Pennyroyal and gastrointestinal cells: multi-target protection of phenolic compounds against t-BHP-induced toxicity

Abstract

Herbal teas are consumed for their valuable content of bioactive phytochemicals. This work represents the first attempt to establish a linkage between the chemical composition of pennyroyal (Mentha pulegium L.) infusion and its potential to prevent cellular oxidative stress. The phenolic profile was established by HPLC-DAD-ESI/MSⁿ, twelve out of fifteen compounds being identified for the first time in this species. The infusion presented a phenolic content of 122.92 mg g⁻¹ (lyophilized extract) and a remarkable antiradical activity. A reduced glutathione detoxification mechanism was demonstrated to be involved in human epithelial colorectal adenocarcinoma (Caco-2) cell resistance against tert-butyl hydroperoxide-induced toxicity, while the same was not observed for human epithelial gastric adenocarcinoma (AGS) cells. For the first time, the presence of these phenolic compounds inside the cells was confirmed by 2-aminoethyl diphenylborinate (DPBA) phenol fluorescence dye staining, suggesting a direct antioxidant effect. M. pulegium infusion seems to provide bioactive compounds able to maintain a proper antioxidant balance in gastrointestinal cells.

---

Introduction

Natural products are still a valuable source of diverse bioactive compounds, many of them capable of remarkable pharmacological activities, being directly or indirectly involved in the design of novel drugs.¹ New approaches to Traditional Medicine, based on reverse-pharmacology, seem to allow the perfect linkage between popular knowledge, science and the positive health outcomes for consumers.²

Mentha pulegium L., commonly known as pennyroyal, is an herbaceous species within the Lamiaceae family and, similarly to the other mints, is strongly aromatic with a characteristic menthol flavor, being used in various culinary preparations.³ In phytotherapy, the infusion of the flowered aerial parts and the essential oil are used to treat and prevent diverse gastrointestinal disorders like lack of appetite, intestinal cramps, low intestinal motility, flatulence and dyspepsia, and to alleviate cold and flu related symptoms.⁴ However, the possible mechanisms underlying M. pulegium therapeutic applications continue to be unknown.

Different activities, such as insecticidal,⁵ neuroprotective,⁶ anti-acetylcholinesterase,⁷ antioxidant and anti-genotoxic,^8,9 have already been reported for this species. Most of the existing works focused on M. pulegium essential oil or alcoholic extracts. In fact, the chemical profile of the essential oil is well established.¹⁰

As for the phenolic profile, data is scarce and lacks consistency, as most of the studies available are directed to the quantification of total phenolics and not to their structural characterization, leaves being the most studied material. Phenolic acids, such as caffeic, vanillic, ferulic, rosmarinic and lithospermic, and the flavonoids luteolin, apigenin, naringenin, (+)-catechin, diosmetin 7-O-rutinoside, jaceosidin, pectolinarigenin and pedalitin are some examples of the phenolic compounds already identified in M. pulegium alcoholic extracts.^11–16

Dietary intake of polyphenols has revealed to provide innumerous benefits to the pathophysiological development of neurodegenerative and cardiovascular diseases, diabetes, osteoporosis and cancer, probably due to their antioxidant action.¹⁷ The combination of these exogenous antioxidants with the endogenous ones, like reduced glutathione (GSH) and GSH-related enzymes, provides a cellular environment able to oppose against adverse oxidative stress, caused either by the normal cellular activities or xenobiotics.¹⁸

Our work aimed at the first phenolics characterization of M. pulegium aqueous extract and to establish a correlation between the chemical composition and the physiologic and toxicological responses of two cellular lines, human colon carcinoma cells (Caco-2) and human gastric carcinoma cells (AGS), representative of the human gastrointestinal tract, under quiescent conditions and subjected to oxidative stress induced by tert-butyl hydroperoxide (t-BHP).

Experimental

Standards and reagents

Sodium pyruvate, 1-(4,5-dimethylthiazol-2-yl)-3,5-diphenylformazan (MTT), β-nicotinamide adenine dinucleotide reduced form (NADH), t-BHP, sodium nitroprusside dehydrate (SNP), phenazine methosulfate (PMS), nitroblue tetrazolium chloride (NBT), N-(1-naphthyl)ethylenediamine dihydrochloride, sulfanilamide, amphotericin B (250 μg mL⁻¹), transferrin (4 mg mL⁻¹), non-essential amino acids, methanol, formic acid, sodium hydroxide, acetonitrile, formaldehyde, perchloric acid, dimethyl sulfoxide (DMSO), 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI), 2-aminoethyl diphenylborinate (DPBA), 2,2-diphenyl-1-picrylhydrazyl (DPPH), GSH, glutathione oxidized form (GSSG), glutathione reductase (EC 1.6.4.2), 96-well plates and caffeic, chlorogenic and rosmarinic acids were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ferulic acid, luteolin 7-O-glucoside and apigenin 7-O-glucoside were from Extrasynthese (Genay, France) and 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) from Alfa Aesar (Karlsruhe, Germany). Dulbecco's Modified Eagle Medium (DMEM), Dulbecco's phosphate buffered saline (DPBS), Hank's balanced salt solution (HBSS), fetal bovine serum (FBS), Pen-Strep solution (Penicillin 5000 units per mL and streptomycin 5000 mg mL⁻¹) and trypsin-EDTA were purchased from Gibco (Invitrogen, Paisley, UK). Water was deionized using a Milli-Q water purification system (Millipore, Bedford, MA, USA). Human epithelial colorectal adenocarcinoma (Caco-2) and human epithelial gastric adenocarcinoma (AGS) cell lines were acquired from American Type Culture Collection (Manassas, Virginia, USA).

Vegetal material and infusion preparation

M. pulegium aerial parts were kindly provided by Direção Regional de Agricultura de Entre Douro e Minho. The material was dried at 30 °C for 24 h, followed by powdering, and stored protected from light and humidity. In order to mimic the herbal tea usually prepared for human consumption, 200 mL of boiling water were added to 4.00 g of the powdered material and left to stand for 15 min. After filtration the obtained extract was frozen at −20 °C and then lyophilized in a Labconco 4.5 Freezone apparatus (Kansas City, MO, USA). For all the experiments, an aliquot was collected and redissolved in accordance with the specific procedures.

