1. INTRODUCTION

Coffee is the third most consumed beverage worldwide, following water and tea [1]. The genus Coffea includes about 125 different species, with the most commonly consumed being Coffea arabica, widely appreciated for its high quality and typical aroma and taste, and Coffea canephora var. Robusta, known to impart a stronger and bitterer taste and exert greater health properties, as it contains higher amounts of chlorogenic acid derivatives and caffeine [2–7]. In addition to these compounds, the other bioactive components present in coffee beans are theobromine, theophylline, diterpenes (i.e., cafestol and kaweol) and micronutrients, such as magnesium, potassium, and vitamin B3, which are derived from trigonelline present before the roasting process [8,9]. The concentrations of these compounds and the biological activities ascribed to drinking coffee depend on several factors such as variations in the plant itself (species, origin, and genetic characteristics), cultivation and harvesting techniques, storage conditions, degree and type of roasting process, type of coffee, and its method of preparation (boiled, unfiltered, filtered, or espresso). Drastic roasting may lead to the loss of 95% of the chlorogenic acids. Despite this, coffee is the only beverage richest in chlorogenic acids with a single cup of espresso coffee containing between 24 and 423 mg of these polyphenols [10].

The antioxidant activity of coffee is well known, but scientific literature is poor in in vivo studies demonstrating the antioxidant effects of coffee consumption. To the best of our knowledge, only six in vivo studies on the antioxidant activity of coffee have been published during the past decade, of which only one was a clinical trial. A 2008 study on C57BL/6J mice evaluated the antioxidant effects of coffee (present in drinking water containing coffee at a concentration of 35 g/L) and its components [caffeine (400 mg/L), chlorogenic acid (150 mg/L) and caffeic acid (10 mg/L)] by measuring lipid peroxidation through the quantification of total hydroxyoctadecadienoic acid in mouse plasma and liver samples by gas chromatography–mass spectrometry. The study showed that a significant antioxidant effect was found only in the group treated with a normal diet with the addition of coffee [11]. In 2011, Abreu et al. demonstrated the antioxidant effects of coffee and caffeine on the brain of rats treated with a diet supplemented with 3% and 6% coffee and 0.04% and 0.08% caffeine for 80 days against control. In rats that received diets supplemented with coffee and caffeine, authors observed a decrease in thiobarbituric acid reactive substances of 38% and an increase in the levels of glutathione reductase, superoxide dismutase, and glutathione (56%, 54%, and 60% increase, respectively) [12]. Recently, Choi et al. [13] published a research article on the antioxidant and anti-inflammatory activities of coffee with four different degrees of roasting in C57BL/6 mice. Mice were pretreated with coffee (300 mg/kg body weight) by oral gavage for 10 days, followed by an intraperitoneally administered dose of lipopolysaccharide (LPS – 15 mg/kg body weight). To assess the antioxidant effect of coffee, glutathione levels were determined in liver samples and the mRNA expression levels of enzymes involved in glutathione synthesis were quantified. Before the administration of LPS, mice pretreated with light roasted coffee showed the highest concentrations of glutathione levels, which decreased with the increasing degrees of roasting. These results have been confirmed with an analysis of mRNA expression levels related to the enzymes involved in glutathione synthesis, such as glutathione synthetase, glutamate-cysteine ligase catalytic and modified subunit, and glutathione peroxidase (GPX) enzyme. Contrary to the authors’ expectations, the results obtained following the administration of LPS did not reveal a significant beneficial effect of coffee pretreatment, considering all the degrees of roasting [13]. Three other research articles investigating the effects of coffee supplementation on antioxidant enzyme expression levels in rats have been published during the last decade, of which two articles showed a significant increase in superoxide dismutase, catalase, and GPX synthesis in hepatic tissue after both a single dose (2 mL) or multiple doses (2 mL/die for 28 days) of coffee, whereas the third article showed an insignificant increase in antioxidant enzyme synthesis in the anterior tibialis muscular tissue after strenuous exercise in rats treated with 3.8 mL/kg/die of coffee for 21 days [14–16].

