Frontiers | An Overview of the Evidence and Mechanisms of Herb–Drug Interactions (original) (raw)

REVIEW article

Front. Pharmacol., 30 April 2012

Sec. Drugs Outcomes Research and Policies

1 Division of Pharmacology, Faculty of Health Sciences, University of Stellenbosch, Cape Town, South Africa
2 Division of Medical Microbiology, Faculty of Health Sciences, University of Stellenbosch, Cape Town, South Africa
3 Synexa Life Sciences, Montague Gardens, Cape Town, South Africa

Despite the lack of sufficient information on the safety of herbal products, their use as alternative and/or complementary medicine is globally popular. There is also an increasing interest in medicinal herbs as precursor for pharmacological actives. Of serious concern is the concurrent consumption of herbal products and conventional drugs. Herb–drug interaction (HDI) is the single most important clinical consequence of this practice. Using a structured assessment procedure, the evidence of HDI presents with varying degree of clinical significance. While the potential for HDI for a number of herbal products is inferred from non-human studies, certain HDIs are well established through human studies and documented case reports. Various mechanisms of pharmacokinetic HDI have been identified and include the alteration in the gastrointestinal functions with consequent effects on drug absorption; induction and inhibition of metabolic enzymes and transport proteins; and alteration of renal excretion of drugs and their metabolites. Due to the intrinsic pharmacologic properties of phytochemicals, pharmacodynamic HDIs are also known to occur. The effects could be synergistic, additive, and/or antagonistic. Poor reporting on the part of patients and the inability to promptly identify HDI by health providers are identified as major factors limiting the extensive compilation of clinically relevant HDIs. A general overview and the significance of pharmacokinetic and pharmacodynamic HDI are provided, detailing basic mechanism, and nature of evidence available. An increased level of awareness of HDI is necessary among health professionals and drug discovery scientists. With the increasing number of plant-sourced pharmacological actives, the potential for HDI should always be assessed in the non-clinical safety assessment phase of drug development process. More clinically relevant research is also required in this area as current information on HDI is insufficient for clinical applications.

Introduction

There is increasing consumptions of medicinal herbs and herbal products globally, cutting across social and racial classes, as it is observed both in developing and developed countries (Cheng et al., 2002; Bodeker, 2007; Mitra, 2007). Medicinal plants were the major agents for primary health care for many centuries before the advent of modern medicine (Sheeja et al., 2006). Their use however declined in most developed western countries during the last century’s industrialization and urbanization (Ogbonnia et al., 2008). In the past two decades however a new resurgence in medicinal plants consumption was observed. According to the WHO, about 70% of the world population currently uses medicinal herbs as complementary or alternative medicine (Wills et al., 2000). It is estimated that over 40% of the adult American population consume herbal products for one medical reason or the other (Tachjian et al., 2010). A recent study involving 2055 patients in the US also reveals that the consumption pattern of traditional medications has no significant gender or social difference (Kessler et al., 2001). Consumption rate has also been particularly exponential in Canada (Calixto, 2000), Australia (Bensoussan et al., 2004), as well as Europe where the highest sales of herbal products have been reported in Germany and France (Capasso et al., 2003). In Africa, there is continuous addition to the list of medicinal herbs while consumption rate is also increasing. Between 60 and 85% native Africans use herbal medicine usually in combination (Van Wyk et al., 2009).

The indications for herbal remedies are diverse as they are employed in the treatment of a wide range of diseases (Ernst, 2005). Studies have shown that 67% of women use herbs for perimenopausal symptoms, 45% use it in pregnancy, and more than 45% parents give herbal medications to their children for various medical conditions (Ernst, 2004). Regulations in most countries do not require the demonstration of therapeutic efficacy, safety, or quality on the part of herbal remedies as most of them are promoted as natural and harmless (Homsy et al., 2004; Routledge, 2008). It is pertinent however, that herbs are not free from side effects as some have been shown to be toxic (Déciga-Campos et al., 2007; Patel et al., 2011). Recent study has shown habitual pattern of concomitant consumption of herbal and prescription medication. Kaufman et al. (2002) reported that 14–16% of American adult population consume herbal supplements often concomitantly with prescribed medications. Also, 49.4% of Israeli consumers of herbal remedies use them with prescription drugs (Giveon et al., 2004). This is significant bearing in mind that less than 40% of patients disclose their herbal supplement usage to their health care providers coupled with the fact that many physicians are unaware of the potential risks of herb–drug interactions (HDI; Klepser et al., 2000).

