Antioxidant and Anti-Inflammatory Profiles of Spent Coffee Ground Extracts for the Treatment of Neurodegeneration

Spent coffee grounds (SCGs), waste products of coffee beverage production, are rich in organic compounds such as phenols. Different studies have demonstrated phenol beneficial effects in counteracting neurodegenerative diseases. These diseases are associated with oxidative stress and neuroinflammation, which initiates the degeneration of neurons by overactivating microglia. Unfortunately, to date, there are no pharmacological therapies to treat these pathologies. The aim of this study was to evaluate the phenolic content of 4 different SCG extracts and their ability to counteract oxidative stress and neuroinflammation. Caffeine and 5-O-caffeoylquinic acid were the most abundant compounds in all extracts, followed by 3-O-caffeoylquinic acid and 3,5-O-dicaffeoylquinic acid. The four extracts demonstrated a different ability to counteract oxidative stress and neuroinflammation in vitro. In particular, the methanol extract was the most effective in protecting neuron-like SH-SY5Y cells against H2O2-induced oxidative stress by upregulating endogenous antioxidant enzymes such as thioredoxin reductase, heme oxygenase 1, NADPH quinone oxidoreductase, and glutathione reductase. The water extract was the most effective in counteracting lipopolysaccharide-induced neuroinflammation in microglial BV-2 cells by strongly reducing the expression of proinflammatory mediators through the modulation of the TLR4/NF-κB pathway. On these bases, SCG extracts could represent valuable nutraceutical sources for the treatment of neurodegeneration.


Introduction
The food industry generates considerable amounts of waste products that require to be appropriately managed to reduce their negative sustainability impacts. An appropriate waste management helps to reduce not only the negative effects on the environment but also has got an important economic impact, since there is less production of nonrenewable resources and less energy is used in the production of new goods. Among food industry wastes, coffee by-products have been extensively taken into consideration for recycle [1][2][3][4][5].
Coffee is made by roasting and grinding coffee beans to produce a powder that is extracted with hot water or brewed. During the preparation of coffee beverages, a solid residue known as spent coffee grounds (SCG) is produced and this is the most abundant coffee waste (55−67%) [6].
About 650 kg of SCG are produced from 1000 kg of green coffee beans, and nearly 2 kg of wet SCG are obtained by the preparation of 1 kg of soluble coffee [7]. SCG is a nonedible resource, which is not entering into the food chain, and its disposal in the environment is dangerous since SCG contains caffeine, tannins, and polyphenols that make it a toxic residue [6,7]. On these bases, numerous authors have suggested different ways to recycle SCG, to manage and reduce its disposal [8][9][10]. SCG can be used as a source of oil for biodiesel production [11][12][13] or as a source of recoverable sugars which can be employed as food addictive or for bioethanol production [13][14][15][16]. Moreover, different papers focused on SCG constituents and their application in the food and nutraceutical industry [1,[17][18][19]. The main constituents of SCG are polysaccharides, proteins, and lipids, as well as minerals, caffeine, melanoidins, and phenols [20]. Phenols of SCG are mainly represented by different highly bioavailable and bioactive phenolic acids such as chlorogenic, caffeic, ellagic, trans-ferulic, gallic, p-hydroxybenzoic, p-coumaric, protocatechuic and tannic acids, and flavonoids such as catechin, epicatechin, rutin, and quercetin [1,21,22]. Phenolic compounds are well known for their beneficial effects on human health, e.g., in the prevention of different chronic degenerative diseases such as cancer, cardiovascular, and neurodegenerative diseases [23][24][25]. Neurodegenerative diseases, mainly including Parkinson's and Alzheimer's diseases, are a health problem primarily affecting the elderly. These disorders share common cellular and molecular events such as oxidative stress, abnormal protein deposition, damaged mitochondrial function, induction of apoptosis, impairment of proteostasis, and neuroinflammation [26]. Neuron cells are particularly vulnerable to oxidative damage due to their high polyunsaturated fatty acid content in membranes, high oxygen consumption, and weak antioxidant defenses [27]. Oxidative damage results in an increase in reactive oxygen species (ROS), which leads to further oxidative damage and feeds this self-propagating cycle. ROS may also trigger protein misfolding, potentially leading to protein aggregation, which is a classical hallmark of neurodegenerative diseases such as Alzheimer's and Parkinson's diseases [28].
In addition to oxidative damage, in recent years, the immune system is emerging as a key determinant in the onset and progression of neurodegeneration [29,30] as it triggers modification of cytokine signaling, immune cell proliferation and migration, impaired phagocytosis, and reactive gliosis [31]. Neuroinflammation, caused by the activation into proinflammatory states of the brain immune cells, namely, microglia and astrocytes, represents a fundamental defense system that protects neurons from toxic substance and microorganisms. In normal physiological conditions, this is commonly a positive mechanism aimed at preserving the brain integrity by removing threats and reestablishing homeostasis [32]. However, chronic neuroinflammation can stimulate a series of events that induce progressive neuronal damage that characterizes many neurodegenerative disorders [33]. Unfortunately, currently, no drugs capable of slowing down or blocking the progression of these debilitating pathologies have been identified. This is why the research is turning its attention to the identification of natural compounds with a preventive/protective activity against neurodegenerative disorders. As we previously demonstrated that extracts obtained by coffee silverskin, another coffee by-product, are rich in bioactive compounds with antioxidant and antibacterial activities, we assumed that also SCG could be rich in bioactive phytochemicals with potential neuroprotective activity [5,34].