Growing evidence suggests that coffee exerts epigenetic effects. Four research articles on the epigenetic activity of coffee and its antioxidant bioactive compounds were published between 2016 and 2017. These articles include two cohort studies on potential links between coffee consumption and blood DNA methylation levels with conflicting results [17,18]. A research article on Caco-2 human colon carcinoma cell cultures elucidates the epigenetic mechanisms of coffee in the prevention of colon carcinoma, via the upregulation of expression of two microRNAs (miR-30c and miR-96), which target Kristen rat sarcoma oncogene gene, inhibiting the synthesis of the proto-oncogene involved in cancer development in a dose-dependent manner [19]. Wang et al. [20] found that chlorogenic acid inhibits schistosomiasis-induced liver fibrosis in both hepatic stellate LX2 cell lines and in schistosoma-infected mice, through interleukin 13/miR-21/Smad7 signaling interactions.

In view of the above, the increasing body of evidence suggesting that coffee acts through epigenetic mechanisms indicates that epigenetic studies may be a valid approach to investigate the mechanisms of the antioxidant activity of coffee, both in in vitro and in vivo systems. In the last decade, only one study on the in vivo antioxidant activity of coffee, indirectly underlying its epigenetic activities, was published in the scientific literature. This was an in-depth study by Kalthoff et al. investigating the mechanism by which coffee exerts protective effects against oxidative stress induced by benzo[α]pyrene (BaP), a polycyclic aromatic hydrocarbon with cytotoxic, carcinogenic, and mutagenic activities, found in large concentrations in tobacco smoke. Mice were treated with coffee (150 mL of coffee beverage prepared with 6 g of ground coffee) for 18 days (the form of administration was left unspecified) with BaP or corn oil (125 mg/kg body weight) administered intraperitoneally during the last 3 days of treatment. On the 19th day, mice were culled and organs collected. The results demonstrated that coffee exerts protective effects against BaP-induced ROS- and H₂O₂-production by upregulation of UGT1A expression. An epigenetic approach was used to better understand the mechanisms underlying the results obtained in vivo. Hepatocarcinoma and esophageal squamous carcinoma cell lines were cultured in the presence of BaP. The coffee beverage was diluted into the growth medium at a concentration of 12%. Part of the cell cultures were treated with siRNA to induce UGT1A silencing. Results show that the coffee-mediated reduction of ROS production was significantly impaired following siRNA treatment, compared with cell cultures with no siRNA treatment. The in vitro epigenetic analyses confirmed the results obtained in vivo, as per the authors’ hypothesis; furthermore, the epigenetic approach better demonstrates that UGT1A expression levels are directly modulated by coffee under oxidative stress induced by BaP [21].

Because of the poor evidence underling the in vivo antioxidant activity of coffee and its mechanisms of action, 36 healthy C57BL/6 mice were used to investigate the epigenetic effect of Robusta coffee (Coffea canephora Pierre ex A. Froehner var. Robusta) supplementation (both regular and decaffeinated) on the enzymes involved in antioxidant defenses. Low-MW (MW < 3500 Da) coffee fractions obtained from decaffeinated and regular coffee were chosen to focus the attention on polyphenols, on which the antioxidant activity of foods and beverages are generally ascribed, and caffeine, excluding the possible influence of melanoidins known to possess antioxidant activity but have low bioavailability due to their high molecular weight. Coffee fractions, incorporated in the mice feed pellets, were previously characterized by HPLC–PAD–ESI–MSn to find a correlation between the registered in vivo effects and the composition of the standardized coffee extract.

2. RESULTS

2.1. HPLC–PAD–ESI–MSn Analysis

To evaluate the in vivo epigenetic effects, regular and decaffeinated low-MW Robusta coffee fractions (MW < 3500 Da) were prepared according to the protocol described in the “Materials and Methods” section to obtain a polyphenol-rich fraction.

