Ready-to-Use Nutraceutical Formulations from Edible and Waste Organs of Algerian Artichokes

Edible, plant-derived foodstuffs are recognized as precious sources of polyphenol compounds, whose consumption has proven to have multiple beneficial effects on human health. However, the awareness that cooking processes are able to induce quali-quantitatively changes in their native occurrence and that their bioavailability after food ingestion is poor led the research to move toward the preparation of nutraceutical supplements aimed at maximizing their content by effective extractive techniques and protecting them from degradation. The present work fits into this context, proposing a green, ready-to-use formulation of capitula, stems, and leaves of Algerian artichokes, in which natural deep eutectic solvents were exploited as extracting solvents but not removed at the end of the process. MTT test on the Caco-2 cell line highlighted that mitochondrial redox activity inhibition was absent below the 50 µg/mL tested dose. Simulated in vitro digestion was used as a predictive model for formulation bioaccessibility, where the joint approach with UHPLC-HRMS techniques allowed to define the release of each polyphenol from the investigated matrices. The capitula-based sample was the richest one in flavonoids, especially luteolin and apigenin glycosides, which survived in the intestinal digesta. On the contrary, simple phenols characterized the stem sample, whose release was mainly in the gastric chyme.


Introduction
The close connection between health and a correct lifestyle led the consumers to an increased awareness about several benefits provided by the daily intake of foods rich with bioactive and specialized metabolites within the diet. These compounds could be able to provide functional support to the human organism apart from nutrition, so much so that they could be claimed as nutraceutical substances. Among them, phenols and polyphenols have been recognized as main actors in exerting several biological activities (e.g., antioxidant, anti-inflammatory, anti-microbic, chemo-preventive, etc.) [1].
Polyphenols are ubiquitously biosynthesized in plants, and for this reason, they can be ingested through the diet from edible plants and derived foodstuffs. It has been widely demonstrated that agro-wastes and by-products also represent a rich but often undervalued source of these compounds [2]. Polyphenols from edible and not-edible parts of Cynara cardunculus var. scolymus (L.) Benth. (globe artichoke), previously simply known as C. scolymus L. (Asteraceae family), have been the focus of many phytochemical investigations [3][4][5][6] also related to the multiple beneficial effects for human health deriving from its consumption and reported in folk medicine, such as hepato-protective, choleretic, cardioprotective, and anti-cholinesterase [7]. Artichoke leaves have been recently included in a European Union herbal monograph highlighting the main chemical constituents claimed for their medicinal properties [8].
However, the use of artichoke for culinary purposes involves the cooking processes of edible parts, and it has been demonstrated that temperature is able to induce both qualitative and quantitative alterations in (poly)phenol composition [9], resulting in a reduced opportunity to exploit their health-related properties after ingestion.
In the last years, the development of dietary supplements containing nutraceutical polyphenols has aroused a huge interest in the scientific community. In this context, the enrichment of these bioactive ingredients has been achieved by means of efficient extractive methods from plant matrices using proper organic solvents (e.g., alcohols, hydro-alcoholic solutions, acetone). However, toxicity and environmental concerns dictated the need to move towards more sustainable extractants, among which natural deep eutectic solvents (NaDES) have gained particular attention. Their chemistry relies on the formation of hydrogen bonds between a donor and an acceptor that are natural substances, such as organic acids, sugars, quaternary ammonium salts, or betaine [10]. The NaDES extraction of plants materials, food samples, and food by-products has been the focus of an increasing number of published papers [11][12][13][14] (just to mention a few), proving to be responsible also for enhanced compound stability, bioactivity, and bioavailability [15], which make them suitable for applications in the nutraceutical sector besides pharmaceutical and cosmeceutical ones. To the best of our knowledge, the only investigation regarding deep eutectic solvents applied to C. cardunculus L. was aimed at improving the extraction yield of the sesquiterpene lactone cynaropricrin from the plant leaves [16].
The aim of the present work was to prepare ready-to-use food supplements containing bioactive compounds from three different organs (capitula, stems, and leaves) of an Algerian artichoke, taking advantage of NaDES extracting power towards polyphenols and avoiding their removal from the end product. To the best of our knowledge, this kind of approach has never been applied to globe artichoke. Capitula (or head) represent the immature inflorescence and is the common edible plant part, whereas stems and leaves are the main by-products deriving from industrial processing [17].

