Safety evaluation of naringenin upon experimental exposure on rat gastrointestinal epithelium for novel optimal drug delivery

Abstract

Objective: To assess the effect of naringenin on the intestinal biochemical composition, function and histology for gastrointestinal toxicity since it has not yet been adequately exploited for safety through standard assays.

Methods: Here, we describe naringenin (1 mM, 10 mM and 100 mM, respectively) or sodium deoxycholate (10 mM) effects on isolated brush border membrane from intestinal segments with single pass intestinal perfusion using lactate dehydrogenase, alkaline phosphatase and protein assays. MTT assay was used for cytotoxicity studies. Everted gut sac studies were used for evaluating the transport of nutrients across the intestinal segments. Lucifer yellow was used for paracellular permeability, followed by histological changes and surface characteristic studies of intestinal sacs.

Results: The results indicated no significant alterations with naringenin, although significant (p < 0.01) changes were noticed with sodium deoxycholate in the activity of the rat intestinal brush border associated enzymes such as LDH, followed by intact cell viability with marked decrease in the villi height of the intestinal segments.

Conclusions: These observations indicate that naringenin was harmless upon exposure to rat gastrointestinal epithelium, clearly demonstrating the potential use of naturally occurring bioflavonoid as safe and novel pharmaceutical adjuvant in oral dosage forms as P-gp inhibitor.

• Bioflavonoid
• cytotoxic
• intestine
• naringenin
• P-gp inhibition

Introduction

P-glycoprotein (P-gp), a membrane transporter pumps out a wide range of xenobiotics out of cells (Ueda et al., 1999) limiting drug absorption in gastrointestinal tract consequently oral bioavailability (Terao et al., 1996; Greiner et al., 1999).

Although P-gp decreases drug absorption of all of its substrates, it quantitatively does not have an equal significant impact on overall drug absorption for all of them. In the case of fast absorbing drugs having larger doses, efflux by P-gp poses less impact on drug absorption and therefore, it is not important in terms of bioavailability or pharmacokinetic properties. This is because the transport activity of P-gp becomes saturated due to high concentrations of drug in the intestinal lumen. In the case of drugs requiring a very small dose for their pharmacological actions or the drugs that have very slow dissolution and diffusion rates, P-gp-mediated drug efflux greatly interferes with their delivery. As it decreases drug absorption, those small amounts of drugs cannot reach the blood circulation in sufficient quantity and at times, can be life threatening. Moreover, it can make the sustained release dosage forms of the substrates completely ineffective by limiting their absorptions (Amin, 2013).

These effects have restricted the development of new chemical entities (NCEs) which are substrates of P-gp. Thus, there is considerable interest in trying to enhance their absorption and oral bioavailability by inhibiting the P-gp-mediated drug efflux.

It has been demonstrated that a number of excipients, such as polyethylene glycols (PEGs), non-ionic surfactants, fatty acids and bile salts (Nerurkar et al., 1996, 1997; Lo & Huang, 2000; Arima et al., 2001; Johnson et al., 2002; Hugger et al., 2002a,2002b, 2003) could inhibit the function of P-gp, thereby increasing the absorption of P-gp substrate drugs (Wagner et al., 2001).

Alternatively, plants offer inexpensive and plentiful of phytochemicals. Among them, the flavonoids from the major class of natural compounds commonly present in fruits, vegetables and plant derived beverages, such as tea, red wine and beer (Zhang & Morris, 2003), that have been recognized as P-glycoprotein modulators (Shapiro & Ling, 1997; Conseil et al., 1998; Bansal et al., 2009) for a long time. Therefore, it is important to find a P-gp inhibitor of natural origin, which does not possess toxicological actions with higher margin of safety.

Naringin (glycoside) is a “bioflavonoid” derivative of grapefruit peel and related citrus species, responsible for the sour flavor and bitter taste of the fruit (Jagetia et al., 2003). Upon oral administration, flavonoid glycoside (naringin) is metabolized to yield a major metabolite-naringenin, which is the main absorbable form in intestine, by human intestinal microflora through enterobacterial enzymes, such as α-rhamnosidase and β-glucosidas (Kim et al., 1998). The average intake of naringenin has been estimated as 8.3 mg per day in Finland. This constituted to 15% of the total flavonoid intake (Erlund et al., 2001).

Such bioflavonoids, naringin or naringenin show inhibition of P-gp-mediated cellular efflux of P-gp substrates, which implies that they might affect the absorption, distribution and elimination of P-gp substrates, with increase in intestinal absorption and oral bioavailability (Lassoued et al., 2011; Gurunath et al., 2013b).

The antioxidant, free radical scavenging and metal chelating properties of naringin demonstrates that naringin reduced the genotoxic effects of bleomycin and consequently increased the cell survival, and therefore may act as a chemoprotective agent in clinical situations (Jagetia et al., 2007). Its cytoprotective effect against ethanol injury in the rat appears to be mediated by non-prostaglandin-dependent mechanisms (Martin et al., 1994).

It has many advantages compared to traditional chemotherapeutic drugs, such as low toxicity, and therefore was found to be an ideal surgical adjuvant therapy for breast cancer patients as a potential anticarcinogenic drug (Qin et al., 2011; Camargo et al., 2012) and also protects against radiation-induced chromosomal damage and decline in the cell proliferation, as observed in mouse bone marrow cells (Jagetia et al., 2003; Camargo et al., 2012) and lomefloxacin-induced genomic instability (Jagetia & Reddy, 2002).

However, the available literature is scanty regarding the interaction of naringenin with intestinal epithelial cells and needs further exploration (Fasinu et al., 2013). Recently, much interest has been focused on developing in vitro toxicity tests evaluating their usefulness in predicting toxicities.

Although, a study on the in vitro cytotoxicity for various flavonoids in a number of cell lines (Kim et al., 1998) has reported that the IC₅₀ of naringenin for cell growth was greater than 1 mM for naringenin in the human hepatoma cell line HepG2, Macacus' rhesus monkey kidney cell line MA-104, and human lung cancer cell line A549, thereby indicating that naringenin has relatively low toxicity in cell culture, but this study is planned to investigate the higher concentrations (10 or 100 mM, respectively) of naringenin on normal epithelial cells of gastrointestinal regions.

