Fetal calf serum (FCS), L-glutamine and phosphate buffered saline (PBS, without Ca and Mg), penicillin-streptomycin (P/S) solution, L-glutamine solution, RPMI-1640, DMEM, William’s Medium E + GlutaMAX™, human recombinant insulin zinc solution (4 mg/ml), trypsin-EDTA 10x (5 mg/ml and 2.2 mg/ml), Trypsin (0.5%) solution and phosphate buffered saline (PBS, without Ca²⁺ and Mg²⁺), and penicillin/streptomycin solution (10,000 U/ml and 10,000 μg/ml) were purchased from Gibco™, Life Technologies GmbH (Darmstadt, Germany). Hydrocortisone 21-hemisuccinate sodium salt, omeprazole (≥99.0%), rifampicin (≥97%), ketoconazole (99%), naringenin (98%), menadione (≥98.0%), and acetylsalicylic acid (100%) were purchased from Sigma-Aldrich (Taufkirchen, Germany). Troglitazone (≥98%) was purchased from Santa Cruz (Heidelberg, Germany). 1-hydroxy-3-methoxy-carbonyl-2,2,5,5-tetramethylpyrolidine hydrochloride (CMH), diethyldithio-carbamate trihydrate (DETC), deferoxamine (DFO), and Krebs-HEPES buffer were purchased from Noxygen Science Transfer & Diagnostics GmbH, Elzach, Germany).

The human hepatic cell line HepaRG was obtained from Biopredic International® (Rennes, France). The cell line was cultured in William’s Medium E + GlutaMAX™, supplemented with 10% FCS, 100 U/mlµ penicillin, and 100 μg/ml streptomycin, 50 µM hydrocortisone 21-hemisuccinate sodium salt, and 5 μg/ml human insulin. The maintenance and differentiation of the cell line was performed according to Biopredic International® instructions, as previously described (Le et al., 2021). Cells were maintained at 37°C in a humidified incubator with a 5% CO₂ and 95% air atmosphere.

Salix cortex extracts were prepared and standardized as previously described (Le et al., 2021). S. pentandra clone PE1 (extract S6), originally collected in 2006 in Eggersdorf (Brandenburg, Germany), was cultivated in Wriezen (in northeastern Berlin, Brandenburg, Germany). One-year-old branches of the clone were cut off in August 2016 and bark was peeled at a height from 10–100 cm. Afterwards, the bark material was frozen (−80°C) and immediately lyophilized. Hardwood cuttings of PE1 were also planted in a clone collection at Humboldt-Universität zu Berlin (Germany) to guarantee the availability and conservation of the Salix clone. The bark was extracted using a solution of 70% methanol and 0.1% formic acid. Extract B refers to a willow bark reference used for phytopharmaceutical production, which was provided from Bionorica SE (Neumarkt, Germany). Both extracts (B and S6) were standardized to 10 mg/ml phenolic content using high performance liquid chromatography (HPLC). Based on the reported pharmacological potential and knowledge of characteristic compounds in different Salix species, the following phytochemicals were used to standardize the extracts: salicylates (salicin, salicortin, 2′-O-acetylsalicin, 2′-O-acetylsalicortin, and tremulacin), flavan-3-ols (catechin and epicatechin), flavonoids (two isomers of naringenin-5-O-glucoside, naringenin-7-O-glucoside, luteolin-7-O-glucoside, quercetin-hexoside, and isosalipurposide), other phenolic compounds (triandrin, two caffeic acid derivatives, and syrengin). S6 extract contained 8.15 mg/ml total salicylates and 0.08 mg/ml salicin, and extract B contained 5.35 mg/ml total salicylates and 2.26 mg/ml salicin (Le et al., 2021).

The enzyme lactate dehydrogenase (LDH) is involved in energy production and found in almost all cells of the human body. Upon damage, it is released from the cell in the medium and can thus be used as a marker for cytotoxicity. After 24 h extract exposure, LDH was quantified using an LDH-Glo™ Cytotoxicity Assay kit (Promega GmbH, Mannheim, Germany) according to the manufacturer’s instructions. As positive control, cell exposure to 0.2% triton-X for 15 min was used.