The present study was undertaken to evaluate the phenolic content of 4 different SCG extracts and their ability to counteract oxidative stress and neuroinflammation in neuron-like SH-SY5Y and microglial BV-2 cells.

Spent Coffee Ground Sample and Extract Preparation.
Roasted beans of 100% Coffea arabica L., Ethiopian origin, were supplied by Simonelli Group S.p.A. (Belforte del Chienti, Italy). Roasted beans were grinded by Mythos 1 grinder (Simonelli Group S.p.A.), and spent coffee ground (SCG) was obtained after a series of replicates of espresso coffee preparations using a VA833 Black Eagle espresso coffee machine (Victoria Arduino, Simonelli Group S.p.A., Belforte del Chienti, Italy). The extraction of espresso coffee was carried out as follows: 7 ± 0:05 g of roasted and ground (R&G) coffee per cup, 25 ± 1 s of extraction, water pressure and temperature 9 bar and 92.0°C, respectively, and 25 ± 2 g in cup. SCG samples were collected and oven-dried at 50°C until constant weight (about 48 h). Dried SCG sample was stored at 4°C up to use. The extract preparation was carried out following a previous work [5] with some adjustments. For the current research, four extracts were selected on the base of their high performance in terms of bioactive compound recovery and extraction yield [5,21]. Briefly, 10 g of SCG were extracted with 50 mL of solvent assisted by a FALC ultrasonic bath (FALC, Treviglio, Italy) at a frequency of 40  The injection volume was 2 μL, and the column was set at 30°C. The drying gas in the ionization source was at 350°C. The flow rate of the gas was 10 L min -1 , the nebulizer pressure was 25 psi, and the capillary voltage was 4000 V. The dynamic "multiple reaction monitoring" (dynamic MRM) mode was used for detection, and the quantification was realized by integrating the dynamic MRM peak areas. The most abundant product ion was used for quantitation, and the other to confirm the analyte. In Table 1, the selected ion transitions and the mass spectrometer parameters comprising the definite time window for each compound (Δ retention time) are listed.

Total Phenolic and Flavonoid Contents and DPPH
Radical Scavenging Activity. The total phenolic content (TPC) was measured spectrophotometrically according to the method developed by Siatka and Kašparová [36] with some modifications. In particular, 0.5 mL of extract solution (1 mg mL -1 in methanol), 2.5 mL of Folin-Denis reagent solution, and 7 mL of Na 2 CO 3 (7.5% w/w in water) solution were added to the test tubes. The reaction mixture was maintained at 25°C for 2 h in the dark, and the absorption was measured at 765 nm. Gallic acid was used as a reference compound, and the TPC in the extracts was calculated using gallic acid calibration curve and expressed as mg of gallic acid equivalents (GAE) per g of dry weight of SCG extract. The total flavonoid content (TFC) of each extract was evaluated as reported in [37] with some modification. 0.5 mL of extract solution (1 mg mL -1 ), 0.15 mL of NaNO 2 (0.5 M), 3.2 mL of methanol (30% v/v), and 0.15 mL of AlCl 3 ·6H 2 O (0.3 M) were added in a 15 mL test tube. 5 min later, 1 mL of NaOH (1 M) was added and the solution was mixed well before measuring the absorbance at 506 nm. Rutin (0 to 100 mg L -1 ) was used to make the standard calibration curve for TFC following the procedure described above. TFC was reported as mg of rutin equivalents (RE) per g of dried extract.
The in vitro antioxidant activity of the extracts was measured as ability to scavenge the radical 2,2-diphenyl-1-picrydrazyl (DPPH) as reported in [38] with some modifications. Briefly, 0.5 mL of extract solution (1 mg mL -1 in methanol) was added to 4.5 mL of ethanolic solution of DPPH (0.1 mM) in a 15 mL test tube and allowed to stand for 30 min in the dark at 25°C. The DPPH reduction was evaluated spectrophotometrically at 517 nm. The % of DPPH scavenging was obtained following the formula: %I = ½ðA control − A sample Þ/A control × 100. A control and A sample indicate the absorbance obtained in the absence and presence of antioxidants, respectively. The scavenging activity of the extracts was reported as the IC 50 value (μg mL -1 ), the extract concentration which causes a 50% DPPH inhibition. The IC 50 value was calculated by interpolation from the linear regression analysis. Trolox® (1-50 μg mL -1 ) was considered as a reference antioxidant.

Cell Cultures and Treatments.
The SH-SY5Y cell line was purchased from Sigma-Aldrich (ECACC 94030304) (St. Louis, MO, USA) and was grown in high-glucose DMEM supplemented with 10% (v/v) of FBS, 2 mM L-glutamine, 50 U/mL of penicillin, and 50 μg/mL of streptomycin, as previously reported [39]. Cells were used for experiments after inducing their differentiation with all-trans retinoic acid (10 μM) for 7 days.