To the best of our knowledge, naringenin has not previously been evaluated for safety through standard in-vitro gastrointestinal toxicological studies. Therefore, there is a pressing need to elucidate the safety profile of naringin or naringenin to be employed as a pharmaceutical adjuvant for oral dosage forms (Lassoued et al., 2011; Gurunath et al., 2013a,b) by inhibition of P-gp in the intestinal regions (Jejunum and Ileum; Tian et al., 2002; Cao et al., 2005; Attia, 2008) using standard assay procedures (Deferme et al., 2008; Shirasaka et al., 2008).

In this study, the investigation was conducted on the effects of naringenin, an active metabolite of naringin, at different concentration levels on the structure and function of rat intestinal segments, the uptake of end-product nutrients like d-glucose and l-histidine, the assessment of cell viability and the analysis of the intestinal brush border membrane intrinsic proteins, enzymatic activities of the intestinal marker enzymes, qualitative and quantitative histological changes by morphometric analysis.

The results of this study will provide an insight for further use of naringin or naringenin in clinical trials as a medication or as a food supplement for humans.

Materials and methods

Materials

Naringenin, Lucifer yellow, glucose, Na₂HPO_4, NaOH, KCl, NaCl, CaCl₂, MgCl₂ and NaHCO₃ NaH₂PO_4, EDTA (Ethylene diamine tetra-acetic acid), bovine serum albumin, dithiothreitol, sodium carbonate, l-histidine, Dimethyl sulphoxide (DMSO), formaldehyde, picric acid and acetic acid were obtained from Sigma Aldrich (Hyderabad, India). The purity of naringenin was authenticated on a Shimadzu HPLC system (LC-20A series) equipped with a UV detector, micro-volume double plunger pump (10 μl/stroke) and manual injector type model 7725-port sample injection valve. The mobile phase was prepared by a 60/40 (v/v) mixture of acetonitrile/water and the pH was adjusted to 3.2 using orthophosphoric acid. The injection volume was 20 μl. The UV detector was set at a wavelength of 285 nm. The HPLC purity of naringenin was determined to be ≥98% by external standard method. Glucose estimation and MTT assay kits from Hi-Media Lab, Mumbai (India). Alkaline phosphatase (ALP), lactate dehydrogenase (LDH), protein assay kits were purchased from Siemens Ltd. (Gujarat, India). All other chemicals reagents were of analytical grade. All drug solutions were freshly prepared before use.

Animals

Sprague–Dawley (SD) rats (250–270 g; 10–15 weeks old, n = 8) purchased from National Institute of Nutrition (NIN), Hyderabad. They were housed in large polypropylene cages with a 12-h light/12-h dark cycle with free access to water and maintained on commercial feed obtained locally. The rats were fasted overnight before experimentation and had access to water ad libitum. All the animal experiments were performed in accordance with the guidelines established by the committee for the purpose of control and supervision on experiments on animals (CPCSEA, New Delhi) and approved by our Institutional Animal Ethics Committee of Vaagdevi Institute of Pharmaceutical Sciences, Bollikunta, Warangal with registration number 1533/PO/a/11/CPCSEA. The study protocol was approved with the certification number as VIPS/IAEC/01/19.

Single-pass intestinal perfusion model (A)

Surgical procedures

Rats were anesthetized with pentobarbital sodium (20 mg/kg body weight i.p.). Abdomen was exposed through the mid-line incision.

After ligating the bile duct, the whole small intestine was isolated and gently an intestinal loop (10 cm in length) was made at two regions (upper jejunum and ileum) by cannulation with a silicone tube (3 mm i.d.), and then the intestinal contents were removed by slow infusion of Tyrode solution (37 °C) using a peristaltic pump (Pharmatek Scientific Systems, Hyderabad, India). The small intestine was returned to the abdominal cavity to maintain its viability without disrupting blood vessels.

The animal was maintained at 37 °C throughout the experiment using an overhead lamp. The exposed area was covered with gauze and saline (37 °C) was applied to keep it warm and moist during the experiment.

Following the above procedure, the blank (Tyrode solution, adjusted to pH 7.4) was perfused with an infusion pump at a flow rate of 0.1 ml/min. The intestinal preparation was flushed with a blank physiological solution (37 °C) until the effluent perfusates were clear. The physiological solution contains in mM: 15 Glucose, 11.90 NaHCO₃, 136.9 NaCl, 4.2 NaH₂PO₄, 2.7 KCl, 1.2 CaCl₂ and 0.5 MgCl₂; pH 7.4.

The intestinal segment was perfused with test solutions containing each compound (naringenin: 1 mM, 10 mM and 100 mM or sodium deoxycholate: 10 mM) about 30 min to achieve absorption equilibrium and stable outflow rates. Subsequently, sampling was made in pre-weighed 5-ml glass vials with a lid every 30 min for a perfusion period of 120 min after 30 min equilibration.

The segments of jejunum and ileum from each animal between two cannulas were excised without dragging, its length was measured using silk thread and subjected to further analyses as described below. Water flux was quantified based on direct measurement of the volume at the outlet (Li, 2005; Bermejo et al., 2004).

Preparation of rat intestinal epithelial cell membrane

The brush border membrane (BBM) was prepared with some modifications as described by Upreti et al. (2006). A known weight of each portion of the intestine (Jejunum and Ileum) was everted and the epithelial layer was scraped off with the help of a blunt scalpel.

The cell scraping was weighed, placed in 75 ml of 5.0 mmol/L EDTA buffer (7.4) and homogenized in ultra-sonicator (Siskin Instruments, Bangalore, India). The homogenate was centrifuged at 3000g for 15 min at 4 °C. The sediment was washed thrice with 5 ml of EDTA buffer. The washed sediment of the crude brush border fraction was then suspended in 2 ml of 90 mmol/L sodium chloride and 0.8 mmol/L EDTA buffer, mixed thoroughly and kept aside for 20–30 min until a well-defined sediment was formed.