HPLC-DAD-ESI/MSⁿ qualitative analysis of phenolic compounds

HPLC-DAD-ESI/MSⁿ chromatographic analysis was developed on a Kinetex column (5 μm, C18, 100 Å, 150 × 4.6 mm; Phenomenex, Macclesfield, UK). The mobile phase consisted of two solvents: water–formic acid (1%) (A) and acetonitrile–formic acid (1%) (B), starting with 10% B and using a gradient to obtain 30% B at 20 min. The flow rate was 1 mL min⁻¹ and the injection volume 20 μL. Spectral data from all peaks were accumulated in the range of 240–400 nm and chromatograms were recorded at 330 nm. The analyses were carried out in an Agilent HPLC 1100 series equipped with a diode array detector and mass detector in series (Agilent Technologies, Waldbronn, Germany). The HPLC consisted of a binary pump (model G1312A), an auto sampler (model G1313A), a degasser (model G1322A) and a photodiode array detector (model G1315B). The HPLC system was controlled by ChemStation software (Agilent, v. 08.03). The mass detector was an ion trap spectrometer (model G2445A) equipped with an electrospray ionization interface and was controlled by LCMSD software (Agilent, v. 4.1). The ionization conditions were adjusted at 350 °C and 4 kV for capillary temperature and voltage, respectively. The nebulizer pressure and flow rate of nitrogen were 65.0 psi and 11 L min⁻¹, respectively. The full scan mass covered the range from m/z 100 up to m/z 1000. Collision-induced fragmentation experiments were performed in the ion trap using helium as the collision gas, with voltage ramping cycles from 0.3 up to 2 V. Mass spectrometry (MS) data were acquired in the negative ionization mode. MSⁿ was carried out in the automatic mode on the more abundant fragment ion in MS⁽ⁿ⁻¹⁾.

HPLC-DAD quantitative analysis of phenolic compounds

HPLC-DAD conditions were the same as those described above, except for the Waters Spherisorb ODS2 (25.0 × 0.46 cm, 5 μm particle size; Milford, Massachusetts, USA) column. 20 μL of the lyophilized extract dissolved in water (25 mg mL⁻¹) were injected on a HPLC-DAD unit (Gilson), spectral data were acquired with a Gilson DAD and processed on Unipoint system software (Gilson Medical Electronics, Villiers le Bel, France). Phenolic compounds quantification was achieved by the absorbance recorded at 320 and 350 nm for phenolic acids and flavonoids, respectively, in relation to external calibration curves (Table 1). Since commercial standards were not available for all the compounds, caffeic acid glucosides, trimers and tetramer were quantified as caffeic acid, ferulic acid glucosides as ferulic acid, luteolin and apigenin derivatives as luteolin 7-O-glucoside and apigenin 7-O-glucoside, respectively. Caffeoylquinic acids were quantified as 5-O-caffeoylquinic acid. Rosmarinic acid and its derivatives were quantified as rosmarinic acid.

Table 1 Regression equations, LOD and LOQ for phenolic compounds quantification

Compounds^a	Regression equation	R^2	Range of concentrations (mg mL^−1)	LOD^b (mg mL^−1)	LOQ^c (mg mL^−1)
a Caffquin: caffeoylquinic; Ac: acid; Lut: luteolin; Apig: apigenin; Gluc: glucoside.b Limit of detection.c Limit of quantification.
5-Caffquin Ac	7.2 × 10^8x − 3.1 × 10^6	0.994	1.2 × 10^−1 − 1.5 × 10^−2	4.3 × 10^−4	1.3 × 10^−3
Caffeic Ac	1.5 × 10^9x + 7.5 × 10^6	0.992	1.6 × 10^−1 − 8.0 × 10^−3	6.4 × 10^−4	1.9 × 10^−3
Ferulic Ac	1.4 × 10^9x + 6.2 × 10^6	0.994	1.3 × 10^−1 − 6.0 × 10^−3	4.7 × 10^−4	1.4 × 10^−3
Rosmarinic Ac	8.2 × 10^8x + 6.1 × 10^6	0.994	2.1 × 10^−1 − 1.1 × 10^−2	1.3 × 10^−3	1.4 × 10^−3
Lut-7-Gluc	4.6 × 10^8x + 2.0 × 10^6	0.999	1.20 × 10^−1 − 6.0 × 10^−3	1.1 × 10^−4	3.4 × 10^−4
Apig-7-Gluc	8.5 × 10^8x − 1.5 × 10^6	0.997	1.4 × 10^−1 − 1.7 × 10^−2	4.3 × 10^−4	1.3 × 10^−3

---

Antiradical activity

DPPH˙ scavenging. The scavenging activity was determined spectrophotometrically.¹⁹ The assay was performed in 96-well plates and a set of lyophilized extract dilutions was prepared. The reaction mixture consisted on extract (redissolved in water) and DPPH˙ methanolic solution. The plates were incubated for 30 min protected from light and the absorbance at 515 nm was determined in a Multiskan Ascent plate reader (Thermo Electron Corporation, Vantaa, Finland).

Superoxide radical scavenging. The procedure was in accordance with that previously reported.¹⁹ The reaction mixture was composed by the extract (redissolved in phosphate buffer 19 mM), NADH, NBT and PMS. The plate reader was set in kinetic function and the absorbance was determined for 2 min after PMS addition, at 562 nm. All components were dissolved in phosphate buffer (19 mM, pH 7.4).

Nitric oxide scavenging. NO radical was generated by a SNP solution at pH 7.4 with posterior reaction with oxygen to produce nitrite, which was determined by Griess reagent in 96-well plates.¹⁹ Each well received extract (redissolved in phosphate buffer 100 mM) and SNP, followed by 60 min of incubation at room temperature under light exposure. Griess reagent (1% sulfanilamide and 0.1% naphthylethylenediamine in 2% H₃PO₄) were then added and the absorbance of the chromophore produced was measured at 540 nm.