HDI is one of the most important clinical concerns in the concomitant consumption of herbs and prescription drugs. The necessity of polypharmacy in the management of most diseases further increases the risk of HDI in patients. The ability of intestinal and hepatic CYP to metabolize numerous structurally unrelated compounds, apart from being responsible for the poor oral bioavailability of numerous drugs is responsible for the large number of documented drug–drug and drug–food interactions (Quintieri et al., 2008). This is more so, considering that oral drug delivery is the most employed in the management of most disease conditions in which case, drug interaction alters both bioavailability and pharmacokinetic disposition of the drug. This alteration and the resulting poor control of plasma drug concentrations would particularly be of concern for drugs that have a narrow therapeutic window or a precipitous dose–effect profile (Aungst, 2000; Perucca, 2006). The risk of pharmacokinetic drug interaction poses two major extremity challenges – pharmacotoxicity and treatment failure. The former can result from the inhibition of the metabolic enzymes responsible for the metabolism and clearance of the drugs while the latter may be the consequence of enzymatic induction leading to faster drug metabolism. This is in addition to the intrinsic pharmacodynamic actions of the herbal products themselves which may include potentiating, additive, antagonism, or neutralization effects.

Until recently, HDI was often unsuspected by physicians for several reasons. Most trained physicians lack adequate knowledge on herbal drugs and their potentials for drug interactions (Clement et al., 2005; Ozcakir et al., 2007; Fakeye and Onyemadu, 2008); herbal products also vary considerably in compositions depending on the source and package (Liang et al., 2004; Sousa et al., 2011); most patients do not consider it necessary to disclose their herbal consumptions to physicians who themselves hardly inquire such (Cassidy, 2003; Howell et al., 2006; Chao et al., 2008; Kennedy et al., 2008). Further challenges with herbal medications include scientific misidentification, product contamination and adulteration, mislabeling, active ingredient instability, variability in collection procedures, and failure of disclosure on the part of patients (Boullata and Nace, 2000). A fairly recent systematic review by Izzo and Ernst (2009) on the interactions between medicinal herbs and prescribed medications provide some more details on these.

Herbal products are made of complex mixture of pharmacologically active phytochemicals (Mok and Chau, 2006), most of which are secondary metabolites generated through the shikimate, acetate–malonate, and acetate–mevalonate pathways. These constituents include phenolics (such as tannins, lignins, quinolones, and salicylates), phenolic glycosides (such as flavonoids, cyanogens, and glucosinolates), terpenoids (such as sesquiterpenes, steroids, carotenoids, saponins, and iridoids), alkaloids, peptides, polysaccharides (such as gums and mucilages), resins, and essential oils which often contain some of the aforementioned classes of phytochemicals (Wills et al., 2000; Wang et al., 2008). This complexity increases the risk of clinical drug interactions.

Aim, Search Strategy, and Selection Criteria

The current review was therefore aimed at providing an overview of known and recently reported HDI with interest in the evidence available and the mechanism thereof. The review was systematically conducted by searching the databases of MEDLINE, PUBMED, EMBASE, and COCHRAINE libraries for original researches, and case reports on HDI using the following search terms or combinations thereof: “drug–herb,” “herb–drug,” “interaction,” “cytochrome P450,” “plant,” “extract,” “medicinal,” “concomitant administration,” “herbal and orthodox medicines.” Relevant search terms were employed to accommodate the various individual medicinal herbs employed in Africa, America, Asia, Europe, and Australia. The reported interactions and their mechanisms, with orthodox medications were searched and collated. Searches were not limited by date or place of publications but to publications available in English language.

