Effect of Rooibos (Aspalathus linearis) extract on atorvastatin‐induced toxicity in C3A liver cells

Abstract Rooibos (Aspalathus linearis) has various health benefits. Two case studies have associated chronic Rooibos consumption with conventional prescription medications, including atorvastatin (ATV), with hepatotoxicity. Statins act by inhibiting hydroxymethylglutaryl‐coenzyme A reductase, a rate‐limiting enzyme in cholesterol synthesis. Although rare, statins are potentially hepatotoxic. The aim was to investigate interactions between aspalathin‐rich Rooibos extract GRT™ and ATV‐induced hepatotoxicity in C3A liver cells cultured with and without palmitate. Effects of co‐treatment of GRT + ATV on cell viability, oxidative stress, apoptosis, mitochondrial integrity, and cellular reactive oxygen species (ROS) production were assessed. Significantly increased ROS production was observed in cells exposed to ATV and palmitate. Combination therapy of GRT + ATV also showed significant increases in ROS production. Under palmitate‐treated conditions, ATV‐induced significant apoptosis which was not ameliorated by GRT + ATV co‐treatment. Despite studies purporting hepatoprotection from Rooibos, our study showed that GRT was unable to modulate ATV‐induced hepatotoxic effects in this model.

presented with clinically significant complications after concomitant use of herbal products, including green tea. Anecdotally, Rooibos tea consumption is safe. However, two case studies exist that associate treatment of rituximab and daily administration of prednisolone with Rooibos (Sinisalo, Enkovaara, & Kivistö, 2010), or a Rooibos and Buchu tea combination taken with oral steroids and long-term atorvastatin (ATV) use (Engels, Wang, Matoso, Maidan, & Wands, 2013), with hepatotoxicity. Although, causality could not be attributed to Rooibos use, the potential risk for herb-drug interaction needs to be investigated ATV is considered a "blockbuster drug" as the best-selling prescription drug in history, with lifetime sales of USD 148,744 million between 1996 and 2016 (editorial in King, 2013;The Lancet, 2011).
Statins are major chronic prescription drugs worldwide that are administered to lower increased cholesterol levels in patients with increased risk of developing cardiovascular disease. They act by competitively inhibiting hydroxymethylglutaryl-coenzyme A (HMG-CoA) reductase, the first and key rate-limiting enzyme of the cholesterol biosynthetic pathway (Björnsson, 2016;Björnsson, Jacobsen, & Kalaitzakis, 2012;Pal, Ghosh, Ghosh, Bhattacharyya, & Sil, 2015;Schaefer & Asztalos, 2006). Although rare, statin usage has resulted in serious side-effects in some patients, the most severe of which include new-onset diabetes, myalgia and myopathy, as well as the potential of rhabdomyolysis or hepatotoxicity (Björnsson, 2016;Clarke & Mills, 2006;Pal et al., 2015). The presence of these sideeffects or even the potential thereof, often causes patients to cease statin treatment and pursue alternative treatment options. A retrospective study (between 2000 and 2008) by Zhang et al. (2013) showed that, in a cohort of 107,835 patients, 17.4% experienced statin-related side-effects, and 59.2% of which discontinued statin treatment as a result (Zhang et al., 2013). These patients may be selfmedicating with supplements that exert their own health and riskprofile modulating effects, Rooibos extract supplementation, as an example. A review by Kraft (2009) showed that the prevalence of patients making use of complementary or alternative medicine ranges from 24% to 70% in various studies in the United States and Canada (Kraft, 2009). A study by Wazaify, Alawwa, Yasein, Al-Saleh, and Afifi (2013) showed that 11.6% of the 700 participants made use of complementary and alternative medicines, 27.2% of who presented with dyslipidaemia (Wazaify et al., 2013 Canda et al., 2014;Kucharská et al., 2004, respectively), but its effect on ATV-induced hepatotoxicity is not known. It is therefore of clinical importance to understand whether these interactions could pose an added risk for hepatotoxicity, or, contrastingly, whether Rooibos is able to ameliorate ATV-induced hepatotoxic damage.
This study aimed to induce acute hepatotoxicity using ATV in C3A liver cells under a normal and simulated hyperlipidaemic condition, to measure the extent of hepatotoxic damage in terms of various cellular parameters, and to investigate the potential hepatoprotective effects of Rooibos in this context. Exposure of C3A liver cells to high-dose ATV was used as a model to assess whether ATV-induced hepatotoxicity could be ameliorated by the hepatoprotective effects of an aspalathin-rich Rooibos extract (Afriplex GRT™).