The supernatant and sediment were passed through a pad of glass wool successively to remove aggregated particles. The glass-wool pad was washed with a further 20 mL of 5 mmol/L EDTA buffer. The brush borders from the total washings were sedimented by centrifugation at 1000g for 15 min.

The sediment was once again washed with 2.5 ml of 2.5 mmol/L EDTA buffer. The final pellet was suspended in a small volume of 2.5 mmol/L EDTA buffer and used for the estimation of enzymatic and biochemical studies.

Effect of naringenin on protein and cytosolic release of enzymes (enzyme assays)

The liberation of the enzymes such as LDH (lactate dehydrogenase), alkaline phosphatase (ALP) and protein release was determined as reported previously (Gao et al, 2008a,b) and as described by the manufacturer's protocols.

The release of enzymes is an indicator (Li et al., 2011) to determine the degree of cell damage caused to the intestinal membranes. LDH activity was determined using LDH kit. The assay is based on the conversion of lactate to pyruvate in the presence of LDH with parallel reduction of NAD.

The formation of NADH from the above reaction results in a change in absorbance at 340 nm. The results were calculated as U/L/cm² of sac area (1 unit reduces 10⁻⁶ mol pyruvate per min at pH of 7.4). The activity of alkaline phosphatase was assayed using Actopack kit from Siemens Ltd. (India) according to the method of Bergmeyer et al. (1978) where p-nitrophenyl phosphate was used as the substrate, which was hydrolyzed by the enzyme to yield p-nitrophenol to be measured at 405 nm.

Protein release was determined according to Lowry et al. (1951) using bovine serum albumin as standard. The amount of protein released was measured using a protein assay kit. Enzyme units were defined as micromoles of the products formed or liberated per minute under the assay conditions. Specific activity was expressed as U/L/cm² of sac area.

Isolation of rat intestinal epithelial cells (B)

Intestinal epithelial cells were prepared by the method as described by Weiser (Weiser, 1973; Fasco et al., 1993; Zhang et al., 1999; Mac Donal et al., 2008). In brief, the small intestines were flushed gently with normal saline containing 1.0 mmol/L dithiothreitol. The cecal end of the intestine was ligated and solution “A” containing 1.5 mmol/L KCl, 96 mmol/L NaCl, 27 mmol/L sodium citrate, 8 mmol/L KH₂PO₄, 5.6 mmol/L Na₂HPO₄; pH 7.4 was filled and the other end, i.e. the proximal part of jejunum was clamped with artery forceps.

The intestine was immersed in solution “A” and incubated at 37 °C for 30 min in a constant temperature shaker bath. The intestine was emptied; fluid was discarded and filled with solution “B” containing 1.5 mmol/L EDTA and 0.5 mmol/L dithiothreitol in buffer solution (pH 7.4) and immersed in solution “A” for incubation. After incubation, the contents were emptied at different time points of incubation (5, 10, 15 and 20 min, respectively) into a plastic centrifuge tube to recover the first epithelial cell population. Fractions 1–4 were pooled and designated as “Upper Villus”.

Effects of naringenin on viability of isolated intestinal epithelial cells

A single cell suspension of the intestinal epithelial cells (2 × 10⁶ cells/mL) was set up in triplicate with or without naringenin or sodium deoxycholate of 10 µL containing 1, 10 and 100 mM (naringenin) or 10 mM (sodium deoxycholate) in 96-well plates and incubated for 24 h at 37 °C. After incubation 20 µL of MTT (5 mg/mL PBS) was added to each well. Following incubation for 3 h at 37 °C the medium was aspirated and 200 µL of DMSO was added to all wells and mixed thoroughly.

After a few minutes the plates were read on a Micro-Elisa reader at the wavelength of 570 nm (Mosmann, 1983; Wilson, 2000; Berridge et al., 2005). Cell viability in response to treatment with naringin was calculated as: % of dead cells = 100 − (OD treated/OD control) × 100.

Effect of naringenin on the transport of glucose and histidine in the intestinal sacs (C)

Preparation of everted intestinal sacs

After dissecting the rat, the abdomen was exposed by a midline incision. Intestinal segments were isolated and each site was defined as described below. First, 5 cm of the top of the small intestine was regarded as the duodenum.

The jejunum was obtained from the next 10 cm portion of the small intestine. The ileum was obtained from the final 10 cm portion of the intestine. However, jejunum and ileal segments were selected for this study. Segments were flushed with physiological solution for cleaning.

The everted sac was prepared by inserting a narrow glass rod with thickening at one end into the lumen and eversion of the gut was done carefully by reaching to the end and folded over it carefully and then pushing the rod through the everted end of the emerging intestinal segment. One end of the segment was clamped and tied with a silk thread forming a sac, while the other end was attached to a sampler, after the blank solution [1 ml, Tyrode solution; (containing in mM: 15 glucose, 11.90 NaHCO₃, 136.9 NaCl, 4.2 NaH₂PO₄, 2.7 KCl, 1.2 CaCl₂ and 0.5 MgCl₂), pH 7.4] was introduced into the everted sac (serosal side), a 10-cm-long everted gut sac. Care was taken that both ends remain firm and tight enough to prevent any leakage, but not too tight to damage the tissue.

The sacs were placed in 15 ml degassed oxygenated Tyrode solution in an oscillating water bath at 37 °C in 20 ml capacity test tubes. One ml each of four samples – control (without naringenin) and naringenin (1, 10 or 100 mM) along with glucose (10 mg/ml) or histidine (10 mg/ml) solution were placed in respective test tubes to ensure full exposure of the whole sac to naringin and maintained at 37 °C using shaking water bath.

At each time point, i.e. 30, 60, 90 and 120 min, the transport of glucose/histidine from mucosa to serosa in each group was measured by sampling 100 μl of serosal medium. All experiments were carried out in triplicate.

l-Histidine uptake

Histidine uptake studies were carried out as described by Plummer (Plummer, 1988). 100 μl aliquots were taken at predetermined intervals at every 30 min up to 120 min from the mucosal and serosal sides. A weak acetic acid solution (0.5 mol/L) was added to deproteinize the above solutions from mucosal and serosal regions.