Cellular assays

Cell culture and treatments. Cells were maintained in DMEM with 10% FBS and 2% Pen-Strep in an incubator (Toreuse model 2428; Saint Louis, Missouri, USA) at 37 °C, with 5% CO₂ and humidified air, in 75 cm² flasks.^20,21 Additionally, Caco-2 cell medium was supplemented with amphotericin B (1%), transferrin (0, 15%), and non-essential amino acids (1%). The incubation conditions were maintained for all the cellular assays. Once confluence was achieved, cells were washed twice with 10 mL HBSS and 3 mL of trypsin-EDTA were added, followed by 8 min of incubation.^20,21 200 μL of the prepared cellular suspension (150 000 cells per mL) was seeded in 96 well-plates and incubated for 2 days before carrying out the viability assays.^20,21 Preliminary assays were performed to determine the appropriate t-BHP concentration and exposure time to assess the activity of the infusion (data not shown). A set of t-BHP dilutions was prepared (0.25, 0.50, 1, 2 and 4 mM) and cells were exposed for 2, 4 and 6 h. By the combination of the results obtained, and in order to achieve a proper viability decrease, the exposure conditions were fixed at 6 h with t-BHP 0.5 mM. Cells were seeded under the same conditions described above. After 24 h, the medium was completely removed and t-BHP was added to each well. The MTT and lactate dehydrogenase (LDH) assays were then carried out to evaluate the effect of the infusion against the induced toxicity.

MTT assay. Cells were incubated with different concentrations of the extract for 24 h. After medium removal, MTT was added to each well, followed by 30 min of incubation. The extent of reduction to formazan was then quantified spectrophotometrically by measuring the absorbance at 510 nm and compared to controls.²⁰

LDH assay. LDH activity was spectrophotometrically determined at 340 nm, by following NADH oxidation during the conversion of pyruvate to lactate.²⁰ The medium was collected after 24 h of cellular exposure to the infusion. The reaction mixture consisted of infusion, NADH and pyruvate, all solutions being prepared in phosphate buffer (0.1 M, pH 7.4).

Determination of total and oxidized glutathione. Total glutathione levels were determined by the DTNB-GSSG reductase recycling assay, after protein precipitation with perchloric acid (0.5%), and GSSG was assessed after sample pre-treatment with 2-vinylpyridine for 1 h at 4 °C, with agitation.²¹ The absorbance was read at 405 nm in kinetic function for 3 min. The experiments were conducted with and without exposure to t-BHP after incubation with the infusion. Proteins were quantified spectrophotometrically at 595 nm with Bradford reagent, using bovine serum albumin as standard.

Intracellular polyphenols' staining and fluorescence microscopy. The procedure was based on a previous work,²² with some modifications. Caco-2 and AGS cells were seeded at 75 000 cells per mL, in 24-multiwell plates, and treated with extract concentrations of 0.31 and 0.16 mg mL⁻¹ respectively, for 24 h. Two positive controls (apigenin and rosmarinic acid, 25 μM) were also evaluated under the same conditions. Cells' fixation procedure consisted on removing the culture medium and adding 600 μL of pre-warmed 3.7% formaldehyde solution, with 10 min of incubation at room temperature. The solution was then removed and each well was rinsed with 600 μL HBSS. Nuclear staining was achieved with 1 μg mL⁻¹ DAPI (with excitation at 364 nm and emission at 454 nm) for 30 min, followed by 0.1% (w/v) DPBA (excitation at 490 nm, emission at 530 nm) staining for 1 min. Fluorescence was detected by microscopy (Nikon TS100 microscope, Tokyo, Japan) and digital images were generated with a Nikon DS-Fi1 camera and NIS-Elements D 3.2 software (Nikon Instruments INC, New York).

Statistical analysis. Statistical analysis was performed using Graphpad Prism 6 Software (San Diego, CA, USA). Quantification of phenolic compounds was achieved from three determinations. The EC₅₀ values were calculated from three independent assays, each of them performed in triplicate. Cellular based assays consisted of six independent experiments. Data from different groups was compared using one-way ANOVA Tukey's multiple comparisons test. A level of statistical significance at p < 0.05 was used.

Results and discussion

Phenolic compounds characterization

The identification of all phenolic compounds was achieved not only by the MS fragmentation patterns, but also according to their UV spectra, chromatographic behavior, and by comparison with existing bibliography. When commercial standards were available, identity was further confirmed by analyzing them under the same conditions used for the sample.

The HPLC-DAD-ESI/MSⁿ analysis of the aqueous extract of M. pulegium at 330 nm (Fig. 1) showed the presence of diverse compounds, the majority of them with an ultra violet (UV) spectrum characteristic of cinnamoyl derivatives (1–3, 7, 11–13 and 15) (Table 2). Compounds 1 and 3 exhibited the deprotonated molecular ion at m/z 353 (caffeoylquinic acid) but different MS fragmentations, corresponding to 3-caffeoylquinic acid and 4-caffeoylquinic acid, respectively (Table 2), according to that reported before.²³ Compound 2, with [M − H]⁻ ion at m/z 311 and MS fragmentation showing the loss of 132 amu (pentosyl moiety) originating the ion at m/z 179 ([caffeic acid − H]⁻), can possibly be labelled as caffeoylpentoside. The MS² of compound 7 presented losses of −162 (hexosyl moiety) and −(162 + 18), corresponding to an interglycosidic linkage, and of −(162 + 132) (hexopentosyl moiety) over a caffeic acid (179, [caffeic acid − H]⁻) (Table 2). Moreover, its MS³[(M − H) → (M − H − 162)]⁻ was similar to the MS² of compound 2; as so, compound 7 should be a derivative of compound 2 with an additional hexose over the pentose, being here characterized as a caffeoyl hexosylpentoside. Compound 12 exhibited an UV spectrum similar to that of 7. Both the deprotonated molecular ion and the fragmentation pattern of compound 12 differed from those of compound 7 in 14 amu: since [feruloyl acid − H]⁻ (193 amu) was observed instead of the ion at m/z 179 [caffeic acid − H]⁻, compound 12 was identified as feruloyl hexosylpentoside.

Fig. 1 HPLC-UV chromatogram (330 nm) of M. pulegium infusion. Ac: not identified cinnamoyl derivatives; 1: 3-caffeoylquinic acid; 2: caffeoylpentoside; 3: 4-caffeoylquinic acid; 4: salvianolic acid H; 5: luteolin 7-O-rutinoside; 6: luteolin 7-O-glucuronide; 7: caffeoyl hexosylpentoside; 8: salvianolic acid E isomer; 9: chrysoeriol/diosmetin 7-O-rutinoside; 10: apigenin 7-O-glucuronide; 11: rosmarinic acid; 12: feruloyl hexosylpentoside; 13: lithospermic acid; 14: salvianolic acid C isomer; 15: feruloyl rosmarinic acid derivative.

