Herbal remedies are an important and often-used commodity in developing countries,
such as those in Africa. There is a long-standing belief that these medicinal
preparations are more effective and safe than allopathic medications due to their
natural origins. However, very little information is available to describe the toxicological
nature of African herbal remedies, especially with regards to their hepatotoxic effects,
or ability to alter the pharmacokinetic profiles of other compunds. The aim of this in
vitro study was to assess the hepatotoxic potential of a panel of selected African herbal
remedies, as well as their potential to induce drug-herb interactions.
Crude hot water and methanol extracts were prepared from seventeen African plants
using brewing and ultrasonic maceration techniques, respectively. Phytochemical
screening was done to determine the broad constituency of the extracts using thin
layer chromatography, biochemical reactions and free radical scavenging. Cytotoxicity
against the HepG2 hepatocarcinoma and Caco-2 colon carcinoma cell lines were
determined using the sulforhodamine B staining assay. Acokanthera oppositifolia,
Boophane disticha, Moringa oleifera, Solanum aculeastrum, Tabernaemontana
elegans, Terminalia sericea and Ziziphus mucronata were selected for further
hepatotoxic assessment using a mixture of spectrophotometric, fluorometric,
chemiluminescent and flow cytometric assays. Oxidative stress (reactive oxygen
species, glutathione and lipid peroxidation levels), mitochondrial membrane potential,
fatty acid accumulation and caspase-3/7 activation was assessed using fluorometric
assays, while adenosine triphosphate levels were assessed using a chemiluminescent
assay. The effect of the extracts on cellular kinetics and mode of cell death was
determined using flow cytometric techniques. Drug-herb interactions of the crude
extracts were assessed by measuring their effect on P-glycoprotein activity, nevirapine
permeability and cytochrome P450 (CYP2B6, CYP2D6 and CYP3A4) enzyme activity.
Phytochemical analyses successfully identified the broad constituency and antioxidant
profiles of the crude extracts, which correlated well with that already described in
literature. Phytochemical classes that were most prominent in the majority of extracts
were the alkaloids, flavonoids, glycosides, phenolic acids and saponins. These results were used to determine potential contributing factors during the hepatotoxicity and
drug-herb interactions assays.
Thirteen extracts displayed activity against the HepG2 cell line, while twelve were
active against the Caco-2 cell line. Cytotoxicity was in generally more potent against
the Caco-2 cell line, indicating a potential susceptibility of the intestinal tract towards
these herbal remedies. A high risk of cytotoxicity was identified for extracts such as A.
oppositifolia, S. aculeastrum and T. elegans. Further hepatotoxic assessment
indicated that the majority of extracts depolarised the mitochondrial membrane,
however oxidative stress was rarely induced. Steatotic changes were evident as
shown by the increased retention of fatty acids. Cytotoxicity was mostly a mixture of
antiproliferative effects, and worryingly, the induction of necrotic cell death. Although
the methanol extracts tended to be more potent than the hot water extracts, the latter
induced detrimental effects as well, albeit at a higher relative concentration. The most
prominent hepatotoxic effects were observed after exposure to T. elegans, with a halfmaximal
inhibitory concentration of 3.07 ?g/mL.
Only the methanol extract of S. aculeastrum displayed prominent P-glycoprotein
inhibitory activity (HepG2 = 2.92-fold, Caco-2 = 1.29-fold). Five extracts (hot water
extracts of Burkea africana and Senecio latifolius, and the methanol extracts of
Mundulea sericea, Rauvolfia caffra and Solanum aculeastrum) were selected for
assessment of their ability to modulate nevirapine transport across the Caco-2 cell
line. All five extracts decreased nevirapine efflux, indicating a propensity for increasing
its bioavailability. Due to the lack of P-glycoprotein inhibitory activity from the majority,
it appears that this reduced efflux is not necessarily P-glycoprotein-dependent. Altered
membrane fluidics and inhibition of other membrane transporters are suggested as
potential contributing factors.
The majority of extracts displayed prominent CYP450 inhibitory activity
(CYP3A4>CYP2B6>CYP2D6). Most extracts displayed a higher selectively towards
CYP3A4, which highlights the caution required. More than 50% of drugs currently used
on the market are metabolised by the CYP3A4 isoform, and thus the risk is high when
comparing it to the low concentrations required to elicit an effect. Both extracts of B.
africana, the hot water extract of T. sericea and the methanol extract of Z. mucronata
displayed non-selective inhibition across all three isoforms, indicating a possible affinity for a structurally-conserved site across CYP450 enzymes. Due to the genetic
makeup of the African populace, which has a significant proportion of CYP450 alleles
displaying reduced or inactived enzyme activity, risk is thus high for attenuating their
metabolic functions. The methanol extract displayed the highest inhibitory activity
against CYP3A4, with a half-maximal inhibitory concentration <1.2 ?g/mL.
It is evident throughout the study that there is a high risk of hepatotoxicity or drug-herb
interactions. This is specifically observed when taking into account the low amount of
extract required to reach the inhibitory concentrations determined in vitro. Inhibition of
intestinal P-glycoprotein transporters and CYP450 enzymes appear to be more at risk
than that of the liver. This is one of the few studies delving into the toxicological natures
of African herbal remedies, specifically in light of hepatotoxicity and herb-drug
interactions, and thus serves as a foundation for further assessment.