Plant Collection, NaDES Preparation, and Extraction
Plants of Cynara cardunculus var. scolymus (L.) Benth. were collected in April 2019 in the Béjaïa Province (Algeria) and identified by Dr. Adjir (Mentouri University, Constantine, Algeria). After harvesting, the plants were dissected into capitula, stems, and leaves; each organ was dried in the oven at 40 • C until a stable weight was achieved (after 7 days); and then, they were pulverized by a rotating knife homogenizer and stored at room temperature until use.
For the preparation of NaDES, choline hydrochloride (ChCl) and citric acid (CA) were anhydrified in the oven at 45 • C for one hour before use. Their mixture at a 1:1 molar ratio was heated under continuous stirring in a water bath at 80 • C for 30 min until a transparent solution was obtained. The solution was then diluted with distilled water (20%) and used as extractant of dried and pulverized artichoke organs (1:10, w:w). The extraction was accelerated by means of an ultrasonic bath (Ultrasonics TM Bransonic TM M3800-E, Danbury, CT, USA). Three cycles were performed, 30 min each, followed by dilution (1:1, w:w) and centrifugation at 4800 rpm for 20 min in an Avant TM J-25 centrifuge (Beckman Coulter, Brea, CA, USA), equipped with a JA-14 rotor in order to remove the plant matrix. The free water was finally evaporated under vacuum until a constant weight was reached. Jelly-like formulations were obtained and stored at 4 • C until use.
The cells were seeded in the medium in a 96-well plate at the density of 2.5 × 10 4 cells/well. The day after, the cells were treated with the NaDES-artichoke samples at different concentrations in serum-free fresh medium (5, 10, 50, 100, and 200 µg/mL, final concentration levels) for 24 h. At the end of each incubation time, the MTT assay was performed as previously described [18]. Two independent experiments were carried out, performing in each six replicate measurements for three samples (n = 3) of the extract (in total, 6 × 3 measurements). Data were expressed as mean ± standard deviation (SD).

In Vitro Digestion Protocol
The simulation of in vitro digestion processes was conducted following the static method protocol proposed by the COST action INFOGEST network [19]. According to the recommendations, three saline solutions were prepared; designed to simulate salivary (SSF), gastric (SGF), and intestinal (SIF) fluids; and formed by the same saline constituents but mixed in different proportions and stored at −20 • C until use. The protocol consists of three phases, each one involving the addition of specific enzymes, as follows: Briefly, 5 g of NaDES-artichoke extracts was mixed with 3.5 mL of simulated salivary fluids (SSF) and 0.5 mL of α-amylase (prepared in SSF; final concentration 75 U/mL) previously heated at 37 • C, followed by 25 µL of a 0.3 M calcium chloride solution and 975 µL of distilled water. The pH was adjusted to 7 with NaOH 1 M, and the mixture thus formed was incubated at 37 • C for 2 min.
In a second experiment, carried out in parallel, in order to simulate the gastric phase, 7.5 mL of simulated gastric fluids (SGF) and 1.6 mL of a pepsin solution (prepared in SGF; final concentration 2000 U/mL) were added to 10 mL of oral bolus deriving from the simulated oral phase. Then, 5 µL of 0.3 M calcium chloride solution was further added and also 6 M HCl to bring the pH to 3, and finally, distilled water was added to obtain a final volume of 10 mL. The mixture was then incubated at 37 • C for 2 h in continuous stirring. The pH value was checked every 30 min in order to obtain the desired value if necessary.
In the third experiment, the digestion went further, and in order to simulate the intestinal phase, 11 mL of simulated intestinal fluids (SIF), 5 mL of pancreatin 800 U/mL (prepared in SIF; final concentration 100 U/mL), 2.5 mL of bile salts 160 mM (final concentration 10 mM), and 40 µL of CaCl 2 0.3 M were added to the gastric chyme. Subsequently, the pH was re-neutralized by NaOH 1 M, and finally, water was added to reach a 1:1 (v:v) ratio with gastric chyme. The mixture was then incubated at 37 • C for 2 h in continuous stirring. The pH value was checked every 30 min in order to obtain the desired value when necessary.
At the end of each stage of simulated in vitro digestion, the sample was immersed in liquid nitrogen to block any enzymatic activity and centrifuged at 4500 rpm for 10 min in an Avant TM J-25 centrifuge (Beckman Coulter, Brea, CA, USA) equipped with a JA-14 rotor; the obtained supernatants underwent instrumental chemical analysis, as described below.