Differentiated SH-SY5Y were treated with different concentrations of the SCG extracts for 24 h and then exposed to 700 μM H 2 O 2 for 1.0 h in 1% FBS DMEM.
BV-2 murine microglial cells were a kind gift of Prof. Elisabetta Blasi (University of Modena and Reggio Emilia, Modena, Italy). Cells were cultured in high-glucose DMEM supplemented with 10% (v/v) of low-endotoxin FBS (Euroclone, Milano), 2 mM L-glutamine, 50 U/mL of penicillin, and 50 μg/mL of streptomycin. The cells were maintained in a humidified incubator at 37°C with 5% CO 2 and subcultured using Trysin-EDTA.
BV-2 cells were pretreated with the SCG extracts at different concentrations for 24 h before the addition of 100 ng mL -1 LPS for 24 h.
2.6. Cell Viability Assay. Cell viability was evaluated by measuring MTT reduction as previously reported [40]. Briefly, at the end of each experiment, the cell medium was removed from 96-well tissue culture plates, and the cells were incubated with 0.5 mg mL -1 of MTT solution. The incubation time was 30 min for BV-2 cells and 90 min for SH-SY5Y cells. After removing the MTT solution, DMSO was added to lyse the cells. The presence of formazan was evaluated spectrophotometrically at 570nm using a microplate spectrophotometer (VICTOR3 V Multilabel Counter; PerkinElmer, Wellesley, MA, USA). Data are reported as percentage with respect to controls. Control cells are considered as 100% cell viability.  2.8. DCFH-DA Assay. Intracellular ROS levels were evaluated using the fluorescent DCFH-DA probe as previously reported [41]. Briefly, at the end of each experiment, 10 μM DCFH-DA solution in DMEM 1% FBS without phenol red was added to the cells allowed to stand for 30 min. PBS was added after removing DCFH-DA solution. Cell fluorescence was measured using 485 nm excitation and 535 nm emission on a microplate spectrofluorometer (VICTOR3 V Multilabel Counter, PerkinElmer).
2.9. RNA Extraction. RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany) was used to extract total RNA. The quality and quantity of RNA were evaluated by a NanoVue Spectrophotometer (GE Healthcare, Milano, Italy).

Real-Time Polymerase Chain Reaction (PCR).
Reversetranscription of 1 μg of the extracted RNA to cDNA was performed using iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA), according to the supplier's instructions.
In Tables 2 and 3, the primers used (Sigma-Aldrich, Milan, Italy) are listed. Two different reference genes were used: GAPDH rRNA for microglial cells and RPS18 for neuronal cells. cDNA amplification was started at 95°C for 30 s to activate the polymerase, followed by 40 cycles of 5 s at 95°C and 30 s at 60°C. Normalized expression levels were calculated relative to control cells according to the 2 −ΔΔCT method.

Flow Cytometry.
To evaluate the surface expression of TLR4 receptor on BV-2 cells, 1 × 10 5 cells were seeded in 12-well tissue culture plates. At the end of each experiment, cells were washed with PBS and detached with accutase solution. The cells were centrifuged at 300 g for 5 min. The cell pellet was washed twice by centrifugation and resuspension in washing buffer (0.2% BSA-PBS), in 1.5 mL tubes. After removing the supernatant, the cells were resuspended with FITC-conjugated rabbit anti-TLR4 antibody (Stressmarq, cat. no. SPC-200), 1 : 100 dilution in 0.2% BSA-PBS, then incubated for 30 min in the dark at 37°C according to the manufacturer's instructions. After antibody incubation, the cells were washed twice as above. After supernatant aspiration, the samples were appropriately diluted to 5 × 10 5 cells mL -1 and finally resuspended in BSA 0.1% PBS for flow cytometry reading. Guava® easyCyte™ 5 HT instrument These product ions were used for quantification; the others to confirm the analytes. 5 Oxidative Medicine and Cellular Longevity was used to collect all raw data. FlowJo software was used to analyze the mean fluorescence intensity (MFI). Unstained samples were used as negative controls.
2.13. Immunofluorescence Confocal Microscopy. BV-2 cells were cultured directly on glass coverslips in 6-well plates. Cells were then fixed with 2% paraformaldehyde in PBS for 15 min at room temperature and permeabilized with Triton 0.1% for 10 min, after which they were treated with a polyclonal antibody (1 : 500) against NF-κB p65 overnight. Following extensive washing with PBS, cells were incubated with a secondary Alexa Fluor 488-conjugated anti-rabbit IgG antibody diluted 1 : 1000 in PBS for 1 h at room temperature. Nuclei were stained with 1 μg mL -1 of 4 ′ -6-diamidino-2-phenylindole (DAPI). Slides were analyzed with a C2 Plus confocal laser scanning microscope (Nikon Instruments, Firenze, Italy). Images were processed using NIS Element Imaging Software (Nikon Instruments, Firenze, Italy).
2.14. Statistical Analysis. The experiments were carried out at least in triplicate, and values were reported as mean ± standard error. The differences among groups were evaluated by one-way ANOVA followed by Dunnett's or Bonferroni's test (Prism 5; GraphPad Software, San Diego, CA) (for cell culture data). Differences at the level p < 0:05 were considered statistically significant.

