Fermented and Unfermented Rooibos (Aspalathus linearis) Exhibit Selective Protection Against Hepatic Stress in Rats Exposed to Fumonisin B1

TL;DR

Rooibos tea, both fermented and unfermented, helps reduce liver stress in rats exposed to a harmful toxin, but doesn't fully protect the liver.

Contribution

The study reveals distinct protective effects of fermented and unfermented rooibos against fumonisin B1-induced liver stress in rats.

Findings

Both rooibos extracts reduced oxidative damage and modulated stress regulators in rats exposed to fumonisin B1.

Fermented rooibos uniquely increased glutathione peroxidase and decreased IL-1β, while unfermented rooibos enhanced Nrf2 and Sirt3.

Neither extract fully restored liver phospholipid profiles or serum cholesterol altered by fumonisin B1.

Abstract

Daily exposure to high doses of fumonisin B1 (FB1) induced liver toxicity, wherein the oxidation, mitochondrial, and cellular stress markers were elevated; inflammatory markers and serum enzyme activities were heightened; and phospholipid acyl chain composition was distorted, while histopathological modifications were not substantially severe.

Rooibos extracts compacted the FB1 toxic effect by facilitating the antioxidant state (increased glutathione peroxidase) and decreasing both levels of malondialdehyde and protein carbonyls.

Rooibos extract administration also enhanced mitochondrial responses (Nrf2, SOD2, PGC1-α, and Sirt-3), while decreasing the cytokine IL-1β and the cellular stress marker HSP70.

Rooibos extracts did not suppress FB1-induced hepatocellular lipids’ distortions, nor serum enzyme activities.

Efficacy variability between rooibos extracts (fermented and unfermented) is attributed to their different polyphenolic contents and capacities.

The exploration of natural redox-modulating agents to mitigate/reduce oxidative damage triggered by toxins is a major area of interest in nutritional and pharmacological fields. Aspalathus linearis (rooibos), traditionally consumed in South Africa for health benefits, was assessed for hepatoprotective effects against fumonisin B1 (FB1)-induced damage. This study involved 24 male rats (n = 6/group) that received FB1 (50 mg/kg diet equivalent, i.p. for 5 days) with or without oral exposure to unfermented (GR) or fermented (FR) rooibos extracts. Alongside somatic records, we assessed blood biochemicals, as well as liver histology, antioxidative stress markers (GSH, GPx, MDA, and carbonylation), regulatory proteins (Nrf2, Sirt3, PGC-α, TRX1, HSP70, and LONp1), inflammation (cytokines), and phospholipid fatty acid profile. Based on results, FB1 suppressed growth, compromised liver function, altered redox status, and elevated stress markers. Both rooibos extracts decreased oxidative damage (↓MDA, ↓carbonylation) and modulated stress regulators (↑Nrf2, ↓HSP70). FR uniquely increased GPx and TRX1 while decreasing IL-1β and PGC-α concentrations, whereas GR strongly increased Nrf2 and Sirt3, reflecting distinct bioactivities linked to their differing polyphenolic profiles. Neither extract compensated for FB1-induced alterations in the liver total phospholipid fatty acid profile or serum cholesterol. In conclusion, GR and FR improved redox potential and inflammatory/stress response; however, this effect was selective, as it did not translate into comprehensive hepatoprotection. These findings support the potential role of rooibos as a dietary modulator of endogenous antioxidant defenses, although clinical translational trials are needed.

Figures

Full text

1. Introduction

Mycotoxins are fungal secondary metabolites that seriously threaten food security because of their toxicity to humans and animals. The infestation of staple crops with Fusarium spp. is particularly concerning, as it is a dominant producer of fumonisin B1 [1]. Several lines of evidence indicate the toxic impact and pathological effects of fumonisin B1 (FB1) on livestock, including pigs, lambs, and horses [1,2,3]. FB1 has been etiologically linked to esophageal cancer in several countries, including Iran [4], China [5,6], and South Africa [7], while other studies have suggested that FB1 is hepatocarcinogenic [6,8].

The disruption of sphingolipid metabolism is a key contributor to the toxicity and carcinogenicity related to FB1 through competitive inhibition of ceramide synthase [9]; however, recent studies have highlighted other possible mechanisms of toxicity, including epigenetic dysregulation, DNA disruption, mitochondrial dysfunction, and oxidative stress [10,11,12,13]. Oxidative stress results from elevated levels of reactive oxygen species (ROS) due to their increased production or decreased detoxification [14]. These ROS have been shown to target and damage macromolecules in cells exposed to FB1 [13,15,16]. Oxidative stress likely contributes to FB1 toxicity and is associated with damage to cellular lipids; however, the mechanisms underlying its augmentation are not fully understood and may involve several events. Distortion of liver membrane lipids by FB1 has been documented in multiple studies, attributed to various toxicokinetic modes of action, primarily the disruption of ceramide synthase and, consequently, the whole sphingolipid metabolism pathway [17]. There is growing appreciation that targeting oxidative stress may present a viable strategy to ameliorate the cell injury commonly observed in systems exposed to FB1 [17,18,19,20].

The plant Aspalathus linearis (Burm.f.) R.Dahlg., commonly known as “rooibos”, is endemic to the Cederberg region of South Africa and is generally consumed as a tisane for its health benefits [21]. There are varieties of rooibos products, which merely depend on the processing, as, for example, the fermentation process impacts its phytochemical structure. Unfermented rooibos, also known as green rooibos, is typically known to be rich in dihydrochalcone C-glucoside, aspalathin, and its precursor, nothofagins. Fermentation, an oxidative process, converts a large portion of these dihydrochalcones into flavonol derivatives such as quercetin, luteolin, and isoorientin, while also reducing the overall polyphenol content [22]. Hence, the rooibos phytomedicinal value is enormously attributed to its rich bioactive polyphenolic content, which is strongly linked to in vivo antioxidant effects [23,24] that are primarily mediated through the modulation of endogenous antioxidant defense systems rather than direct radical scavenging. A previous report by Sheik Abdul and colleagues proposed that rooibos can prevent cancer induction by improving the antioxidant response in FB1-treated models [25]. However, rooibos compositional differences are expected to influence their respective bioactivities and mechanisms of action.

Since mycotoxins cannot be eliminated from the food chain, innovative solutions are needed to reduce their damage. The rat model is a well-established and frequently employed model in the toxicological field (including areas focusing on FB1-induced hepatotoxicity), as it largely mimics pathological/finding responses in other mammals, making this in vivo model suitable for evaluating the potential protective effects of dietary interventions like rooibos. Therefore, this study mainly aimed to investigate the use of rooibos extracts (i.e., unfermented and fermented) as a potential intervention strategy. Accordingly, we hypothesized that rooibos would modulate cellular defense mechanisms to prevent and/or mitigate oxidative stress and inflammatory responses that are critical secondary drivers of FB1 toxicity in a rat model, a model that reflects mammalian biology. This study further investigated the potential of rooibos to preserve the structure and composition of membrane lipids by diminishing alterations triggered by FB1 exposure. Notably, this approach has not been previously investigated, highlighting the novelty of our study. In brief, this study contributes to the understanding of how plant-derived antioxidants can modulate redox homeostasis in vivo, providing views with respect to the fields of food safety, functional foods, and preventative strategies against environmental toxicants.