Our study started with the characterization of sample composition by means of the HPLC–PAD–ESI–MSn method. In agreement with the evidence cited in the “Introduction” section, regarding the in vivo antioxidant activity of coffee and its components, the chemical characterization of the regular and decaffeinated coffee fractions is crucial in understanding whether the biological activity of coffee can be ascribed to caffeine, or to polyphenols. Retention times, UV–Vis and MS spectra were used to identify the main polyphenol compounds occurring in the coffee fractions: 17 hydroxycinnamic acid derivatives were identified in the regular coffee fraction, together with an organic acid (dihydroxybenzoic acid), as shown in Figure 1a and Table 1; the decaffeinated coffee fraction only yielded 12 hydroxycinnamic derivatives, again with dihydroxybenzoic acid Figure 1b and Table 2.

Figure 1

Chromatogram resulting from RP-HPLC–UV–PAD analysis shows the polyphenolic profile of the fraction with a MW < 3500 Da obtained from: (a) regular coffee; (b) decaffeinated coffee.

Peak	Retention time (min)	m/z [M − H]−	m/z fragment ions (% of the base peak)	λmax (nm)	Proposed structure
1	37.18	153	191 (100), 82 (2), 48 (2)	258	Dihydroxybenzoic acid
2	43.98	353	179 (100), 135 (20)	324	Cis-3-CQA
3	48.10	353	191 (100), 179 (58), 135 (10)	325	3-CQA
4	56.81	353	191 (100), 179 (10)	325	4-CQA
5	60.00	353	191 (100), 179 (10), 135 (5)	325	5-CQA
6	62.00	367	193 (100), 191 (2)	324	3-FQA
7	63.48	533	359 (100), 191 (20), 255 (10), 370 (5)	325	3-H-caffeoyl-4-caffeoylquinic
8	66.81	515	353 (100), 335 (25)	325	1,3-diCQA
9	70.17	337	173 (100), 160 (38)	324	4-pCoQA
10	72.83	367	191 (100), 173 (2)	324	5-FQA
11	74.90	367	173 (100), 191 (85)	324	4-FQA
12	80.16	193	149 (100), 178 (50), 134 (15)	292	Ferulic acid
13	91.49	515	353 (100), 203 (5), 191 (5)	324	4,5-diCQA
14	94.87	515	353 (95), 335 (8), 299 (5), 255 (2), 2103 (2), 179 (10), 173 (10)	324	3,4-diCQA
15	100.33	515	353 (100), 335 (8), 317 (10), 299 (18), 255 (15), 203 (25), 179 (5)	325	1,4-diCQA
16	104.75	529	367 (95), 335 (8), 349 (5)	324	CFQA-2
17	107.42	543	349 (100), 367 (20)	270	3,4-Di-O-feruloylquinic acid
18	107.92	543	367 (100), 349 (25)	270	3,5-Di-O-feruloylquinic acid

CFQA, caffeoyl-feruloylquinic acid; CQA, caffeoylquinic acid; FQA, feruloylquinic acid; pCoQA, p-coumaroylquinic acid.

Table 1

Chromatographic and spectrophotometric data obtained by HPLC analysis of regular coffee fraction (MW < 3500 Da)

Peak	Retention time (min)	m/z [M − H]−	m/z fragment ions (% of the base peak)	λmax (nm)	Proposed structure
1	37.31	153	109 (100), 82 (2), 67 (2)	258	Dihydroxybenzoic acid
2	44.02	353	179 (100), 173 (10), 135 (20)	324	Cis-4-CQA
3	44.67	353	191 (100), 179 (5)	325	4-CQA
4	47.85	353	191 (100), 179 (80), 135 (10)	325	3-CQA
5	56.48	353	191 (100), 179 (5), 135 (2)	325	5-CQA
6	61.90	367	193 (100), 191 (2)	324	3-FQA
7	70.20	337	173 (100), 163 (15)	324	4-pCoQA
8	72.77	367	191 (100), 173 (2)	324	5-FQA
9	74.60	367	173 (100), 191 (15)	324	4-FQA
10	80.20	193	149 (100), 178 (30), 134 (15)	292	Ferulic acid
11	94.88	515	353 (100), 335 (5), 299 (2), 255 (5), 203 (3), 173 (8)	324	3,4-diCQA
12	99.52	515	353 (100), 317 (2), 299 (20), 255 (2), 203 (20), 179 (5)	330	1,4-diCQA
13	104.80	529	367 (100), 335 (15), 349 (5)	324	CFQA-2

CFQA, caffeoyl-feruloylquinic acid; CQA, caffeoylquinic acid; FQA, feruloylquinic acid; pCoQA, p-coumaroylquinic acid.