Table 2 Rt, UV and MS: [M − H]⁻, MS²[M − H]⁻ and MS³[(M − H) → base peak]⁻ data of phenolic compounds from M. pulegium infusion^a

Compounds^b		Rt (min)	UV (nm)	[M − H]^−, m/z	MS^2[M − H]^−, m/z (%)	MS^3, m/z (%)
a Main observed fragments. Other ions were found but they have not been included.b Ac: acid; Caff: caffeoyl; Fer: feruloyl; Quin: quinic; Pent: pentosyl; Hex: hexosyl; Gluc: glucuronoyl; Lut: luteolin; Apig: apigenin; Chrys: chrysoeriol; Diosmt: diosmetin; Salv: salvianolic; Rosm: rosmarinic; Lith: lithospermic; Rut: rutinoside.c MS^3(537 → 359): 197(40), 179(40), 161(100).d MS^3(535 → 359): 197(20), 179(35), 161(100).
1	3-CaffQuin Ac	3.1	300sh, 328	353	191(100), 179(45)
2	CaffPent	4.0	300sh, 328	311	179(65), 149(100)
3	4-CaffQuin Ac	4.9	300sh, 328	353	173(100)
4	Salv Ac H	10.3	252, 286, 314sh, 344	537	493(15), 339(100)	295(80), 229(100)
5	Lut-7-Rut	11.1	256, 266sh, 348	593	285(100)
6	Lut-7-Gluc	12.0	256, 266sh, 348	461	285(100)
7	CaffHexPent	12.3	300sh, 328	473	311(100), 293(60), 179(10)	179(50), 149(100)
8	Salv Ac E isomer	13.3	256, 284, 316, 348	717	519(100), 475(35)	475(100)
9	Chrys/Diosmt-7-Rut	14.4	256, 268sh, 348	607	299(100), 284(30)
10	Apig-7-Gluc	14.8	267, 337	445	400(10), 269(100)
11	Rosm Ac	15.0	300sh, 328	359	197(30), 179(40), 161(100)
12	FerHexPent	16.1	300sh, 328	487	325(100), 307(57), 293(85), 193(30)	193(100), 149(4)
13	Lith Ac	17.1	298sh, 325	537	493(100), 359(35)^c	359(100)
14	Salv Ac C isomer	19.2	262, 320	491	311(100)
15	Fer Rosm Ac	20.0	300sh, 328	535	359(90), 177(100)^d

---

With a similar UV spectrum (Table 2) and a deprotonated molecular ion at m/z 359, compound 11 presented in its MS² fragmentation losses of 3,4-dihydroxyphenyllactic acid (danshensu (DSS), −198 amu), caffeic acid (−180 amu) and caffeic acid − 18 (−162 amu), with formation of ions at m/z 161, 179 and 197, respectively. These are characteristic of dimers, trimers and tetramers of caffeic acid.²⁴ This fragmentation pattern is in agreement with the one reported for rosmarinic acid,²⁵ and the co-elution of the extract with a standard allowed the confirmation of compound 11 as rosmarinic acid.

Compound 15 also displayed a similar UV spectrum (Table 2) and its MS fragmentation showed a loss of 176 amu (feruloyl moiety), with formation of the ion at m/z 359. As the fragmentation (MS³[535 → 359]⁻) observed matched with the one of rosmarinic acid, compound 15 was tentatively identified as a feruloyl rosmarinic acid derivative.

Compounds 4, 13 and 14 are trimers. Compound 4 presented [M − H]⁻ ion at m/z 537, losses of DSS (−198, m/z 339 base peak) and of a carbonyl group (−44) coincident with the fragmentation observed for salvianolic acid H (SAH).²³ In addition, its UV spectrum, different from those of the other caffeic acid derivatives above described, has already been reported.²⁶

As for compound 13, an isomer of compound 4, it was observed a characteristic UV spectrum of caffeic acid derivatives (Table 2). Both the similar MS fragmentation, showing a base peak corresponding to the loss of a carbonyl group (−44, m/z 493), and chromatographic mobility compared to rosmarinic acid were reported for lithospermic acid;²⁴ therefore, it was tentatively labelled as lithospermic acid. Compound 14, exhibited the same deprotonated molecular ion of salvianolic acid C (m/z 491), also with the same chromatographic behavior, eluting at the end of the chromatogram. However, the loss of the DSS (−198) fragment with formation of the base peak reported before was not observed;²⁴ instead, the fragmentation originated the base peak at m/z 311 by loss of a caffeic acid moiety (−180), as reported previously,²⁶ either caused by different analytical conditions or by the presence of an isomer. Thus, it was tentatively labelled as salvianolic acid C isomer.

Compound 8 is a tetramer ([M − H]⁻ at m/z 717). Its MS fragmentation showed the loss of DSS, giving rise to the base peak (m/z 519), as reported for the tetramers salvianolic acids E, B and L.²⁷ Nevertheless, taking into account its lower chromatographic mobility, compared with that of rosmarinic acid, and strong resemblance with the fragmentation pattern and UV spectrum of a salvianolic acid E isomer described before,²⁶ compound 8 may correspond to this molecule.

Other compounds with UV spectra of cinnamoyl derivatives were noticed at Rt 6.0 and 15.3 min (Ac, Fig. 1), though their structures could not be determined.

