Potential inhibition of major human cytochrome P450 isoenzymes by selected tropical medicinal herbs—Implication for herb–drug interactions

Abstract Background Increasing use of medicinal herbs as nutritional supplements and traditional medicines for the treatment of diabetes, hypertension, hyperlipidemia, and malaria fever with conventional drugs poses possibilities of herb–drug interactions (HDIs). The potential of nine selected widely used tropical medicinal herbs in inhibiting human cytochrome P450 (CYP) isoenzymes was investigated. Materials and methods In vitro inhibition of eight major CYP isoenzymes by aqueous extracts of Allium sativum, Gongronema latifolium, Moringa oleifera, Musa sapientum, Mangifera indica, Tetracarpidium conophorum, Alstonia boonei, Bauhinia monandra, and Picralima nitida was estimated in human liver microsomes by monitoring twelve probe metabolites of nine probe substrates with UPLC/MS‐MS using validated N‐in‐one assay method. Results Mangifera indica moderately inhibited CYP2C8, CYP2B6, CYP2D6, CYP1A2, and CYP2C9 with IC 50 values of 37.93, 57.83, 67.39, 54.83, and 107.48 μg/ml, respectively, and Alstonia boonei inhibited CYP2D6 (IC 50 = 77.19 μg/ml). Picralima nitida inhibited CYP3A4 (IC 50 = 45.58 μg/ml) and CYP2C19 (IC 50 = 73.06 μg/ml) moderately but strongly inhibited CYP2D6 (IC 50 = 1.19 μg/ml). Other aqueous extracts of Gongronema latifolium, Bauhinia monandra, and Moringa oleifera showed weak inhibitory activities against CYP1A2. Musa sapientum, Allium sativum, and Tetracarpidium conophorum did not inhibit the CYP isoenzymes investigated. Conclusion Potential for clinically important CYP‐metabolism‐mediated HDIs is possible for Alstonia boonei, Mangifera indica, and Picralima nitida with drugs metabolized by CYP 2C8, 2B6, 2D6, 1A2, 2C9, 2C19, and 3A4. Inhibition of CYP2D6 by Picralima nitida is of particular concern and needs immediate in vivo investigations.

In various settings, about 40%-57% of patients fail to disclose herbs used concomitantly with their medications to the physician (Robinson & McGrail, 2004;Fakeye, Tijani, & Adebisi, 2008;Kennedy, Wang, & Wu, 2008). Some of these herbs especially in sub-Saharan Africa, India, and in some other parts of the world are consumed as seasoning agents (garlic-Allium sativum), vegetables (Gongronema latifolium and Moringa oleifera), and as edible fruits and seeds (Musa sapientum, Mangifera indica, and Tetracarpidium conophorum). Patients who are on medications and consumes these medicinal herbs may not be aware of potential herb-drug interactions (HDIs) that may occur. Others such as Alstonia boonei, Bauhinia monandra, and Picralima nitida are frequently used in sub-Saharan Africa and India in the management of chronic diseases such as hypertension, diabetes, asthma, peptic ulcer, and cancer, as antimalarials and antimicrobials and other minor ailments (Mahomoodally, 2013;Ezuruike & Prieto, 2014;Iwu, 2014).
Sometimes, in the treatment of these diseases, herbs may be used singly or in combination and prepared as dried powder or decoction. There are many examples of such practice. For example, for rheumatism, bark of Picralima nitida, leaves of Allium ascalonicum, Calliandra portoricensis, and Xylopia aethiopica are grounded, cooked, and 15 ml of the mixture is mixed with corn pap and taken once daily (Olorunnisola, Adetutu, & Afolayan, 2015). For anemia, a mixture of the bark of Mangifera indica and fruits of Aframomum melequeta is dried and powdered with one tablespoonful administered daily (Gbadamosi, Moody, & OYekini, 2012). A decoction of dried bark of Alstonia boonei is also taken two-to three times daily by diabetic patients to lower blood glucose (Adotey, Adukpo, Opoku Boahen, & Armah, 2012).
Many diseases are usually managed or treated with conventional medicines which are mostly metabolized by cytochrome P450 (CYP) isoenzymes (Rendic & Guengerich, 2014). These isoenzymes are responsible for the metabolism of over 70% of prescription and over-the-counter medications (Rendic & Guengerich, 2014). Concomitantly administered drug may modulate these CYP isoenzymes activities leading to clinically significant drug-drug interactions. Such interaction may result in serious adverse drug reactions requiring hospitalization, especially drugs with narrow therapeutic index such as carbamazepine, theophylline, digoxin warfarin, and phenytoin (Greenblatt & von Moltke, 2005;Perucca, 2006;de Leon, 2007;Patel, Rana, Suthar, Malhotra, & Patel, 2014). It is thus rational to expect herbs to also elicit similar HDIs via modulation of cytochrome P450 isoenzymes when concomitantly administered with conventional drugs.
Several studies have documented the occurrence of serious adverse events as a result of concurrent administration of herbs and drugs. Such interactions included bleeding experienced with coadministration of Allium sativum and nonsteroidal anti-inflammatory drugs (Fugh-Berman, 2000), manic episode with Panax ginseng and phenelzine (Fugh-Berman, 2000), and fatal seizure with concomitant administration of Ginkgo biloba and valproate (Kupiec & Raj, 2005).
Little is known of the effect of most tropical medicinal herbs on the metabolic capacity of CYP isoenzymes.
The possibility of occurrence of HDIs when herbs are consumed as seasoning agents, vegetables, fruits and as herbal medicines with conventional drugs informed this study. Since inhibition of CYP isoenzymes is the first step toward elucidating potential HDIs, this study evaluated nine widely used tropical medicinal herbs with a view to identifying those with potentials for CYP-metabolismmediated HDIs.