2. Materials and Methods

2.1. Fumonisin B1

Fumonisin B1 of analytical purity (100% by qNMR) was purchased from Fumizol Ltd. (Szeged, Hungary). For intraperitoneal (i.p.) injection, 65.625 µg FB1 was dissolved in 0.5 sterile 0.9% saline to achieve the desired concentration based on the target dietary equivalent dose of 50 mg/kg diet, which was calculated based on ∼3.5% intestinal absorption rate when applying oral dosage [26]. A daily recovery check of FB1 from the stock solution was conducted using a Shimadzu 2020 LC/MS system (Shimadzu, Kyoto, Japan). The results demonstrated an accuracy within ±0.2%.

2.2. Rooibos Plant Material and Preparation

Commercially available fermented (Superior, PO6/59/II) and green rooibos (Fine cut, U15-03-JJ) plant materials were obtained from Rooibos Ltd. (Clanwilliam, South Africa). The rooibos herbal tea extract was prepared at a concentration traditionally used for brewing, as previously described by Ajuwon et al. [27]. Freshly boiled tap water was added to the plant material at a concentration of 2 g/100 mL, and the mixture was allowed to steep for 30 min before being filtered through Whatman number 3 filter paper. The aqueous extracts were allowed to cool in the refrigerator before they were dispensed into the water bottles. Fresh rooibos tea extracts were prepared daily.

2.3. Total Antioxidant Content and Capacity of Rooibos Extracts

2.3.1. Total Polyphenol Content

The method described by Singleton and Rossi [28] was used to assess the total polyphenol content (TPC) in the extracts of unfermented (green) and fermented rooibos aqueous (GR and FR, respectively). All reagents were freshly prepared, and a standard gallic acid stock solution was prepared by dissolving 0.040 g of gallic acid (Sigma, Cape Town, South Africa) in 50 mL of 10% ethanol. This solution was used to prepare the standard series (0, 25, 50, 100, 250, and 500 mg/L), with double distilled water (ddH_2_O) used as the diluent. The assay working solution, Folin Ciocalteu reagent (FCR), was prepared by diluting 1 mL of FCR (Merck, Johannesburg, South Africa) with 9 mL of ddH_2_O in a 15 mL tube. A thawed plasma sample or study beverage aliquot was equilibrated to room temperature (RT) and diluted (10× dilution) with ddH_2_O. The assay was performed in triplicate in a clear 96-well microplate, and the reaction mixture in each well consisted of 25 μL of standard/sample and 125 μL of FCR. The microplate was incubated for 5 min at RT, after which 100 μL of 7.5% sodium carbonate (Na_2_CO_3_) was added to each well. The microplate was further incubated for 2 h at RT before the absorbance was read in a SpectraMax i3x platform plate reader (Molecular devices, Shanghai, China) set at 25 °C and 765 nm. Data processing and calculations were performed via the Microsoft^®^ Excel 2016 program based on a calibration curve plotted using the standard series and absorbance readings. Results are expressed as mg gallic acid equivalents (GAE) per mL rooibos (mg GAE/mL).

2.3.2. High-Performance Liquid Chromatography Analysis of Aqueous Rooibos Extracts

The aqueous rooibos extracts were filtered (Whatman no. 4) and chromatographically separated on an Agilent Technologies (Santa Clara, CA, USA) 1200 series HPLC system according to an adapted method described by Bramati et al. [29]. The HPLC system consisted of a G1315C diode array and multiple wavelength detector, a G1311A quaternary pump, a G1329A autosampler, and a G1322A degasser. A 5 µm YMC-Pack Pro C18 (YMC, Kyoto, Japan) (150 mm × 4.6 mm i.d.) column was used for separation, and the acquisition wavelength was set at 287 nm for aspalathin and 360 nm for the other major rooibos polyphenolic constituents. The mobile phases consisted of water (A) containing 300 µL/L trifluoroacetic acid and methanol (B) containing 300 µL/L trifluoroacetic acid. The gradient elution started at 95% A, changed to 75% A after 5 min and to 20% A after 25 min, and returned to 95% A after 28 min. The flow rate was set at 0.8 mL/min, the injection volume was 20 µL, and the column temperature was set at 23 °C. Peaks were identified based on the retention time of the standards and confirmed by comparison of the wavelength scan spectra (set between 210 nm and 400 nm).

2.3.3. Antioxidant Capacity of the Aqueous Rooibos Extracts

Three assays were used to determine the antioxidant capacity: the ferric reducing antioxidant power (FRAP) assay, the Trolox equivalent antioxidant capacity (TEAC) assay, and the oxygen radical absorbance capacity (ORAC) assay.

For the FRAP assay, the method described by Benzie and Strain [30] was followed, and 300 mM acetate buffer solution (pH 3.6) was prepared by dissolving sodium acetate (1.627 g) in 16 mL of glacial acetic acid (Merck, SA) with ddH_2_O added to reach a volume of 1 L, and a stock standard solution of ascorbic acid was prepared by dissolving 0.0088 g of ascorbic acid (Sigma, SA) in 50 mL of ddH_2_O. This solution was used to prepare the standard series (0, 50, 100, 250, 500, and 1000 µM), with ddH_2_O used as a diluent. The FRAP working reagent (straw-colored solution) was prepared in a 50 mL tube by mixing 30 mL of 300 mM sodium acetate buffer, 3 mL of TPTZ solution (0.0093 g TPTZ dissolved in 3 mL of 0.1 M HCl), 3 mL of ferric (III) chloride (FeCl_3_) solution (0.053 g ferric chloride dissolved in 10 mL of ddH_2_O), and 6 mL of ddH_2_O. The rooibos herbal tea extracts prepared from green and fermented rooibos plant material were analyzed in triplicate, with each well of a clear 96-well microplate containing 10 µL of standard/sample and 300 µL of FRAP reagent. The microplate was incubated at RT for 30 min before the absorbance was read in a SpectraMax i3x platform plate reader (Molecular devices, China) set at 25 °C and 593 nm. Data capture and calculations were executed in the Microsoft^®^ Excel 2016 program based on a calibration curve plotted using the standard series. Results are expressed as µmole ascorbic acid equivalents per mL of rooibos extract.

The method described by Re et al. [31] was followed for the TEAC assay. All reagents were freshly prepared; briefly, 0.0192 g of 2,2′-azino-di-3-ethylbenzthialozine sulphonate (ABTS) was dissolved in 5 mL of ddH_2_O, while 0.1892 g of potassium peroxodisulphate (K_2_S_2_O8) was also dissolved in 5 mL of ddH_2_O. The ABTS•+ radical was prepared by adding 88 µL of prepared K_2_S_2_O8 solution to 5 mL of prepared ABTS solution. This mixture was heated in an oven at 70 °C for 30 min, and 1 mL of the dark green heated ABTS solution was diluted with 19 mL of ddH_2_O or further diluted until 1.90–2.10 absorbance at 734 nm was read; the resulting mixture was designated the ABTS•+ working solution. The rooibos herbal tea extracts were diluted (10× dilution) with ddH_2_O. The Trolox standard solution was prepared by dissolving 0.0125 g of Trolox in 50 mL of ethanol and used as a stock solution to prepare a standard series (0, 50, 100, 150, 250, 500 µM) with ddH_2_O as the diluent. The assay was performed in triplicate in a clear 96-well microplate. The reaction mixtures in each well consisted of 25 µL of standard/sample and 275 µL of ABTS working solution. The microplate was incubated at RT for 30 min before reading the absorbance in a SpectraMax i3x platform plate reader (Molecular devices, China) set at 25 °C and 734 nm. Data analyses were performed using the Microsoft^®^ Excel 2016 program on the basis of a calibration curve plotted via a standard series. Results are expressed as µmole of Trolox equivalents per mL rooibos extract (µM TE/mL).