The tubes were covered with aluminium foil and placed in boiling water bath for 10 min. After cooling, an aliquot of the deproteinized solution taken and to it added sodium nitrite (50 g/L prepared fresh).

The tubes were left for 5 min with occasional stirring and added sodium carbonate solution (75 g/L) with vigorous shaking. To it was added ethanol and distilled water, mixed thoroughly and after 30 min, the optical density (OD) was read at 498 nm.

d-Glucose uptake

The viability and the integrity of the gut sacs were further demonstrated by analyzing the glucose concentrations both in the mucosal and serosal sides. Viable enterocytes transport glucose against a concentration gradient, so in a non-leaking metabolically active membrane it should be possible to measure a glucose gradient between the external medium and the serosal fluid (Barthe et al., 1999; Ballent et al., 2006). Samples of incubation medium were collected at predetermined times as stated above and glucose concentrations were measured using a commercially available glucose estimation kit.

Effect of naringin on the transport of Lucifer yellow

Lucifer yellow was chosen as a fluorescent hydrophilic paracelluar marker to evaluate epithelial cell tightness in this experiment (Lacombe et al., 2004; Konsoula & Barile, 2007).

Fluorescent indicators were used at concentrations of 1 mg/ml in Tyrode solution and applied to the apical side of the gut sacs. At the end of incubation, the baso-lateral medium was collected and fluorescence intensity was measured by a direct spectrofluorimetric method with the BioTek FL600 fluorescence microplate reader and quantified with a calibration curve prepared in a buffer solution (pH 6.8).

Mucosal (1 mL) and serosal (100 μl) samples were centrifuged for 15 min at 5000 × g to precipitate mucus and other solid matter and were then diluted suitably. The excitation and emission wavelengths for Lucifer yellow were 418 and 512 nm, respectively. Relative cell permeability was expressed as a percent of untreated control groups. Experimental and process (blank) controls were monitored simultaneously.

Histopathological studies (D)

Small intestinal pieces obtained after in situ intestinal studies were washed in ice cold saline and fixed in Bouin's fixative (picric acid, glacial acetic acid and formaldehyde) for 24 h. After fixation, the tissues were carefully processed in alcohol, xylene and embedded in paraffin wax (58–60 °C).

Paraffin sections were cut at 5–7 µm thickness and then subjected to Delafield hematoxylin–eosin histological staining technique. Microscopic examinations were performed on all tissues from all animals in the control and naringenin treatment groups.

If treatment-related changes were noted in a particular organ or tissue in 100 mM narigenin treatment, extended examination was conducted on the corresponding tissues from lower concentration groups. The resulting slides were viewed independently by two observers, at a magnification of 10× and 250×, respectively.

A morphological score was designed to compare the histological lesions (Table 1). The scoring system included both morphometry and lesional informations. The following criteria were included in the score: the number of villi and the number of crypts (0–3 points each), the length of the villi, the morphology of enterocytes, and the degrees of villi coalescence and autolytic changes of the tissue (edema, necrotic debris and apoptotic cells).

Table 1. Endpoints used to assess sections histologically as a morphological score (maximal score of 15 was used for control).

	Endpoint	Score
Number of villi (No. per 500 µm)	0	0
	<5	1
	5–10	2
	>10	3
Number of crypts (No. per 450 µm)	0	0
	<5	1
	5–10	2
	>10	3
Villi length (µm)	0	0
	+	1
	++	2
	+++	3
Enterocytes morphology	No enterocytes	0
	+ flattened epithelium	1
	++ Cuboid epithelium	2
	+++ Columnar epithelium	3
Lesions of the tissue	Lysis, necrosis	0
	Coalescence of villi, odema, apoptosis	1
	Weak coalescence, weak odema, apoptosis	2
	No lesion, slight flattening of the villi	3

Each score value was the result of 2–4 intestinal sections from the same rat at each naringenin concentration. In addition, the length of the villi was measured in all the sections scored and the result for each section was the mean of all the villi measured (Olympus Image Analysis Software, Olympus India).

Data analysis

Data analysis was computed using GraphPad Prism 5 software (GraphPad, San Diego, CA). The results were represented as mean ± standard deviation (±SD) for all the experiments. The apparent permeability average values (P_app) were compared for each sample using one way analysis of variance (ANOVA) test followed by Dunnett's test or Bonferroni test for multiple comparisons. The difference was considered significant at p ≤ 0.05. D'Agostino and the Pearson omnibus normality test was used to assess the homogeneity of variances in data. If the variance was not homogeneous, the Kruskal–Wallis test was applied.

Results

Effect of naringenin on protein and cytosolic release of enzymes (A)

We examined the effect of naringenin on intestinal membrane toxicity by evaluating the total amount of LDH, ALP and protein released from the intestinal membranes with or without naringenin.

Table 2 shows the effects of naringenin at different concentrations on the various intestinal marker enzymes in the BBM isolated from jejunum and ileum segments of the rat intestine. From the results in Table 2, it can be seen that naringenin showed no significant changes in the activity of alkaline phosphatase, lactate dehydrogenase and protein release from the intestinal tissues.

Table 2. The effect of naringenin on the rat small intestinal brush border membrane enzymes.

