Effect of Extraction Solvent and Temperature on Polyphenol Profiles, Antioxidant and Anti-Inflammatory Effects of Red Grape Skin By-Product

A fully-detailed LC-MS qualitative profiling of red grape skin, extracted with a mixture of ethanol and water (70:30 v:v) has permitted the identification of 65 compounds which can be classified into the following chemical classes: organic and phenolic acids (14 compounds), stilbenoids (1 compound), flavanols (21 compounds), flavonols (15 compounds) and anthocyanins (14 compounds). The extraction yield obtained with water at different temperatures (100 °C, 70 °C, room temperature) was then evaluated and the overall polyphenol content indicates that EtOH:H2O solvent is the most efficient and selective for polyphenol extraction. However, by analyzing the recovery yield of each single polyphenol, we found that water extraction under heating conditions is effective (extraction yield similar or even better in respect to the binary solvent) for some polyphenolic classes, such as hydrophilic procyanidins, phenolic acids, flavonol glucosides and stilbenoids. However, according to their lipophilic character, a poor yield was found for the most lipophilic components, such as flavonol aglycones, and in general for anthocyanins. The radical scavenging activity was in accordance with the polyphenol content, and hence, much higher for the extract obtained with the binary solvent in respect to water extraction. All the tested extracts were found to have an anti-inflammatory activity in the R3/1 cell line with NF-kb reporter challenged with 0.01 µg/mL of IL-1α, in a 1 to 250 µg/mL concentration range. An intriguing result was that the EtOH:H2O extract was found to be superimposable with that obtained using water at 100 °C despite the lower polyphenol content. Taken together, the results show the bioactive potentialities of grape skin extracts and the possibility to exploit this rich industrial waste. Water extraction carried out by heating is an easy, low-cost and environmentally friendly extraction method for some polyphenol classes and may have great potential for extracts with anti-inflammatory activities.


Introduction
Grapes are amongst the most cultivated fruits in the world. In Europe alone, around 3.5 million hectares are dedicated to grape cultivation with a production of almost 27 million tons of fruit [1]. It is also estimated that 14.5 million tons of grape by-products are generated annually [2]. The main component of this waste is the grape pomace, mainly composed (50-65%) of grape skin [3]. Even though it is mostly used as compost or animal fertilizer, the phenolic-rich composition of the skin is what supports its use a source of bioactive phytochemicals. It contains anthocyanins alongside various members of the flavonoid family (flavan-3-ols, flavonols and flavanones), which have shown healthy activity as antioxidant and anti-inflammatory agents [4][5][6].
Hence, grapes skin represent a valuable source of bioactive polyphenols and would represent a valuable industrial source to satisfy the growing demand. A recent study conducted by Transparency Market Research, a global market intelligence group, has predicted a further increase in the polyphenol market owing to increasing demand and market size. This study indicates that the global demand for polyphenols in the global polyphenol market was valued at USD 761.9 million in 2020 and is expected to reach USD 969.2 million by the end of 2026, growing at a CAGR (Compound Annual Growth Rate) of 3.5% between 2021-2026 [7].
The appropriate industrial evaluation of red grape skin needs a suitable and green extraction method which should have the least experimental set-up, a low cost and have environmental and user-friendly characteristics.
Although many technological advancements in polyphenol extraction have been proposed, such as supercritical fluid extraction (SFE), ultrasound-assisted extraction (UAE), microwave assisted extraction (MAE), pressurized liquid extraction (PLE) and pressurized hot water extraction (PHWE) [8,9], the solid-liquid extraction (SLE), which simply consists of solvent application and leaching, remains the most popular. Since the structure of phenolic compounds determines their solubility in solvents of different polarity, the type of extraction solvent may have a significant impact on the yield of extraction polyphenols from plants material [10].
There are some reports concerning optimization of extraction conditions of the phenolic compound content and antioxidant activities of some plant foods, and the optimal procedure is usually different for different plant matrices and depends on the polyphenol composition [11][12][13]. For instance, acetone has been proven efficient in polyphenol extraction from lychee flowers compared to methanol, water and ethanol [14]. However, another study reported water as the best solvent for polyphenol extraction from walnut green husks [15]. In general, binary solvents such as ethyl acetate, acetone, methanol and ethanol with a differing water content are used as suitable extraction solvents of polyphenols from raw material.
However, legal limitations for solvent residues and restrictions on the use of conventional organic solvents are becoming more and more rigorous, especially in the fields of food and pharmaceutical, and together with the use of flammable and costly solvents, they represent a limit for industrial extraction.
In this work, a selected red grape skin by-product was extracted using water extraction at different temperatures and the metabolomic, antioxidant and anti-inflammatory profiles compared in respect to those obtained using EtOH:H 2 O 70:30 (% v:v). Our aim is to fully characterize the profile and activities of the polyphenol fractions isolated by a water-based extraction of red grape skin and evaluate the temperature effect thus to understand if this easy, low-cost and environmentally friendly extraction method may prompt industrial interest. Polyphenol profiling was carried out by targeted LC-MS analyses and the antioxidant and anti-inflammatory activities of the extracts tested by in vitro and cell methods, respectively.