The method described by Ou et al. [32] with some modifications was used for the ORAC assay. Briefly, 12 µL of the diluted sample or Trolox standard was mixed with 138 µL of fluorescein (14 µM), and 50 µL of AAPH (4.8 mM) was added to initiate a free radical attack. Fluorescence (excitation 485, emission 538) was recorded every 1 min for 2 h in a Fluoroskan Ascent plate reader (Thermo Fisher Scientific, Waltham, MA, USA). Results are expressed as µM Trolox equivalents (TE)/mL rooibos extract.

2.4. Experimental Animals and Treatments

The experimental protocol was reviewed and authorized by the Food Chain Safety and Animal Health Directorate of the Somogy County Agricultural Office of Hungary, under the permission number: SO/31/00232-7/2023 (approval date: 25 April 2023). This allowance is in line with the EU Directive 2010/63 for the protection of animals used for scientific purposes [33]. Methods were applied in accordance with pertinent regulations and guidelines that are established by ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines (https://arriveguidelines.org (accessed on 10 February 2026)).

Twenty-four adult healthy male HAN:Wist SPF rats (250–275 g body weight) were acclimatized for five days and then randomly assigned (to eliminate potential bias) to each of the four experimental treatment groups (n = 6 animals/group). The rats were kept one animal per cage, exposed to a 12-h light and a 12-h dark daily rhythm at 20 °C in a rodent room (50% relative air humidity) at the university, and fed commercial feed (Ssniff GmbH, Soest, Germany) offered ad libitum.

All rooibos-treated animals received 15 mL of tea extract/day, diluted to 30 mL with drinking water. The control and FB1-treated animals received 30 mL/day of drinking water. Both the feed and fluid intakes were recorded daily. During the experiment, the animals were weighed, and their health and well-being were monitored daily while all clinical signs were recorded. Final body weights were recorded at the end of the 10-day co-exposure experiment. After blood was collected from the retro-orbital plexus, animals were sacrificed by cervical dislocation, exsanguinated, and dissected, and pathological changes were recorded. Livers were excised, weighed, immediately frozen in liquid nitrogen, and stored at −80 °C until biochemical analyses. The four experimental treatments included a negative control group that consumed water and was administered a physiological saline injection intraperitoneally (i.p.), a group that received a daily dose of 50 ppm (mg/kg feed) FB1 was expressed as a feed dose equivalent, i.e., 65.625 µg FB1 in 0.5 mL sterile physiological salt solution administered i.p., and two intoxicated groups that received the same daily i.p. doses of FB1 and consumed either GR or FR (FB1 + GR and FB1 + FR groups, respectively).

2.5. Serum Clinical Chemistry

The serum clinical biochemical parameters, including triglycerides, total cholesterol (tCHOL), high-density (HDL) and low-density (LDL) lipoprotein cholesterols, aspartate transaminase (AST), alanine transaminase (ALT), gamma-glutamyl transferase (GGT), alkaline phosphatase (ALP), and lipase, were measured on a Technicon RA 1000 automated analyzer (Technicon Corporation, Marburg, Germany) from the extracted serum after centrifugation of the native blood samples.

2.6. Biomarkers for Liver Oxidative Damage (Lipid and Protein) and Redox Status

2.6.1. Malondialdehyde, and Conjugated Dienes and Trienes (Lipid Damage)

A small amount (0.5 g) of the thawed liver samples was homogenized in a nine-fold volume of isotonic saline (0.9% w/v NaCl). To investigate the initial phase of the lipid peroxidation process, the amounts of conjugated dienes (CD) and trienes (CT) were measured by the AOAC method [34], which uses trimethyl-pentane, and the absorption at 232 nm for CD and 268 nm for CT was measured. The terminal phase of lipid peroxidation was followed by the determination of thiobarbituric acid reactive substances (TBARS), based on their color complex formation with 2-thiobarbituric acid [35] and expressed as malondialdehyde (MDA), with 1,1,3,3-tetraethoxypropane used as a standard.

2.6.2. Determination of Protein Carbonyls (Protein Damage)

The protein carbonyl (PC) assay was completed as previously described [36]. Briefly, liver tissue samples were homogenized in an ice-cold protein extraction buffer, followed by centrifugation to collect the supernatant for protein concentration determination using a bicinchoninic acid (BCA) assay. The derivatization of protein carbonyls was achieved by incubating equal volumes of the supernatant and a 10 mM solution of 2,4-dinitrophenylhydrazine (DNPH) for 1 h at RT. Subsequent precipitation with 20% trichloroacetic acid (TCA) and washing with ice-cold ethanol/ethyl acetate removed impurities, and the resulting protein carbonyl pellet was resuspended in a 6 M guanidine hydrochloride solution. Spectrophotometric measurement of the resuspended sample at 370 nm facilitated the quantification of protein carbonyls. Results were expressed as nmol protein carbonyls per mg protein using a molar absorption coefficient of the DNP (22,000 M^−1^cm^−1^) [37].

2.6.3. Glutathione and Glutathione Peroxidase

The concentration of reduced glutathione (GSH) was determined in blood plasma and in the 10,000 g supernatant fraction of liver homogenates by the method described by Sedlak and Lindsay [38]. The activity of glutathione peroxidase (GPx) was measured in the same samples as described by Matkovics [39], where the loss of GSH was measured via 5,5′-dithio-bis-2-nitrobenzoic acid (DTNB). GSH content and GPx activity were calculated from the protein content of the 10,000 g supernatant fraction of the samples, which was determined by the method of Lowry et al. [40].

2.7. Fatty Acid Composition of Total Phospholipids of the Liver

Liver samples (ca. 100 mg, thawed just before extraction) were homogenized (IKA T18 Digital Ultra Turrax, Staufen, Germany) in a 20-fold volume of chloroform: methanol (2:1 v/v), and the total lipid content was extracted according to Folch et al. [41]. Solvents were ultrapure-grade (Merck-Sigma-Aldrich, St. Louis, MO, USA), and 0.01% w/v butylated hydroxytoluene was added to prevent fatty acid oxidation. In the frame of lipid fractionation, extracted total lipids were transferred to glass chromatographic columns containing 300 mg silica gel (230–400 mesh) for 10 mg of lipids, according to Leray et al. [42]. Neutral lipids were eluted with 10 mL of chloroform for the above fat amount, then 15 mL of acetone: methanol (9:1 v/v) was added, while 10 mL of pure methanol eluted the total phospholipids. This latter fraction was evaporated to dryness under a nitrogen stream and trans-methylated via a base-catalyzed NaOCH_3_ method [43]. Fatty acid methyl esters were extracted into 300 μL ultrapure n-hexane for gas chromatography (AOC 20i automatic injector; Shimadzu 2030, Kyoto, Japan) equipped with a Phenomenex Zebron ZB-WAXplus capillary GC column (30 m × 0.25 mm ID, 0.25 μm film, Phenomenex Inc., Torrance, CA, USA) and a flame ionization detector. Characteristic operating conditions were injector temperature: 220 °C; detector temperature: 250 °C; helium flow: 28 cm s−1. The oven temperature was graded as follows: from 60 (2 min hold) to 150 °C, from 150 to 180 °C: 2 °C min−1 and 10 min at 180 °C, and from 180 to 220 °C: 2 °C min^−1^ and 16 min at 220 °C. The make-up gas was nitrogen. The calculation was performed with the LabSolutions 5.93 software, using the Post Run module (Shimadzu, Kyoto, Japan), with manual peak integration. The identification of fatty acids was performed based on the retention time of a CRM external standard (Supelco 37 Component FAME Mix, Merck-Sigma Aldrich, CRM47885, Darmstadt, Germany). Fatty acid results were expressed as the weight percentage of total fatty acid methyl esters.