Results and Discussion
3.1. Bioactive Compounds in Different SCG Extracts. Before extract analysis, the analytical method has been validated testing linearity, limit of detection (LOD), limit of quantification (LOQ), and repeatability. The calibration curves were plotted on seven points by injecting seven different concentrations of mixtures of 30 analytes, and the respective determination coefficients (R 2 ) were calculated. The R 2 for each monitored molecule was ≥0.9937, which implied good linearity. LOD and LOQ were evaluated by injecting gradually lower concentration of standard mixtures, and the concentration with signal-to-noise ratio (SNR) of 3 was assigned to LOD that with SNR of 10 was assigned to LOQ. The LODs obtained ranged from 0.3 to 50 μg L -1 , while the LOQs were between 1 and 200 μg L -1 . The repeatability has been tested by injecting five replicates of three different concentrations of the standard solutions on the same day (run-to-run precision) and on three consecutive days (day-to-day precision). Relative standard deviation (RSD) % was utilized to define the intraday repeatability or run-to-run precision and interday repeatability or day-to-day precision. Run-to-run precision was between 1.7% and 3.9%, whereas day-to-day precision was between 4.3% and 7.4%.
Four SCG extracts were prepared, i.  Table 4. All extracts were prepared using ultrasound-assisted extraction (UAE), and the analytes were quantified using an HPLC-MS/MS system. The higher content of bioactive compounds was found in EtOH : H 2 O extract (71629:19 ± 3025:85 μg g -1 ) followed by MeOH : H 2 O (69891:35 ± 3102:12 μg g -1 ), MeOH (58796:31 ± 2756:32 μg g -1 ), and H 2 O (56792:60 ± 2521:98 μg g -1 ). Therefore, the solvent type significantly influenced the analyte extraction, and the EtOH : H 2 O and MeOH : H 2 O were shown to be the most efficient. Similar outcomes were reported in another recent work [21] which dealt with the chemical composition and some biological properties of different SCG and coffee silverskin (CS) extracts. Caffeine . Andrade et al. [42] have reported similar levels of caffeine in SCG, using UAE with different solvents, finding the best results with dichloromethane (38200.00 μg g -1 ) and ethanol (25700.00 μg g -1 ). Considering that the use of dichloromethane should be discouraged since it is associated with both acute and chronic toxicity in humans, including respiratory, central nervous system, and cardiovascular toxicity, carcinogenicity, and genotoxicity [43], the use of ethanol was  . Such molecules were also the most abundant in different SCG extracts prepared using a filter coffeemaker preceded by a defatting process with petroleum ether, as reported by Monente et al. [44]. At slightly lower concentrations, we identified gallic acid ( . Interestingly, cyanidin 3-glucoside, an anthocyanin that occurs in 7 Oxidative Medicine and Cellular Longevity coffee skin and pulp [47], has been found in all extracts ranging from 1.02 to 2:03 ± 0:05-0.09 μg g -1 but not delphinidin 3,5-diglucoside. As already reported by , iridoids and secoiridoids did not occur in spent coffee and probably in coffee beans. On the other hand, an alkaloid first isolated from the Cinchona tree known as a potent antimalarial agent, namely, quinine (1.44-3:23 ± 0:07 -0.12 μg g -1 ) [48], and a xanthone of Gentian plant [49], namely, isogentisin (1.12-1:65 ± 0:04-0.06 μg g -1 ), were detected in all SCG extracts.

Total Phenolic and Flavonoid Contents and DPPH Radical
Scavenging Activity of SCG Extracts. Table 5 reports the content of the phenolic and flavonoid compounds and the radical scavenging activity of different SCG extracts. The TPC has been spectrophotometrically measured, and data are reported as mg of gallic acid equivalents (GAE) per g of dry weight of SCG extract. The highest levels of phenolic compounds were found in E4 (112:65 ± 4:53 mg GAE/g) followed by E3 (95:12 ± 3:56 mg GAE/g), E1 (88:75 ± 2:13 mg GAE/g), and E2 (69:32 ± 2:11 mg GAE/g) extract. These levels were higher than those reported by other works when a simply solid-liquid extraction was employed [50][51][52]. For instance, Bravo et al. found SCG extracts with TPC of 17:44 ± 0:26 mg GAE/g using water for analyte extraction [52]. On the other hand, Al-Dhabi et al., who performed UAE at different conditions, obtained higher levels of TPC (32.81-36.23 mg GAE/g) [53]. The use of ultrasound during the extraction process increases the mass transfer due to acoustic cavitation effect generated by ultrasonic waves [54], and this can be the reason together with the coffee variability of higher TPC obtained in the current research. The total content of chlorogenic acids, one of the most important class of phenolic compounds in coffee, measured by the HPLC system, was characterized by the same abovementioned ranking, i.e., EtOH : H 2 O (18512:04 ± 895:32 μg g -1 ) followed by MeOH : H 2 O (17869:34 ± 925:26 μg g -1 ), MeOH (17252:40 ± 823:12 μg g -1 ), and H 2 O (10795:92 ± 772:65 μg g -1 ). In contrast, the highest level of TFC, expressed as mg of rutin equivalents (RE) per g of dried extract, was obtained in MeOH : H 2 O (6:29 ± 0:23 mg RE/g) followed by MeOH . The radical scavenging activity of SCG extracts has been evaluated by DPPH assay, and it was expressed as the IC 50 value (μg mL -1 ) which is the concentration of the extract necessary to cause 50% of DPPH inhibition. The solvent type influenced the antioxidant capacity of the extracts, and the highest radical scavenging activities were obtained with EtOH : H 2 O (196:25 ± 6:87 μg mL -1 ) and MeOH (215:35 ± 7:42 μg mL -1 ). Notably, the H 2 O extract (585:32 ± 25:32 μg mL -1 ) was the worst in terms of antioxidant capacity and it was characterized by lower content of bioactive compounds as well. The latter together with an inefficient extraction of low-polar com-pounds could be the reason of lower antioxidant activity. In fact, some lipophilic compounds which usually occur in coffee, e.g., diterpenes and tocopherols, are known as powerful antioxidants [55,56].