Adenosine triphosphate (ATP) is a key indicator of cellular activity and has been used as another marker of cytotoxicity upon 24 and 48 h treatment using the CellTiter-Glo® 2.0 Cell Viability Assay (Promega GmbH, Mannheim, Germany) according to the manufacturer’s instructions. As positive control, cell exposure to 0.2% triton-X for 24 or 48 h was used. In both assays, 0.5% distilled water was used as solvent control (SC).

The production of reactive oxygen species (ROS) by Salix cortex extracts in hepatocyte-like HepaRG cells was detected using electron paramagnetic resonance (EPR) spectroscopy as described by Lamy et al. (2013) and adapted by Odongo et al. (2017). Differentiated HepaRG cells were treated with different concentrations of Salix cortex extracts or 0.5% distilled water (solvent control) for 1 or 24 h. Cell exposure to 200 µM menadione for 30 min was used as positive control. Afterwards, for ROS detection, 200 µM 1-hydroxy-3- methoxy-carbonyl-2,2,5,5-tetramethylpyrolidine hydrochloride (CMH, Noxygen Science Transfer & Diagnostics GmbH, Elzach, Germany), 25 µM deferoxamine (DFO), and 5 µM DETC were used in Krebs-HEPES buffer for 30 min (Odongo et al., 2017). Supernatants were then measured by EPR spectroscopy for ROS production evaluation. The instrument setting and the number of scans used were set as previously described (Lamy et al., 2013).

The effects of Salix cortex extracts on CYP1A2 and CYP3A4 enzyme activity were evaluated at 1 and 48 h treatment (Bernasconi et al., 2019; FDA, 2020) using the cell based P450-Glo™ Induction/Inhibition Assay Systems (Promega, Walldorf, Germany) according to the manufacturer’s protocol. In brief, HepaRG cells were differentiated as described above in white 96 well plates. After differentiation, cells were incubated for 1 or 48 h with the Salix cortex extracts, or 0.5% distilled water (solvent control). For 48 h treatment, the medium was exchanged after 24 h with the addition of fresh extract. To analyze enzyme activity of CYP2C19, the biochemical P450-Glo™ CYP2C19 Assay and Screening System (Promega, Walldorf, Germany) was used according to the manufacture’s protocol.

50 µM omeprazole (induction) and 320 µM naringenin (inhibition) were used as positive controls (PC) in the CYP1A2 assay.10 µM Rifampicin (induction) and 10 µM ketoconazole (inhibition) were used as positive control in the CYP3A4 assay. 10 μg/ml (22.6 µM) troglitazone was used as positive control for CYP2C19 inhibition (Wishart et al., 2018).

CYP2C19 mRNA expression was quantified using qRT-PCR. In brief, differentiated HepaRG cells were treated with different concentrations of Salix cortex extracts or 0.5% distilled water (solvent control) for 6 or 48 h. Total RNA from HepaRG cells was isolated using the RNeasy mini Isolation kit from Qiagen (Hilden, Germany) followed by DNA purification step using the RNase-free DNase kit from Qiagen (Hilden, Germany) according to the manufacturer’s instructions. RNA quality and quantity were measured using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Freiburg, Germany). Isolated RNA was resuspended in 10 µL of RNAse-free water. Each sample was treated twice with 2 µL RNAse-free DNAse 1unit/μL (Qiagen, Hilden, Germany) for 10 min at 37°C to eliminate remaining DNA. The prepared RNA was reverse-transcribed as previously described (Helmig et al., 2009). For quantitative comparison of CYP2B6, CYP2C19 and CYP2D6 mRNA levels real-time PCR was performed using SYBR-green fluorescence in a LightCycler^® System (Roche Diagnostic GmbH). After optimization of PCR conditions, amplification efficiency was tested in standard curves using serial cDNA dilutions. The correlation coefficient had to be above 0.9 and the slope around −3.5. Amplification specificity was checked using melting curves. Gene expression was related to the mean expression of the three housekeeping genes (HSK) beta-2-microglobulin (B2M), hypoxanthine-guanine phosphoribosyltransferase (HPRT) and glycerinaldehyd-3-phosphat-dehydrogenase (GAPDH) (Vandesompele et al., 2002). Calculations of expression was performed with the 2^−ΔΔCT method (Pfaffl, 2001). The sequences of the used specific primers are listed in Table 1 (Zhang et al., 2005; Chen et al., 2014).