| Preparation of aqueous extracts of plant parts
Nine medicinal herbs widely used in the tropics for the management of chronic diseases such as hypertension, diabetes, dyslipidemia, chronic kidney disease were selected for this study with the plant parts used and listed in Table 1 Three hundred gram each of dried and powdered Alstonia boonei stem bark, Mangifera indica stem bark, Musa sapientum unriped fruits, Bauhinia monandra leaves, Tetracarpidium conophorum seeds, Allium sativum bulb, and Gongronema latifolium leaves were macerated in 1.5 L of distilled water for 24 hr according to extraction procedure practiced locally by users of these herbs.
Dry powdered Moringa oleifera leaves (50 g) and powdered dry Picralima nitida seeds (70 g) were each macerated in 300 ml of distilled water. Each mixture was filtered, concentrated, and freeze-dried. The freeze-dried extracts were stored at −20°C until needed for in vitro analysis.

| CYP inhibition experiments
Human liver microsomes were obtained from a pool of liver samples from 25 male and female donors and contain 20 mg protein/ ml (Lot No: 99268 from BD Biosciences Labware, Bedford, MA). This was used for the metabolite profiling and CYP inhibition study. N-inone approach (cocktail-approach) for elucidating inhibition toward CYP-specific model reactions was conducted with minor changes as described in earlier studies (Turpeinen, Jouko, Jorma, & Olavi, 2005;Tolonen, Petsalo, Turpeinen, Uusitalo, & Pelkonen, 2007;Showande, Fakeye, Ari, & Hokkanen, 2013). Briefly, the incubation mixture was made up of 0.3 mg microsomal protein/ml, 0.1 M phosphate buffer (pH 7.4), 1 mM NADPH, and nine probe substrates representing the major drug-metabolizing CYPs. The specific probe substrate used for each CYP isoenzymes studied and the respective final concentration in the incubation mixture were as follows: acetaminophen (CYP1A2,  Table S1), CYP isoenzymes inhibited, and concentration used in the N-in-one assay are as reported in the validated methods (Turpeinen et al., 2005;Tolonen et al., 2007;Showande et al., 2013). Spiked standard samples were not used for quantification of probe metabolites, but quantification based on relative peak areas was used (solvent control = 100%).

| Liquid chromatography-mass spectrometry conditions
The analysis of probe metabolites from CYP-specific marker reactions was conducted with a LC/MS/MS method modified from earlier works (Turpeinen et al., 2005;Tolonen et al., 2007) and also reported by (Showande et al., 2013). Briefly, a Waters Acquity UPLC system (Waters Corp., Milford, MA) was used together with a Waters HSS C18 column (2.1 mm × 50 mm; 1.8 μm particle size) and an online filter at 35°C. The injection volume was 4 μl, and UPLC eluents were aqueous 0.1% acetic acid (pH 3.2, A) and acetonitrile (B). The gradient elution from 2%-65%-95% B was applied in 0-2.5-3.5 min, followed by column equilibration, giving a total time of 4.5 min/injection. The eluent flow rate was 0.5 ml/min. Data were acquired using a Thermo TSQ Endura triple quadrupole MS. Multiple reaction monitoring (MRM) mode using positive ion mode. For all compounds, the spray voltage was 4500 V, vaporizer temperature and transfer tube temperature were 400°C and 350°C, respectively. The CID argon pressure was set to 2.0 mTorr. The MRM transitions were as previously described (Tolonen et al., 2007;Turpeinen et al., 2005).