UHPLC-HRMS Analyses of Parental Extracts and Digestate Therefrom
The chromatographic separation of samples (2 µL injection volume), carried out on a NEXERA UHPLC system (Shimadzu, Tokyo, Japan), was achieved on a Luna ® Omega C18 column (150 × 2.1 mm, 1.6 µm; Phenomenex), using 0.1% aqueous formic acid (solvent A) and acetonitrile (solvent B) as mobile phase. The linear gradient started at 2% B, was kept constant for 2 min, and then was led to 15% in 5 min, maintained for 5 min, and to 45% in the following 2 min. Then, after 0.5 min, the initial conditions were restored for re-equilibration. The flow rate was set at 0.5 mL/min. High-resolution mass spectrometry (HRMS) analyses were carried out by using the AB SCIEX TripleTOF ® 4600 spectrometer (AB Sciex, Concord, ON, Canada) equipped with a DuoSpray™ ion source operating in negative electrospray ion mode. The APCI probe was used for automated mass calibration in all scan functions using the Calibrant Delivery System (CDS). A non-targeted approach was developed combining TOF-MS and MS/MS with Information Dependent Acquisition (IDA), consisting of a full scan TOF survey (accumulation time 250 ms, 100-1500 Da) and eight IDA MS/MS scans (accumulation time 100 ms, 80-1300 Da). Other source and analyzer parameters were the following: curtain gas (CUR) 35 psi, nebulizer and heated gases (GS 1 and GS 2) 60 psi, ion spray voltage (ISVF) −4500 V, interface heater temperature (TEM) 500 • C, declustering potential (DP) 70 V, collision energy (CE) 35 V, and collision energy spread (CES) 25 V. For triterpenes, stronger DP and CE values were applied, which were 120 V and 100 (CES 25) V, respectively. The instrument was controlled by Analyst ® TF 1.7 software, while data processing was carried out using PeakView ® software version 2.2.
ABTS radical cation was generated as previously reported [20]. The ABTS •+ solution was diluted with phosphate-buffered saline (PBS; pH 7.4) until an absorbance of 0.7 at 734 nm was reached. The extracts at different doses were directly dissolved in the ABTS •+ solution, and after 6 min, the absorbance was measured by a Victor3 spectrophotometer (Perkin Elmer/Wallac, Waltham, MA, USA) in reference to a blank, in which the samples were replaced with only NaDES. DPPH • scavenging capability was estimated as previously reported [20], and the absorption at 517 nm was measured on the Victor3 spectrophotometer in reference to a blank, in which the samples were replaced with only NaDES.
Trolox ® (2, 4, 8, 16, 32 µM) was used as the positive standard, and Trolox equivalent antioxidant capacity (TEAC) of samples was calculated based on both ABTS and DPPH tests. For each antiradical test, three replicate measurements for three samples (n = 3) of the extract (in total, 3 × 3 measurements) were performed. All data were expressed as mean ± standard deviation (SD).

Results and Discussion
NaDES were formed by CA and ChCl (1:1), as hydrogen bonds donor and acceptor, respectively. They are considered food-grade ingredients, and therefore, their removal is not necessary from extracts developed as food products for human consumption [21]. The preparation of potential ready-to-use formulations was achieved in four steps: (1) mixing of NaDES with the plant organs (capitula, stems and leaves); (2) sonication; (3) centrifugation; and (4) free water evaporation under vacuum. The final products obtained appeared as jelly-like formulations. The idea to consider them as ready-to-use food supplements with functional properties for human health cannot ignore their safety as well as their chemical composition (responsible for putative health-relative effects) and their fate during digestion, which could compromise their bioaccessibility.

MTT Test on Caco-2 Cell Line
In light of the above, the NaDES-based samples were at first tested on the human colorectal adenocarcinoma (Caco-2) cell line to obtain preliminary information about their capability to affect the cell mitochondrial redox activity (RAI), which represents the first sign of cell suffering due to cytotoxicity. Indeed, this cell line has been widely employed as a model of the intestinal epithelial barrier, especially in studies dealing with the absorption of dietary components [22].
Caco-2 cells were exposed to increasing dose levels (5-200 µg/mL) of NaDES-based samples prepared from the three different artichoke organs (capitula, stems, and leaves). After 24 h MTT test was performed, and the results are reported in Figure 1. The cell mitochondrial redox activity appeared mildly compromised in a dose-dependent manner for all the extracts, and below the 50 µg/mL tested concentration, its inhibition could even be considered absent. Among the organs, the stem-based sample appeared more effective in the inhibition, whereas the leaves and capitula showed almost superimposable dose-response curves, with the latter being responsible for a maximum RAI, equal to 40.2 ± 0.06%, only at the highest tested dose. The comparison with data obtained for NaDES alone, which were used as control, highlighted their poor contribution to the exerted effects. The preliminary results paved the way for considering the artichoke-based supplements as putative ready-to-use formulations once their chemical composition, which is responsible for the observed activity, are unraveled.