Results

Clinical Presentation of Herb–Drug Interactions

Clinical presentations of HDI vary widely depending on the herbs and the drugs concerned. Typical clinical presentation of HDI include the potentiation of the effects of oral corticosteroids in the presence of liquorice (Glycyrrhiza glabra; Liao et al., 2010); potentiation of warfarin effects with resultant bleeding in the presence of garlic (Allium sativum; Borrelli et al., 2007), dong quai (Angelica sinensis; Nutescu et al., 2006), or danshen (Salvia miltiorrhiza; Chan, 2001); decreased blood levels of nevirapine, amitriptyline, nifedipine, statins, digoxin, theophylline, cyclosporine, midazolam, and steroids in patients concurrently consuming St John’s wort (SJW; Hypericum perforatum; De Maat et al., 2001; Henderson et al., 2002; Johne et al., 2002; Mannel, 2004; Borrelli and Izzo, 2009), decreased oral bioavailability of prednisolone in the presence of the Chinese herbal product xiao-chai-hu tang (sho-saiko-to; Fugh-Berman, 2000); ginseng (Panax ginseng)-induced mania in patients on antidepressants (Engelberg et al., 2001); production of extrapyramidal effects as a result of the combination of neuroleptic drugs with betel nut (Areca catechu; Huang et al., 2003; Coppola and Mondola, 2012); increased blood pressure induced by tricyclic antidepressant-yohimbe (Pausinystalia yohimbe) combination (Tam et al., 2001), increased phenytoin clearance and frequent seizures when combined with Ayurvedic syrup shankhapushpi (Patsalos and Perucca, 2003), among other clinical manifestations. These clinical presentations depend on the mechanism of HDI.

Evidence-Based HDI Studies and Clinical Relevance

Herb–drug interactions have been reported through various study techniques. While these reports usually give evidence of potential interactions, the level of evidence varies often failing to predict the magnitude or clinical significance of such HDI. Apart from the specific limitations attributable to study methods employed, major draw-back in deducting relevant conclusions from reported HDI include misidentification and poor characterization of specimen, presence and nature of adulterants (some of which may be allergens), variations in study methodologies including extraction procedures, source location of herbs involved, seasonal variation in the phytochemical composition of herbal materials, under-reporting and genetic factors involved in drug absorption, metabolism, and dynamics. Table 1 provides some limitations of the study methods.

Table 1. Comparison of study methods available for HDI.

Recently, structured assessment procedures are emerging in an attempt to provide levels of evidence for drug interactions. In addition to evidence of interaction, such assessment take into consideration clinical relevance of the potential adverse event resulting from the interaction, the modification- and patient-specific risk factors, and disease conditions for which the interaction is important. Van Roon et al. (2005) developed a system of hierarchical evidence-based structured assessment procedure of drug–drug interaction. This can be applicable to HDI. This method particularly allows the extraction of HDIs that have been well established and those that are merely inferred from certain phytochemical characteristics. A modified form of this method as presented in Table 2 is applied in this paper to provide the nature and level of evidence for the HDIs mentioned.

Table 2. Quality of HDI evidence for clinical risk assessment.

Mechanisms of Herb–Drug Interactions

The overlapping substrate specificity in the biotransformational pathways of the physiologic systems is seen as the major reason for drug–drug, food–drug, and HDI (Marchetti et al., 2007). The ability of different chemical moieties to interact with receptor sites and alter physiological environment can explain pharmacodynamic drug interactions while pharmacokinetic interactions arise from altered absorption, interference in distribution pattern as well as changes and competition in the metabolic and excretory pathways (Izzo, 2005). The major underlying mechanism of pharmacokinetic HDI, like drug–drug interaction, is either the induction or inhibition of intestinal and hepatic metabolic enzymes particularly the CYP enzyme family. Additionally, similar effect on drug transporters and efflux proteins particularly the p-glycoproteins in the intestines is responsible in most other cases (Meijerman et al., 2006; Nowack, 2008; Farkas et al., 2010). The pre-systemic activity of CYP and efflux proteins often influence oral bioavailability, thus the modulating activity of co-administered herbal products has been shown to result in pronounced reduction or increase in the blood levels of the affected drugs (Brown et al., 2008).

Potential for in vivo drug interactions are often inferred from in vitro studies with liver enzymes. The correlation of in vitro results with in vivo behavior has yielded reliable results in certain cases in terms of in vivo predictability although the extent of clinical significant is poorly inferable (Rostami-Hodjegan and Tucker, 2007; Iwamoto et al., 2008; Xu et al., 2009; Umehara and Camenisch, 2011). Thus most of the well established HDIs, as will be seen in subsequent sections, were initially demonstrated through in vitro studies.