Sample Preparation
Extraction from dried red grape skin derived from controlled and selected wineries in central Italy was carried out under the following conditions: 2 g of raw material in 40 mL of EtOH-H 2 O 70:30 (% v/v) under magnetic stirring for 24 h at room temperature (extract A); 2 g of raw material in 40 mL of EtOH-H 2 O 70:30 (% v/v) at 50 • C for one hour followed by 24 h at room temperature under magnetic stirring (extract A50); 2 g of raw material in 40 mL of H 2 O (100%) at 100 • C for one hour followed by 24 h at room temperature under magnetic stirring (extract B); 2 g of raw material in 40 mL of H 2 O (100%) at 70 • C for one hour followed by 24 h at room temperature under magnetic stirring (extract C); 2 g of raw material in 40 mL of H 2 O (100%) under magnetic stirring for 24 h at room temperature (extract D).
Samples were then centrifuged at 5000× g for 10 min, filtered on 0.45 µm filters and dried overnight under a vacuum. The conditions used combined the idea of a lowcost and simple method easily transferrable to industrial scale. Room temperature was selected as the low-cost condition for 24 h, since it is demonstrated that a prolonged extraction period lead to the degradation of polyphenols [16]. A limited period of warming (1 h) was applied to avoid polyphenols degradation (especially anthocyanins [17]) at two different temperatures: boiling point (100 • C) and 70 • C, which is a mean of the most used temperature for water extract (60-80 • C) applied in the literature.

Determination of Total Polyphenol Content
Determination of total polyphenol content was conducted spectrophotometrically by a slight modification of the method described by Dewanto et al. [29]. Briefly, 12.5 µL of diluted extract (1000 µg/mL) was spiked with 50 µL of water, 12.5 µL of Folin-Ciocalteu reagent and 125 µL of Na 2 CO 3 (7%) in a 96-well plate. The mixture was incubated at room temperature for 90 min in the dark, then the absorbance was read at 760 nm in a spectrophotometric microplate reader (BioTek's PowerWave HT, Winooski, VT, USA). The results were then compared with a curve of gallic acid (0-1000 µg/mL) prepared with the same protocol of samples and expressed as % w/w (grams of polyphenols per one gram of starting material or dry extract).

Determination of Tannin Content
Determination of total tannin content was carried out by the vanillin assay [30] in 96-plate wells. Briefly, 150 µL of vanillin (4% solution in methanol) and 75 µL of concentrated HCl were added to 2.5 µL of diluted extract (1000 µg/mL). The mixture was mixed at room temperature for 15 min and then the absorbance measured against the blank at 500 nm using a microplate reader (BioTek's PowerWave HT, Winooski, VT, USA). The readings were compared to standards containing known amounts of (+)-catechin (0-1000 µg/mL) and prepared with the same protocol as the samples. The results were then expressed as % w/w (grams of polyphenols per one gram of starting material or dry extract).

Antioxidant Activity
The antioxidant capacity was evaluated by the DPPH radical-scavenging method [32], with a few modifications. An aliquot of 100 µL of the extract solution at different concentrations (1-25 µg/mL) was spiked with 750 µL of ethanol and 400 µL of acetate buffer (0.1 M, pH 5.5), mixed and spiked with 250 µL of DPPH ethanolic solution (0.5 × 10 −3 M). After 90 min at room temperature and in the dark the absorbance at 515 nm was measured for each sample analyzed in triplicate with a UV reader Shimadzu™ UV 1900 (Shimadzu, Milano, Italia). The percentage of inhibition was calculated as expressed by Equation (2) and the results expressed as mean ± SD.