MTT Test on Caco-2 Cell Line
In light of the above, the NaDES-based samples were at first tested on the human colorectal adenocarcinoma (Caco-2) cell line to obtain preliminary information about their capability to affect the cell mitochondrial redox activity (RAI), which represents the first sign of cell suffering due to cytotoxicity. Indeed, this cell line has been widely employed as a model of the intestinal epithelial barrier, especially in studies dealing with the absorption of dietary components [22].
Caco-2 cells were exposed to increasing dose levels (5-200 μg/mL) of NaDES-based samples prepared from the three different artichoke organs (capitula, stems, and leaves). After 24 h MTT test was performed, and the results are reported in Figure 1. The cell mitochondrial redox activity appeared mildly compromised in a dose-dependent manner for all the extracts, and below the 50 μg/mL tested concentration, its inhibition could even be considered absent. Among the organs, the stem-based sample appeared more effective in the inhibition, whereas the leaves and capitula showed almost superimposable doseresponse curves, with the latter being responsible for a maximum RAI, equal to 40.2 ± 0.06 %, only at the highest tested dose. The comparison with data obtained for NaDES alone, which were used as control, highlighted their poor contribution to the exerted effects. The preliminary results paved the way for considering the artichoke-based supplements as putative ready-to-use formulations once their chemical composition, which is responsible for the observed activity, are unraveled.

UHPLC-HRMS Profiling of Artichoke-Based Food Supplements
All the samples underwent UHPLC-HRMS and MS/MS analyses to disclose their bioactive constituents. Apart from signals ascribed to citrate in the first 2 min (herein not discussed), 38 specialized metabolites were tentatively identified as 15 simple phenols, 21 polyphenols, and 2 triterpenes, all differently distributed. In Table 1, chromatography and mass spectrometry data that are useful for their tentative identification are summarized, whereas details are reported below for the compounds grouped in sub-classes.

UHPLC-HRMS Profiling of Artichoke-Based Food Supplements
All the samples underwent UHPLC-HRMS and MS/MS analyses to disclose their bioactive constituents. Apart from signals ascribed to citrate in the first 2 min (herein not discussed), 38 specialized metabolites were tentatively identified as 15 simple phenols, 21 polyphenols, and 2 triterpenes, all differently distributed. In Table 1, chromatography and mass spectrometry data that are useful for their tentative identification are summarized, whereas details are reported below for the compounds grouped in sub-classes.
Compounds 6 and 10 showed the molecular formula C 17 H 20 O 9 with an RDB value of 8. Although from the TOF-MS spectra, they could resemble feruloylquinic acid isomers, their TOF-MS/MS allowed us to discard this hypothesis. In fact, as depicted in Figure 2, fragment ions deriving from the collision-induced dissociation of the precursor ion were in accordance with the presence of a caffeoyl moiety. Thus, taking into consideration the neutral loss of 188 Da ascribable to dehydrated methylquinic acid, these metabolites were tentatively identified as caffeoyl methylquinate isomers.
Analogously, compounds 32 and 35 were putatively identified as dicaffeoyl methylquinate isomers. It is worthy of note that when the carboxylic group of quinic acid was not free, the corresponding deprotonated fragment ion (methylquinate, at m/z 205.0718, calc. mass) was never generated. This observation could be pivotal for the straightforward discrimination of this molecular skeleton.
Compounds 6 and 10 showed the molecular formula C17H20O9 with an RDB value of 8. Although from the TOF-MS spectra, they could resemble feruloylquinic acid isomers, their TOF-MS/MS allowed us to discard this hypothesis. In fact, as depicted in Figure 2, fragment ions deriving from the collision-induced dissociation of the precursor ion were in accordance with the presence of a caffeoyl moiety. Thus, taking into consideration the neutral loss of 188 Da ascribable to dehydrated methylquinic acid, these metabolites were tentatively identified as caffeoyl methylquinate isomers. Analogously, compounds 32 and 35 were putatively identified as dicaffeoyl methylquinate isomers. It is worthy of note that when the carboxylic group of quinic acid was not free, the corresponding deprotonated fragment ion (methylquinate, at m/z 205.0718, calc. mass) was never generated. This observation could be pivotal for the straightforward discrimination of this molecular skeleton.