Finally, four other peaks (5, 6, 9 and 10) found in trace amounts and with UV and MS spectra of flavonoids were observed. Compounds 5, 6 and 9 presented an UV spectrum characteristic of flavones with di-substitution at the B ring (Table 2). The MS fragmentation of compounds 5 and 6 showed the ion at m/z 285 ([luteolin − H]⁻) as base peak, thus being identified as luteolin derivatives. Compound 5 fragmentation, with loss of 308 amu (rhamnohexosyl moiety) and without ions characteristic of interglycosyl linkage cleavage, suggested the presence of a rhamnosyl(1→6)hexoside linkage. As so, it was tentatively labelled as luteolin 7-O-rhamnosyl(1→6)glucoside (luteolin 7-O-rutinoside). Moreover, the loss of 176 amu (glucuronoyl moiety) by fragmentation of compound 6 led to its identification as luteolin 7-O-glucuronide. As it happened for compound 5, in the MS fragmentation of compound 9 it was observed the loss of 308 amu, resulting in the formation of the deprotonated aglycone ion (m/z 299), pointing to a methyl-luteolin, probably chrysoeriol (5,7,4′-OH-3′-OMe-flavone, 3′-Me-luteolin) or diosmetin (5,7,3′-OH-4′-OMe-flavone, 4′-Me-luteolin). Thus, compound 9 was tentatively labelled as chrysoeriol/diosmetin 7-O-rutinoside. Compound 10 was identified as apigenin 7-O-glucuronide due to the UV spectrum characteristic of an apigenin derivative (Table 2) and because the fragmentation of its deprotonated molecular ion ([M − H]⁻, 445) resulted in the deprotonated molecular ion of apigenin ([apigenin − H]⁻, 269), the base peak, by loss of a 176 amu fragment (glucuronoyl moiety).

With the exceptions of rosmarinic and lithospermic acids and of diosmetin 7-O-rutinoside, all the other twelve compounds are identified for the first time in M. pulegium.

Phenolic compounds quantification

The phenolic fraction of the infusion corresponded to 122.92 mg g⁻¹, distributed by the fifteen identified compounds (Table 3). Phenolic acids represented 94% of the determined compounds. Rosmarinic acid (compound 11) was the main compound (54% of the quantified compounds), as expected for a Lamiaceae species, being followed by other cinnamoyl derivatives, namely caffeoyl hexosylpentoside (compound 7, 13%) and the trimers salvianolic acid H (compound 4) and lithospermic acid (compound 13, 6%, each) (Table 3).

Table 3 Phenolic composition of M. pulegium infusion

Compounds^a		mg g^−1 Lyophilized infusion^b
a Ac: acid; Caff: caffeoyl; Fer: feruloyl; Quin: quinic; Pent: pentosyl; Hex: hexosyl; Gluc: glucuronoyl; Lut: luteolin; Apig: apigenin; Chrys: chrysoeriol; Diosmt: diosmetin; Salv: salvianolic; Rosm: rosmarinic; Lith: lithospermic; Rut: rutinoside.b Results are expressed as means ± standard deviations of three independent determinations.c Sum of the quantified compounds.
1	3-CaffQuin Ac	4.16 ± 0.05
2	CaffPent	2.44 ± 0.02
3	4-CaffQuin AC	2.64 ± 0.04
4	Salv Ac H	7.60 ± 0.76
5	Lut-7-Rut	0.64 ± 0.02
6	Lut-7-Gluc	1.40 ± 0.06
7	CaffHexPent	16.20 ± 1.36
8	Salv Ac E isomer	1.88 ± 0.12
9	Chrys/Diosmt-7-Rut	3.40 ± 0.13
10	Apig-7-Gluc	1.32 ± 0.09
11	Rosm Ac	66.08 ± 5.19
12	FerHexPent	3.48 ± 0.10
13	Lith Ac	7.52 ± 0.69
14	Salv Ac C isomer	1.84 ± 0.03
15	Fer Rosm Ac	2.24 ± 0.07
Σ^c		122.92

---

As for the identified flavonoids, chrysoeriol/diosmetin 7-O-rutinoside (compound 9) was the most abundant one (3%).

Antiradical activity

There is not a single accepted methodology to fully assess the antiradical activity. The data obtained with in vitro assays cannot be simply extrapolated to in vivo models, thus being interpreted as screening tools. Among other variables, they largely depend upon the amount of radicals that can be scavenged by antioxidants, their chemical nature and reaction mediums.²⁸ For this work, three different reactive species were considered, to have a more wide idea about the infusion's antiradical potential: one synthetic radical (DPPH˙), a reactive oxygen species (O₂˙⁻) and a reactive nitrogen species (˙NO). Furthermore, both O₂˙⁻ and ˙NO are reactive species with biological/physiological significance, being produced in the human body; the reaction between them originates peroxynitrite, another important and deleterious reactive species.

Concentration-dependent responses were observed, although total scavenging activity was not reached for ˙NO radical, due to insolubility interference of high infusion's concentrations. The EC₅₀ values obtained were 23, 39 and 226 μg mL⁻¹ against DPPH˙, O₂˙⁻ and ˙NO radicals, respectively (data not shown). A similar result was obtained in a previous work, in which the hot water extract revealed to be very effective against DPPH˙.¹⁰ As for the other radicals, no previous results were found. As referred above, phenolic compounds are extremely effective in neutralizing radical species. Therefore, the results here obtained can be, at least partially, explained by the phenolic content of the tested infusion (Table 3).

It is well known that a homeostatic cell environment is necessary to maintain proper physiological functions. Dietary polyphenols intake may display an important role on balancing potential toxic environments, caused by a vast number of free radicals: reactive oxygen species (ROS), including either oxygen radicals like superoxide (O₂˙⁻), hydroxyl (˙OH), and peroxyl (RO^˙₂) radicals, or non-radical oxidizing agents, such as hydrogen peroxide (H₂O₂) and hypochlorous acid (HOCl), easily converted into radicals; reactive nitrogen species (RNS) and reactive sulphur species (RSS), which are formed by thiols' reaction with ROS. Besides the antioxidant activity, it is accepted that these compounds can display other combined complex bioactivities. A deeper understanding over the ability to activate phase II enzymes, to act as anti-inflammatory agents, and to interact with the intestinal flora will certainly explain and sustain the importance of a phenolic-rich diet to prevent diverse human pathologies.^29,30

Mitochondrial activity, membrane integrity and protection against t-BHP-induced toxicity

Both Caco-2 and AGS cell lines were selected for this study as models of cellular response to xenobiotics since, when consumed, M. pulegium infusion directly contacts with gastric and intestinal epithelia.

A preliminary experiment was conducted in order to assess the range of concentrations for which the exposure to the infusion was not able to affect cellular viability (Fig. 2), since phenolic compounds can exert dose-dependent cytotoxicity.³¹

Fig. 2 Effect of M. pulegium infusion on Caco-2 and AGS cell lines viability after 24 h of exposure, assessed by MTT reduction and LDH leakage assays. Values show mean ± SEM of six independent assays performed in triplicate (**p < 0.01, ****p < 0.0001 and #p < 0.05 compared to the respective controls).