The interaction of herbal products with hepatic enzymes can also result in pharmacodynamic effects (van den Bout-van den Beukel et al., 2008; Nivitabishekam et al., 2009; Asdaq and Inamdar, 2010; Dasgupta et al., 2010; Kim et al., 2010a.) Specific liver injury inducible by phytochemical agents includes elevation in transaminases (Zhu et al., 2004; Saleem et al., 2010), acute and chronic hepatitis (Stedman, 2002; Pierard et al., 2009), liver failure (Durazo et al., 2004), veno-occlusive disorders (DeLeve et al., 2002), liver cirrhosis (Lewis et al., 2006), fibrosis (Chitturi and Farrell, 2000), cholestasis (Chitturi and Farrell, 2008), zonal or diffusive hepatic necrosis (Savvidou et al., 2007), and steatosis (Wang et al., 2009). Mechanism of liver injury may include bioactivation of CYP, oxidative stress, mitochondrial injury, and apoptosis (Cullen, 2005).

Induction and inhibition of metabolic enzymes

The CYP superfamily is generally involved in oxidative, peroxidative, and reductive biotransformation of xenobiotics and endogenous compounds (Nebert and Russell, 2002; Hiratsuka, 2011). It is conventionally divided into families and subfamilies based on nucleotide sequence homology (Fasinu et al., 2012). There is a high degree of substrate specificity among the various families. CYP belonging to the families 1, 2, and 3 are principally involved in xenobiotic metabolism while others play a major role in the formation and elimination of endogenous compounds such as hormones, bile acids, and fatty acids (Norlin and Wikvall, 2007; Amacher, 2010). The most important CYP subfamilies responsible for drug metabolism in humans are 1A2, 2A6, 2C9, 2C19, 2D6, 2E1, 3A4, and 3A5 (Ono et al., 1996; Wang and Chou, 2010).

CYP1A1 and 1A2 are the two major members of the human CYP1A subfamily. CYP 1A1 is mainly expressed in extra-hepatic tissues such as the kidney, the intestines, and the lungs while CYP1A2 constitutes about 15% of total hepatic CYP (Martignoni et al., 2006). CYP2B6 is involved in drug metabolism while most other members of the CYP2B subfamily play less significant metabolic roles (Pavek and Dvorak, 2008). The subfamily 2C is the second most abundant CYP after 3A representing over 20% of the total CYP present in the human liver. It comprises three active members: 2C8, 2C9, and 2C19 all of which are also involved in the metabolism of some endogenous compounds including retinol and retinoic acid (Lewis, 2004). Few clinically relevant drugs including paracetamol, chlorzoxazone, and enflurane are metabolized by CYP2E1, the most active of the 2E subfamily (Leclercq et al., 2000). CYP3A subfamily constitutes over 40% of the total CYP in the human body (although the levels may vary 40-fold among individuals) with CYP3A4 being the most abundant of all isoforms highly expressed in the liver and the intestines and participates in the metabolism of about half of drugs in use today (Ferguson and Tyndale, 2011; Singh et al., 2011). The specificity and selectivity of substrates and inhibitors for these enzymes are particularly useful in pharmacokinetic and toxicological studies.

Induction is the increase in intestinal and hepatic enzyme activity as a result of increased mRNA transcription leading to protein levels higher than normal physiologic values. When this happens, there is a corresponding increase in the rate of drug metabolism affecting both the oral bioavailability and the systemic disposition. In the formulation and dosage design of oral medications, allowance is often made for pre-systemic metabolism in order to achieve predictable systemic bioavailability. A disruption in this balance can result in significant changes in blood concentrations of the drugs. Certain herbal products have been shown to be capable of inducing CYP. Concomitant administration of enzyme-inducing herbal products and prescription drugs can therefore result in sub-therapeutic plasma levels of the latter with therapeutic failure as a possible clinical consequence.

Apart from enzyme induction, herbal products can also inhibit enzyme activities. The inhibition of CYP and other metabolic enzymes is usually competitive with instantaneous and inhibitor concentration-dependent effects (Zhang and Wong, 2005). Most inhibitors are also substrates of CYP (Zhou, 2008). This phenomenon alters pharmacokinetic profiles of xenobiotics significantly. As a result of the suppression of the anticipated pre-systemic intestinal and hepatic metabolism, unusually high plasma levels of xenobiotics are observed. Toxic manifestation could be the ultimate effect of this observation. An equally clinically important consequence of enzyme inhibition is drug accumulation due to subdued hepatic clearance. These effects will be of particular concerns in drugs with narrow therapeutic window or steep dose–response profiles.