Table 2

Chromatographic and spectrophotometric data obtained by HPLC analysis of decaffeinated coffee fraction (MW < 3500 Da)

Then, using a calibration curve prepared using standard solutions of caffeine and 5-caffeoylquinic acid in a concentration range between 50 and 250 μg/mL [22], the quantification of the chlorogenic acid derivatives, expressed as the 5-caffeoylquinic acid and caffeine occurring in the samples, was carried out. The total amount of chlorogenic acids in the regular coffee were found to be 236 mg/g of the lyophilized fraction (MW < 3500 Da) whereas their concentration in decaffeinated coffee was found to be 270 mg/g. The concentrations of caffeine were 52 and 1.7 mg/g in freeze-dried regular and decaffeinated coffee fractions, respectively Table 3.

Coffee fraction	Chlorogenic acid derivatives (mg of 5-caffeoylquinic acid equivalents/g)	Caffeine (mg/g)
Regular	236	52
Decaffeinated	270	1.7

Table 3

Quantification of chlorogenic acid derivatives, expressed as 5-caffeoylquinic acid and caffeine in lyophilized coffee fractions (MW < 3500 Da)

2.2. Real-Time PCR of miRNA and mRNA

Once their composition was determined, the freeze-dried fractions were mixed into the mouse bolus (0.2% w/w). The coffee supplementation dose was chosen to mimic a moderate coffee consumption in humans (about three cups/day). More specifically, coffee dry residues were 8.70 and 11.25 g/1000 mL, for decaffeinated and regular coffee, respectively. The average consumption of coffee was considered to be three cups a day (300 mL/day of coffee obtained with the infusion method). So, the daily consumption ranges from 2610 mg/day (for decaffeinated coffee) to 3375 mg/day (for regular coffee). The human equivalent dose (HED) ranges from 37 to 48 mg/kg, for a 70 kg person. The extrapolation of the animal dose to a human dose was performed through normalization to body surface area (BSA) using the following formula:

Animal dose=HED×Human KmAnimal Km

where human K_m factor is 37 for a human and animal K_m factor is 3 for a mouse [23].

In this way, the animal doses should range from 460 to 594 mg/kg. The average body weight of a C57BL/6 male mouse (8 weeks of age) is 22 g [24]. Therefore, the daily intake of coffee should range from 10.12 to 13.07 mg. In our experimental conditions, mice received a daily coffee fraction intake of 10 mg [5 g of pellet/day/mouse containing 0.2% (w/w) of coffee fraction], lower than those corresponding to a moderate regular coffee consumption. As far as bioactive components are concerned, mice receiving the regular coffee fraction-enriched diet, received about 2.36 mg of chlorogenic acids and 0.52 of caffeine, whereas mice treated with the decaffeinated coffee fraction-enriched diet took a similar amount of chlorogenic acids (2.70 mg) and 0.017 mg of caffeine.

Mice were finally culled for the collection of their blood, from which RNA was extracted for the final qPCR. First of all, the expression levels of miR-124-3p were evaluated. A significant decrease in miR-124-3p was detected in mice fed with regular coffee compared with the other two groups (p < 0.01) (Figure 2). This miRNA targets the mRNA coding for GPX enzyme, involved in the maintenance of the redox equilibrium in the cell. Consequently, a reduction in miR-124-3p could suggest a possible antioxidant effect for regular coffee.

Figure 2

Expression levels of miR-124-3p are given in terms of ΔCq in rats fed with bolus enriched with decaffeinated coffee extract (MW < 3500 Da) or regular coffee extract, compared with the control group. Statistical analysis reveals a significant reduction of miR-124-3p levels following daily assumption of regular coffee (p < 0.01).