2.8. Liver Histology

Following macroscopic external and internal examination, liver samples were fixed in 10% buffered formaldehyde (pH 7.4) and embedded in paraffin. After deparaffinization and dehydration, tissues were processed by standard histologic techniques to produce 5 μm thick sections, which were subsequently stained with hematoxylin and eosin (HE). The main pathological alterations were described and scored according to their extent and severity: 0 = no alteration, 1 = slight/small scale/few, 2 = medium degree/medium scale/medium number, and 3 = pronounced/extensive/numerous.

2.9. Quantification of Inflammatory Cytokines in Liver Tissue

To evaluate the hepatic inflammation, the liver tissue was rinsed in ice-cold PBS to remove residual blood. Tissue samples (0.5 g) were homogenized in 4.5 mL of ice-cold PBS (1:9 tissue to buffer ratio) using a homogenizer (Omni International, Kennesaw, GA, USA). The homogenates were subjected to sonication for 5 min to disrupt the cells and then centrifuged at 5000× g for 10 min at 4 °C to precipitate the cellular debris. The resulting supernatants were carefully pipetted into new tubes and stored at −80 °C until subsequent analysis. Samples were thawed, and the levels of liver tissue inflammatory cytokines IL-1β, TNF-α, IL-6, and IL-10 in the liver tissue were quantified using a sandwich ELISA. Commercially available ELISA kits from Elabscience (USA) were used according to the manufacturer’s protocol; E-EL-ROO12 (IL-1β), E-EL-R2856 (TNF-α), E-EL-R0015 (IL-6), and E-EL-ROO16 (IL-10). The assay involved the sequential addition of samples or standards, biotinylated detection antibodies, and avidin-HRP conjugate to ELISA microplates pre-coated with an antibody specific to rat IL-1β, TNF-α, IL-6, and IL-10, respectively. The colorimetric reaction was initiated via a substrate reagent and was terminated by adding a stop solution, with absorbance measured by a spectrophotometer at 450 nm. A standard curve was used to calculate the cytokine concentrations.

2.10. Liver Gene Quantitation by Quantitative Polymerase Chain Reaction

To study gene expression levels, total RNA was extracted using the RNeasy Plus Mini extraction kit (Qiagen, Cat. No. 74104) as per the manufacturer’s instructions. Approximately 30 mg of liver tissue was excised from the whole liver tissue and homogenized in 600 µL RPL buffer. To effectively identify differentially expressed genes (Table 1), rat samples within each group were pooled (n = 6) to diminish individual variability and ensure proper assessment of the effects of co-exposure to FB1 mycotoxin and rooibos extracts on liver toxicity [44]. RNA concentrations were determined using a Jenway Genova Nano micro-volume spectrophotometer (Life Science). First-strand cDNA was synthesized from the total RNA extracted. This synthesis was performed using the QuantiNova Reverse Transcription Kit (Qiagen, Cat No. 205413), following the manufacturer’s detailed protocol, which ensures high-quality and reliable cDNA synthesis. Post cDNA synthesis, quantitative polymerase chain reaction (qPCR) analyses were conducted to quantify the expression levels of the target genes (Table 1). The qPCR reactions were carried out utilizing the QuantiNova SYBR Green PCR Kit (Qiagen, Cat No. 208056), which includes SYBR Green I dye for real-time monitoring of PCR amplification. The thermal cycling conditions were set according to the kit guidelines, and each sample was run in triplicate to ensure reproducibility and accuracy of the results. The quantification of gene expression was performed relative to that of the GAPDH housekeeping gene, and the data was analyzed using the ^ΔΔ^Ct method to derive fold changes in expression.

2.11. Liver Protein Expression by Western Blot

To determine protein expression, total protein was extracted from rat liver tissues (rats: 1–24) using a cytobuster extraction reagent (Merck 71009) supplemented with protease and phosphatase inhibitors according to the manufacturer’s instructions. Proteins were quantified via the BSA assay and standardized to 1 mg/mL. Equal amounts of protein were resolved on 10% SDS-PAGE gels to optimize the detection of lower molecular weight proteins. Due to the sample size, the samples were distributed across three separate gels, which were run simultaneously under identical electrophoresis and transfer conditions. The samples were then boiled (5 min) in Laemmli sample buffer, electrophoresed, transferred, and immunoprobed as described by Abdul and Marnewick [10]. Membranes were incubated with primary antibody against LONp1, NRF2 (E5F1A, Cell signalling), SIRT3 (D22A3, Cell Signalling, Johannesburg, South Africa), SOD2 (D3X8F, Cell Signalling, Johannesburg, South Africa) and PGC1-α (sc-518025, Santa Cruz, TX, USA) overnight (4 °C) before incubation at RT with an HRP-conjugated secondary antibody for 2 h (1:5000 dilution). The membranes were stripped and re-probed with proteins of interest as well as with an anti-beta-actin antibody. Protein bands were visualized using the Clarity Western ECL chemiluminescent substrate, and images were captured with the iBright Imaging System (Thermo Fisher, Waltham, MA, USA). Protein expression was analyzed using Image J Software (version 1.54k+), and the results are expressed as relative band density (RBD). Representative bands were used for illustrative purposes without any image enhancement or alteration.

2.12. Statistical Analyses

To avoid bias during data analysis, no exclusion criteria were employed and datasets have been analyzed blindly. For redox status markers measured in the liver samples, protein carbonyls, clinical chemistry, lipid composition, and histological assessment datasets were tested for normality (Shapiro–Wilk test), whereas the extent of standard deviation was compared between groups with Levene’s F test. After this, the univariate analysis of variance (ANOVA) was used on the control and other group means, with Tukey’s “post hoc” test for detailed inter-group differences. In the case of a dataset that is not normally distributed, a non-parametric univariate analysis (Kruskal-Wallis test) was performed. Within this study, the primary outcome measure was liver MDA concentration, which was used to determine the sample size for detecting differences among treatment groups. The software that performed data evaluation was IBM SPSS version 20. For the significance level identification, the calculated probability of a p-value < 0.05 was set for all tests. The IBM SPSS version 20 was used as the primary tool to assess the qPCR and western blotting data. The unpaired t-test with Welch’s correction was employed to determine the significance of the observed differences among the experimental groups. A p-value threshold of 0.05 or less was used for statistical significance. Principal component analysis (PCA), sparse principal loading analysis (sPLA), and cluster analysis were performed on several data via R project version 4.1.2 (2017) and the mixOmics package (6.18.1.) and the Metaboanalyst GUI [45].

3. Results

3.1. Antioxidant Capacity of the Extracts and the Major Rooibos Polyphenolic Constituents

The dihydrochalcone, aspalathin, was the dominant flavonoid in GR, while lower quantities of flavone analogues, orientin and iso-orientin, flavonol glycosides, rutin/isoquercitrin, and flavone analogues of dihydrochalcone, nothofagins, vitexin and isovitexin were also quantified (Table 2). In FR, aspalathin concentration was substantially low, resulting in relatively high concentrations of iso-orientin, orientin, and rutin. Furthermore, FR revealed unique flavonoids, mainly luteolin and quercetin, which were absent or below the detection limit compared to GR.