Polyphenols
All polyphenol compounds were flavonoids, whose putative identity is discussed below, grouping them into three subclasses referring to the aglycone core. The detected 14 flavones were luteolin and apigenin derivatives, differently glycosylated. Compound 7 was identified as a hexosyl-hexuronyl derivative of apigenin. In fact, its TOF-MS/MS spectrum showed sequential neutral losses of 162.05 and 176.03 Da, leading to the disclosure of the deprotonated aglycone ion (at m/z 269.04) and the corresponding radical (at m/z 268.0378) ( Table 1), whose intensity ratio allowed us to hypothesize a 7-Ohexuronidation and a 4′-O-hexosylation ( Figure S2). On the contrary, TOF-MS/MS of metabolite 23 showed a reversed ratio between the two aglycone ions, suggesting an apigenin 7-O-hexosyl derivative (C21H20O10). Moreover, apigenin 7-O-hexuronide (24) and 7-O-rutinoside (25) were putatively identified, the latter being characterized by a neutral

Polyphenols
All polyphenol compounds were flavonoids, whose putative identity is discussed below, grouping them into three subclasses referring to the aglycone core. The detected 14 flavones were luteolin and apigenin derivatives, differently glycosylated. Compound 7 was identified as a hexosyl-hexuronyl derivative of apigenin. In fact, its TOF-MS/MS spectrum showed sequential neutral losses of 162.05 and 176.03 Da, leading to the disclosure of the deprotonated aglycone ion (at m/z 269.04) and the corresponding radical (at m/z 268.0378) ( Table 1), whose intensity ratio allowed us to hypothesize a 7-O-hexuronidation and a 4 -O-hexosylation ( Figure S2). On the contrary, TOF-MS/MS of metabolite 23 showed a reversed ratio between the two aglycone ions, suggesting an apigenin 7-O-hexosyl derivative (C 21 H 20 O 10 ). Moreover, apigenin 7-O-hexuronide (24) and 7-O-rutinoside (25) were putatively identified, the latter being characterized by a neutral loss of 308 (162 + 146) Da ( Figure S2) [26]. Finally, compound 36 could be in accordance with apigenin methylhexuronide. Indeed, the absence of a fragment ion at m/z 299, allowing us to exclude the possibility of a methylluteolin as aglycone, seemed to confirm the structural hypothesis. Instead, this latter fragment was observed in the TOF-MS/MS spectrum of compound 31, putatively identified as methylluteolin (e.g., chrysoeriol; [27]) hexuronide, and of its N-derivative 15, whose complete identity still remains unknown.
Flavones also occurred as malonylhexosides. They could be distinguished from the detection of a neutral loss of 248.05 Da (malonylhexose-H 2 O) from the deprotonated molecular ion [29] besides decarboxylation ( Figure 3).  [28].
Flavones also occurred as malonylhexosides. They could be distinguished from the detection of a neutral loss of 248.05 Da (malonylhexose-H2O) from the deprotonated molecular ion [29] besides decarboxylation ( Figure 3). Flavanone and flavonol glycosides were also detected. In particular, based on the occurrence of characteristic fragment ions at m/z 151, 135, and 107 ( Figure S3), the aglycone of compounds 9, 11, and 12 was recognized as eriodictyol [30]. The saccharidic moiety, derived from neutral losses from deprotonated molecular ions, corresponded to hexose, hexuronic acid, and rutinose, respectively. Finally, TOF-MS/MS of compound 14 was in accordance with a quercetin 4′-O-hexoside due to the intensity ratio of deprotonated and radical aglycone ions (at m/z 301.035 and 300.027; [31]), whereas compound 13 was identified as the flavonol 3-O-hexuronide derivative [32]. Flavanone and flavonol glycosides were also detected. In particular, based on the occurrence of characteristic fragment ions at m/z 151, 135, and 107 ( Figure S3), the aglycone of compounds 9, 11, and 12 was recognized as eriodictyol [30]. The saccharidic moiety, derived from neutral losses from deprotonated molecular ions, corresponded to hexose, hexuronic acid, and rutinose, respectively. Finally, TOF-MS/MS of compound 14 was in accordance with a quercetin 4 -O-hexoside due to the intensity ratio of deprotonated and radical aglycone ions (at m/z 301.035 and 300.027; [31]), whereas compound 13 was identified as the flavonol 3-O-hexuronide derivative [32].