| Prediction of possibility of in vivo herb-drug interactions from in vitro data
Determination of molar IC 50 values in herbal extracts is challenging thus representing the in vitro IC 50 values in Liter/dose as described by Strandell, Neil, and Carlin (2004) gives insight into possibility of in vivo inhibition. The volume to which an estimated human unit dose of the aqueous extract of each herb should be diluted in vivo to give similar in vitro inhibitory concentration for each CYP isoenzymes was calculated using the equation (2)
Aqueous extract of Mangifera indica stem bark showed weak inhibitory activities on CYP2A6, CYP2C19 and CYP3A4 but exhibited moderate inhibition of CYP1A2, CYP2B6, CYP2C8, CYP2C9, and CYP2D6 (Table 2). The plant extract displayed 100% inhibition of all the CYP isoenzymes investigated at a dose of 1000 μg/ml ( Figure 1b).
In order to estimate the possibility of in vivo inhibitory potential from the in vitro data, IC 50 in L/dose was calculated according  Figure 2d,e,f, respectively.

| D ISCUSS I ON
Drug-drug interaction or herb-drug interaction studies could be conducted as in vitro or in vivo studies, case reports, and human studies using human subjects. In vitro herb-drug interactions studies are mostly conducted for commonly used herbs to evaluate and predict potentially significant in vivo herb-drug interactions and to help design appropriate in vivo herb-drug interaction studies (Fasinu, Bouic, & Rosenkranz, 2012;Awortwe, Bouic, Masimirembwa, & Rosenkranz, 2013  study, only two water-soluble components of aged garlic were able to moderately inhibit CYP3A4 (Greenblatt et al., 2006). Also, aged garlic extract produced no significant inhibition of CYP isoenzymes in humans (Markowitz, 2003). Though we did not report any inhibition of CYP isoenzymes by Allium sativum, the discrepancies in these reports and ours may be due to the extraction procedure, assay method, concentration and type of the extract used and enzyme sources. In this study, aqueous extract of oven-dried Allium sativum bulbs was used.
There is no conclusive report of a significant in vitro inhibition of CYP isoforms by Allium sativum both from this study and others. Thus, the possibility of Allium sativum aqueous extract in vivo HDIs mediated by CYP isoenzymes may be remote as supported by our study and that of Markowitz (2003). were 3.12 L/dose and 1.43 L/dose, respectively. These are higher than 0.88 L/dose estimated by Strandell et al. (2004) as the minimum required to elicit further investigation of in vivo inhibitory activity of the plant extract on CYP isoenzyme. However, in vivo study in human showed that M. oleifera did not affect the pharmacokinetic parameters of Nevirapine, a substrate of CYP2C9, CYP2D6, and CYP3A5 (Monera-Penduka et al., 2017). Nonetheless, couse of Moringa oleifera leave extract with substrates of CYP1A2 and CYP2C9 should be further investigated for possibility of in vivo inhibition of these isoenzymes which may affect therapeutic outcome of coadministered medication.
Stem bark aqueous extract of Mangifera indica moderately inhibited CYP1A2, CYP2B6, CYP2C8, CYP2C9, and CYP2D6 (IC 50 < 100 μg/ml) while others CYPs such as 2A6, 2C19, and 3A4 were weakly inhibited (IC 50 > 100 μg/ml) by the extract. These results are similar to the inhibition of CYP1A2, CYP2C9, CYP2D6, and CYP3A4 by aqueous extract of Mangifera indica stem bark using human liver microsomes and primary hepatocytes reported by Rodeiro et al. (2008Rodeiro et al. ( , 2013. Mangifera indica fruit and stem bark extract are commonly used by patients because of the folkloric claims and documented pharmacological activities in ameliorating the signs and symptoms of diabetes, treatment of malaria fever, menorrhagia, hypertension, and hypercholesterolemia (Ezuruike & Prieto, 2014;Awortwe, 2015;Gururaja et al., 2017). Most of the drugs used in the treatment and management of these diseases are metabolized by the major CYP isoenzymes, namely CYP1A2, CYP2C9, CYP2C19, CYP2D6, and CYP3A4 (Ogu & Maxa, 2000;Lynch & Price, 2007). This interaction was suggested to be due to the inhibitory effect of Vitamin A, contained in the fruit, on CYP2C19-an enzyme responsible for the metabolism (7-hydroxylation) of the warfarin R-isomer (Yamazaki, 1999). Aqueous extract of Mangifera indica bark also showed a weak inhibitory effect on CYP2C19 in the present study.