St John’s wort is one of the most widely used herbal antidepressants (Lawvere and Mahoney, 2005; Høyland, 2011). It is a potent inducer of CYP3A4 and depending on the dose, duration and route of administration, it may induce or inhibit other CYP isozymes and P-gp (Roby et al., 2000; Markowitz et al., 2003b; Tannergren et al., 2004; Madabushi et al., 2006). Studies from case reports indicate that, due to its inducing effects on CYP3A4, it significantly reduces the plasma levels of CYP3A4 substrates including cyclosporine, simvastatin, indinavir, warfarin, amitriptyline, tacrolimus, oxycodone, and nevirapine (Henderson et al., 2002; Johne et al., 2002; Nieminen et al., 2010; Vlachojannis et al., 2011). It has also been reported that the alteration in the blood serum concentration of cyclosporine due to SJW has led to organ rejection in patients (Ernst, 2002; Murakami et al., 2006). Reports of breakthrough bleeding and unplanned pregnancies due to interaction between SJW and oral contraceptives have also been documented (Hu et al., 2005). The group of drugs with the highest potential for clinically significant pharmacokinetic drug interaction with SJW is the antidepressants as SJW itself is consumed by patients with depression. Its concomitant use with SSRI like sertraline and paroxetine has been reported to result in symptoms of central serotonergic syndrome (Barbenel et al., 2000; Dannawi, 2002; Spinella and Eaton, 2002; Birmes et al., 2003; Bonetto et al., 2007). It has also been said to increase the incidence of hypoglycemia in patients on tolbutamide without apparent alteration in the pharmacokinetic profile of tolbutamide (Mannel, 2004). It also inhibits the production of SN-38, an active metabolite of irinotecan, in cancer patients.

Amitriptyline is a substrate to both CYP3A4 and intestinal P-gp. The risk of therapeutic failure is thus high due to induction of CYP3A4-dependent metabolism activities resulting in poor oral bioavailability. In a study by Johne et al. (2002), a 21% decrease in the area under the plasma concentration–time curve of amitriptyline was observed in 12 depressed patients who were concomitantly administered with extracts of SJW and amitriptyline for 2 weeks.

Other CYP and P-gp substrates whose pharmacokinetic profile have been reportedly altered by SJW include anticoagulants like phenprocoumon and warfarin; antihistamines like fexofenadine; antiretroviral drugs including protease inhibitors and reverse transcriptase inhibitors; hypoglycemic agents such as tolbutamide; immunosuppressants like cyclosporine, tacrolimus, and mycophenolic acid; anticonvulsants such as carbamazepine; anti-cancer like irinotecan; bronchodilators like theophylline; antitussive like dextromethorphan; cardiovascular drugs like statins, digoxin, and dihydropyridine calcium channel blockers; oral contraceptives; opiates like methadone and loperamide; and benzodiazepines including alprazolam and midazolam (Greeson et al., 2001; Di et al., 2008; Hojo et al., 2011). Following a single dose administration of 300 mg standardized extracts of SJW containing 5% hyperforin in humans, a maximum plasma concentration of 0.17–0.5 μM hyperforin yielding a [_I_]/_K_i > 0.22, in vivo extrapolation suggests a high possibility of in vivo pharmacokinetic drug interaction (Agrosi et al., 2000). Bray et al. (2002) confirmed through animal studies that SJW modulates various CYP enzymes. Dresser et al. (2007) demonstrated that SJW is capable of inducing CYP3A4 in healthy subjects through the observation of increased urinary clearance of midazolam. Thus animal and human studies further confirm SJW as containing both inhibitory and inducing constituents on various CYP isozymes. These effects may depend on dosage and duration of administration, and may also be species- and tissue-specific. While the individual phytochemical constituents of SJW have elicited varying effects on the metabolic activity of the CYP isozymes, whole extracts and major constituents especially hyperforin have been reported to inhibit the metabolic activities of CYP1A2, 2C9, 2C19, 2D6, and 3A4 via in vitro studies and in vivo studies (Lee et al., 2006; Madabushi et al., 2006; Hokkanen et al., 2011).