Triterpenes
The presence of two triterpene saponins was highlighted only in the artichoke capitula sample (37 and 38, Table 1). Both molecules appeared deprotonated in the TOF-MS spectrum, suggesting the presence of an oxidized sugar on the molecular skeleton, in accordance with previously published guidelines by HR-MS/MS tools [33]. To obtain additional information useful for their identification, stronger TOF-MS/MS declustering and collision energy potentials were applied. Both compounds shared the same saccharidic moieties, represented by a hexose bound to the C-28 carboxylic group and a (pentosyl)hexuronidyl residue linked at the C-3 position. The aglycone ion of saponin 37 showed a further hydroxyl group, likely at the C-21 position, enhancing its polarity and thus decreasing its retention time. In Figure 4, the HR-MS/MS spectra are reported, with tentative structures based on the variability of ring E, whereas in Figure S4, their hypothesized fragmentation pathways were depicted taking into account cynarasaponin H and J.

Occurrence and Relative Content of Detected Bioactive Metabolites
Apart from cynarasaponins, which peculiarly characterized the capitula sample (Table 1, Figure 5a), simple phenols and flavonoids were differently distributed among the obtained products both qualitatively and quantitatively. As a general observation, the relative amount of bioactive (poly)phenols decreased in the order C > S > L (Figure 5b).

Occurrence and Relative Content of Detected Bioactive Metabolites
Apart from cynarasaponins, which peculiarly characterized the capitula sample (Table 1, Figure 5a), simple phenols and flavonoids were differently distributed among the obtained products both qualitatively and quantitatively. As a general observation, the relative amount of bioactive (poly)phenols decreased in the order C > S > L (Figure 5b).
The capitula-based formulation was the richest one in flavonoids, with the only exception of apigenin methylhexuronide (36), whose content was about 2.5-fold higher in the leaf sample. Flavone glycosides were the most representative, followed by flavanone and flavonol derivatives. The latter ones (13,14) were extracted only from artichoke heads, together with eriodictyol hexuronide (11) and rutinoside (12).

Occurrence and Relative Content of Detected Bioactive Metabolites
Apart from cynarasaponins, which peculiarly characterized the capitula sample (Table 1, Figure 5a), simple phenols and flavonoids were differently distributed among the obtained products both qualitatively and quantitatively. As a general observation, the relative amount of bioactive (poly)phenols decreased in the order C > S > L (Figure 5b). The capitula-based formulation was the richest one in flavonoids, with the only exception of apigenin methylhexuronide (36), whose content was about 2.5-fold higher in the leaf sample. Flavone glycosides were the most representative, followed by flavanone and flavonol derivatives. The latter ones (13,14) were extracted only from artichoke heads, together with eriodictyol hexuronide (11) and rutinoside (12). On the contrary, most of the phenols accounted for stems, in particular, caffeoyl-and dicaffeoyl-methylquinic acids and dicaffeoyl-quinic acids.