Anti-Inflammatory Activity
To assess the in vitro anti-inflammatory activity of the four different extracts, and evaluate the influence of the different extraction methods, the same cell model used by Baron et al. was deployed [18]. Extracts ranging from 1 to 250 µg/mL were incubated with R3/1 cells transduced with the NF-kb reporter gene and challenged with 0.01 µg/mL IL1α. The cell viability of the extracts on the same cell line was evaluated for the same concentrations used in the anti-inflammatory assay by MTT assay as reported by Baron et al. [18].

Absolute Quantification of Polyphenols, Tannins and Anthocyanins
As shown in Table 2, the % of dry extract in respect to the starting material was the highest for 100% water at 100 • C and reduced on the basis of the extraction temperature, reaching the lowest residue amount for EtOH:H 2 O and EtOH:H 2 O at 50 • C. By contrast, the extraction yield of polyphenols calculated in respect to the dry extract was much higher (almost 4 fold) in EtOH:H 2 O in respect to water extracts. The % yield of polyphenols in respect to the starting material was at 4% for EtOH:H 2 O and reduced by almost 50% in water and further reduced at 70 • C and room temperature. The data well indicate that EtOH:H 2 O solvent is the most efficient and selective for polyphenol extraction, while water is able to extract not only polyphenols but also many other constituents in a temperature dependent manner as already observed by González-Centeno et al. [33]. We observed no difference between the two hydroalcoholic extracts.  Table 3 reports the % of tannins and anthocyanins in respect to the dry extract and starting material. For EtOH:H 2 O, the % in respect to the starting material was 3.28 and 0.271, respectively, an amount almost two times and one and a half times higher in respect to water at 100 • C. The extraction yield of tannins and anthocyanins reduced on the basis of the temperature for the water extracts, while for the two hydroalcoholic extracts overlapped.  (Table 4), as expected we can generally observe a higher anthocyanins yield when acid is added in the extraction solvent and the temperature is lower, but their content is highly variable depending on the grape variety [35,36]. On the other hand, tannin yield seems not to be affected by acid or temperature. Observing the results, where a higher tannins/polyphenols ratio is found, there is a lower anthocyanins/polyphenols ratio due to the different physico-chemical properties of the two classes.  The comparison between the extract A and extract A50 profiles (which are almost overlapping) is reported in Figure S1 of the Supplementary Materials. The peak eluting at 1.5 min and detected in all the extracts was assigned to tartaric acid (peak 1). The TIC showing the highest number of peaks is that referred to EtOH-H 2 O (panel A); eluted peaks of the extracts can be clustered into three main groups based on their RT: group (A), which includes peaks eluting between 3 and 20 min, group (B) medium lipophilic compounds, eluting between 20 and 40 min, and group (C) the most lipophilic compounds, eluting after 40 min. TICs relative to groups A, B and C are reported in Figures 2-4, respectively. The LC-ESI-(+)-MS (positive ion mode) profiles are reported in Figure 5. The comparison between the extract A and extract A50 positive profiles (which are almost overlapping) are reported in Figure S2 of the Supplementary Materials. Compound identification was carried out on the basis of the accurate MS and MS/MS fragmentation and the lists of identified peaks in all the extracts are reported in Table 5 (relative to the acquisition in negative ion mode) and Table 6 (relative to the acquisition in positive ion mode). The 65 identified compounds can be classified into the following chemical classes: organic and phenolic acids (14 compounds), stilbenoids (1 compound), flavanols (21 compounds), flavonols (15 compounds) and anthocyanins (14 compounds).     Table 5.  Table 6.  As expected, most of the procyanidins (18 out of 21) eluted in group A, as the most hydrophilic compounds of the extract. A total number of 21 procyanidins were identified from monomers (catechin, epicatechin, gallocatechin, epigallocatechin, epicatechin-gallate) to oligomers, such as hexamer gallate.

Compound Identification, Absolute and Relative Content
The relative content of procyanidins calculated for the EtOH:H 2 O extract on the basis of the area of the peaks reconstituted by setting the [M-H] − values as filter ion is summarized in Table 7. Dimers and the corresponding gallate esters, reaching more than 40% of the total procyanidins, represent the main species, followed by monomers (catechin, epicatechin and gallates), more than 25%, and trimers, with a relative content of 21%, followed by oligomers such as tetramer, pentamer and hexamer. Table 7. Relative content of flavanols in the EtOH-H 2 O extract. The % was determined by measuring the peak area of each compound in respect to the sum of the peak areas of all the identified flavanols.

