Epigenetic-based antioxidant effect of an ethanolic extract of Corylus avellana L. on THLE-2 human primary hepatocytes

Abstract The ethanolic extract of Corylus avellana L hazelnut, prepared in our laboratories, has been previously characterized by liquid chromatography coupled to high resolution mass spectrometry. We here aimed at testing the antioxidant effect of such extract in H2O2-challenged THLE-2 human primary hepatocytes and verified whether it might be based on microRNA-34b/c expression changes. We here demonstrate that miR-34b/miR-34c undergo significant stimulation (≥2-fold change, p < 0.05) in THLE-2 when treated for 72h with not-toxic hazelnut concentrations (0.04–0.4 mg/ml), when compared with 0.06% ethanol control. When administered with H2O2 (1000–2000 µM, 24h), THLE-2 are significantly protected from oxidative stress if pre-treated with hazelnut, the H2O2-driven cytotoxicity and reactive oxygen species generation being recovered by hazelnut extract, through miR-34b/c stimulation. Although preliminary, our findings pave the way for further preclinical studies aimed at validating the possible health-related application of hazelnut matrix, and/or its metabolites, as powerful epigenetic-based drugs, food supplements or nutraceuticals. Graphical Abstract


Introduction
The preparation of an ethanolic extract from the hazelnut kernel of Corylus avellana L. (cultivar Tonda Gentile Romana) was carried out by our group and previously characterized by liquid chromatography coupled to high resolution mass spectrometry (Cappelli et al. 2018).In brief, we detected 85 primary (9 acids, 4 amino acids, 9 sugars, 7 vitamins) and secondary (3 alkaloids, 42 phenylpropanoids, 5 taxanes) metabolites, the phenylpropanoids representing the most abundant metabolic class (49.4% of the detected molecules), followed by acids and sugars (10.6% each).Within primary metabolites, a very high accumulation of oxaloglutarate (acid), Bis-D-fructose 2 0 ,1:2,1 0 -dianhydride (sugar) and vitamin B2 was reported.In terms of secondary metabolism, 3 alkaloids (C 26 H 28 N 2 O 5 , C 26 H 28 N 2 O 6 , and C 26 H 28 N 2 O 9 ) were recognized, and a large group of phenylpropanoids, highly over-represented in the ethanolic hazelnut extract, with flavonoids being the most abundant sub-class (Supplementary material Figure S1, Cappelli et al. 2018).A recent work reported thirteen phenols identified in the hazelnut kernels, including 7 flavanols, 2 hydroxybenzoic acids, 3 flavonols, and one dihydrochalcone, the flavanols being the major phenolic groups (more than 50% of all phenolic compounds identified) accumulated in the kernel, independently of the place of cultivation (Solar et al. 2022).Another key factor in determining the hazelnut quality is represented by the lipid fraction (Bacchetta et al. 2013;Sciubba et al. 2014) which affects both the taste and the nutritional properties, and provides several beneficial effects in terms of human health (Granada et al. 2016;Silvestri et al. 2021).
The antimicrobial effect of our extract against Staphylococcus aureus has been proven in human macrophages (Cappelli et al. 2018); we also demonstrated its functional ability to stimulate the expression of the low-density lipoprotein receptor (LDLR) in both HepG2 human hepatocarcinoma cells and THLE-2 human primary hepatocytes by a promoter DNA methylation-based epigenetic mechanism (Benassi et al. 2019(Benassi et al. , 2021)), thus highlighting the first nutri-epigenomic property of a hazelnut matrix (Bacchetta et al. 2019;Fratantonio et al. 2021).