In Vitro Digestion of Artichoke-Based Food Supplements
The bioaccessibility of phenols and polyphenols within the artichoke-based food supplements was evaluated by means of a joint approach between in vitro digestion and UHPLC-HRMS techniques. For this purpose, the protocol proposed by the COST action INFOGEST network [19] was applied, the relative quantitation of each compound was performed at the end of each phase (oral, gastric, and intestinal), and their bioaccessibility was calculated in terms of release index from the matrix based on peak areas (Table S1). The amount of phenols and polyphenols released from the food matrix during the simulated digestion process has been commonly measured determined by means of colorimetric assays (e.g., Folin-Ciocalteu method for total phenol content-TPC) and also indirectly measured in terms of antioxidant activity [34]. However, it should be noted that these tests are not properly specific for phenolic compounds, as they are based on the chemical reactivity of different compounds also other than phenols [35]. Just to give an idea, suffice it to say that the Folin-Ciocalteu assay was first developed to detect the aromatic amino acid tryptophan, which, although it is not a non-phenolic compound, is able to reduce the reagent and increase the measured absorbance value. Thus, the mentioned method could be useful to compare samples in a preliminary screening, but in the case of in vitro assessment of food bioaccessibility, where some enzymes are added to the simulated fluids, it could give rise to overestimated results. On the contrary, chromatographic analyses coupled with MS detection are able to show a real picture of the digesta chemical composition without interferences.
In Figure 6a, the release index from the matrices under study is reported, grouping the compounds into phenol and flavonoid classes. Moreover, the results were also analyzed by principal component analysis (PCA), further dividing the two main classes into sub-classes based on the number of acyl residues linked to quinic acid (mAQ, monoacyl quinic acid; dAQ, diacylquinic acid) or on the aglycone skeleton (LUT, luteolin; API, apigenin; ERI, eriodictyol; QUE, quercetin; MeLUT, methylluteolin) (Figure 6b).
In Figure 6a, the release index from the matrices under study is reported, grouping the compounds into phenol and flavonoid classes. Moreover, the results were also analyzed by principal component analysis (PCA), further dividing the two main classes into sub-classes based on the number of acyl residues linked to quinic acid (mAQ, monoacyl quinic acid; dAQ, diacylquinic acid) or on the aglycone skeleton (LUT, luteolin; API, apigenin; ERI, eriodictyol; QUE, quercetin; MeLUT, methylluteolin) (Figure 6b).  It appears clear that phenols from capitula, stem, and leaf NaDES-based food formulations are mainly released in the gastric chyme, whereas, among flavonoids, luteolin and apigenin derivatives from capitula were positively correlated to the final digesta (Ci). The data obtained are in accordance with previous results that emphasized the increase of artichoke monocaffeoylquinic acids in the gastric environment, whereas the transition to the intestinal conditions causes a significant increase in the total polyphenol recovery [36]. Indeed, the changes in the chemical composition of ingested foods are closely connected with the processes that take place at the digestive level, which modify the structures of the bioactive compounds and at the same time maintain or modify their biological activity [37]. In fact, bioactivity as well as cell absorption are closely related to their chemical structure. The bioaccessibility of (poly)phenols is influenced by several factors, such as chemical structure, the food matrix containing them, interaction with other components, or the presence of suppressors or cofactors [38]. Therefore, it is reasonable to assume that the peculiar composition of each organ as a whole is able to influence the bioaccessibility of individual compounds. As an example, regarding capitula samples, among apigenin derivatives, demethylation reactions of the methylhexuronidyl moiety of compound 36 could explain the enhanced relative content of compound 24 (hexuronide derivative) at the end of the digestion. Moreover, the massive release in the simulated gastric environment of apigenin hexosyl-hexuronide (7) and its successive degradation could be in line with the observed enhanced availability of compound 23 and, again, of metabolite 24, which maintain only one of the glycosidic moieties (hexose or hexuronic acid, respectively). Quercetin glycosides were the only flavonols detected in the capitula-based formulation, whose integrity was massively affected in the gastric phase, so much so that they were almost absent in the final digestion step (Table S1).

DPPH and ABTS + Radical Scavenging Capacity
The radical scavenging capacity of the undigested ready-to-use food supplements was evaluated by DPPH and ABTS + radical tests and compared to the digested samples in the three stages. NaDES alone were tested at the same final doses to exclude their influence in the tests. The dose-dependent capability was verified in all cases, with capitula formulations before simulated digestion exerting the highest radical scavenging capacity of both radical probes, followed by leaf-and stem-based samples, respectively (Table S2). Following the three phases of simulated digestion, different responses were recorded, which likely remarked the peculiar quali-quantitative composition in terms of compounds' release percentage. The PCA biplot reported in Figure 7a, able to describe the 77.94% of data variability, underlines the correlation between the investigated samples and the observed RSCs also in relation to the occurrence of different (poly)phenol sub-classes. Capitula gastric and intestinal digesta preserved their efficacy, which likely positively correlated to total flavonoid content, estimated in particular as the sum of apigenin and luteolin glycosides detected after digestion (Figure 7) beyond 1,3-dicaffeoylquinic acid (cynarin). On the contrary, stem-based samples, which mostly accounted for simple phenols, were located on the left-hand side of the graph, suggesting an inverse correlation to both radical tests. In particular, the lowest efficacy against the two probes, shown by So and Si samples, could be explained considering the overall very poor metabolite release compared to Sg.  The antiradical behavior of (poly)phenol-enriched samples was in accordance with Jiménez-Moreno et al. [39], who underlined a high and significant correlation between the phenolic composition of artichoke waste extracts evaluated by HPLC and the antioxidant capacity exerted against radical probes. In line with previous research, the DPPH and ABTS methods were herein employed to further corroborate chemical composition data before and after simulated digestion, as they are able to provide accurate results strongly related to the sample chemical features. In fact, these tests are based on a mixed mode of action of a sample in transferring a single electron (SET) and/or a hydrogen atom (HAT) to the radical probe, resulting in its neutralization [22,40]. Thus, only compounds that possess active groups (e.g., catechol moieties in simple phenols and polyphenols) are able to react.