Enzyme activity	Intestinal segment	30	60	90	120
Negative control
ALP (U/L/cm^2)	Jejunum	17.35 ± 0.49	18.54 ± 0.74	18.72 ± 0.44	19.91 ± 0.97
	Ileum	16.45 ± 0.38	16.53 ± 0.19	16.42 ± 0.43	16.91 ± 0.31
	LDH (U/L/cm^2)	Jejunum	17.49 ± 0.54	18.34 ± 0.46	19.02 ± 0.56	19.45 ± 1.13
Ileum		15.12 ± 0.24	15.04 ± 0.15	15.02 ± 0.38	14.45 ± 0.56
Positive control (Sodium deoxycholate, 10 mM)
ALP (U/L/cm^2)	Jejunum	37.15 ± 0.29^a	38.34 ± 0.54^a	48.12 ± 0.24^a	47.12 ± 0.17^a
	Ileum	35.85 ± 0.48^a	36.13 ± 0.39^a	47.52 ± 0.23^a	46.71 ± 0.71^a
	LDH (U/L/cm^2)	Jejunum	48.39 ± 0.64^a	48.44 ± 0.56^a	49.32 ± 0.16^a	48.75 ± 1.53^a
Ileum		45.32 ± 0.54^a	45.14 ± 0.35^a	45.22 ± 0.48^a	44.25 ± 0.16^a
Naringenin (1 mM)
ALP (U/L/cm^2)	Jejunum	16.67 ± 0.43^bc	17.06 ± 0.42^bc	18.67 ± 0.52^bc	17.63 ± 0.43^bc
	Ileum	16.87 ± 0.51^bc	16.06 ± 0.13^bc	16.17 ± 0.34^bc	17.03 ± 0.23^bc
	LDH (U/L/cm^2)	Jejunum	17.07 ± 0.42^bc	18.13 ± 0.51^bc	18.81 ± 0.38^bc	20.02 ± 1.12^bc
Ileum		14.97 ± 0.28^bc	14.83 ± 0.31^bc*	14.88 ± 0.27^bc	14.52 ± 0.34^bc
Naringenin (10 mM)
ALP (U/L/cm^2)	Jejunum	17.58 ± 0.43^bc	18.54 ± 0.38^bc	17.91 ± 0.31^bc	18.15 ± 0.43^bc
	Ileum	16.12 ± 0.34^bc	16.34 ± 0.36^bc	16.91 ± 0.45^bc	17.15 ± 0.41^bc
	LDH (U/L/cm^2)	Jejunum	18.34 ± 0.37^bc	18.76 ± 0.41^bc	19.6 ± 0.61^bc	17.39 ± 0.56^bc
Ileum		14.54 ± 0.27^bc	14.76 ± 0.41^bc	15.06 ± 0.46^bc	15.12 ± 0.46^bc
Naringenin (100 mM)
ALP (U/L/cm^2)	Jejunum	18.06 ± 0.54^bc	18.54 ± 0.42^bc	18.91 ± 0.57^bc	18.63 ± 0.46^bc
	Ileum	16.03 ± 0.21^bc	16.24 ± 0.31^bc	17.11 ± 0.43^bc	16.43 ± 0.37^bc
	LDH (U/L/cm^2)	Jejunum	17.92 ± 0.42^bc	19.18 ± 0.51^bc	18.23 ± 0.31^bc	18.66 ± 0.39^bc
Ileum		14.92 ± 0.31^bc	14.68 ± 0.37^bc	14.23 ± 0.31^bc	15.66 ± 0.35^bc

Values are mean ± SD for eight rats in each group.

^ap < 0.01 showing statistically significant when compared with negative control.

^bp > 0.05 showing statistically insignificant when compared with negative control.

^cp < 0.01 showing statistically significant when compared with the positive control.

LDH activity, at 30 min in the incubation media of jejunal and ileal segments with naringenin (1, 10 or 100 mM) was 17.07 ± 0.42; 14.97 ± 0.28, 18.34 ± 0.37; 14.54 ± 0.27, 17.92 ± 0.42; 14.92 ± 0.31 U/L/cm², respectively, and was not significantly different from the LDH release at 120 min [p = 0.4147 (Jejunum); 0.9359 (ileum)].

Furthermore, the ALP activity in jejunal and ileal segments was 16.67 ± 0.43; 16.87 ± 0.51, 17.58 ± 0.43; 16.12 ± 0.34, 18.06 ± 0.54; 16.03 ± 0.21 U/L/cm², respectively, at 30 min and no significant change (p = 0.1544; 0.0930) was noticed at 120 min at 1 or 10 or 100 mM of naringenin concentrations.

As shown in Figure 1(a and b), naringenin did not produce any effect on the release of total protein content from the segments of intestines [p = 0.3503 (jejunum); 0.1704 (ileum)] as compared to control at 120 min. However, significant changes were noticed with sodium deoxycholate in the activity of LDH, ALP or protein release at 10 mM concentration.

Figure 1. The effect of naringenin on the rat small intestinal brush border membrane protein release. (A) Jejunum. Values are mean ± SD for eight rats in each group. ^#Statistically insignificant compared to control (p = 0.9860); *(p < 0.05) when compared to sodium deoxycholate. (B) Ileum. Values are mean ± SD for eight rats in each group. ^#Statistically insignificant compared to control (p = 0.4914); *(p < 0.05) when compared to sodium deoxycholate.

Effect of naringenin on cell viability of intestinal epithelial cells (B)

The percentage of cell viability of epithelial cells following MTT assay revealed a similar concentration-dependent cell death percentage for both control and naringenin groups.

Figure 2 shows a comparison of cell death percentages between control and naringenin treated cells that incubated for 24 h after plating. At no time, the numbers of cells were different between the two groups (control and naringenin at all tested concentrations). A p value was more than 0.05 (p = 0.0782).

Figure 2. Cell viability of epithelial cells following 24 h in vitro exposure to naringenin. Values are mean ± SD for eight rats in each group. ^#Statistically insignificant compared to control (p = 0.9743); *(p < 0.05) when compared to sodium deoxycholate.

Thus, we found that the percentage of cell viability was not significantly different with naringenin at all tested concentrations when compared with control. This was true even at the highest concentration of naringenin (100 mM).

From the results, it can be noticed that naringenin had no effect on cell death in view of the fact that it possess a protective outcome (Conseil et al., 1998; Bansal et al., 2009) in reference to the control as previously reported, even at higher tested (10 or 100 mM) concentrations. In contrast, sodium deoxycholate significantly reduced the cell viability when compared to control and naringenin tested groups.

Uptake of d-glucose and l-histidine (C)

The ability of everted sacs to concentrate glucose and amino acids (Barthe et al., 1998; Lodish et al., 2000; Bröer, 2008) by active transport across the intestine wall was used as an indicator of tissue viability.

Both in the absence and in the presence of the P-gp-modulating compound (naringenin), the ratios between the glucose/histidine concentration measured in the sac content and incubation medium increased gradually with time up to 120 min of incubation (Table 3A and B) which indicates an adequate viability of the everted intestinal tissue.