When comparing the two cell lines, AGS cells demonstrated to be more sensitive, since viability started to significantly decrease when exposed to 0.31 mg mL⁻¹ of infusion, while for Caco-2 this happened only at 1.25 mg mL⁻¹ (Fig. 2). Moreover, the results obtained by the MTT reduction assay were more expressive than the ones by LDH leakage for both cell lines, thus suggesting that mitochondrial damage happens prior to membrane's damage.

The following step was to evaluate the potential of M. pulegium infusion to protect the cells against the toxicity caused by t-BHP. This hydroperoxide is frequently applied to induce oxidative stress, the toxicity to hepatocytes being a result from either the metabolism through cytochrome P450 or reduction of GSH.³² Thus, several biomarkers can be monitored, like the increased levels of alanine transaminase, aspartate aminotransferase, LDH and malondialdehyde, and also GSH depletion.³²

In this work, cells were treated with M. pulegium infusion for 24 h prior to t-BHP exposure (0.5 mM, 6 h). Cellular viability was again determined by MTT reduction and LDH leakage assays (Fig. 3). The infusion clearly exerted a dose-dependent protective effect in the MTT reduction assay. In fact, total protection, when compared to controls, was almost achieved with the infusion at the concentrations of 1.25 and 0.63 mg mL⁻¹ for Caco-2 and AGS cells, respectively (Fig. 3), despite that at this concentrations the infusion induced cytotoxicity by itself (Fig. 2).

Fig. 3 Caco-2 and AGS cellular viability assessed by MTT reduction and LDH leakage assays, after exposure to M. pulegium infusion, with and without t-BHP-induced toxicity. Cells were pre-treated with the infusion for 24 h. Insulted cells were further exposed to t-BHP (0.5 mM) for 6 h. Values show mean ± SEM of six independent experiments performed in triplicate (***p < 0.001, ****p < 0.0001 and ##p < 0.01, ####p < 0.0001 compared to the respective controls).

Hence, it can be assumed that from the pre-treatment with the infusion and post exposure to the toxicant did not result a synergic toxic effect.

Contrarily to the expected, no increase of LDH leakage was observed in either cell lines (Fig. 3). In a previous work, using LDH leakage as a biomarker of necrosis and DNA fragmentation as a biomarker of apoptosis, it was reported that a concentration of t-BHP of 0.4–0.5 mM provides a transition point, below which apoptosis is favored and beyond which necrosis is favored.³³ Taking this into consideration, it can be inferred that for the experimented t-BHP exposure conditions (0.5 mM, 6 h) cell death was triggered by apoptosis, thus not being observed the increase of extracellular LDH, which is related with cellular membrane damage.

The phenolic composition of M. pulegium infusion here reported (Table 2) can be implicated in the antioxidant protective effect observed. Since apoptosis can be induced by a ROS-dependent mitochondrial pathway, phenolic compounds may also display a protective role by reducing toxic levels of reactive species.³⁴ Other flavonoids, like rutin and quercetin, also tested in a similar pre-treatment experimental model, were very effective at preventing Caco-2 DNA damage induced by t-BHP.³⁵

The authors attributed such results to the metal iron chelating and free radical scavenging properties of those compounds. With HepG2 cells, but using a co-exposition model, it was demonstrated that other phenolics like luteolin, quercetin, luteolin 7-O-glucoside, caffeic and rosmarinic acids also protected the cells against t-BHP toxicity, by preventing lipid peroxidation and GSH depletion.³⁶

Total and oxidized GSH

Attempting to disclosure a possible mechanism behind the antioxidant activity observed, total glutathione (GSx) and GSSG levels were quantified under the same model of pre-treatment with the infusion, followed by t-BHP-induced toxicity (Fig. 4). The results revealed a different response of the two cell lines. In Caco-2 cells, no increase of GSx levels was observed when compared to the respective controls without infusion treatment. However, when post-exposed to t-BHP, GSx levels increased very significantly in a concentration-dependent response for infusion concentrations above 0.04 mg mL⁻¹ (Fig. 4). The observed increase may be the result of a synergic induction of antioxidant response elements (ARE) genes through the nuclear factor erythroid 2-related factor 2 (Nrf2).³⁷

Fig. 4 Total glutathione (GSx) and oxidized glutathione (GSSG) determination. Caco-2 and AGS cells were pre-treated with M. pulegium infusion for 24 h. GSx and GSSG levels were quantified with and without t-BHP exposure. Values show mean ± SEM of three independent assays performed in triplicate (**p < 0.01, ***p < 0.001 and ****p < 0.0001 compared to respective controls and ####p < 0.0001 comparison with and without t-BHP).

This known regulator mechanism of cellular resistance to oxidants may ultimately lead to the enhancement of the γ-glutamylcysteine synthetase enzyme transcription, the rate limiting enzyme responsible for glutathione synthesis, which is known to be more active in response to cellular treatments with phenolic compounds, leading to de novo synthesis.¹⁸

As expected, after t-BHP exposure, higher GSSG levels were observed when compared to the respective control, and it was demonstrated that the infusion was able to prevent GSH oxidation (Fig. 4).

Taking together these results with the ones obtained for the MTT reduction assay (Fig. 3), it can be noticed that the infusion prevented GSH oxidation even at concentrations at which it was not able to protect against the induced toxicity (0.04 to 0.16 mg mL⁻¹). As so, this mechanism is certainly important, but not the only one behind the protective effect observed. No statistical significance was found for the direct action of the infusion over the GSSG variation levels (Fig. 4).