Neuroprotective Activity of SCG Extracts
3.3.1. Antioxidant Activity. The in vitro antioxidant activity of E1, E2, E3, and E4 has been investigated in neuron-like SH-SY5Y cells differentiated with retinoic acid. To study the potential cytotoxicity of E1, E2, E3, and E4, cells were treated with 1-200 μg mL −1 of the four extracts for 24 h and MTT assay was used to measure cell viability (Figures 1(a)-1(d)). The extracts were not cytotoxic up to 200 μg mL −1 except the E1 extract that induced a significant reduction of cell viability at 200 μg mL −1 . Interestingly, the treatment with the extracts led to a significant increase of cell viability. As MTT evaluates cell viability as the enzymatic conversion of the tetrazolium compound to water-insoluble formazan crystals by dehydrogenases occurring in the mitochondria of living cells [57], we can suppose that this cell viability increase could be caused by an intensification of mitochondrial respiration. All extracts are rich in caffeine that has been associated to an increased mitochondrial content due to the upregulation of peroxisome proliferatoractivated receptor gamma coactivator 1-alpha (PGC-1α) that modulates the nuclear respiratory factors 1 and 2 (NRF1/2) and mitochondrial transcription factor A (TFAM) [58][59][60][61]. Moreover, the treatment with caffeine of isolated human muscle fibers showed a direct effect on the mitochondrial activity by increasing the respiration rate and concomitantly decreasing the mitochondrial membrane potential [62]. In a previous study, we observed a significant increase of cell viability of differentiated SH-SY5Y cells treated with extracts containing caffeine, and these new data reinforce the hypothesis that caffeine could be the compound responsible for this effect [5]. Of course, further investigations are needed to determine a direct involvement of caffeine in the enhancement of mitochondrial respiration in SH-SY5Y cells. To clarify if the observed increase in cell viability is only linked to an increase in mitochondrial activity, we measured cell viability  Oxidative Medicine and Cellular Longevity with a different viability assay. Differentiated SH-SY5Y cells were treated with 50 μg mL −1 of each extract for 24 h, and cell viability was measured by the trypan blue assay (Figure 1(e)) that is based on the principle that living cell membranes are intact and exclude trypan blue, whereas the dead cells are permeable to the dye. Of note, all treatments did not increase cell viability suggesting that the increase in cell viability measured by MTT assay is related to an increase in mitochondrial activity. Another hypothesis to explain the observed increase in cell viability could be a corresponding increase in cell proliferation. To investigate this aspect, differentiated SH-SY5Y cells were treated with 50 μg mL −1 of each extract for 24 and the total cell number was counted (Figure 1(f)). Interestingly, the treatments did not modify the cell number, confirming that the observed increase in cell viability measured by MTT assay is related to a higher rate of mitochondrial respiration and not to an increased proliferation.
The antioxidant activity of the extracts has been evaluated pretreating SH-SY5Y cells with 1-100 μg mL −1 of the extracts for 24 h before exposing the cells to 700 μM H 2 O 2 to induce oxidative stress (Figure 2). At the lowest concentrations, only the E4 extract significantly increased cell viability with respect to H 2 O 2 -treated cells, meanwhile, at 10 μg mL −1 , E1 was also able to significantly increase cell viability. E3 significantly counteracted oxidative stress at 50 μg mL −1 and E2 only at the highest tested concentrations. Of note, at 50 μg mL −1 , E1 increased cell viability with respect to peroxide-treated cells by about 22%, meanwhile E3 and E4 by about 16%, evidencing a higher ability of E1 in counteracting oxidative stress-induced damage in SH-SY5Y cells.
To further investigate the antioxidant activity of the extracts, SH-SY5Y cells were treated with 1-100 μg mL −1 of each extract and the DCFH-DA assay was used to evaluate the effect on intracellular ROS production (Figure 3). The   Figure 2: Cytoprotective activity of the extracts in SH-SY5Y cells exposed to H 2 O 2 . Cells were pretreated with 1-100 μg mL −1 of each extract for 24 h, exposed to 700 μM H 2 O 2 for 1 h before measuring cell viability by MTT assay. Each bar represents means ± SEM of at least four independent experiments. Data were analyzed by one-way ANOVA followed by Bonferroni's test.°p < 0:05 with respect to CTRL; * p < 0:05 with respect to H 2 O 2 . results showed that E1 and E4 were the most effective ones in reducing ROS levels, meanwhile E3 reduced ROS levels only at the 100 μg mL −1 and E2 did not influence this parameter. These results confirm that E1 and E4 are the extracts with the strongest antioxidant activity.