Table 3. Effect of naringenin concentration on d-glucose and l-Histidine transport by small intestinal segments in control and naringenin treated rats.

Sample	AP -BL	Fold increase in AP–BL	BL–AP	% decrease in BL–AP	Efflux ratio BL-AP/AP–BL
(A) Transport of glucose across the intestinal segments using everted gut sac studies.
Jejunum
Control	8.06	0	6.19	0	0.76
Naringenin (1mM)	15.4	1.91	13.3	53.4	0.86^a
Naringenin (10mM)	16.2	2	14.4	57	0.88^a
Naringenin (100mM)	16.8	2.08	15.8	60	0.94^a
Ileum
Control	7.15	0	5.43	0	0.75
Naringenin (1 mM)	12.14	1.69	10.34	47.4	0.85^a
Naringenin (10 mM)	13.65	1.9	12.28	55.7	0.89^a
Naringenin (100 mM)	14.4	2.01	13.3	59.1	0.92^a
(B) Transport of histidine across the intestinal segments using everted gut sac studies.
Jejunum
Control	2.88	0	1.68	0	0.56
Naringenin (1 mM)	2.79	0.96	1.78	5.6	0.63^a
Naringenin (10 mM)	3.05	1.05	2.36	28.8	0.77^a
Naringenin (100 mM)	3.24	1.12	2.88	41.6	0.89^a
Ileum
Control	2.09	0	1.33	0	0.636
Naringenin (1 mM)	2.34	1.11	1.78	25.2	0.76^a
Naringenin (10 mM)	2.66	1.27	2.13	37.5	0.80^a
Naringenin (100 mM)	2.83	1.35	2.3	42.1	0.81^a

Values are mean ± SD for eight rats in each group.

^aStatistically significant compared to control (p < 0.05).

Uptake of glucose or histidine in jejunal sacs

The integrity of the sac (jejunum) was confirmed by the transport of glucose and histidine across the membrane. AP–BL (apical to basal) permeability was found to be 15.7, 12.5 and 6.3% more than BL–AP (basal to apical) permeability across the jejunal membrane at various concentrations of (1 or 10 or 100 mM) naringenin, respectively, for glucose molecules, while AP–BL permeability was found to be 36, 22.6 and 11.1% more than BL–AP permeability for histidine at all dose levels of naringin, respectively

P-gp inhibition by naringenin (1 or 10 or 100 mM) significantly increased AP–BL permeability (p < 0.01), and decreased BL–AP permeability (p < 0.01) for both glucose and histidine molecules across the jejunal segments in a dose-dependant manner.

AQ and SQ (calculated using AP–BL and BL–AP permeabilities in the presence of naringenin at all concentrations of glucose/histidine in the absorptive and secretory directions, respectively) indicated that about 1.91- (1 mM), 2- (10 mM) and 2.08 (100 mM)-fold diminished in glucose transport by P-gp-mediated efflux in A–B direction as compared to control, and P-gp enhanced secretory permeability by 2.19-, 2.37- and 2.06-fold respectively for all concentrations of naringenin.

For histidine concentrations across jejunal membranes, AQ indicated a reduction of 1.05- and 1.12-fold for 10 and 100 mM of naringenin, respectively, whereas no change in AQ was noticed with 1 mM of naringenin as compared to control. Nevertheless, P-gp enhanced secretory permeability by 1.09-, 1.44- and 1.76-fold for all naringenin concentrations of 1, 10 and 100 mM, respectively.

Increase in naringenin concentrations diminished polarized transport by further increasing the absorptive transport and reducing the secretory transport. Significant differences were found in the glucose/histidine transport between naringenin treated and control groups across the jejunal membranes (p < 0.01). The efflux ratios (BL–AP/AP–BL) of glucose and histidine in the absence and presence of naringenin are illustrated in Table 3.

It can be noticed that the ratios were gradually increased and reached a factor of about 1.12, 1.16 and 1.22 at various concentrations naringenin (1 or 10 or 100 mM) for glucose levels over control while the ratios of histidine concentrations reached a factor of about 1.12, 1.16 and 1.2 in jejunum at 120 min, respectively.

Uptake of glucose or histidine in ileal sacs

Ileum sac integrity was also confirmed by the transport of glucose and histidine across the membrane.

AP–BL permeability was found to be 14.8, 10 and 7.6% more than BL–AP permeability across the ileal membranes at various concentrations of naringenin (1 or 10 or 100 mM), respectively, for glucose molecules, while AP–BL permeability was found to be 23.9, 19.9 and 18.7% more than BL–AP permeability for histidine at all concentrations of naringenin, respectively.

P-gp inhibition by naringenin (1 or 10 or 100 mM) significantly increased AP–BL permeability (p < 0.01), and decreased BL–AP permeability (p < 0.01) for both glucose and histidine molecules across the ileal segments in a dose-dependent manner.

AQ and SQ (calculated using AP–BL and BL–AP permeabilities in the presence of naringenin at all concentrations of glucose/histidine in the absorptive and secretory directions, respectively) indicated that about 1.69- (1 mM), 1.9- (10 mM) and 2.01 (100 mM)-fold diminished in glucose transport by P-gp-mediated efflux in A–B direction as compared to control, and P-gp-enhanced secretory permeability by 1.9-, 2.26- and 2.4-fold, respectively, for all concentrations of naringenin.

For histidine concentrations across ileal membranes, AQ indicated a reduction of 1.11-, 1.27- and 1.35-fold for 1, 10 and 100 mM of naringenin, respectively, as compared to control. Nevertheless, P-gp-enhanced secretory permeability by 1.33-, 1.60- and 1.72-fold for all naringenin concentrations of 1 or 10 or 100 mM, respectively. The efflux ratios (BL–AP/AP–BL) of glucose and histidine in the absence and presence of naringenin are illustrated in Table 2.