3.2. Daily Intake from Liquids, and Cumulative Rooibos and Polyphenolic Intakes

FB1 alone did not affect the daily total liquid intake. However, from day 3, rats in the FB1 + GR and FB1 + FR groups displayed reduced consumption of rooibos extracts, with FB1 + GR exhibiting a greater decline (Table 3). Despite this, rats in the FB1 + GR group had significantly higher intake of total polyphenolic constituents, including aspalathin, compared to the FB1 + FR group over the 10-day study period. Based on the daily total liquid intake and analyzed polyphenolic content, the estimated cumulative total polyphenolic content was 89.4 ± 1.14 mg GAE for the FB1 + GR group and 68.9 ± 6.07 for the FB1 + FR group. The cumulative aspalathin intake was estimated at 26.3 ± 3.37 mg for FB1 + GR and 0.52 ± 0.005 for FB1 + FR.

3.3. Body Weight Change, Feed Intake, and Absolute and Relative Liver Weight

Exposure to FB1 significantly (p < 0.05) decreased body weight, and rooibos extract did not prevent this effect. Although fermented extract slightly attenuated the loss, the change was not significant (Figure 1). Reduced growth in all FB1-treated rats was accompanied by reduced feed intakes at the study’s end (Table 4). Similarly, liver weights (absolute and relative) decreased in all FB1-treated rats.

3.4. Serum Clinical Chemistry Parameters Reflecting Liver Function

Compared to the control, FB1 exposure substantially increased concentrations of total cholesterol, HDL, and LDL, while triglyceride changes remained insignificant (Table 5). Notably, supplementation with GR or FR did not ameliorate this effect on cholesterol profiles. Concerning enzyme activities, except for lipase (unaffected), all other liver function marker enzymes (AST, ALT, and GGT) and ALP were altered by FB1, as elevated (3-, 5-, 3-, and 2-fold, respectively). Critically, the co-administration of rooibos extracts failed to prevent the rise in these key clinical markers of hepatocellular injury. Notably, these enzyme activities in FB1-treated rats exceeded those in the control.

3.5. Histological Assessment

Histopathological examination revealed vacuolar degeneration in hepatocytes across all groups, especially in the portal region (Figure 2b), with no significant difference in the severity degree (Figure 3). Overall, the FB1 group showed the highest average lesion score value, mainly marked by moderate degeneration and vacuolation. Rooibos groups provided low lesion score tendencies that ranged between low and moderate degeneration, as well as mild vacuolation. Notably, severe vacuolar degeneration (hydropic change) was most frequent in the FB1 group (3 animals = 50% of rats) (Table 6), which, in most cases, was observed diffusively. In contrast, no severe score occurred in the FB1 + GR group, while one case (16.7% of rats) in the FB1 + FR group displayed severe degeneration and vacuolation.

3.6. Oxidative Stress, Redox Status, and Protein Damage Markers

When the liver redox markers were determined, all FB1-treated rats showed significant increases in their liver GSH levels, which were similar to each other in magnitude (see Table 7). However, GPx activity varied: it was significantly higher in the FB1 + FR group activity compared to the FB1 alone, while the FB1 + GR group revealed no significant difference but remained elevated relative to the control. Regarding MDA, its concentration increased significantly with FB1, whereas GR or FR supplementation markedly minimized the FB1 effects (Figure 4a). Among rooibos extracts, the group treated with GR showed the lowest MDA level, below even the control level. However, irrespective of the treatment, levels of CD and CT (early oxidation markers) remained unaltered. Concerning protein carbonyls (biomarkers for damaged proteins), FB1 exposure significantly increased their levels, which substantially decreased to control level by rooibos extracts (Figure 4b).

The observations from Figure 5, based on cluster analysis, indicate clear separation between FB1-treated rats and those given rooibos extracts, with one misclassified case in the FB1 + FR group. Rooibos supplementation enhanced antioxidant defense (increased GPx activity and GSH concentrations) and decreased oxidative damage, as indicated by low concentrations of MDA and protein carbonyls. These results collectively highlight the effectiveness of rooibos extracts in mitigating FB1-induced oxidative stress.

3.7. Liver Cytokines and mRNA Expression Profiles

3.7.1. Inflammatory Markers

Effects of FB1 and its combination with rooibos on the hepatocellular cytokines and their mRNA expression are presented in Figure 6. Compared to control, FB1 significantly increased pro-inflammatory interleukin-6 (IL-6, Figure 6g) and, contrastingly, decreased anti-inflammatory interleukin-10 (IL-10, Figure 6c,d) level. The FB1 + FR group showed a marked reduction in the pro-inflammatory interleukin-1β (IL-1β, Figure 6a,b) level compared to FB1 alone.

No significant changes were found in the levels of TNF-α across groups; however, its mRNA expression was upregulated over 3-fold in the FB1 + GR group compared to the control (Figure 6e,f). IL-1β expression was substantially elevated in the FB1-treated rats, peaking at about a 4.5-fold increase in the FR-supplemented group (Figure 6b). Conversely, the IL-10 expression significantly decreased by FB1 treatment and was not restored by rooibos supplementation (Figure 6d).

3.7.2. Oxidative Stress

As shown in Figure 7b, FB1 treatment markedly upregulated the nuclear factor erythroid 2-related factor 2 (Nrf2) expression, with even greater increases in rats treated with FB1 and rooibos extracts (GR or FR). In contrast, the superoxide dismutase 2 (SOD2) expression remained unchanged; however, its protein concentration was significantly elevated in the FB1 + GR group compared to both the control and FB1 + FR groups, while FB1 + FR mitigated the effect of FB1-induced changes (Figure 7a,c).

3.7.3. Stress-Induced Responses in Mitochondrial and Cell

As shown in Figure 8a, the mRNA expression of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) was significantly increased in hepatocellular mitochondria from all FB1-treated rats compared to the control group. Exposure to FB1 alone did not alter mitochondrial sirtuin 3 (Sirt3) mRNA expression (Figure 8b). However, a substantial difference was observed between the FB1 + GR and FB1 + FR groups, with FR administration resulting in greater expression (Figure 8b). Focusing on the protein synthesis of both PGC-1α and Sirt3, elevated levels were observed in the FB1 and FB1 + GR groups compared to the control. In contrast, the FB1 + FR group showed suppressed synthesis of these proteins, with levels similar to those observed in the control group (Figure 8c,d).

As illustrated in Figure 9, FB1 substantially upregulated the expression of cellular stress response markers: thioredoxin 1 (TRX1) and heat shock protein 70 (HSP70). While rooibos supplementation did not restore TRX1 expression, both extracts normalized HSP70 expression to the control levels (Figure 9a,b). The mitochondrial stress response lon protease homolog (LONp1) activity was significantly elevated in the FB1 and FB1 + GR groups, with a slight reduction (yet not significant) observed by FR supplementation (Figure 9c).

The principal component results in Figure 10a demonstrate effective separation of treatments along the principal component 1 (PC1) axis, with the FB1 + FR and FB1 + GR groups displaying the greatest variations. The IL-1β, Nrf2, and TNF-α were major contributors to these variations (Figure 10b), indicating that FR or GR notably influence the expression/activity of inflammatory and oxidative stress markers.