The potential for Mangifera indica stem bark aqueous extract to cause in vivo HDIs specifically with drugs with narrow therapeutic index is supported by the estimated IC 50 values in L/dose which are higher than 5 L. As reported by Strandell et al. (2004), herb extracts with IC 50 in L/dose greater than 0.88 L may cause in vivo inhibition of same CYP isoenzymes similar to the observed in vitro inhibitory potential of the herb on the same CYP isoenzymes.
Several studies have used Strandell et al. (2004) methods to predict possibilities of in vivo inhibition of CYP isoenzymes from in vitro data.
For example, in vitro study by Sevior et al. (2010) predicted the possible absence of in vivo inhibition of CYP 2D6 and 3A4 by black cohosh and aqueous extract of valerian and inhibition of CYP2D6 by goldenseal.
Earlier studies by Gurley et al. (2005Gurley et al. ( , 2006Gurley et al. ( , 2008 confirmed these predictions. Strandell et al. (2004) predicted the likelihood of potent in vivo inhibition of CYP3A4 by hypericum (St John's Wort) preparations; this also confirmed the report of Markowitz et al. (2000).
Currently, no study seems to be available on the in vitro or in vivo inhibitory activity of six of the herbs studied-aqueous extracts of Picralima nitida seeds, Gongronema latifolium leaves, Tetracarpidium conophorum seeds, Musa sapientum unripe fruits, Bauhinia monandra leaves, and Alstonia boonei stem bark. Our study may be the first report of the in vitro inhibitory potential of these herb extracts.
Gongronema latifolium and Alstonia boonei may produce in vivo inhibitory activity on CYP1A2, CYP2C19 and CYP3A4 since the converted in vitro IC 50 values are greater than 0.88 L/dose despite the weak in vitro inhibitory potential on these isoenzymes. Aside from its medicinal properties, Gongronema latifolium is a popular delicacy in African cuisine (Akinsanmi & Nwanna, 2015;Nwanna et al., 2016 (Strandell et al., 2004).
Picralima nitida seem to be the most potent of the nine herbs studied with IC 50 < 100 μg/ml for CYP2C19 and CYP3A4 and IC 50 < 10 μg/ml for CYP2D6. The IC 50 value converted into L/dose for CYP2C19, CYP3A4, and CYP2D6 far exceeded the 0.88 L/ dose unit cutoff point by Strandell et al. (2004). With these values, this herb possesses the ability to elicit clinically significant in vivo HDIs, if used concomitantly with substrates of CYP2D6 such as amitriptyline, metoprolol, clozapine, flucainide, and retonavir (Lynch & Price, 2007). This isoenzyme is phenotypically expressed and the effect of HDIs may vary from one person to another. Thus, it is important to counsel patients not to coadminister this herb with any substrates of CYP2D6 until conclusive clinical data are available.
It should be borne in mind that the IC 50 values are apparent values because of the complex mixture of herbs. The preparation of the herb extracts and the dose used also vary between localities.
These may affect the computation of the IC 50 values in L/dose unit and should be considered when interpreting these results. Also, despite the use of human liver microsomes in the in vitro study, extrapolation of the in vitro findings to in vivo may be limited by the in vivo bioavailability of the phytochemicals in the herbs. In spite of these result limitations, this study assisted in rapidly identifying herb F I G U R E 2 Inhibition of cytochrome P450 isoezymes by herbs with IC 50 < 1,000 µg/ml-aqueous extracts of Alstonia boonei (

ACK N OWLED G M ENT
Authors would like to appreciate the kind gift of Picralima nitida seeds by Dr James Adetunji Bello, The Chief Medical Director, Oyo state Hospital, Ogbomoso.

D I SCLOS U R E
Authors have no conflict of interest to declare.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing.