Assessment of the in vitro potential of natural products to interact with drugs especially when taken concurrently has important implications for predicting the likelihood of natural product-drug interactions or a possible mechanism of interaction. Garcinia kola seed, also commonly called bitter kola, is a natural product relevant in Africa traditional medicine, as well as in cultural and social ceremonies in many parts of West and Central Africa [1]. The seeds or seed extracts (alone or in combination with other phytochemicals) are available as proprietary dietary supplements [2]. Owing to its acclaimed health benefits relevant to the management and chemoprevention of some life-threatening diseases related to the liver and lungs, and as an anti-infective agent, the seeds are usually eaten as snack [1,2].

In spite of being a common masticatory nut, not much has been reported on its possible interaction with prescription drugs in humans. Previous in vitro and in vivo studies have suggested that G. kola seeds or its extracts have the potential to modulate the activities of some drug metabolising enzymes [3-5]. Some authors have reported that G. kola reduced ciprofloxacin absorption in vitro [6], and reduced both the C_max and AUC when co-administered with ofloxacin in humans [4,5]. Some authors have also reported a herb - drug interaction in which a Kampo preparation which contains various herbal medicines and metal cations such as calcium, magnesium and aluminum caused significant reduction in plasma tetracycline and ciprofloxacin concentrations in healthy volunteers through complex formation [8,9].

A significant pharmacokinetic interaction between quinine (an antimalarial medicine) and G. kola in healthy volunteers has earlier been reported and the results were suggestive of interference with quinine absorption since parameters indicative of interference with elimination were not significantly altered [7]. This report, coupled with the fact that knowledge of the mechanism by which food-drug interactions occur may contribute to predicting and preventing unwanted alterations in the pharmacokinetic profile of drugs [10], formed the rationale for the present study which investigated a possible mechanism of interaction between quinine and G. kola through in vitro interaction experiments.

Chemicals and reagents

Quinine sulphate powder was purchased from BDH Chemicals (Poole, UK). HPLC grade acetonitrile, methanol, perchloric acid, hydrochloric acid, potassium phosphate monobasic, and sodium hydroxide were from Sigma-Aldrich (Steinheim, Germany). Fresh Garcinia kola seeds were purchased locally from a retail market in Ile-Ife, Nigeria and were identified by Mr G. Ademoriyo, a taxonomist in the Herbarium of the Department of Botany, Obafemi Awolowo University, Nigeria. A voucher specimen (voucher number IFE-17478) was deposited at the Herbarium.

HPLC analysis of quinine

Quinine analysis was performed using a validated HPLC method as previously reported [7]. The HPLC equipment (Agilent 1200 series, Agilent Technologies, Santa Clara, California) was fitted with an isocratic pump (model G1341A) coupled to a variable wavelength detector [Agilent Technologies; standard version (model G1341B)]. Mobile phase consisted of methanol: acetonitrile: 0.02 M potassium dihydrogen phosphate (volume ratio 15:15:70), and 74 mmol/L perchloric acid (0.64 ml), adjusted to pH 2.6 with 10 M NaOH. It was pumped through the column (Eclipse XDB-C18 reverse phase; 5μm particle size and 150 x 4.6 mm, i.d., Agilent Technologies, Santa Clara, California) at a flow rate of 1.6 ml/min at ambient temperature while monitoring the effluent at 254 nm. Quinine stock (1 mg/ml) was prepared by dissolving 50 mg of the pure quinine powder in 50 ml of 0.1 M HCl. Further dilutions were made to obtain calibration concentration ranging from 2.5 - 80 µg/ml. A volume of 20 µl of each calibration concentration was injected in triplicate, and the mean of the peak area was used to generate a calibration curve.

Preparation of G. kola suspension

The husks of G. kola seeds (50 g) were removed and the seeds were sliced into small bits, oven-dried at 40 ^oC for about 24 h and subsequently blended into coarse powder in an electric blender. Thereafter, the  powder was transferred into a beaker and a small quantity of  0.02 M KH₂PO₄ (pH 5.8) buffer solution was added to make a paste which was later transferred into a 2 L volumetric flask and made up to 2 L with the same buffer solution resulting in 2.5 % w/v G. kola suspension.

Determination of equilibration time for adsorption of quinine on G. kola

Previous methods [11-13] on adsorption of drug by antacids and G. kola were adapted. Stock solution of quinine (8 ml) containing 1 mg/ml of quinine was pipetted into a 100 ml volumetric flask and made up to mark with the 2.5 % w/v suspension of G. kola. This was mixed in a vortex mixer (Parloworld Scientific Ltd, Straffordshire, UK) and incubated in a water-bath with a shaker at 37 ± 0.1 °C.  Thereafter, the suspension (5 ml) was taken at 0.25, 0.5, 1, 2, 3, 4, 5, 6 and 8 h and centrifuged at 3000 g for 5 min. The concentrations of quinine in the supernatants were then determined using HPLC. From the peak area obtained at the various time points, the corresponding concentrations of the drug in the supernatants were estimated from a calibration curve. Subsequently adsorption equilibration time was determined from a plot of the concentration of the quinine in supernatant against time.

Adsorption of quinine on G. kola

Duplicate series of 100 ml flasks containing aliquots (0.25, 0.5, 1.0, 2.0 and 4.0 ml) of 1 mg/ml quinine solution were prepared. G. kola suspension (2.5 % w/v) in 0.02 M KH₂PO₄ buffer was added to the respective flask to a final volume of 100 ml mixture. All flasks were incubated in a thermostated water bath with shaker at 37 ± 0.1 °C for 5 h (equilibration experiment showed that after 5 h, no further adsorption changes occurred in the quinine - G. kola suspension system). Thereafter, 5 ml was taken from each of the flasks, centrifuged and the concentration of quinine in the respective supernatant determined using HPLC. The amount of quinine adsorbed was determined by subtracting the amount of quinine in supernatant from the initial amount of quinine in the mixture. Adsorption percentage (%) was derived from the difference of the initial concentration of quinine (C_o) and equilibrium concentration of quinine in the supernatant (C_e) using the equation:    

Adsorption %= (C_o-C_e)/C_o × 100 % ……….. (1)

An adsorption isotherm was also generated.  The adsorption isotherm of quinine on G. kola at 37 ± 0.1 °C expressed by a double logarithm plot based on Freundlich model was derived from the equation:

log x/m = log K + 1/n (logC_e) ……………….. (2)

where x is the mass of the quinine adsorbed by m grams of G. kola (adsorbent), K is the Freundlich constant representing the quinine adsorbed per gram of the G. kola at a µg/ml quinine concentration, log K is the intercept, 1/n is the slope which represents the amount of quinine adsorbed for a given concentration change and C_e is as previously defined.

Desorption experiments

Elution of the adsorbed quinine by G. kola was performed by preparing a concentration of 80 μg/ml solution of quinine in 2.5 % w/v G. kola suspension (100ml). This was equilibrated in a thermostated water bath at 37 ± 0.1 °C for 5 h. Thereafter, the suspension was centrifuged, and the concentration of quinine in the supernatant measured using HPLC. Quinine concentration in the sediment was calculated from a difference in the initial concentration of quinine (80 μg/ml) and the concentration in the supernatant. The sediment was then dispersed in 100 ml of 0.1 M HCl; the flasks were shaken and aliquots of 5 ml were removed at 0, 0.25, 0.5, 1, 2, 4, 6, and 8 h and centrifuged. The amounts of quinine in the new supernatants were determined from the calibration curve, and the percentage of quinine desorbed was calculated as the ratio of the concentration of quinine in the supernatant and the concentration of quinine in the sediment multiplied by 100.

Adsorption and desorption of quinine on G. kola

The standard curve gave a coefficient of determination (R²) of 0.9999 for quinine in 0.1 M HCl at concentrations between 1.25 and 80 µg/ml with the regression equation:

y = 41.628x - 10.858.

A plot of the concentration of quinine in the supernatant against time () clearly revealed that the adsorption equilibration time of quinine in the quinine - G. kola suspension was between the 4th and 5th hour.

Quinine was adsorbed on G. kola and the amount adsorbed increased with increasing concentration of quinine (). At a concentration of 10 µg/ml of quinine, the highest percentage of quinine adsorbed by G. kola was 63.66 %.

The adsorption isotherm of quinine onto G. kola at 37.1 ^oC is shown in . The Freundlich constant (K) and the slope (1/n) which represents the amount of quinine adsorbed for a given concentration change were 52.66 µg/g and 0.69 respectively.

Desorption data of quinine on 2.5 % G. kola is presented in .The amount of quinine eluted increased from 30.12 % at 0.25 h and peaked to 37.51 % at 1 h. Thereafter, the amount desorbed decreased from the 2 h sample and remained relatively constant (35.77 to 35.426 %) through the remaining sample time point.

Interaction of quinine with G. kola has been demonstrated in this in vitro study to occur as a result of capacity-limited adsorption of the drug unto G. kola. Numerous factors may have aided G. kola in its capacity to absorb quinine. In the first place, G. kola is known to contain flavonoids [14] which have functional groups that may favour complex formation with some compounds [15]. Secondly, G. kola also contains trace elements or minerals such as calcium, aluminum, magnesium, potassium, sodium, zinc and copper [16], some of which are known to cause drug interactions through chelate formation [8,17]. It is known that quinine may also act as a ligand by forming a stable five-membered ring with some metals through the quinuclidinic nitrogen and the hydroxyl oxygen, or by binding through the quinolinic aromatic nitrogen [18].

From a plot of the adsorption isotherm, 0.69 and 52.66 µg/g were obtained for slope (1/n) and K, respectively. The slope of the adsorption isotherm is dependent on the linearity of the isotherm, usually varies between 0 and 1 [11] and provides information on the adsorption intensity between quinine and G. kola.  These isotherm values decrease with increasing intensity of drug adsorption, while large values suggest that the adsorbent has a high capacity for the adsorbate [19,20].  In the present study, the Freundlich linear isotherm profile suggests a slope indicative of weak intensity in interaction between quinine and G. kola. Although a close observation of the isotherm plot showed initial linearity at lower values of log x/m suggestive of intense interaction at concentrations less than 2.41 µg quinine per gram of G. kola (), the observed K for the overall data, however, indicates a mild adsorption capacity of the G. kola for quinine [11]. Desorption data () showed that no appreciable change occurred after the second hour during which the elution of quinine attained equilibrium in the system. On the basis of these results and what is known about the constituents of G. kola [8,17,18], it appears that the adsorption of quinine on G. kola may have occurred by physical adsorption and perhaps by some degree of chemisorption.

The findings from this in vitro study corroborate our earlier in vivo report of a pharmacokinetic drug interaction between quinine and G. kola when ingested concurrently. In this study, we aimed at using the same G. kola concentration as in our in vivo study since one goal of an in vitro study is to simulate experimental conditions as close as possible to what is obtained in vivo. In order to obtain an in vitro experimental condition close to the earlier in vivo study in fasting healthy volunteers, the estimated volume of intestinal fluid and the amount of G. kola used in the in vivo study [7] were also put into consideration in determining concentration of G. kola used.  About 12.5 g of G. kola was used in our human studies. It has been reported that the average volume of intestinal fluid in a fasted individual is about 500 ml [21,22]. Thus 2.5 % w/v of G. kola suspension was employed so as to simulate the dose used in the in vivo study. In the in vivo study, the time lag required to reach maximum plasma concentration (T_max) increased by about 48 %, with a significant reduction in maximum plasma concentration (C_max) and a non-significant decrease in area under the plasma concentration-time curve, AUC [7]. These in vivo findings were suggestive of interferences with the absorption of quinine. Indeed, the results of this present adsorption-desorption experiments make it apparent that quinine is adsorbed on G. kola. Since G. kola is a popular snack in the tropics, especially in West African region where malaria is also endemic, the interaction between quinine and G. kola may have decrease bioavailability of quinine and compromise its anti-malarial efficacy.

The results of this in vitro study clearly demonstrate that quinine is adsorbed on G. kola and it may be necessary to avoid concurrent administrations of quinine and G. kola.