Conclusions
The work reported herein showed that both edible and waste organs from artichokes could be exploited in the development of nutraceutical formulations. The use of NaDES as green extractive solvents promoted their enrichment in bioactive phenols and polyphenols, peculiarly occurring based on the plant organ. At the same time, the possibility of avoiding NaDES removal from the end products without affecting cell health status made them ready-to-use formulations. The joint approach of UHPLC-HRMS  Table S3). C, not digested capitula; Co, capitula oral phase; Cg, capitula gastric phase; Ci, capitula intestinal phase; S, not digested stems; So, stems oral phase; Sg, stems gastric phase; Si, stems intestinal phase; L, not digested leaves; Lo, leaves oral phase; Lg, leaves gastric phase; Li, leaves intestinal phase.
The antiradical behavior of (poly)phenol-enriched samples was in accordance with Jiménez-Moreno et al. [39], who underlined a high and significant correlation between the phenolic composition of artichoke waste extracts evaluated by HPLC and the antioxidant capacity exerted against radical probes. In line with previous research, the DPPH and ABTS methods were herein employed to further corroborate chemical composition data before and after simulated digestion, as they are able to provide accurate results strongly related to the sample chemical features. In fact, these tests are based on a mixed mode of action of a sample in transferring a single electron (SET) and/or a hydrogen atom (HAT) to the radical probe, resulting in its neutralization [22,40]. Thus, only compounds that possess active groups (e.g., catechol moieties in simple phenols and polyphenols) are able to react.

Conclusions
The work reported herein showed that both edible and waste organs from artichokes could be exploited in the development of nutraceutical formulations. The use of NaDES as green extractive solvents promoted their enrichment in bioactive phenols and polyphenols, peculiarly occurring based on the plant organ. At the same time, the possibility of avoiding NaDES removal from the end products without affecting cell health status made them ready-to-use formulations. The joint approach of UHPLC-HRMS techniques to in vitro simulated digestion gave a picture of individual compounds' bioaccessibility, whose correlation to antiradical assays highlighted that the capitula-based sample retained their efficacy also after gastric and intestinal phases, likely due to the release of apigenin and luteolin glycosides beyond 1,3-dicaffeoylquinic acid. Future perspectives will involve cell metabolomic approaches aimed at deeply investigating cellular uptake and eventual biotransformations as mechanistic effects on polyphenols bioavailability. Moreover, cell-based antioxidant and anti-inflammatory activities will undergo more research to yield further insights into their functionality.  Table S1: Release index (%) of compounds in DES-artichoke extracts, based on UHPLC-HRMS analyses. Co, capitula oral phase; Cg, capitula gastric phase; Ci, capitula intestinal phase; So, stems oral phase; Sg, stems gastric phase; Si, stems intestinal phase; Lo, leaves oral phase; Lg, leaves gastric phase; Li, leaves intestinal phase; n.a., not applicable; n.q., not quantifiable; Table  S2: Radical scavenging capacity (RCS %) of ready-to-use artichoke formulations before and after simulated digestion protocol, evaluated by DPPH and ABTS tests. Values are reported as mean ± SD. C, not digested capitula; Co, capitula oral phase; Cg, capitula gastric phase; Ci, capitula intestinal phase; S not digested stems; So, stems oral phase; Sg, stems gastric phase; Si, stems intestinal phase; L, not digested leaves; Lo, leaves oral phase; Lg, leaves gastric phase; Li, leaves intestinal phase. Table S3. TEAC (Trolox®Equivalent Antioxidant Capacity, µM) values of ready-to-use artichoke formulations before and after simulated digestion protocol, evaluated by DPPH and ABTS tests. Values are reported as mean ± SD. C = not digested capitula; Co = capitula oral phase; Cg = capitula gastric phase; Ci = capitula intestinal phase; S = not digested stems; So = stems oral phase; Sg = stems gastric phase; Si = stems intestinal phase; L = not digested leaves; Lo = leaves oral phase; Lg = leaves gastric phase; Li = leaves intestinal phase.