3.8. Liver Total Phospholipid Fatty Acid Profile

The fatty acid composition of the hepatic total phospholipids was altered by all treatments, with relatively similar patterns in all FB1-treated rats (Table 8). Among the saturated fatty acids, myristic acid (C14:0) proportion significantly increased in FB1 + GR, whereas stearic acid (C18:0) and total saturation proportions decreased in all FB1-treated groups. In contrast, the oleic acid (C18:1n9) proportion and overall monounsaturation (MUFA) level were significantly elevated. Polyunsaturated fatty acids showed mixed modifications and inconsistent alteration patterns among n3 fatty acids. Among n6 fatty acids, the adrenic acid (C22:4n6) proportion was significantly elevated in all FB1 groups, especially in the FB1 + GR group. Similarly, the docosahexaenoic acid (C22:6n3, or DHA) proportion substantially increased with FB1 exposure, while, contrastingly, the eicosapentaenoic acid (C20:5n3, or EPA) proportion declined, unaffected by any of the rooibos supplementation. Despite proportional alterations in individual n6 fatty acids, their overall levels remained stable; however, the level of total n3 fatty acids increased in FB1-intoxicated rats, lowering the n6 to n3 ratios (n6:n3). Notably, overall polyunsaturation (PUFA) remained unchanged across groups, while unsaturation index (UI) values rose in groups exposed to FB1.

Results of the sparse principal loading analysis can be seen in Figure 11. Figure 11a illustrates a clear separation between the control group and the FB1-treated groups (FB1, FB1 + GR, and FB1 + FR) along component 1. The loadings plot reveals that the major fatty acids contributing markedly to the variance between treatments are C22:6n3, C18:1n9, C20:5n3, C22:4n6, and C18:0 (Figure 11b).

4. Discussion

4.1. Feed Intake and Body Weight

FB1 i.p. administration significantly reduced body weight and overall feed intake in rats, consistent with findings reported by Szabó et al. [46] (i.p. ≡ 50 mg/kg feed for 10 days) and Kócsó et al. [47] (50 mg FB1/kg diet for 5 days), who reported up to 12% and 45% loss in body weight and feed intake of approximately 12% and 45%, respectively. According to Kócsó et al. [47], feed intake dropped by day 2, whereas body weight loss (ca. 5%) was evident by day 5. It is important to note that the body weight loss may result from the FB1 toxicity itself, rather than merely a consequence of decreased feed intake. According to Gbore et al. [48], FB1 at a relatively lower concentration (20 mg/kg feed) in rats decreased the relative weight gain (a 34% drop compared to the control) without affecting feed intake, leading to reduced nutrient utilization and an increased feed conversion ratio. In our study, both reduced feed intake and the hepatotoxic effect of FB1 appeared to contribute to the weight loss, supported by decreased relative liver weight and other liver- and kidney-specific parameters, including serum biochemicals and histopathological findings. However, the present study primarily addresses the hepatotoxic effect; consequently, an in-depth discussion of the nephrotoxic effects has been avoided.

4.2. Hepatotoxicity Assessment via Clinical Chemistry and Histopathology

Mycotoxin target organs vary across animal species; however, the liver has been regarded as a common target organ due to its role in toxin retention or xenobiotic transformation. In rats, i.p. FB1 exposure (≡20, 50, or 100 mg FB1/kg diet) dose-dependently increased total cholesterol levels [46]. This disruption might be a consequence of elevated ABCA1 expression (being a cholesterol efflux promoter [49,50,51]) and altered nuclear factor LXR activity (which regulates hepatic cholesterol metabolism [52]). FB1 also raised HDL and LDL levels, likely due to disrupted sphingolipid metabolism (a typical event caused by FB1 exposure), with sphingosine-1-phosphate influencing cholesterol efflux and the function of HDL cholesterol [53]. ALT activity, a specific hepatic damage/stress biomarker, increased alongside AST, ALP, and GGT activities, reflecting oxidative stress and inflammation [54]. Our findings corroborate those reported in rats by Szabó et al. [46]. The rooibos extracts, in combination with FB1, did not prevent the FB1-induced elevation in serum parameters (ALT, AST, GGT, ALP, and cholesterol), nor did they mitigate the FB1-induced toxicological damage (histological and membrane fatty acid profile). A critical finding of this study is that the primary mechanism driving hepatocyte lesion and necrosis in our model may be independent of, or precede, the oxidative stress pathways that were ameliorated by rooibos. The redox-modulating potential of rooibos, while clearly demonstrated (Section 4.3), may therefore be insufficient to counteract these more cytotoxic effects within the short-term duration of this study. However, these toxic events are most likely consequences driven by the well-known primary action of FB1 on sphingolipid metabolism [17]. Thus, the clinical significance of reducing hepatic oxidative stress with no functional improvement warrants further research in a long-term model, where possibly the cumulative effect of oxidative stress plays a major role. Alongside serum marker analysis, FB1 demonstrated histopathology revealed an increase in the mean vacuolar degeneration score of the FB1 group. This lesion has been reported in rats upon exposure to relatively lower, exact, and higher doses (≡100 mg FB1/kg feed) compared to the dose used in our study [46], demonstrating a parallel increase with FB1 dose and exposure duration. Typically, this hepatic vacuolar degeneration is closely linked to disturbances of cellular water and/or lipid metabolism. Membrane lipid compositional modifications and signs of early necrosis, especially in the portal region, were observed in our study. Apoptosis was also evident by day 5, aligning with findings of Bondy et al. [55] on male Sprague–Dawley rats, wherein apoptosis occurred within 4 to 6 days upon i.p. FB1 exposure (0.75 mg/kg body weight). With respect to rooibos groups, the degeneration scores did not differ from those attained with FB1 alone, suggesting their hepatoprotective effects likely depend on the dose, duration, and compound bioactivity. According to Ajuwon et al. [27], 8 weeks of rooibos supplementation, which is a relatively longer period than ours, protected rat liver against tert-butyl hydroperoxide (tBHP)-induced toxicity. In brief, it is important to note that the findings present a dual picture. While rooibos extracts did not prevent FB1-induced elevation of serum enzyme activities or correct serum dyslipidemia, they did demonstrate a substantial protective effect against other toxicological parameters, especially oxidative stress and certain inflammatory markers, as detailed in the following sections (Section 4.3 and Section 4.4). However, long-term studies are imperative to determine its protective potential.

4.3. The Antioxidant Defense System and Oxidative Damage

Among the results of the redox system, early-stage lipid peroxidation (CD and CT levels) indicators did not increase in the liver following FB1 exposure, corroborating previous findings in male Wistar rats treated with relatively higher FB1 doses for 5 consecutive days [46]. However, late-stage lipid peroxidation was confirmed by the elevated MDA concentrations, proposing a phenomenon driven by rapid metabolism (clearance or biotransformation) of early peroxidation products or activation of non-classical alternative enzymatic pathways (e.g., lipoxygenase and/or cyclooxygenase) that generate signaling molecules (e.g., leukotrienes, prostaglandins) with less CD or CT bulk [56], warranting further lipidomic research. GSH is the major intracellular antioxidant in the liver, which was increased following FB1 exposure, which was unexpected as oxidative stress is known to deplete GSH reserves. It may suggest a compensatory regulation of GSH synthesis, potentially through the Nrf2 pathway (which was increased (Figure 7b)), as an adaptive response to the FB1-induced oxidative challenge. Notably, the literature reports inconsistent effects of FB1 on GSH, likely depending on dose and duration ([46]: unchanged GSH but GPx decreased; [57]: GSH decreased but no change in GPx). Importantly, GSH elevation alone did not limit lipid peroxidation rate, pointing out the vital role of GPx for effective antioxidant defense. Rooibos extracts, especially the fermented, elevated GPx activity compared to the control and FB1 groups, supporting earlier findings of rooibos-induced detoxification enzyme activation (e.g., GSH, GPx, glutathione S-transferase, and uridine diphosphate glucuronosyltransferase) in rat and human livers [27,58,59,60].