These biological results on the antioxidant activity of the extracts are in agreement with the results obtained by DPPH assay ( Table 2). In particular, in SH-SY5Y cells, E1 and E4 extracts were the most effective ones in terms of antioxidant activity, meanwhile E2 showed the lowest activity. As previously underlined, the low antioxidant capacity of E2 could be caused by an inefficient extraction of low-polar compounds, which usually occur in coffee and are known for their elevated antioxidant activity [55,56].

Oxidative Medicine and Cellular Longevity
Evidence is cumulating which shows that many phytochemicals exert antioxidant activity through an indirect antioxidant mechanism, i.e., enhancing the expression of antioxidant enzymes and cytoprotective proteins [63][64][65][66]. To verify if the extracts modulate the endogenous antioxidant system, we treated SH-SY5Y cells with 50 μg mL −1 of each extract before analyzing the expression of the antioxidant enzymes glutathione peroxidase (GR), heme oxygenase 1 (HO1), NADP(H) oxidoreductase 1 (NQO1) and thioredoxin reductase by RT-PCR (Figure 4). All the extracts significantly upregulated HO1, NQO1, and TRX, meanwhile GR expression was significantly increased only by E1, E3, and E4.
We also evaluated the expression of these antioxidant enzymes in the presence of H 2 O 2 . In particular, SH-SY5Y cells were pretreated with 50 μg mL −1 of each extract and then exposed to H 2 O 2 before analyzing mRNA levels of GR, HO1, NQO1, and TRX ( Figure 5). H 2 O 2 exposure significantly reduced the expression of all the tested genes in respect to control cells. Considering the short H 2 O 2 exposure, the observed downregulation of these genes could be probably ascribed to the H 2 O 2 -induced oxidation of the corresponding mRNA. E1 was able to significantly upregulate all the four genes with respect to both H 2 O 2 and controls. E2 treatment did not influence GR and TRX expressions with respect to H 2 O 2 -treated cells, meanwhile slightly but significantly upregulated HO1 and NQO1 expressions. E3 significantly increased mRNA levels of HO1, NQO1, and TRX with respect to H 2 O 2 -treated cells and upregulated HO1 and TRX with respect to control cells. E4 significantly increased the expression of HO1, NQO1, and TRX with respect to both H 2 O 2 and controls, meanwhile significantly upregulated GR only with respect to H 2 O 2.
Considering the strong upregulation of HO1 with respect to the other tested genes, we performed an immunoblotting analysis to confirm HO1 induction also at a protein level. SH-SY5Y cells were pretreated with 50 μg mL −1 of each extract and then exposed to H 2 O 2 before western blot analysis ( Figure 6). H 2 O 2 exposure reduced HO1 protein levels with respect to control cells, even if not significantly. On the contrary, E1 strongly and significantly increased the expression of HO1, confirming the expression data.
Interestingly, E1, with respect to the other extracts, showed a marked ability to upregulate the four antioxidant enzymes both in the absence and in the presence of H 2 O 2 suggesting that the higher ability of E1 to protect SH-SY5Y cells against oxidative stress could be ascribed to its ability to strongly upregulate the endogenous antioxidant system.
Considering the characterization of the extracts in terms of bioactive compound content (Table 1), E1 showed the highest content of (-)-epicatechin and isogentisin with respect to the other extracts. Of note, no correlation was evidenced between (-)-epicatechin content and the different parameters tested to analyze the antioxidant activity of the extracts. On the other hand, isogentisin content was positively correlated with the protection against H 2 O 2 (r = 0:9745, p < 0:05) and GR expression (r = 0:9575, p < 0:05) and inversely correlated with ROS levels (r = −0:9604, p < 0:05). The xanthone isogentisin is a characteristic constituent found in plants such as Gentianaceae [67]. Very few studies investigated its bio-logical activity. In particular, isogentisin has been shown to counteract smoking-caused injury in human umbilical vein endothelial cells (HUVECs) [68] and to inhibit monoamine oxidase types A and B in rat brains [69,70]. Monoamine oxidase inhibitors are considered important agents for the treatment of depression, anxiety, and neurodegenerative disorders, including Alzheimer's and Parkinson's diseases [71,72]. From this point of view, further studies will be necessary both to investigate the antioxidant activity of pure isogentisin in neuron cells and to verify if the E1 extract also can act as a monoamine oxidase inhibitor.

3.3.2.
Anti-Inflammatory Activity. The in vitro antiinflammatory activity of the extracts has been investigated in microglial BV-2 cells. Microglia are equivalent to macrophages in the brain and represent the first and most important line of defense in the central nervous system. Under physiological conditions, microglia have a key role in neuronal survival through the production of neurotrophic factors and the phagocytosis of dead cells, cellular debris, protein aggregates, and invading pathogens [73]. However, excessively activated microglia can lead to neurotoxicity through the production of proinflammatory mediators such as tumor necrosis factor alpha (TNF-α), nitric oxide, interleukin-1β (IL-1β), IL-6, and ROS [74]. Different studies have shown that microglia play an important role in the onset and progression of neurodegenerative diseases such as Parkinson's disease and Alzheimer's disease [75][76][77].
