities were significantly lower than those of the NeC, but similar to normal and PC groups. It, therefore, appears that the inhibition of the absorption of glucose from the intestines may not be the mechanism of action of the extracts.

The use of different concentrations of AEVA and MEVA for this study clearly makes the comparisons between the two extracts difficult. It was, however, borne out of our experience with extracts of the leaves of V. amygdalina. Over the years, we have noticed that methanol extracts of the said leaves are usually more potent. The pilot study conducted in preparation for this study corroborated the above observation. We, therefore, necessarily used lower concentrations of MEVA.

In conclusion, the mechanism(s) through which AEVA and MEVA exert their reported anti-obesity effects were studied. It appears that the weight-loss potentials of AEVA and MEVA are as a result of their ability to restore insulin and leptin sensitivity to the test rats, and the down-stream modulation of ghrelin and adiponectin concentrations in serum. A combination of these apparently resulted in the reversal of hyperphagia, which ordinarily drives obesity in the studied model. The reduced appetite ultimately would lead to a negative energy balance and the desired weight-loss.

Disclosure statement

The authors have no real or potential conflicts of interest to declare. CAE was responsible for performing the experiments, contributed to study design and revision of the initial draft of the manuscript. III conceived the study, participated in study design, supervised the study, and contributed to the revision of the manuscript. PNO participated in supervising the study. CECCE contributed to study design and supervision, analyzed the data, plotted the graphs and wrote the manuscript. All authors read and approved the eventual manuscript.

Funding

The authors would like to acknowledge research support from TETFund Nigeria 2012/2013 grant administered through the MOUAU-DURA, which supported this study.