| MTT assay
Cellular metabolic activity was assessed using an MTT cell viability assay (Mosmann, 1983). C3A cells were seeded at a density of 11 × 10 4 cells per ml in 200 µl growth medium per well in a 96-well plate and cultured and treated as per Figure 1. For the assay, cells were washed with prewarmed phosphate-buffered saline before being incubated with 50 µl 2 mg/ml MTT for 30 min at 37°C. Following incubation, the MTT was aspirated and 200 µl DMSO and 25 µl Sorenson's Buffer (pH 7.4) was added. Spectrophotometric measurements were recorded at OD 570 using a BioTek ELx800 absorbance microplate reader. Results were generated using Gen5 (RRID:SCR_017317) version 1.05. Three replicates per treatment were assessed in three independent experiments.

| DCF assay
Oxidative stress was assessed by DCF staining (Kalyanaraman et al., 2012). C3A cells were seeded at a density of 11 × 10 4 cells per ml in 200 µl growth medium per well in a black 96-well clear bottom plate and cultured and treated as per Figure 1. ROS production was determined using an endpoint fluorogenic DCF assay incubated for 30 min (37°C in 5% CO 2 in humified air). Fluorescent readings were recorded using a SpectraMax i3x multi-mode microplate reader at an excitation wavelength of 485 nm and an emission wavelength of 535 nm. Results were generated on SoftMax Pro Data Acquisition and Analysis Software (RRID:SCR_014240), version 7.0.2. Three replicates per treatment were assessed in three independent experiments. Reactive oxygen species production was calculated relative to cell viability which was assessed by an MTT assay.

| JC-1 assay
Mitochondrial membrane integrity was assessed by utilizing the JC-1 stain (Di Lisa et al., 1995;Reers, Smith, & Chen, 1991). C3A cells were seeded at a density of 11 × 10 4 cells per ml in 1 ml growth medium per well in a 24-well plate and cultured and treated as per

| Annexin V/PI assessment
Apoptosis was assessed with annexin V/PI staining (Van Engeland, Ramaekers, Schutte, & Reutelingsperger, 1996;Vermes, Haanen, Steffens-Nakken, & Reutelingsperger, 1995). C3A cells were seeded at a density of 11 × 10 4 cells per ml in 1 ml growth medium per well in a 24-well plate and cultured and treated as per Figure 1. Apoptosis status was assessed with dual staining of annexin V and PI, incubated at 37°C for 30 min. Flow cytometry was performed using a BD Accuri™ C6. Two replicates per treatment were assessed in three independent experiments.
Cells in late apoptosis, not treated with palmitate (Figure 5c), showed inverse results to the viable cells, not treated with palmitate ( Figure 5a), and showed similar statistical comparisons between treatment groups. The ATV1 group (36.50 ± 10.93%) showed significantly (p < .001) more late apoptotic cells compared the vehicle control (7.83 ± 2.79%), as did the ATV2 group (81.67 ± 5.57%; p < .001), which showed the greatest percentage of late apoptotic of all the treatment groups in the normal condition. Compared with ATV1 + GRT1 (35.17 ± 19.28%), GRT1 (8.00 ± 9.90%) showed a significant (p < .01) decrease in the prevalence of late apoptotic cells.