Prior to investigating the effect of the extracts on BV-2 microglia-mediated neuroinflammation, we assessed the potential cytotoxicity of E1, E2, E3, and E4 on BV-2 microglial cells using MTT assay (Figure 7). The extracts were not cytotoxic up to 100 μg mL −1 , meanwhile all extracts were cytotoxic at 200 μg mL −1 as demonstrated by a significant   Figure 6: Protein level of HO1 in SH-SY5Y cells treated with the extracts and exposed to H 2 O 2 . Cells were treated with E1, E2, E3, and E4 (50 μg mL −1 ) and after 24 h exposed to H 2 O 2 700 μM for 1 h. Immunoblotting was performed using anti-HO1. Data are expressed as fold over CTRL and normalized by β-actin. Each bar represents mean ± SEM of three independent experiments. Data were analyzed with a one-way ANOVA followed by the Bonferroni's test.°p < 0:05 vs. CTRL; * p < 0:05 vs. H 2 O 2 .

12
Oxidative Medicine and Cellular Longevity The anti-inflammatory activity of the extracts was evaluated pretreating BV-2 cells with different concentrations (1-100 μg mL −1 ) of the extracts for 24 h before exposing the cells to lipopolysaccharide (LPS) to induce inflammation ( Figure 8). LPS is the most widely used inflammatory mediator to activate microglial cells in vitro and triggers the proinflammatory signaling cascade [78,79]. LPS treatment significantly reduced cell viability with respect to control cells by~40%. Interestingly, E1, E2, and E4 were able to significantly increase cell viability with respect to LPS-treated cells, and at 50 μg mL -1 , all of them were able to maintain cell viability to a value comparable to control cells. On the other hand, E3 did not show any protective effect against LPS-induced damage.
As it has been shown that LPS induces oxidative stress [80,81], we measured intracellular ROS levels in BV-2 cells pretreated with 50 μg mL −1 of the extracts for 24 h and then activated by LPS ( Figure 9). As expected, LPS significantly increased intracellular ROS levels with respect to controls. E1, E3, and E4 significantly reduced ROS levels compared to LPS-treated cells, meanwhile E2 did not modify ROS levels with respect to LPS-treated cells, in agreement with the results obtained in SH-SY5Y cells.
Since proinflammatory cytokines and enzymes including tumor necrosis factor α (TNF-α), interleukin 1β (IL-1β), cyclooxygenase 2 (COX-2), and inducible nitric oxide synthase (iNOS) are crucial mediators of neuroinflammation, we next measured the effects of the extracts on these inflammatory mediators in LPS-stimulated BV-2 microglial cells  Figure 10: Expression of proinflammatory cytokines and enzymes in BV-2 cells treated with the extracts. Cells were treated with E1, E2, E3, and E4 (50 μg mL −1 ) for 24 h, exposed to 100 ng mL-1 LPS for 24 h and TNF-α, IL1-β, iNOS, and COX-2 mRNA levels were measured by RT-PCR. Data are expressed as relative abundance compared to untreated cells. Each bar represents the mean ± SEM of three independent experiments. Data were analyzed with a one-way ANOVA followed by Bonferroni's test.°p < 0:05 vs. CTRL; * p < 0:05 vs. LPS.
14 Oxidative Medicine and Cellular Longevity ( Figure 10). BV-2 microglial cells were pretreated with 50 μg mL -1 of E1, E2, E3, and E4, followed by LPS for 24 h. Total RNA was isolated, and proinflammatory cytokine and enzyme expressions were measured using RT-PCR. As expected, LPS significantly increased the expression of TNF-α, IL-1β, COX-2, and iNOS with respect to control cells. In agreement with the MTT data, E3 did not show any ability to inhibit LPS-induced expression of TNF-α and COX-2 and significantly increased the expression of IL-1β with respect to LPS. Moreover, E3 was able to significantly reduce iNOS expression with respect to LPS, but the extent of this reduction was very small, maintaining iNOS expression to levels strongly higher than those measured in control cells. E1 and E4 had no effect on LPS-induced IL-1β expression, meanwhile they were able to significantly reduce iNOS and COX-2 expression with respect to LPS-treated cells.