As the response of human hepatic cells to both injury and external stimuli, including toxins, metabolites and food-derived biomolecules, is regulated by different tissue-specific epigenetic mechanism(s), ranging from DNA methylation, histone posttranslational modifications and microRNAs (miRs) (Brosch et al. 2018;Aloia 2021), we here aimed at testing the ability of our C. avellana ethanolic extract to exert an antioxidant function in human hepatocytes and whether it might be related to microRNA-34b and microRNA-34c (miR-34b/c) pathways.The role of miR-34b/c in liver toxicology is not univocally depicted by experimental evidence.In human cells, miR-34b and miR-34c encoding genes form a cluster on chromosome 11q23, where they are transcribed from a unique primary transcript (pri-miR-34b/c), that further undergoes different processing to release mature single miRs-34b and miR-34c molecules (Agostini and Knight, 2014).MicroRNA-34 family is involved in the response to oxidative stress, especially in the central nervous system (Consales et al. 2018;Huang et al. 2019), whereas, in liver cells, the miR34-redox axis has been mostly described in terms of miR-34a (Wan et al. 2017;Klieser et al. 2019).

Results and discussion
We first screened the response of THLE-2 human cell line to 72h treatment with the ethanolic extract of Corylus avellana L. (HZN) (0.04 and 0.4 mg/ml versus 0.06% ethanol control).Figure 1 shows that 72h exposure triggers a significant dose-dependent increase in pri-miR-34b/c at transcriptional level, thus demonstrating a further level of epigenetic-based functional ability of such natural extract in addition to the previously proven DNA methylation role (Benassi et al. 2019(Benassi et al. , 2021)).A slight divergence between 34b and 34c molecules is reported in terms of miR mature molecules (Figure 1), the miR-34c being stimulated exclusively at the highest HZN dose (0.4 mg/ml).MicroRNA biogenesis entails multiple maturation steps after the pri-miR transcription by RNA polymerase II (Thomson et al. 2006); this is particularly relevant when poly-cistronic miR genes are involved, like the miR-34b/c locus, as the down-stream released mature miRs might play their regulative role by binding and inhibiting specific target genes and cellular pathways.The epigenetic regulation is tissue-specific, we might thus expect a different trend of miR-34b/c expression changes in response to hazelnut treatment if we moved to other cellular models.However, some bioactives composing our hazelnut extract (like the quercetin) have been demonstrated to stimulate miR-34 pathway also in non-small cell lung carcinoma cells (Chai et al. 2021), thus supporting a possible key role played by the miR-34 axis in the cellular response to this food matrix.
We further selected the hydrogen peroxide (H 2 O 2 )-induced oxidative stress as a model of cell injury.The response of THLE-2 cells to H 2 O 2 was preliminarily characterized (Supplementary material, Figure S2).Doses higher than 1000 mM H 2 O 2 impair both cell growth (Figure S2A) and cell viability (Figure S2B) of primary liver cells, concomitantly to a rise in intracellular reactive oxygen species (ROS) (Suppl.Figure S2C).Oxidative stress in THLE-2 is also characterized by a statistically significant decline in pri- miR-34b/c expression levels (Figure S2D).To test the potential antioxidant ability of the C. avellana L. ethanolic extract, we applied the administration schedule reported in Figure S3A.THLE-2 cells were treated with 0.4 mg/ml of the hazelnut extract (or control ethanol 0.06% solution) for 72 h, then washed and administered with the highest H 2 O 2 doses (1000, 1500, and 2000 mM) for the following 24h.Both H 2 O 2 -induced proliferation and viability impairment (at 1000 and 1500 mM) were partially restored in HZN-pretreated cells (Figure 2A and 2B), whereas hazelnut pre-treatment was unable to maintain its potential antioxidant effect at the highest H 2 O 2 dose tested (2000 mM).In accordance, ROS generation was significantly reduced in HZN-pretreated cells exclusively at 1000 and 1500 mM H 2 O 2 concentrations (Figure 2C and Figure S3B), the antioxidant effect being accompanied by a concomitant rise in pri-miR-34b/c expression driven by stimulation with the ethanolic extract of C. avellana L. (Figure 2D).In our experimental model, the miR-34b/c stimulation is involved in the protection of primary hepatocytes from H 2 O 2 , selected as a model of