| DISCUSSION
ATV is pharmacologically classified as a high-intensity statin, and was chosen as it appears to be the predominant statin associated with hepatotoxicity in patients (Björnsson, 2016). Rooibos (Aspalathus linearis) is commonly prepared as a herbal infusion and has a number of health properties including antioxidant, hepatoprotective, and metabolic effects (Ajuwon et al., 2013;Mazibuko-Mbeje et al., 2019;Waisundara & Hoon, 2015). Given the prevalent use of herbal preparations as complementary and even alternative treatments to modern medicine, we explored the use of GRT, an aspalathin-rich extract of Rooibos, alone and in combination with ATV in an in vitro model of palmitate-induced dyslipidaemia. The high-performance liquid chromatography chemical characterization of the aspalathin-rich unfermented Rooibos extract GRT™ revealed an aspalathin content of 12% (Patel et al., 2016) which is much higher than expected from the daily consumption of fermented or unfermented Rooibos tea. Joubert and de Beer (2012) considered the phenolic content of Rooibos at a "cup of tea" concentration as well as at the concentrations of an industrial extract, which is accepted to be the equivalent of six cups of Rooibos tea. The study showed that the average aspalathin content of a "cup of tea" is approximately 0.53%, and typically 7% in industrial extracts . Furthermore, ATV is lipophilic and passively diffuses across the cell membrane. The addition of palmitate pretreatment to the study design to simulate dyslipidaemia in vitro is representative of the clinical condition wherein statin therapy is likely to be prescribed. The presence of these metabolic alterations leaves the liver cells more susceptible to further injury by ATV than normal liver cells (Koh, Sakuma, & Quon, 2011). The palmitate concentrations selected (500 µM) was within a literature-relevant range (Abu Bakar & Tan, 2017;Mazibuko et al., 2015;Zezina et al., 2018).
Mitochondrial membrane potential was assessed using a JC-1 assay, which was confirmed with a DCF assay assessing ROS generation. Positive results in these assays were confirmed in terms of apoptosis activation, with specific consideration on membrane integrity (annexin V/PI staining) and caspase activation.
In cells not pretreated with palmitate, both 10 and 25 µM ATV caused a significant decrease in relative MTT activity, while GRT alone showed no difference at the concentrations used after 24 hr. In combination, GRT did not modulate the toxicity induced by ATV in terms of MTT activity. The same trend was noted in the palmitatetreated cells. Taken together, this suggests that GRT did not have a modulating effect on ATV-induced toxicity in terms of MTT activity.
In the normal condition, as expected GRT did not induce ROS production. The higher concentration of ATV (25 µM) increased ROS production, and co-treatment with 0.1 mg/ml GRT further exacerbated this effect. In the hyperlipidaemic condition, a similar trend was seen: GRT alone did not induce ROS production, however 25 µM ATV treatment, both alone and in combination with GRT, increased ROS. Shu et al. (2016) attributed ATV-induced hepatotoxicity in the hyperlipidaemic condition to increased ROS production. The antioxidant potential of Rooibos has been well documented and has been shown to increase the activity of endogenous antioxidant systems, as well as having ROS scavenging capabilities (Ajuwon et al., 2013(Ajuwon et al., , 2014Kucharská et al., 2004;Marnewick, Joubert, Swart, van der Westhuizen, & Gelderblom, 2003;Ulicná et al., 2003). In this study, GRT was unable to ameliorate the ATV-induced ROS, suggesting that GRT was unable to protect the cells against ATV-induced ROS. Importantly, by inhibiting mevalonate pathway, apart from cholesterol synthesis, statins also suppress nonsterol isoprenoid pathways, involving ubiquinone biosynthesis (CoQ 9/10 ). This is an important redox component of the mitochondrial electron transport chain, responsible for synthesizing ATP. Reducing the biosynthesis of CoQ 9/10 in the liver affects mitochondrial oxygen consumption and increases cellular ROS production, thought to be a major causal factor for statin-induced adverse effects (Rundek, Naini, Sacco, Coates, & DiMauro, 2004).
Along with increased ROS, high-dose ATV decreased mitochondrial membrane potential in both the normal and palmitate-treated C3A cells. GRT could not attenuate the toxicity of the ATV in terms of mitochondrial membrane potential changes. The loss of mitochondrial membrane integrity and subsequent release cytochrome C from the inner membrane of the mitochondrion initiates the intrinsic apoptotic pathway culminating in the activation of caspase 3/7 (Brentnall, Rodriguez-Menocal, De Guevara, Cepero, & Boise, 2013).
As expected, caspase 3/7 activity was significantly increased by ATV, particularly at 25 µM. These results are in agreement with Docrat, Nagiah, Krishnan, Naidoo, and Chuturgoon (2018), who, using a similar assay to assess caspase activity, demonstrated a significant increase in ATV-induced caspase activation in HepG2 cells, of which C3A cells are a sub-clone, treated with 1.2 mM ATV. In vivo, Pal et al. (2015) demonstrated increased caspase activation in healthy rats treated with ATV at concentrations of 10 mg/kg/day. The combination of ATV2 + GRT2 showed significantly less caspase 3/7 activity compared with the ATV2 group alone, suggesting that GRT had a modulating effect on ATV-induced caspase activity under normal conditions. However, a protective effect of GRT could not be confirmed by flow cytometry using annexin V/PI staining for apoptosis.

| CONCLUSION
This study found that ATV was concentration-and time-dependently hepatotoxic to C3A liver cells and this hepatotoxic effect was exacerbated by the addition of palmitate pretreatment. ATV toxicity was associated with decreased mitochondrial membrane potential, increased ROS production, as well as increased caspase 3/7 activity. GRT was unable to protect against the ATV-induced mitochondrial dysfunction and the consequent induced toxicity. The findings from the current study suggest that concurrent supplementation with GRT appears to be ineffective at protecting C3A liver cells against ATV-induced toxicity in vitro. However, given the complexity of bioavailability and pharmacokinetics of complex mixtures, such as plant extracts, it is difficult to accurately extrapolate in vitro results to a clinical setting. Despite this, our findings suggest that it is unlikely that GRT will exhibit sufficient hepatoprotective effects in patients with ATV-related hepatotoxicity.