These two extracts showed opposite behaviors with respect to TNF-α expression: E1 significantly inhibited LPSinduced expression of this cytokine, on the contrary E4 significantly increased its expression with respect to LPStreated cells. Of note, E2 was the most effective extract and significantly and strongly inhibited the expression of all proinflammatory mediators analyzed. These results are partially in agreement with the data on cell viability that showed that E1, E2, and E4 had a similar effect in counteracting LPSinduced damage, and all of them were able to completely protect cells against neuroinflammation at 50 μg mL -1 . This discrepancy could be explained by the different mechanisms by which these extracts counteract LPS-induced damage. The mechanism behind E2 protection is easy to understand as this extract exerts a strong anti-inflammatory activity that significantly and strongly reduces the expression of all proinflammatory mediators investigated. On the contrary, the protective activity of E1 and E4 cannot be explained only in terms of their ability in reducing proinflammatory cytokine and enzyme expression levels. Taking into consideration the results also obtained in SH-SY5Y, we can suggest that E1 and E4 were able to protect BV-2 cells against LPSinduced damaged thanks to their antioxidant activity. In fact, it is widely accepted that LPS generates ROS that trigger oxidative stress and cell damage [82][83][84]. Of note, E1, E3 and E4, but not E2 at 50 μg mL -1 , significantly reduced ROS levels in BV-2 cells. The NF-κB pathway is a key mediator of inflammation and is activated via Toll-like receptors (TLRs) resulting in increased cytokine and chemokine production [85]. It has been observed that the activation of NF-κB and release of its subunits play a crucial role in the onset and progression of neurodegenerative disorders [86,87]. Moreover, transcription of TNF-α, IL-1β, iNOS, and COX-2 is regulated by the transcription factor NF-κB. To further elucidate the mechanisms of the extracts on the inhibition of the expression of these proinflammatory mediators in BV-2 cells, the effect of E1, E2, E3, and E4 on NF-κB activation was investigated by confocal microscopy (Figure 11). BV-2 cells were pretreated with 50 μg mL -1 of the extracts, exposed to LPS for 24 h and immunostained with a primary antibody against NF-κB p65, followed by Alexa Fluor 488-conjugated secondary antibody. LPS induced a strong increase in NF-κB protein levels and triggered its translocation to the nucleus with respect to control cells. In agreement with the previous data, E1, E2, and E4 reduce NF-κB protein levels with respect to LPStreated cells, confirming their anti-inflammatory ability. Of note, E1 and E2 maintained NF-κB protein levels to values comparable to control cells. On the other hand, E3 maintained NF-κB protein levels to a value comparable to LPStreated cells.
One of the main receptors mediating the activation of microglia and release of proinflammatory mediators is Tolllike receptor 4 (TLR4). LPS is a well-characterized ligand of TLR4 [88]. Dimerization of TLR4 induces the downstream activation of NF-κB signaling, triggering the activation of immune cells such as microglia [89]. On these bases, we further investigated the effects of the extracts on TLR4 cell surface expression by cytofluorimetric analysis (Figure 12).  Figure 11: NF-κB nuclear translocation in BV-2 cells treated with the extracts. Cells were treated with 50 μg mL -1 of each extracts for 24 h, activated by LPS for 24 h. Cells were immunostained with a primary antibody against NF-κB p65, followed by secondary Alexa Fluor 488-conjugated anti-rabbit IgG antibody (green), and cell nuclei (blue) were visualized with DAPI. Scale bar: 10 μm. Tests were performed in triplicate.
LPS induced a strong and significant increase of TLR4 surface expression with respect to control cells. According to the previous results, E2 showed the strongest effect in significantly reducing TLR4 surface expression both with respect to LPS-treated cells and control cells. E1, E3, and E4 significantly reduced TLR4 surface expression with respect to LPS-treated cells and, in agreement with the previous results, E3 was the least effective.
Considering the characterization of the extracts in terms of bioactive compound content (Table 2), it is not possible to find any correlation among the anti-inflammatory activity and the presence of specific compounds in the extracts. E2, the most effective extract in counteracting neuroinflammation, does not contain compounds that are not present in the other extracts; moreover, all characterized compounds are present at a lower concentration with respect to the other extracts. Therefore, we can hypothesize that E2 could contain some bioactive compounds that we have not identified. Further researches are needed to better characterize the composition of this extract.

Conclusions
The extract analysis evidenced that the different solvents had a profound impact on the composition of the extracts. In particular, the higher content of potential bioactive compounds was found in EtOH : H 2 O and MeOH : H 2 O extracts. Interestingly, the biological data revealed that the richest extracts in terms of compounds were not the most effective in counteracting oxidative stress and inflammation. The MeOH extract showed the strongest antioxidant activity in neuron-like SH-SY5Y cells, reducing intracellular ROS levels and upregulating endogenous antioxidant enzymes such as NQO1, GR, TRX, and HO1. This effect seems to be related to its higher content of isogentisin with respect to the other extracts. The H 2 O extract elicited the highest antiinflammatory activity, markedly reducing the expression of proinflammatory mediators by the TLR4/NF-κB pathway. Of note, none of the identified compounds in the H 2 O extract can explain its higher anti-inflammatory activity with respect to the other extracts. For this reason, further studies should be carried out to better characterize this extract and identify potential bioactive compounds responsible for its antiinflammatory activity. In conclusion, the antioxidant and anti-inflammatory properties of the extracts suggest that SCGs are a valuable source of nutraceuticals that could be used to prevent/counteract neurodegeneration.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.  Figure 12: Effect of the extracts on cell surface expression of TLR4 in BV-2 cells. Cells were treated with E1, 50 μg mL -1 of each extract for 24 h, exposed to 100 ng mL -1 LPS for 24 h, and TLR4 surface expression was evaluated by flow cytometry. Data are expressed as percentage compared to LPS-activated cells. Each bar represents the mean ± SEM of three independent experiments. Data were analyzed with one-way ANOVA followed by Bonferroni's test.°p < 0:05 vs. CTRL, * p < 0:05 vs. LPS.