Compound
Relative Content (%) Procyanidins were present in all four extracts but with a different relative abundance as shown in Table 8. The relative percentage of each extract (where i = A or A50 or B or C or D) was calculated with the following formula with the area of the peaks identified in extract A set as 100%: Area analyte /Area IS extract i Area analyte /Area IS extract A × 100 (3) For some of the most hydrophilic compounds (procyanidins eluting up to 12.3 min) such as gallocatechin and epigallocatechin, the yield was much higher when the extraction process was carried out with water in respect to both EtOH:H 2 O A and A50. For other hydrophilic components such as procyanidin trimer, catechin, epicatechin and some procyanidins B, the extraction yield in water was slightly but significantly higher and this in accordance with their hydrophilicity. Heating at 50 • C seems not to globally affect the yield of extraction.
The extraction yield in water solvents was found reduced for procyanidins eluting from 12 to 19 min by almost 50% and much lower when the water-based extraction was carried out without a heating step which was also found to significantly increase the extraction of the most hydrophilic components.
In Table 9 are reported the absolute concentrations (µg/mL) of epicatechin in all the extracts.

Flavonols
Flavonols were identified mainly as glycosides (8 out of 15), which eluted in group B. A glucuronide derivative (quercetin 3-glucuronide) and 6 aglycones (group C, except for myricetin) were also detected. Table 10 reports the relative abundance of flavonols in the EtOH:H 2 O. Quercetin is the main flavanol, accounting for almost 40%, followed by its glucuronide (16%), syringetin 3-glucoside (11%), myricetin (9%) and kaempferol (almost 6%). The other constituents have a relative content lower than 5%.  Table 11 shows the relative percentage in the five extracts. For all the identified flavonols both as aglycones and glucosides, the highest extraction yield was obtained when EtOH:H 2 O mixture was used as extraction solvent and the recovery yield followed the following order: extract B > extract C > extract D. By calculating the mean of %, the yield of extraction for the conjugated forms reduced to almost 66% when water was used in respect to EtOH:H 2 O and further decreased for aglycones which reduced by almost 88% in extract B and 94% in extract C to reach an almost negligible yield for H 2 O extraction in RT condition (0.3%). The extraction yield of EtOH:H 2 O after heating is comparable to that without heating both for glycosides and aglycones.  Table 12 reports the absolute concentrations of quercetin 3-galactoside in all the extracts.

Stilbenoids, Phenolic and Organic Acids
Only one stilbenoid was detected in all the extracts, resveratrol glucoside, whose extraction yield was similar in extracts A, A50, B and C but reduced by almost 30% in extract D. A total number of 14 organic and phenolic acids were identified, whose relative abundance in extract A is reported in Table 13. Tartaric acid represents the main organic acid, accounting for more than 40%. Taking into consideration only phenolic acids, ethyl gallate is the main species (30.2%) followed by p-coumaroyl-glucosides (more than 20%) gallic acid (15.4%) and galloyl glucose (14.5%). Table 13. Relative content of phenolic acids in the EtOH-H 2 O extract. The % was determined by measuring the peak area of each compound in respect to the sum of the peak areas of all the identified phenolic acids.

Compound
Relative Content (%) As shown in Table 14, some of the organic acids were extracted in a greater yield with water based solvents, such as protocatechuic acid, caftaric acid and p-coumaroyl hexoside 1, while for some others, such as galloyl glucose, protocatechuic acid hexoside, fertaric acid and ethyl gallate, the extraction yield was greater with EtOH:H 2 O (both extract A and A50). The % mean calculated not considering tartaric acid was higher by almost 40% for extract B in respect to A and reduced in C and D, this latter similar to A.  Table 15 reports the absolute concentrations of protocatechuic acid and ethyl gallate in all the extracts.

Anthocyanins
As summarized in Table 6, two pyroanthocyanins (vitisin A and B) and twelve anthocyanins were identified: 5 glucosides, 1 caffeoyl glucoside, 2 acetyl glucosides and 4 coumaroyl glucosides. The relative content of anthocyanins in the EtOH:H 2 O extract is summarized in Table 16. Malvidin coumaroyl glucoside is the most abundant component, reaching a relative content of almost 55%, followed by malvidin glucoside (14%) and malvidin acetyl glucoside (almost 8%), and petunidin coumaroyl glucoside (6.4%) while the relative content of the other compounds was lower than 5%. Table 16. Relative content of anthocyanins in the EtOH-H 2 O extract. The % was determined by measuring the peak area of each compound in respect to the sum of the peak areas of all the identified anthocyanins.