It can be noticed that the ratios were gradually increased and reached a factor of about 1.12, 1.17 and 1.21 at various dose levels of naringenin (1 or 10 or 100 mM) for glucose concentrations over control while the ratios of histidine concentrations reached a factor of about 1.19, 1.25 and 1.27 in ileum at 120 min, respectively. In addition, a significant differences were found in the glucose/histidine transport between naringenin treated and control groups across the ileal membranes (p < 0.01).

Effect of naringenin on the transport of lucifer yellow across the intestinal membranes (D)

In addition, Lucifer yellow, a membrane integrity marker, is co-incubated to investigate whether naringenin can alter the membrane integrity via the paracelluar route using everted gut sacs.

Lucifer yellow is the marker used to study the paracelluar absorption along the small intestine. No significant effect on the permeability of Lucifer yellow was observed in the presence and absence of naringenin at 1 or 10 or 100 mM (9.54 ± 0.6 × 10⁻⁶ cm/s, 9.65 ± 0.6 × 10⁻⁶ cm/s, 9.65 ± 0.03 × 10⁻⁶ cm/s and 9.61 ± 0.6 × 10⁻⁶, respectively, p = 0.052).

Histological analysis (E)

The morphological changes induced in the different segments of the intestine following naringenin treatment can be measured by qualitative and quantitative histological analysis. The changes that take place in the abnormal surface can be correlated well with the changes in the absorption and transport of various nutrients.

Table 4 shows the results in villi number, villi height, crypt number, enterocytes morphology and lesions in the intestinal segments upon treatment with the naringenin (1 or 10 or 100 mM) or sodium deoxycholate (10 mM).

Table 4. The effect of naringenin on rat small intestinal villi number, villi height, crypt number, enterocyte morphology and lesions by micrometric analysis of the histological slides.

Micrometer parameter	Intestinal segment	Negative control	Sodium deoxycholate (10 mM)	Naringenin (1 mM)	Naringenin (10 mM)	Naringenin (10 mM)
Villi number (No. per 500 µm)	Jejunum	3	0	3	3.0	3.0
	Ileum	3	1	3	2.9	2.9
	Villi height (µm)	Jejunum	3	0	3	3.0	3.0
Ileum		3	1	3	3.0	3.0
Crypt number (No per 450 µm)		Jejunum	3	0	3	3.0	3.0
	Ileum	3	0	3	3.0	3.0
	Enterocytes morphology	Jejunum	3	1	3	2.9	2.9
	Ileum	3	1	3	3.0	3.0
Lesions	Jejunum	3	1	3	2.9	2.9
	Ileum	3	1	3	3.0	3.0
Score	Jejunum	15	02	15^ab	14.9^ab	14.9^ab
	Ileum	15	05	15^ab	14.9^ab	14.9^ab

Values are mean ± SD for eight rats in each group.

^aStatistically insignificant compared to control (p > 0.05).

^b(p < 0.05) when compared with naringenin positive control.

No significant decrease in villi number and villi height due to the naringenin treatment as compared to the control was noticed including the crypt number, which did not register any significant changes.

Figure 3(A–D) shows the intact columnar cell lining along with the lamina propria without shrinking of the villi. The crypts of Liberkühn were not disorganized in the naringenin groups. Along with the intact columnar cell lining, the intact tips of the villi cells were also observed. No pyknotic nuclei in enterocytes, flattening and coalescent villi, edema and necrosis in the lamina propria of intestinal epithelium were noticed in control and naringenin treated groups.

Figure 3. Micrograph of control of small intestine (LS) H & E 10×. (A) Showing the normal histoarchitecture of villi and globet cells. (B) Naringenin (1 mM) showing numerous globet cells along with villi and normal surface characteristics. (C) Naringenin (10 mM) showing numerous globet cells along with villi and normal surface characteristics. (D) Naringenin (100 mM) showing normal morphology of columnar epithelium along with intact brush border and lamina propria. Micrograph of small intestine (LS) H & E 250×. (E) Sodium deoxycholate (10 mM) showing disorganized or damaged histoarchitecture of villous and globet cells. (F) Sodium deoxycholate (10 mM) showing disorganized globet cells with damaged morphology of columnar epithelium along the brush border.

No decrease is observed in the villi number, villi height and no disorganization of the villi surfaces, reflects the constant absorption and transport of various nutrients (glucose and histidine) following the naringenin treatment over a period of 120 min.

Figure 3 (E and F) shows significant changes with sodium deoxycholate in the villi number, villi height and disorganization of columnar lining along the lamina propria suggesting its cytotoxic toxic action at tested concentration.

Discussion

In view of the wide distribution in plant foods, broad spectrum of pharmacological activities, potential health benefits and the lack of toxicological properties, there is a pressing need to elucidate the in vitro toxicological profile of naringenin. Therefore, in our study, we presented a comprehensive safety evaluation on naringenin by performing short-term in vitro gastrointestinal toxicological studies in Sprague–Dawley rats. Following the exposure to any chemical agent, the cell membrane is the first site of interaction as the uptake and exchange of different ions take place through the intestinal mucosa.

Inhibition of membrane transport enzymes like alkaline phosphatase and Ca²⁺–Mg²⁺–ATPase may influence the transport of PO⁴⁻, Ca²⁺, Mg²⁺ and other vital ions in the intestine. In this study, these membrane marker enzymes were analyzed and compared between the intestinal epithelial cells following in vitro naringenin exposure. We found that naringenin was harmless with no change in the protein release, ALP and LDH leakages from the intestinal membranes as compared to sodium deoxycholate, which showed marked damage of intestinal membranes due to significant release of protein, ALP or LDH leakages.

The release of protein is a good index to evaluate the intestinal membrane toxicity, since many toxic substances such as sodium deoxycholate could stimulate the release of protein, one of the most important components of biological membranes from the intestinal membranes (Swenson et al., 1994; Yamamoto et al., 1996). LDH and ALP, marker enzymes of the brush border membrane, are also used as a biological marker for evaluating the membrane toxicity (Yamamoto et al., 1996; Gao et al., 2008a,b).