To further confirm this last hypothesis, the expression levels of GPX mRNA were defined by qPCR. Our results indicate a significant increase in this mRNA in the group fed with regular coffee compared with the control (p < 0.01), whereas no significant differences were found for the decaffeinated coffee, in accordance with the variation of miR-124-3p described earlier Figure 3.

Figure 3

Expression levels of GPX mRNA are presented in terms of ΔCq in rats fed with bolus enriched with regular coffee extract (MW < 3500 Da) is significantly increased compared with the control group (p < 0.01).

3. DISCUSSION

To the best of our knowledge, this is the first study on the in vivo antioxidant effects mediated by an epigenetic mechanism in healthy mice fed with a diet supplemented with a low-MW coffee fractions obtained from decaffeinated or regular coffee beverage. The regular coffee fraction was found to decrease the expression levels of miR-124-3p, and increase the expression levels of mRNA, validated target of miR-124-3p coding for GPX, one of the most important endogenous antioxidant defenses because of its enzymatic removal of oxygen peroxide. These findings could be considered an explanation for the literature data showing that GPX (and other enzymes involved in glutathione synthesis) was significantly increased in mice treated with coffee extracts under physiological conditions (for 10 days by oral gavage at the dose of 300 mg/kg body weight) [25].

It is noteworthy to underline that the expression levels of the selected miRNA and mRNA were found to be insignificantly different from the control, following the intake of the low-MW decaffeinated coffee fraction, containing about the same amount of chlorogenic acid derivatives. These results suggest that caffeine may play a role and may be considered responsible for the in vivo modulation of miR-124-3p expression levels registered after the treatment of mice with regular coffee.

The antioxidant activity of caffeine has long been a topic of investigation. In vitro studies have shown that caffeine exerts an antiradical activity, mainly ascribed to scavenging properties toward the OH^· radical [25]. Prasanthi et al. [26] showed that caffeine (30 mg/die, 12 weeks) reduced ROS levels and reverted glutathione depletion in rabbits treated with a cholesterol-enriched diet. More recently, a clinical trial carried out by Metro et al. showed that the ingestion of pure caffeine, in 15 healthy volunteers, was found to improve plasma levels of antioxidant markers after a 7-day treatment period. Total antioxidant capacity, glutathione, oxidized glutathione, glutathione/oxidized glutathione ratio, lipid hydroperoxides, and malondialdehyde, all considered indicators of oxidative status, were found to have improved from measurements taken prior to administration of caffeine, changing from −41% for oxidized glutathione to −70% for lipid hydroperoxides levels, and +106% for glutathione levels to +249% for the glutathione: oxidized glutathione ratio. These results were taken on a dose of 5 mg/kg body weight/die, corresponding in mice to 61.6 mg/kg body weight/die, using the same formula mentioned above (Animal dose = Human Equivalent Dose × Human K_m/Animal K_m) [27].

In our study, we registered antioxidant activity at a caffeine dose three times lower (20.8 mg/kg body weight, estimating a mean mouse weight of 25 g) than that used in the Metro et al. study.

Our study suggests that it is important to continue investigating the effects of regular coffee on miRNA expression involved in antioxidant defenses to better understand its functional properties and their underlying molecular mechanisms of action.

Finally, these findings, despite being promising, should be confirmed in humans. In fact, more studies are required to confirm the epigenetic effects of coffee in humans and the possible involvement of caffeine in epigenetic mechanisms.

4. MATERIALS AND METHODS

CONFLICTS OF INTEREST

The authors declare they have no conflicts of interest.

AUTHORS’ CONTRIBUTION

MD and MD contributed in conceptualization. VC, MV, MD, IM and MD contributed in data curation. VC and MV contributed in formal analysis. MD and MD contributed in investigation. VC, MV and MD contributed in methodology. MD and MD contributed in project administration and supervision. VC contributed in validation. AB, MD, IM and MD contributed in writing – original draft. AB and MD contributed in writing – review and editing.

ABBREVIATIONS