PCR reactions were carried out in a final volume of 20 µl using 1x ABsolute® QPCR SYBR Green Capillary Mixes (Abgene, Brumath, France), 300 nM of primers and 2 µL cDNA. The PCRs started with a 15 min denaturation phase and at the end, a melting curve was acquired form 40°C to 95°C at a thermal transition rate of 0.1°C for 40 s. Specific conditions for primers were as follows: CYP2B6 45 cycles of 95°C for 10 s, 58°C for 15 s, 72°C for 15 s; CYP2C19 45 cycles of 95°C for 15 s, 61°C for 15 s, 72°C for 20 s; CYP2D6 45 cycles of 95°C for 10 s, 61°C for 15 s, 72°C for 15 s. The PCR conditions for HSKs were as follows: B2M 55 cycles of 95°C for 10 s, 63°C for 10 s, 72°C for 25 s; HPRT 45 cycles of 95°C for 15 s, 61°C for 15 s, 72°C for 15 s; GAPDH 45 cycles of 95°C for 10 s, 61°C for 10 s, 72°C for 25 s. All measurements were made without information about the origin of the samples and were performed in duplicate.

Data were analyzed using GraphPad Prism 6.0 software (La Jolla, CA, United States) and presented as means + SD of at least three independent experiments. When comparing multiple means, the results were analyzed either by one-way ANOVA followed by Dunnett’s multiple comparison tests or two-way ANOVA followed by Tukey’s multiple comparison test.

The experiments tested a concentration range of the extracts that had previously shown pharmacological activity in terms of blocking LPS-induced inflammation in primary human immune cells (Le et al. 2021). To investigate potential drug interaction, CYP450 enzyme activity of three enzymes (CYP1A2, CYP3A4 and CYP2C19) was quantified upon Salix cortex extract exposure. As shown in Figure 1A, narigenin (PC1) reduced CYP1A2 enzyme activity in hepatocyte-like HepaRG cells after 1 h by 70%. Omeprazole (PC2) increased the CYP1A2 enzyme activity after 1 h by more that 2-fold (1.6 at 48 h, Figures 1A,B). The Salix cortex extracts did not affect cellular enzyme activity at that time. After 48 h treatment, both extracts reduced the enzyme activity at high concentrations (25 or 50 μg/ml), while the effect was more pronounced by S6, then. Ketoconazole (PC3) completely abolished CYP3A4 activity after 1 h exposure of HepaRG cells, while rifampicin (PC4) triggered enzyme activity induction by about 16-fold compare to control after 48 h (Figures 1C,D). Acetylsalicylic acid (ASA) did not affect CYP1A2 and CYP3A4 enzyme activity after 1 or 48 h at the tested concentrations (Figures 1A–D). For assessment of CYP2C19 enzyme activity, a cell-free assay was used. After 0.5 h, troglitazone (PC5) reduced CYP2C19 activity by 68%; at ≥25 μg/ml both Salix cortex extracts also reduced enzyme activity by 81% (extract S6) and 31% (extract B) compared to solvent control. Again, the inhibitory effect of S6 on CYP450 enzyme activity was stronger compared to extract B.

Differentiated HepaRG cells were exposed for 6 and 48 h to Salix cortex extracts and mRNA expression of CYP2D6, CYP2B6 and CYP2C19 quantified using qRT-PCR (Figure 2). The baseline mRNA levels of CYP2D6 and CYP2B6 were very low in HepaRG cells and no mRNA expression upon treatment could be seen (data not shown). Baseline CYP2C19 mRNA levels were not reduced after 6 h treatment with Salix cortex extracts. After 48 h treatment with 25 μg/ml extract S6, but not B significantly reduced CYP2C19 mRNA expression by 55%.

As given in Figures 3A–C, neither of the two extracts affected intracellular ATP levels or triggered LDH release in hepatocyte-like HepaRG cells at the tested concentrations (0.25–50 μg/ml). We also tested whether the extracts could elevate the level of intracellular ROS in the cells, which in turn could cause damage to lipids, proteins and DNA. From Figures 3D,E it can be seen that after treatment with Salix cortex extracts for 1 or 48 h, no increase in ROS production could be detected.