Ginkgo biloba have been reported to induce CYP 2C19-dependent omeprazole metabolism in healthy human subjects (Yin et al., 2004). Piscitelli et al. (2002) in a garlic–saquinavir interaction study reported 51% decrease in saquinavir oral bioavailability caused by the presence of garlic and attributable to garlic-induced CYP3A4 induction. Its effects on the warfarin pharmacokinetic has also been reported in animal models (Taki et al., 2012).

Although grapefruit juice is not consumed for medicinal purposes, the discovery of the inhibitory activity of its flavonoid contents on CYP has led to further researches in medicinal herbs which have revealed HDI potentials in flavonoid-containing herbal remedies (Choi and Burm, 2006; Palombo, 2006; Paine et al., 2008; Quintieri et al., 2008; Alvarez et al., 2010). A related CYP inhibitor is rotenone. By interfering with the electron transfer of the heme iron, rotenone, a naturally occurring phytochemical found in several plants such as the jicama vine plant is known to inhibit CYP activity (Sanderson et al., 2004). Resveratrol, a natural polymer, and tryptophan, an amino acid have been documented as potent CYP inhibitors (Rannug et al., 2006). Some herbal medications and their phytochemical constituents capable of interacting with CYP are presented in Table 3. A more detailed involvement of CYP in HDI is detailed in some recently published reviews (Delgoda and Westlake, 2004; Pal and Mitra, 2006; Cordia and Steenkamp, 2011; Liu et al., 2011).

Table 3. Some herbal products known to interact with CYP and efflux proteins.

Phase II metabolic enzymes including uridine diphosphoglucuronosyl transferase (UGT), _N_-acetyl transferase (NAT), glutathione _S_-transferase (GST), and sulfotransferase (ST) catalyze the attachment of polar and ionizable groups to phase I metabolites aiding their elimination. While cytochrome P450-mediated HDI have been extensively investigated in various studies, the effects of herbal extracts on phase II enzymes have not been adequately studied. However, there is sufficient evidence in literature to suggest the potentials of phase II enzymes to induce clinically significant HDI.

In a study carried out in rat models by Sheweita et al. (2002), extracts of hypoglycemic herbs, Cymbopogon proximus, Zygophyllum coccineum, and Lupinus albus reduced the activity of GST and GSH. Curcumin, from Curcuma longa, an herbal antioxidant with anti-inflammatory and antitumor properties increased the activity of GST and quinone reductase in the ddY mice liver (Iqbal et al., 2003). Valerian, an herbal sleeping aid has also demonstrated the potential of inducing HDI through the inhibition of UGT. Up to 87% of inhibition of UGT activity by valerian extract was reported in an in vitro study utilizing estradiol and morphine as probe substrate (Alkharfy and Frye, 2007). Kampo, a traditional Japanese medicine made of a mixture of several medicinal herbs has shown inhibitory effects on some phase II enzymes. In an in vitro study by Nakagawa et al. (2009), nine out of 51components of kampo medicine elicited more than 50% inhibition of UGT2B7-mediated morphine 3-glucuronidation. In the same study, extracts of kanzo (Glycyrrhizae radix), daio (Rhei rhizoma), and keihi (Cinnamomi cortex) elicited more than 80% inhibition of morphine AZT glucuronidation. This result agrees with Katoh et al. (2009) who carried out similar studies on rhei, keihi, and ogon (Scutellariae radix).

Apart from the well-known effects on Ginkgo biloba on CYP enzymes as illustrated earlier, its extracts have demonstrated potent inhibition of mycophenolic acid glucuronidation investigated in human liver and intestinal microsomes (Mohamed and Frye, 2010).

In a study to investigate the influence of 18 herbal remedies on the activity of human recombinant sulfotransferase 1A3 employing dopamine and ritodrine as substrates, extracts of grape seed, milk thistle, gymnema, SJW, ginkgo leaf, banaba, rafuma, and peanut seed coat showed potent inhibition with IC50 values lower than putative gastrointestinal concentration (Nagai et al., 2009). Similarly, Mohamed and Frye (2011b) reported the inhibition of UGT1A4 by green tea derived epigallocatechin gallate; UGT 1A6 and UGT1A9 by milk thistle; UGT 1A6 by saw palmetto; and UGT 1A9 by cranberry. A recent publication presents evidence of potential HDI mediated by UGT (Mohamed and Frye, 2011a).