In contrast, under the same experimental conditions, the infusion was not able to exert the same effect in AGS cells (Fig. 4). GSx levels were not significantly altered by either the infusion or toxicant exposure. However, for the concentration of 0.31 mg mL⁻¹, a significant decrease of the GSSG levels was observed (Fig. 4), in accordance with the verified viability increase (Fig. 3). The results indicate that for AGS cells, t-BHP reduction by GSH was not the most relevant toxicity mechanism. Nevertheless, for both cell lines, taking into account the possible toxicity mechanisms exerted by t-BHP and the antioxidant potential of phenolic compounds, it is possible that the cellular protection was also achieved either by the activity of glutathione peroxidase (GPx), leading to the conversion of the aggressor into a less toxic alcohol,³⁸ by chelation of metallic ions, namely iron that proved to be essential for t-BPH-induced toxicity,³⁹ by down-regulation of the apoptotic ROS-dependent mitochondrial pathway,³⁴ or even by high level of DNA repaired synthesis that demonstrated to be involved in the protection mediated by pretreatment with the flavonoid baicalin.⁴⁰

DPBA staining of intracellular polyphenols

DPBA is capable of binding to flavonoids and phenolic acids in reactive sites after the loss of an ethanolamine chain, with adduct formation and fluorescence emission.⁴¹ One of its applications is related to flavonoids staining in plant cells, to understand the mechanisms of their transport and to follow their localization.⁴²

To sustain whether the phenolic compounds present in M. pulegium infusion were able to enter into the cells, a DPBA staining technique was applied. Nuclear staining with DAPI was also used to check nuclear morphology. Based on the phenolic profile here reported, phenolic acids were the major represented class, rosmarinic acid being the one at the highest concentration (Table 3). So, in addition to the positive apigenin control (25 μM), as described in a previous work,⁴³ a rosmarinic acid control (25 μM) was also tested. The results showed that after 24 h of exposure to the infusion, the phenolic compounds could yet be targeted inside Caco-2 and AGS cells (Fig. 5).

Fig. 5 Fluorescence microscopy of intracellular DPBA staining of phenolic compounds and DAPI nuclear staining (magnification X200). Caco-2 and AGS cells were treated with M. pulegium infusion at the concentrations of 0.31 and 0.16 mg mL⁻¹, respectively, for 24 h. Three independent experiments were performed and a representative field was selected.

Based on the positive controls, the detected fluorescence was derived by DPBA linkage to flavonoids and phenolic acids (Fig. 5). As advanced in a previous work with leukemia cells,⁴³ for the assayed conditions, it is possible that when incubated with apigenin, AGS and Caco-2 cells may undergo an apoptotic process in which chromatin condensation and vacuolization are clearly observed (Fig. 5).

As for rosmarinic acid control, in comparison to the respective negative control (cells and DAPI) some slight morphological alterations can be observed, but in a smaller scale when compared to apigenin (Fig. 5). It has been reported that despite its antioxidant properties, rosmarinic acid can also act as antiproliferative and as an apoptotic inducer agent.^44–46 The viability assays did not demonstrate cellular toxic effects for the experimented infusion concentrations (Fig. 2). However, nuclear morphology alterations may be earlier visible than other toxicological end-points.⁴⁷

This work also sustains that the phenolic compounds are able to pass through the cellular membrane, even without prior hydrolyses.⁴⁸ At the incubation conditions no chemical hydrolyses occurs: the compounds present are O-heterosides, requiring very acid pH conditions and high temperature in order to break the link between the sugar moieties and the aglycones.

Furthermore, no enzymes were present in the culture medium nor added to it. As so, the compounds could only have been taken upon the cell on their natural forms. It is known that when compared with the aglycones, glycosylated compounds are able to enter the cell, even if it occurs in a smaller scale.²²

More important, since the protective effect was observed under a pre-treatment model, the phenolic compounds were still inside the cells when exposed to t-BHP and the 24 h incubation period may have enhanced the induction of Nrf2/ARE-dependent transcription. Besides GSH defenses, phenolic compounds might also act directly over the toxicant, and down-regulate the apoptotic ROS-dependent mitochondrial pathway.

Conclusions

In summary, with this work it was established for the first time the phenolic profile of M. pulegium infusion. Twelve new compounds were identified for the first time in this species. Rosmarinic acid was the major compound. The chemical profile found supports the very effective scavenging activity displayed against DPPH˙, ˙NO and O₂˙⁻ radicals. The infusion demonstrated to protect Caco-2 and AGS cells against t-BHP-induced toxicity, in a concentration-dependent manner, probably due to GSH induction and prevention of GSSG formation. It was also demonstrated that, even being glycosylated, phenolic compounds were able to pass through the cellular membrane and act as direct antioxidants. Overall, these findings sustain that a phenolic-rich diet may display multi-target activities and play important key-roles, with proved human health benefits. In this sense, new techniques should be implemented towards the bioavailability improvement of natural compounds, on behalf of a more clear and accurate knowledge over the Traditional Medicine approach.

Acknowledgements

The authors are grateful for the financial support from the European Union (FEDER funds through COMPETE) and National Funds (FCT, Fundação para a Ciência e Tecnologia) through project Pest-C/EQB/LA0006/2013 and from the European Union (FEDER funds) under the framework of QREN through Project NORTE-07-0124-FEDER-000069, to the CYTED Programme (reference 112RT0460) CORNUCOPIA Thematic Network and project AGL2011-23690 (CICYT).