cell injury.Oxidative stress is triggered by high concentrations of H 2 O 2 , as revealed by the massive ROS generation and concomitant cytotoxicity; interestingly, it is also associated to a significant decline in miR-34b/c expression levels.By pre-loading cells with the hazelnut extract, the oxidative damage induced by H 2 O 2 is recovered, in terms of both ROS production and viability.This is in line with the experimental data obtained by Huang and colleagues in animal models, where miR-34b protects against both focal cerebral-induced oxidative stress and H 2 O 2 -induced oxidative stress in rat neuroblast B35 cells (Huang et al. 2019).Also in an in vivo Parkinson's Disease model, the up-regulation of miR-34b has been shown to improve the oxidative stress injury in animal substantial nigra (Li et al. 2020).We here demonstrate that the miR-34b/c-dependent antioxidant property is specifically activated by Corylus avellana L in liver cells, although the protective effect is efficient over a defined window of oxidative damage (1000 and 1500 mM under our experimental conditions).At the highest H 2 O 2 concentration applied (2000 mM), the pre-treatment with the Corylus avellana L extract cannot drive the functional recovery from the oxidative damage, suggesting the occurrence of an irreversible damage to cell biological structures due to oxidative stress that overcomes the protective ability stimulated by the natural extract.
Multiple mechanism(s) of action might underlie the biological response triggered by hazelnut.According to literature, some of the metabolite classes identified in our extract have been proven to directly interact with specific intracellular proteins, e.g.transcription factors, and to drive downstream microRNA transcription.This is the case of catechins, that upregulate the expression of miR-210 by binding to the Hypoxiainducible factor 1-alpha (HIF-1a) (Zhou et al. 2014).Moreover, in silico analysis identified the properties of some metabolites (such as the caffeic acid and quercetin also identified in our hazelnut ethanolic extract) to act as both activators and inhibitors of specific enzymes that control key liver regulatory pathways (Akbari Kordkheyli et al. 2019;de Almeida Chuffa et al. 2022).In addition, some hazelnut metabolite might also directly exert an antioxidant function, as demonstrated for quercetin that acts as a powerful ROS scavenger (particularly via the catechol group in the B ring and hydroxyl group at positions 3 and 5 in the heterocyclic ring), and activates Nuclear factor-erythroid 2 related factor 2 (Nrf-2) and its downstream targets, such as the Heme oxygenase-1 (HO-1) antioxidant enzyme (Akbari Kordkheyli et al. 2019).
Whatever the mechanism(s) responsible for the hazelnut-driven effect we report in liver cells in vitro, we are aware that additional steps in vivo and in humans might affect the composition and bioavailability of components coming from a food matrix before it can reach the hepatic tissue.For instance, the rate of intestinal absorption and metabolization of the extract, likely involving the microbiota contribute, might affect plasma availability of the active hazelnut biomolecules.Besides, given the function of liver tissue, the metabolization at hepatic level might further alter the composition of bioactives.In this view, our study is just a preliminary evaluation of the possible microRNA-based effect triggered by a whole hazelnut ethanolic extract at hepatic level, and complete the previous epigenetic (DNA methylation-based) characterization carried out by our group in both THLE-2 and HepG2 liver experimental models in response to C. avellana L. extract (Benassi et al. 2019(Benassi et al. , 2021)).In perspective, the single biomolecules might be individually characterized and correlated to the whole hazelnut extract-induced health beneficial properties, to better identify those mostly responsible for the stimulation of epigenetic pathways.

Figure 1 .
Figure 1.Analysis of the pri-miR-34b/c gene expression and both mature miR-34b and miR-34c levels, carried out by real-time PCR in THLE-2 human primary hepatocytes.Cells have been treated with either the C. avellana L. ethanolic extract (HZN, 0.4-0.04mg/ml) or the control ethanol (0.06%) up to the indicated time windows (72 h).Values are expressed as Mean ± S.D., calculated in N ¼ 3 independent experiments.Ã p < 0.05 refers to HZN-treated cells compared to control ethanolexposed cells.