Certain phytochemicals including coumarin, limettin, auraptene, angelicin, bergamottin, imperatorin, and isopimpinellin have also been reported to be capable of inducing hepatic GST activities (Kleiner et al., 2008). While the clinical significance of these findings are yet to be determined, it is noteworthy that phase II metabolic enzymes may play significant roles in HDIs.

Inhibition and induction of transport and efflux proteins

The ATP-binding cassette (ABC) family of drug transporters plays significant roles in the absorption, distribution, and elimination of drugs. P-gp, the most studied member of this family is a 170-kDa plasma glycoprotein encoded by the human MDRI gene. It is constitutively expressed in a number of body tissues and concentrated on the apical epithelial surfaces of the bile canaliculi of the liver, the proximal tubules of the kidneys, the pancreatic ductal cells, the columnar mucosal cells of the small intestine, colon, and the adrenal glands (Marzolini et al., 2004; Degorter et al., 2012). It is actively involved in drug absorption and elimination from the intestines the liver, kidneys, and the brain. Specifically these proteins are involved in the processes of hepatobiliary, direct intestinal, and urinary excretion of drugs and their metabolites (Szakács et al., 2008). Thus, the modulation of P-gp, or competitive affinity as substrates for its binding sites by co-administered herbs presents a potential for alteration in the pharmacokinetic profile of the drug.

Pharmacokinetic interaction occurs when herbal drugs inhibit or decrease the normal activity level of drug transporters through a competitive or non-competitive mechanism. Interactions can also occur through the induction of transport proteins via the increase of the mRNA of the relevant protein. Studies have identified a number of clinically important P-gp inhibitors including phytochemicals – flavonoids, furanocoumarins, reserpine, quinidine, yohimbine, vincristine, vinblastine among others (Krishna and Mayer, 2001; Zhou et al., 2004; Patanasethanont et al., 2007; Iwanaga et al., 2010; Eichhorn and Efferth, 2011; Yu et al., 2011). Borrel et al. (1994) reported that mobile ionophores such as valinomycin, nonactin, nigericin, monensin, calcimycin, and lasalocid inhibit the efflux of anthracycline by P-gp whereas channel-forming ionophores such as gramicidin do not (Larsen et al., 2000). A number of herbal products which interact with CYP also have similar effects on transport proteins (Table 3). The transport proteins are actively involved in the pharmacokinetics of anti-cancer drugs and account for one of the well-known mechanisms of multiple resistance of cancerous cells to chemotherapeutic agents (Bebawy and Sze, 2008; Bosch, 2008; He et al., 2011). The influence of some herbs on transport proteins is presented in Table 4. Clinically relevant interactions between herbal medicine and chemotherapeutic agents are detailed in a recent review by Yap et al. (2010).

Table 4. Influence of herbal products on transport proteins.

Alteration of gastrointestinal functions

Besides their influence on the intestinal metabolic enzymes and efflux proteins, herbal medications can alter the absorption of concomitantly administered medicines through a number of mechanisms. Changes in the gastrointestinal pH and other biochemical factors can alter dissolution properties and the absorption of pH-dependent drugs such as ketoconazole and itraconazole. Complexation and chelation, leading to the formation of insoluble complexes and competition at the sites of absorption especially with site-specific formulations can greatly affect the absorption of medicines. Anthranoid-containing plants – cassia (Cassia senna), Cascara (Rhamnus purshiana), rhubarb (Rheum officinale), and soluble fibers including guar gum and psyllium can decrease drug absorption by decreasing GI transit time. They are known to increase GIT motility. On concomitant use with prescribed medication, significant alteration in the absorption of the latter has been reported due to decreased GI transit time (Fugh-Berman, 2000).