Compound
Relative Content (%) Malvidin coumaroyl glucoside was the most abundant also in the water extracts but its relative content is reduced in respect to malvidin acetyl glucoside as explained by its greater polarity.
The relative percentage of the aqueous extracts in respect to EtOH:H 2 O is reported in Table 17, showing that for most of the identified anthocyanins, the recovery yield followed the following order: extract A > extract A50 > extract B > extract C > extract D and the yields reduced proportionally with lipophilicity (RT). Vitisin B was an exception, since its extraction yield was greater in extracts C and D. By calculating the % mean, the anthocyanin content is almost 40% in B in respect to A, similar to C and reduced to 28% in D. Regarding the two hydroalcoholic extracts, heating reduces anthocyanins content of about 11%.  Table 18 reports the absolute concentrations of malvidin 3-glucoside in all the extracts. Considering the almost overlapping analytical profile of extract A and A50, the activities were evaluated for the extract A only.

DPPH Assay
The radical scavenging activity was measured by the DPPH test as reported in Table 19. By calculating the activity as IC 50 of the dry residue, as expected, extract A was significantly more active than water extracts, according to the polyphenol contents, which represent the main antioxidant components of the extract. The radical scavenging activity of the extracts was almost superimposable when the activity was normalized in respect to the polyphenol content demonstrating that the extracted polyphenols have a quite similar radical scavenging activity.

Anti-Inflammatory Activity
The MTT assay ( Figure 6) indicates that cell viability was not affected by the four extracts at all concentrations tested and up to 250 µg/mL. The anti-inflammatory activity was then tested by measuring the NF-kB dependent luciferase activity induced by IL1-α stimulus. The protective effect of the four extracts is reported as % of inhibition of luciferase activity in respect to control cells (Figure 7). For all the extracts, a dose-dependent effect was observed. All the data were statistically significant (p < 0.01) except for the 1 and 10 µg/mL concentrations of extract C and D and 10 µg/mL for extract A. The anti-inflammatory activity of extract A was found superimposable to B and this latter higher than C and D. The order of anti-inflammatory potency for the water extracts was expected on the basis of the content of polyphenols which, in line of principle, could be considered as the main anti-inflammatory components. However, the superimposable activity of A with B cannot be explained on the basis of the polyphenol content since it is more than three times higher in A in respect to B. This apparent contradiction can be explained in different ways. The first, is that the extracted polyphenols do not have the same anti-inflammatory potency and that those extracted in a similar way in A and B, such as the hydrophilic procyanidins, are the most effective anti-inflammatory components. A second explanation is that beside polyphenols, other constituents present in the water extracts but not present or present to a lesser extent in A, do have an antiinflammatory activity which adds up to those of the polyphenols. A third explanation is that the water-soluble constituents such as macromolecules (proteins, polysaccharides) significantly influence polyphenol bioaccessibility in the cells [37]. Regarding this latter aspect it could be considered that water extraction could preserve the extracellular vesicles (EVs) which are naturally loaded with polyphenols and which could facilitate the cell absorption of polyphenols [38]. Water extraction but presumably not EtOH-H 2 O, would preserve the integrity of such vesicles.

Conclusions
In conclusion, a fully detailed qualitative profiling of red grape skin extracted with ethanol and water mixture has permitted the identification of 65 compounds which can be classified into the following chemical classes: organic and phenolic acids (14 compounds), stilbenoids (1 compound), flavanols (21 compounds), flavonols (15 compounds) and anthocyanins (14 compounds). The extraction yield of water at different temperatures (100 • C, 70 • C, room temperature) was then evaluated and the results indicate that extraction using water under heating conditions is effective in extracting some polyphenolic classes with a good recovery in respect to EtOH based binary extraction. In particular, a good extraction yield was observed for the hydrophilic procyanidins, phenolic acids, flavonol glucoside and stilbenoid. However, according to their lipophilic character, poor absorption was observed for the most lipophilic components, such as flavonol aglycones and in general for anthocyanins. The radical scavenging activity was in accordance with the polyphenol content and hence much higher for the extract obtained with the binary solvent in respect to water extraction. All the tested extracts were found effective as antiinflammatory compounds, and an intriguing result was that the EtOH:H 2 O extract was found superimposable with that obtained using water at 100 • C. Taken together, the results indicate that water extraction carried out with heating condition is an easy, low-cost and environmentally friendly extraction method for some polyphenol classes and may have an industrial application for extracts with anti-inflammatory activity.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available.