Despite FB1’s elevated hepatic MDA level, membrane polyunsaturation levels remained stable, indicating lipid peroxidation, to a lesser extent, involved membrane fatty acids. In addition, hepatic protein damage was also confirmed by the increased level of protein carbonyls in FB1-treated rats, consistent with FB1 toxicity in human and broiler models [15,61]. Interestingly, both rooibos types, especially GR that is rich in polyphenolic content (especially aspalathin), mitigated FB1-induced MDA concentration. Notably, this finding was alongside a decrease in GR intake, but its aspalathin intake remained high (≈50-fold more intake than aspalathin intake from FR), potentiating GR to mediate a strong upregulation of the cellular antioxidant response (high SOD2 and low MDA and carbonylation) even at lower extract intake. This aligns with the known GR’s high content of aspalathin, a dihydrochalcone with potent reported antioxidant properties, which may explain its strong performance despite lower total extract intake in this study. Moreover, both extracts decreased protein carbonyl levels, implying a protective potential against protein damage and corroborating previous findings in human umbilical vein endothelial cells and HepG2 cells [60,62]. Overall, rooibos supplementation effectively modulated hepatic antioxidant response to counter FB1-induced hepatic oxidative stimulus, as supported by the clear separation between rooibos groups and FB1 alone in Figure 5, which was largely driven by increased associated enzymes and reduced oxidative damage markers. However, the translation of this antioxidant effect into full hepatoprotection, as defined by the normalization of serum liver enzymes, was not observed in this model. Therefore, further studies are important to translate these antioxidant findings to clinical settings.

4.4. Inflammatory Responses

In this study, the ELISA outcomes revealed elevated IL-1β protein levels in the FB1 group and decreased levels in the FB1 + FR group, whereas, in contrast, qPCR results showed increased IL-1β mRNA expression in the FB1 + FR group (as compared to the FB1 group). While this study did not mechanistically validate the cause, this pattern is highly suggestive of a potential post-translational regulatory mechanism, probably through the ubiquitin ligase UBE2L3 and HECT-3 (mediate protein degradation [63]) to contribute to the compensatory anti-inflammatory response. Specific flavonoids of FR, like quercetin and luteolin, have been shown to modulate ubiquitin ligase indirectly through modulating the NF-kB pathway [64,65], which could accelerate IL-1β protein degradation even as its transcription is upregulated as a general response to FB1. Thus, this target protein degradation could represent a novel anti-inflammatory mechanism of FR. The IL-1β protein depletion corroborates earlier findings linking FR and quercetin to the declined IL-1β level [66] through MAPK and NF-κB inhibition [67]. However, FR is composed of diverse compounds that may have divergent effects, resulting in our observation of upregulated IL-1β gene expression. This finding may be due to luteolin (a compound characteristic of fermented rooibos), which has a known developmental toxicity risk, in contrast to quercetin [68]. Collectively, the distinct modulation of inflammatory markers by FR, compared to GR, likely stems from its unique profile of fermentation-derived flavanols, which are known to interact with pathways such as NF-κB and protein degradation machinery. Regarding other cytokines, FB1 increased the pro-inflammatory IL-6 level while decreasing the IL-10 (a pleiotropic cytokine), apparently contributing, together with IL-1β, to the hepatotoxicity and atherosclerosis risk facilitated by lipid (cholesterol) dysregulation [69]. FB1’s effect on IL-6 and IL-10 may vary depending on the specific cellular environment and conditions [70,71,72,73]. Our detected IL-6 elevation might be a response to multiple stimuli of IL-1β and TNF-α throughout systemic inflammation [74]. Despite the TNF-α protein level remaining stable under FB1 exposure, its expression showed a slight (yet not significant) upregulation, which was markedly aggravated by GR supplementation, indicating a selective interaction between GR and FB1 that may sustain inflammatory stimuli despite reduced oxidative stress. Rooibos exposure did not significantly alleviate the FB1 effect on IL-6 and IL-10 levels, reflecting their limited modulatory effect to counteract FB1-induced hepatotoxicity, unlike their demonstrated efficacy against LPS-, diesel exhaust particle-, and ultraviolet-induced toxicity/stress [62,66,75,76,77] in various cell types. Hence, further investigations are warranted to clarify the interactions between FB1 and rooibos polyphenolic compounds.

4.5. Mitochondrial and Cellular Stress Responses

FB1’s toxic mechanism has been shown to involve mitochondrial dysfunction [25], primarily achieved via the inhibition of complex I activity, leading to mitochondrial depolarization, increased ROS production, and cellular damage. In response to FB1-induced oxidative stress, cells activate antioxidant defenses, notably upregulating the nuclear factor Nrf2, which, following binding to the antioxidant response element in DNA, enhances transcriptions of antioxidant enzymes and cytoprotective genes. Our study endorsed the upregulation of Nrf2 gene expression in response to FB1 exposure in all treatments, likely corroborating findings in HepG2 cells [15]. Commonly, Nrf2 is recognized as an early adaptive response to oxidative stress; notably, the fact that both rooibos varieties further elevated its expression suggests a potentiation of this endogenous defense mechanism, even if the ultimate functional outcome on clinical enzymes was limited. Among rooibos extracts, FR revealed the most pronounced Nrf2 induction, indicating a strong interaction between the FR extract and this key regulatory pathway. Specific compositions of fermentation-derived flavanols, like luteolin, quercetin, and isoorientin, have been reported to stimulate Nrf2 expression [78,79,80]. According to these findings, these fermentation-derived compounds may exert tissue-specific effects or synergize with FB1 to boost Nrf2 expression, a possibility that necessitates further studies since the functional outcome of this increase may be limited by the severity and nature of the FB1 insult, as reflected in the persistent elevation of hepatotoxicity-related biomarkers. Moreover, Nrf2 is an early stress response biomarker. A further antioxidant response is SOD2; however, its expression did not differ across groups, while its protein level increased in the FB1 + GR group. This unchanged gene expression might indicate a late-phase response in the oxidative stress response or that the feedback mechanism maintained steady-state mRNA levels. While the present study employed a 10-day experimental design, previous research (e.g., [16]) noted delayed SOD2 upregulation after 42 days of FB1 exposure in mice, proposing a late-phase response masked by other antioxidant enzymes. It has been reported that rooibos-induced alterations in SOD2 activities in rat liver and erythrocytes were observed after 7 weeks of exposure [81], likely indicating that the 10-day setting is not adequate to provoke marked modifications. In our model, the marked elevation in SOD2 protein in the FB1 + GR group strongly suggests a compensatory upregulation at the post-transcriptional level that may be uniquely triggered by GR’s polyphenolic signature, potentially involving aspalathin-mediated signaling. This could possibly be driven by enhanced translation, potentially involving mechanistic target of rapamycin (mTOR) signaling pathways (mediated by elevated Nrf2 [82]) to boost ribosomal efficacy and cellular redox balance, or the high stability of protein to prevent its degradation under oxidative stress [83]. Rooibos has also been reported to modulate mTOR level (regulates mitochondrial protein synthesis) in H9C2 cardiac cells [84], although additional investigations are necessary to confirm this hypothesis in the liver. Furthermore, our findings revealed that FR mitigated FB1-induced change in SOD2, corroborating findings in rats’ livers subjected to tBHP-induced oxidative stress [85] and proposing the potential of FR in modulating SOD under certain circumstances.