Izzo et al. (1997) demonstrated that anthranoids could be harmful to the gut epithelium by inhibiting Na+/K+ ATPase and increasing the activity of nitric oxide synthase. This significantly increased intestinal transit due to the alteration in the intestinal water and salt absorption and the subsequent fluid accumulation. In a study conducted by Munday and Munday (1999), a garlic-derived compound was shown to increase the tissue activities of quinone reductase and glutathione transferase in the gastrointestinal tract of the rat. In view of their roles in metabolism, both enzymes are considered chemoprotective especially from chemical carcinogens. In addition to CYP and P-gp mediated mechanisms, the well-known ginseng-induced pharmacokinetic HDI may also be due to its gastrointestinal effects especially its inhibitory effects on gastric secretion (Suzuki et al., 1991). The potential of rhein and danthron to increase the absorption of furosemide, a poorly water-soluble drug, has been demonstrated through in vitro studies (Laitinen et al., 2007). In a study carried out on mice, a Chinese herbal plant, Polygonum paleaceum, showed the potential to depress the motility of the gastrointestinal tract, inhibit defecation reflex and delay gastric emptying (Zhang, 2002). A similar study demonstrated the inhibitory effects of two Chinese traditional herbal prescriptions, Fructus aurantii immaturus and Radix paeoniae alba on gastrointestinal movement (Fang et al., 2009).

The absorption of drugs such as phenoxymethylpenicillin, metformin, glibenclamide, and lovastatin may be reduced by high-fiber herbal products through the sequestration of bile acids (Colalto, 2010). Mochiki et al. (2010) reported the ability of Kampo, a traditional Japanese medicine, to stimulate elevated intestinal blood flow, and to induce increased secretion of gastrointestinal hormones including motilin, vasoactive intestinal peptide, and calcitonin gene-related peptide. Similarly, another traditional Japanese medicine has been shown to increase the intestinal secretion of ghrelin, a hunger-related hormone, leading to delayed gastric emptying (Tokita et al., 2007; Kawahara et al., 2009; Hattori, 2010; Matsumura et al., 2010). Also, Qi et al. (2007) demonstrated the capability of Da-Cheng-Qi-Tang, a traditional Chinese herbal formula, to increase plasma motilin, enhance gastrointestinal motility, improve gastric dysrhythmia, and reduce gastroparesis after abdominal surgery. These effects have the potential of reducing the intestinal transit time of concurrently administered drug, with the risk of reduced absorption.

Alteration in renal elimination

This involves herbal products capable of interacting with renal functions, leading to altered renal elimination of drugs. Such interaction can result from the inhibition of tubular secretion, tubular reabsorption, or interference with glomerular filtration (Isnard et al., 2004). In addition to this group of herbal products are those products consumed as diuretics. The mechanism of herbal diuresis is complex and non-uniform. Certain herbs increase the glomerular filtration rate but do not stimulate electrolyte secretion while some others act as direct tubular irritants (Crosby et al., 2001; Al-Ali et al., 2003). Some herbs capable of interacting with renal functions and drug elimination are presented in Table 5.

Table 5. Some herbal remedies capable of interacting with other drugs via alteration in renal functions.

Pharmacodynamic synergy, addition, and antagonism

Herb–drug interaction can occur through the synergistic or additive actions of herbal products with conventional medications as a result of affinities for common receptor sites (Ma et al., 2009). This can precipitate pharmacodynamic toxicity or antagonistic effects (Table 6). Like most other herbs, SJW contains complex mixture of phytochemicals including phenylpropanes, naphthodanthrones, acylphloroglucinols, flavonoids, flavanol glycosides, and biflavones. Hyperforin is known to inhibit the reuptake of neurotransmitters (dopamine, serotonin, noradrenalin) and is believed to be the bioactive responsible for the antidepressant activity of SJW.

Table 6. Some examples of pharmacodynamic interactions between herbal products and conventional drugs.

Conclusion

Concomitant use of herbs and conventional drugs may present with untoward events. Evidence available in literature indicates various mechanisms through which this can occur. By interacting with conventional medication, herbal remedies may precipitate manifestations of toxicity or in the other extreme, therapeutic failure. A good knowledge of the potential of commonly consumed herbal medicines to interact with prescription medicines, irrespective of the nature of evidence available, will equip health professionals in their practice. Apart from those demonstrated in significant number of human subjects, not all reported HDIs are clinically significant. As such, more clinically relevant research in this area is necessary. This review provides information on commonly used herbs and their potentials for HDI within the levels of evidence currently available.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

The authors will like to acknowledge the support of HOPE Kapstadt-Stiftung (HOPE Cape Town) and the Stellenbosch University Rural Medical Education Partnership Initiative (SURMEPI) for providing funds for this study.

References

Abii, T. A., and Onuoha, E. N. (2011). The chemical constituents of the leaf of Aspilia africana as a scientific backing to its tradomedical potentials. Agric. J. 6, 28–30.