As shown in Table 1 and Figure 1, naringenin did not cause any serious membrane damage or toxicity to the intestinal mucosa at tested concentrations used in this study; however, it negligibly increased the release of protein from the intestinal membrane at 100 mM. At all tested concentrations (1 or 10 or 100 mM), no significant (p > 0.05) change in protein release were noticed between three different concentrations of naringenin, respectively. This suggests that naringenin possess no deleterious effects on GI tract, further indicating the normal integrity of intestinal membranes.

Such effect of enzyme activity might have resulted due to either no substrate affinity (kinetic effect) or no modulation of protein molecular number or activity (metabolic effect), which needs further experimentation to elucidate the mechanism.

The viability pattern of the intestinal epithelial cells revealed no concentration dependent cellular deaths, indicating significant harmless effects of naringenin on intestinal cell viability, as measured by the MTT assay. However, sodium deoxycholate proved to be cytotoxic at the tested concentration. This is also supported by some evidence that sodium deoxycholate may act as a carcinogen not only in human colon, but also in other segments of the gastrointestinal (GI) tract (Burnat et al., 2010; Bernstein et al., 2011). A variety of mechanisms are believed to contribute to the deleterious effects of sodium deoxycholate on the GI tract. Sodium deoxycholate was found to perturb membrane structures by alteration of membrane microdomains (Jean-Louis et al., 2006).

Consequently, the addition of naringenin showed a significant effect on the glucose and histidine gradient over a period of 120 min with gradual increase in their transport across the epithelial cell membrane.

As these substrate molecules were actively taken up in the small intestine, intact and metabolically active sacs will maintain their (glucose and histidine) gradient between the external medium (mucosal) and serosal fluid, as similar to our results demonstrated.

Our results elucidated that the tissue of gut sac was viable and well functioning, even at 120 min, thus clarifying the active transport of the substrate molecules to supply metabolic energy and obviously, if the sacs were not biochemically active, or if they were not physically intact, such a concentration gradient would not be maintained (Lehmann et al., 1981).

The movement of the substrate molecules, particularly across the epithelial absorptive cells depends upon the carrier molecules (transport proteins; Eilam & Stein, 1974).

Such mediated transport depends on the coupled movement of Na⁺ across the membrane using the steep concentration gradient of the ion and thereby called the “secondary active transport system” (Boyd & Lund, 1981).

Thus, the increase in the uptake of nutrients after naringenin treatment could be attributed to the changes in absorption affinity of the transport protein, or also due to the contribution of efflux transporters such as P-gp in the intestine (Seo et al., 2009).

The obtained results for the transport of Lucifer yellow suggest that the paracelluar route, which is a passive diffusion mechanism for small hydrophilic molecules, was intact throughout the small intestine when naringenin was applied at all tested concentrations.

Moreover, the result from the unchanged permeability of Lucifer yellow with narigenin indicates that the tight junctions and integrity of intestinal epithelium remained intact over the entire period of incubation for 120 min.

Qualitative histopathological analysis of the small intestine in control group showed the normal characteristics of the epithelium of villi and crypts, with very small number of cells in lamina propria.

Therefore, this sample may be considered as the normal sample of small intestine. The same findings were observed in the sample upon naringenin treatment and observed changes were not related to concentration-dependent at all tested concentrations.

In the histopathological examination, no lesions in the intestines were noted. However, all of those pathological changes were sporadically detected in controls and the rats administrated with 1 or 10 or 100 mM of naringenin and no consistent histopathological to be spontaneous and/or incidental in nature but not relevant to naringenin treatment.

To the best of our knowledge, this study is the first published report involving rat intestinal epithelium for short-term gastrointestinal toxicity studies using naringenin.

The various tested concentrations of naringenin selected in this study were based on our previously published reports (Gurunath et al., 2013a, 2013b) wherein we demonstrated that naringin at 15 mg/kg was noticed to significantly inhibit the P-gp in the intestine, thus facilitating the intestinal absorption and oral bioavailability of P-gp substrate, when formulated in solid dispersions. This was also supported by other authors using naringenin as P-gp inhibitor in their research work, although with a different P-gp substrate (Surya Sandeep et al., 2013).

Taking into consideration the above information, we were further convinced and encouraged to assess the short-term gastrointestinal in vitro toxicity studies at cellular level, as cell membrane is the first site of interaction of intestinal mucosa following the exposure to any chemical agent subsequent to oral drug administration or ingestion of food.

Hence, the two concentrations (10 or 100 mM), well above 1 mM were selected for the study to examine their effects on the structure and function of rat intestinal segment, the uptake of end-product nutrients like d-glucose and l-histidine, the assessment of cell viability and the analysis of the intestinal brush border membrane intrinsic proteins, enzymatic activities of the intestinal marker enzymes, qualitative and quantitative histological changes by morphometric analysis.

In view of our above findings, it can be indicated that naringenin at all tested concentrations can safely be employed as a potential adjuvant for oral dosage forms of P-gp substrates belonging to Biopharmaceutic Classification System (BCS) II and III drugs for improving oral bioavailability as a P-gp inhibitor.

Our findings stand out as a novel strategy for employing naringin or naringenin as a safe and naturally occurring pharmaceutical adjuvant in the preparation of oral dosage forms as a promising alternative to other techniques of enhancing intestinal absorption and oral bioavailability. Since no in vitro toxicological assessment of naringenin on intestinal epithelium have been reported until today, although further studies are required for comprehensive and multi-disciplinary claims of naringin or naringenin such as their ability to exert various physiologic actions, and the fact that they are regarded as herbal extracted chemicals with non-uniform standards and are not approved for commercial pharmaceutical use by the US FDA, although the flavonoids are commercially available.

Conclusion

Naringin or its metabolite, naringenin, has already been classified as relatively harmless or non-toxic substance based on the classification of the relative toxicity of chemicals. Our comprehensive studies involving preservation of intestine mucosal biochemical composition, structure and function, further clarified naringenin as a safe drug. In our toxicity studies, test concentrations of 1 mM or 10 mM or 100 mM of naringenin were well tolerated and did not cause any toxic injury such LDH and ALP leakages from rat intestines. The above findings clearly demonstrate that naturally occurring naringin or naringenin could serve as safe and novel pharmaceutical adjuvant in the formulation of oral dosage forms.