Notes and references

1. D. J. Newman and G. M. Cragg, J. Nat. Prod., 2012, 75, 311.
2. A. Helmstadter and C. Staiger, Drug Discovery Today, 2014, 19, 4.
3. A. O. Tucker and F. C. Naczi, Mentha: an overview of its classification and relationships, ed. B. M. Lawrance, CRC Press, Florida, 2007, pp. 1–40.
4. A. P. Cunha, Plantas e produtos vegetais em fitoterapia, Fundação Calouste Gulbenkian, Lisboa, 2003, pp. 532–533.
5. P. Kumar, S. Mishra, A. Malik and S. Satya, Ind. Crops Prod., 2011, 34, 802.
6. V. López, S. Martín, M. P. Gómez-Serranillos, M. E. Carretero, A. K. Jäger and M. I. Calvo, Phytother. Res., 2010, 24, 869.
7. A. T. Mata, C. Proença, A. R. Ferreira, M. L. Serralheiro, J. F. Nogueira and M. M. Araújo, Food Chem., 2007, 103, 778.
8. M. Romero-Jiménez, J. Campos-Sánchez, M. Analla, A. Muñoz-Serrano and Á. Alonso-Moraga, Mutat. Res., Genet. Toxicol. Environ. Mutagen., 2005, 585, 147.
9. L. Alpsoy, H. Şahin and S. Karaman, Toxicol. Ind. Health, 2011, 27, 647.
10. B. Teixeira, A. Marques, C. Ramos, I. Batista, C. Serrano and O. Matos, Ind. Crops Prod., 2012, 36, 81.
11. C. Proestos, N. Chorianopoulos, G. J. E. Nychas and M. Komaitis, J. Agric. Food Chem., 2005, 53, 1190.
12. C. Proestos, K. Lytoudi, O. Mavromelanidou, P. Zoumpoulakis and V. Sinanoglou, Antioxidants, 2013, 2, 11.
13. R. Tahira, M. Naeemullah, F. Akbas and M. Masood, Pak. J. Bot., 2011, 43, 151.
14. D. A. A. Kogiannou, N. Kalogeropoulos, P. Kefalas, M. G. Polissiou and A. C. Kaliora, Food Chem. Toxicol., 2013, 6, 152.
15. P. C. Dinis, P. L. Falé, P. J. A. Madeira, M. H. Florêncio and M. L. Serralheiro, Eur. J. Med. Plants, 2013, 3, 381.
16. F. Zaidi, B. Voirin, M. Jay and M. R. Viricel, Phytochemistry, 1998, 48, 991.
17. A. Scalbert, C. Manach, C. Morand, C. Rémésy and L. Jiménez, Crit. Rev. Food Sci. Nutr., 2005, 45, 287.
18. R. Masella, R. Di Benedetto, R. Vari, C. Filesi and C. Giovannini, J. Nutr. Biochem., 2005, 16, 577.
19. F. Ferreres, F. Fernandes, J. M. A. Oliveira, P. Valentão, J. A. Pereira and P. B. Andrade, Food Chem. Toxicol., 2009, 47, 1209.
20. G. Lopes, C. Sousa, L. R. Silva, E. Pinto, P. B. Andrade, J. Bernardo and P. Valentão, PLoS One, 2012, 7, e31145.
21. C. Sousa, H. Pontes, H. Carmo, R. J. Dinis-Oliveira, P. Valentão and P. B. Andrade, Toxicol. in Vitro, 2009, 23, 1131.
22. G. Hostetler, K. Riedl, H. Cardenas, M. Diosa-Toro, D. Arango and S. Schwartz, Mol. Nutr. Food Res., 2012, 56, 558.
23. M. N. Clifford, K. L. Johnston, S. Knight and N. Kuhnert, J. Agric. Food Chem., 2003, 51, 2900.
24. G. Zeng, H. Xiao, J. Liu and X. Liang, Rapid Commun. Mass Spectrom., 2006, 20, 499.
25. M. B. Hossain, D. K. Rai, N. P. Brunton, A. B. Martin-Diana and C. Barry-Ryan, J. Agric. Food Chem., 2010, 58, 10576.
26. F. Ferreres, J. Vinholes, A. Gil-Izquierdo, P. Valentão, R. F. Gonçalves and P. B. Andrade, Food Chem., 2013, 136, 1390.
27. A.-H. Liu, H. Guo, M. Ye, Y.-H. Lin, J.-H. Sun and M. Xu, J. Chromatogr. A, 2007, 1161, 170.
28. M. Takashima, M. Horie, M. Shichiri, Y. Hagihara, Y. Yoshida and E. Niki, Free Radicals Biol. Med., 2012, 52, 1242.
29. G. Chiva-Blancha and F. Visiolib, J. Berry Res., 2012, 2, 63.
30. F. Giampieri, J. M. Alvarez-Suarez and M. Battino, J. Agric. Food Chem., 2014, 62, 3867.
31. V. Mendes, V. Costa and N. Mateus, RSC Adv., 2015, 5, 1.
32. W.-L. Lin, C.-J. Wang, Y.-Y. Tsai, C.-L. Liu, J.-M. Hwang and T.-H. Tseng, Arch. Toxicol., 2000, 74, 467.
33. K. Haidara, I. Morel, V. Abaléa, M. Gascon-Barré and F. Denizeau, Biochim. Biophys. Acta, Mol. Cell Res., 2002, 1542, 173.
34. T. Y. Forbes-Hernández, F. Giampieri, M. Gasparrini, L. Mazzoni, J. L. Quiles, J. M. Alvarez-Suarez and M. Battino, Food Chem. Toxicol., 2014, 68, 154.
35. S. A. Aherne and N. O'Brien, Free Radicals Biol. Med., 2000, 29, 507.
36. C. F. Lima, M. Fernandes-Ferreira and C. Pereira-Wilson, Life Sci., 2006, 79, 2056.
37. Q. Ma, Annu. Rev. Pharmacol. Toxicol., 2013, 53, 401.
38. W. Lin, C. Wang, Y. Tsai, C. Liu, J. Hwang and T. Tseng, Arch. Toxicol., 2000, 74, 467.
39. K. Abe and H. Saito, Pharmacol. Toxicol., 1998, 83, 406.
40. J. Hwang, C. Wang, F. Chou, T. Tseng, Y. Hsieh, J. Hsu and C. Chu, Arch. Toxicol., 2005, 79, 102.
41. P. Matteini, G. Agati, P. Pinelli and A. Goti, Monatsh. Chem., 2011, 142, 885.
42. C. S. Buer, G. K. Muday and M. A. Djordjevic, Plant Physiol., 2007, 145, 478.
43. M. A. Vargo, O. H. Voss, F. Poustka, A. J. Cardounel, E. Grotewold and A. I. Doseff, Biochem. Pharmacol., 2006, 72, 681.
44. Y.-G. Hur, Y. Yun and J. Won, J. Immunol., 2004, 172, 79.
45. D.-O. Moon, M.-O. Kim, J.-D. Lee, Y. H. Choi and G.-Y. Kim, Cancer Lett., 2010, 288, 183.
46. C. P. R. Xavier, C. F. Lima, M. Fernandes-Ferreira and C. Pereira-Wilson, Nutr. Cancer, 2009, 61, 564.
47. A. R. Todd, J. H. William and K. A. Deevendra, J. Pharmacol. Toxicol. Methods, 1997, 37, 215.
48. W. Jiang and M. Hu, RSC Adv., 2012, 2, 7948.

---

Footnote

† Ferreres and Bernardo are co-first authors on this work.

---