FB1 triggers a mitochondrial stress response, modulating ATP production and protein detoxification [15]. The upregulation of PGC-1α might be a compensatory response (facilitating mitochondrial biogenesis and ATP production) to oxidative stress, though its protein concentration decreased in the FB1 + FR group, indicating possible post-translational regulation (e.g., protein stability and degradation [86,87]) or modulation of mitochondrial signaling linked to oxidative stress. Furthermore, FBs (FB1 + 2 + 3: 15 mg/kg feed for 14 days in rabbits) have been reported to enhance glucose uptake in livers [88], indirectly implying bioenergetic modification, although future studies with mitochondrial-specific assays are important to validate this mechanism [89]. The induction of ROS detoxification is critically regulated by transcriptional co-activator PGC-1α through the enhancement of mitochondria-specific enzymes like SIRT3 and SOD2 [90,91]. Our finding showed FB1-induced oxidative stress enhanced both PGC-1α and SIRT3 expressions, likely involving the activation of estrogen-related receptor α (ERR-α). However, the novelty (first report) of our study is that GR supplementation downregulated Sirt3 mRNA, plausibly a feedback mechanism in response to the confirmed reduction in oxidative stress (low MDA level), while its protein may be maintained via a post-transcriptional mechanism involving mRNA stability or high translation efficiency [92]. The stress response protein LONp1, an ATP-dependent protease regulated by SIRT3 at the post-translational phase, increased with FB1 exposure and remained elevated with GR administration, suggesting ongoing mitochondrial protease activity. This protein typically catalyzes the degradation of oxidatively damaged proteins in the mitochondrial matrix to prevent their aggregation, cross-linking, and toxicity [93]. Rooibos extracts (GR and FR) may shift stress responses toward mitochondrial repair, as indicated by reduced HSP70 and an increase in LONp1. In addition, the interaction between Sirt3 and LONp1 seems clear here, as the decrease of Sirt3 activity potentially implies lower deacetylation of LONp1 and, consequently, the upregulation of LONp1 expression [94].

In the current study, FB1 exposure upregulated TRX1 and HSP70, likely driven by oxidative stress. Similar findings have been reported in rats’ lungs and kidneys (HSP70 [47]) and in a murine model (TRX1 [95]). In our study, both rooibos extracts decreased the stress-induced HSP70 protein, a consequence of Nrf2 stimulation [96]. TRX1 was further upregulated in FB1 + FR (plausibly to enhance antioxidant efficacy), an effect probably mediated by quercetin and isoorientin flavonoids, known to upregulate TRX1 via Nrf2 enhancement and keap1 protein suppression, as demonstrated in mouse livers and neural models [97]. Nevertheless, luteolin’s role, a fermented rooibos flavonoid, in TRX1 regulation remains unclear and necessitates further investigations. Overall, FB1-induced stress activated mitochondrial and cellular-protective responses involving PGC-1α, Sirt3, LONp1, TRX1, and HSP70. Rooibos extracts appear to modulate these markers, potentially offering partial and selective protection depending on their polyphenolic composition and antioxidant properties.

4.6. Hepatic Total Phospholipid Fatty Acid Composition

In the present study, FB1 altered the liver total phospholipid fatty acid composition, likely through alteration in lipid metabolism enzymes rather than as a consequence of lipid peroxidation, a process that was probably selective towards non-phospholipidic fractions. Similar findings in pig livers support this [98], with marked decreases in C18:0/C18:1n9 indirectly implying high stearoyl-CoA 9-desaturase (SCD) activity and MUFA compensation for the reduced total saturation. However, further studies are needed to confirm this mechanism through direct measurement of SCD expression or activity. FB1 also dose-dependently increased oleic acid and MUFA levels in hepatic-phosphatidylcholines (major membrane fraction ≡ 50% of the total phospholipids in eukaryotic cells) of rats and pigs [98,99]. Regarding polyunsaturated fatty acids, both DHA (anti-inflammatory) and adrenic acid (pro-inflammatory) were modified, corroborating similar findings in phosphatidylcholines, phosphatidylethanolamines, and phosphatidylinositols in rats’ livers upon exposure to lower (20 mg/kg) or higher (100 mg/kg) FB1 doses for 10 days [99]. The n3 fractions, including DHA, are generally known for their anti-inflammatory and proliferation signals [100]. Adrenic acid, on the other hand, is likely less identified and determined in earlier studies related to FBs and membrane lipids. This fatty acid is generated from arachidonic acid via an elongation reaction that is catalyzed by specific elongases located in the endoplasmic reticulum. However, upon its cleavage from membrane phospholipids (often phosphatidylcholines) by phospholipase A2 (from position sn-2) to contribute to n6 fatty acid-derived eicosanoids [101], consequently, it may affect membrane fluidity, phase behavior, signaling, and membrane-bound proteins. Ultimately, neither rooibos extract was sufficient to restore FB1-induced membrane lipid remodeling (supported by the sPLA in Figure 11a), likely attributed to several factors such as the rooibos dose, exposure duration, and compensatory regulatory processes. Accordingly, these findings underscore that the primary protective mechanism of both rooibos varieties in this model is likely through redox and stress-pathway modulation (as seen in Section 4.3, Section 4.4 and Section 4.5) rather than direct correction of deep metabolic disruptions in lipid metabolism. In this respect, multiple events can contribute to membrane lipid distortions by FB1 [17], whereas rooibos extracts offer only partial protection, mainly against secondary oxidative stress.

5. Conclusions

The study findings highlight the complex picture of rooibos extracts against FB1-induced hepatotoxicity. Importantly, rooibos extract effectively mitigated FB1-induced oxidative stress, as proven by decreased lipid peroxidation (MDA), protein carbonylation, and anti-inflammatory IL-1β, as well as boosted antioxidant defenses (GPx and Nrf2). However, this efficacy did not translate into full hepatoprotection, in which rooibos failed to prevent elevation in clinical liver enzymes (AST, ALT, and GGT) or normalize the hepatocellular-membrane fatty acid compositions, plausibly due to factors like the dose used, the treatment duration, or the complexity of FB-induced metabolic disruptions. Therefore, under these experimental conditions, rooibos acts primarily as a redox-modulating dietary supplement that targets secondary oxidative injury, rather than a complete prophylactic agent against FB1-primary metabolic insult. Apparently, rooibos varieties demonstrated certain distinctions (efficacy was selective and extract-specific), particularly on inflammatory and mitochondrial stress markers: GR revealed stronger antioxidant properties, a marked reduction of MDA, unique upregulation of SOD2 protein, and boosted Sirt3 protein synthesis, while FR displayed a more pronounced effect on Nrf2 mRNA and specific inflammatory markers like IL-1β protein. The differing effects between green and fermented rooibos suggest that their distinguished extract compositions (high aspalathin in GR vs. high luteolin/quercetin in FR) inequivalently modulate stress and inflammatory response pathways, such as the Nrf2-antioxidant and NF-κB systems. These differential actions are consistent with the distinct polyphenolic suites of each extract. In this respect, the lack of molecular validation for certain mechanisms (e.g., mitochondrial function, lipid metabolism) remains a limitation. Also, the limitation of short-term duration may not fully capture the rooibos’s potential protective effects. Furthermore, the study lacks the comprehensive protein-level validation (e.g., all cytokines and Nrf-2), which could limit the mechanistic insight. Therefore, future investigations should target isolating the effects of specific polyphenols, leveraging dosing strategies and assessing the timing of intervention (e.g., pre- vs. post-exposure), and applying a comprehensive proteomic approach to fully elucidate the underlying mechanism of rooibos and determine its protective potential. Future work should also focus on sphingolipid intermediates to provide a more complete image of the interaction between FB1 and rooibos.