Bioaccessibility, bioactivity and cell metabolism of dark chocolate phenolic compounds after in vitro gastro-intestinal digestion

Abstract The bioaccessibility of phenolic compounds after in vitro gastro-intestinal digestion of dark chocolate, dark chocolate enriched with Sakura green tea and dark chocolate enriched with turmeric powder was studied. The phenolic profile, assessed by accurate mass spectrometry analysis, was modified during in vitro gastro-intestinal digestion, with a considerable decrease of total and individual phenolic compounds. Phenolic acids showed the highest bioaccessibility with hydroxycinnamic acids displaying higher bioaccessibility (from 41.2% to 45.1%) respect to hydroxybenzoic acids (from 28.1% to 43.5%). Isomerisation of caffeoyl-quinic acids and galloyl-quinic acids as well as dimerization of (epi)gallocatechin were also observed after in vitro gastro-intestinal digestion. Antioxidant activity increased after the gastric step and rose further at the end of the digestion. Furthermore, in vitro digested phenolic-rich fractions showed anti-proliferative activity against two models of human colon adenocarcinoma cell lines. Cell metabolism of digested phenolic compounds resulted in the accumulation of coumaric and ferulic acids in the cell media.


Introduction 17
Cocoa and cocoa-based products, such as dark chocolate, are widely consumed in several countries 18 and significantly contribute to the daily intake of antioxidants and phenolic compounds in adults 19 and children (Rusconi, & Conti). Recently, our research group comprehensively analysed the 20 phenolic profile of dark chocolate (Martini, Conte, & Tagliazucchi, 2018). More than 140 21 individual phenolic compounds were identified by accurate mass spectrometry analysis.  ols are the most abundant phenolic compounds in dark chocolate, accounting for around the 64% of 23 total phenolics (Martini et al., 2018). 24 There are several in vivo studies suggesting that cocoa-derived polyphenols may have beneficial 25 effects on markers of cardiovascular disease risk (Del Rio et al., 2013). Short-term randomized 26 clinical trials have demonstrated that dark chocolate intake reduced blood pressure, improved flow-27 mediated dilation and ameliorated the lipid profile in healthy and hypertensive subjects (Grassi, 28 Lippi, Necozione, Desideri, & Ferri, 2005a;Grassi et al., 2005b;Lin et al., 2016). These effects 29 have been partially attributed to the high flavan-3-ols content of dark chocolate (Engler et al., 30 2004). Furthermore, dark chocolate intake has been shown to reduce the number of pre-neoplastic 31 lesions in azoxymethane-induced colonic cancer in rats (Hong, Nulton, Shelechi, Hernández, & 32 Nemposeck, 2013;Rodríguez-Ramiro et al., 2011a). The protective effect of dark chocolate against 33 colon cancer may be due to the biological activities of its phenolic compounds through the 34 regulation of several signal transduction pathways and the modulation of gene expression 35 (Carnésecchi et al., 2002;Granado-Serrano et al., 2010;Martín et al., 2010;Rodríguez-Ramiro, 36 Ramos, Bravo, Goya, & Martín, 2011b). 37 The bioavailability of phenolic compounds differs widely among the different classes. Some 38 phenolic compounds are poorly absorbed (Del Rio et al., 2013) and/or are unstable under the gastro-39 intestinal tract conditions (Bouayed, Deußer, Hoffmann, & Bohn, 2012;Juániz et al., 2017). Indeed, 40 dark chocolate phenolic compounds are entrapped in a solid food matrix and only the released 41 compounds are potentially bioavailable and able to exert their beneficial effects in the gastro-42 intestinal tract or at systemic level (Tagliazucchi, Verzelloni, Bertolini, & Conte, 2010;43 Tagliazucchi, Verzelloni, & Conte, 2012). Therefore, studies carried out with cell culture models 44 using pure phytochemicals (Carnésecchi et al., 2002) or cocoa/chocolate extracts (Rodríguez-45 Ramiro, et al., 2011b) are unrealistic unless the bioaccessibility and gastro-intestinal tract stability 46 of the phenolic compounds have been well defined. Furthermore, in vitro studies did not take into 47 account the stability of tested molecules in cell cultures and their metabolic fate within the cells 48 (Aragonès, Danesi, Del Rio, & Mena, 2017). 49 This work aimed to investigate the effect of in vitro gastro-intestinal digestion on the 50 bioaccessibility of phenolic compounds in dark chocolate and dark chocolate functionalized with 51 Sakura green tea leaves or turmeric powder. In addition, the antioxidant and anti-proliferative 52 activities of in vitro digested dark chocolates phenolic compounds against two models of human 53 colonic cell lines were assessed. Finally, the last task was to identify and quantify the main 54 metabolites derived from incubation of in vitro digested dark chocolate phenolic compounds with 55 cells. 56 were collected and analysed by LC-MS/MS to determine the stability and the metabolism of the 157 dark chocolate phenolic compounds in the cell media. Cell media were extracted according to Sala 158 et al. (2015) and investigated according to Martini, Conte, & Tagliazucchi (2017). Briefly, samples 159 were analysed using a HPLC Agilent 1200 Series system equipped with an Agilent 6300 ion trap 160 mass spectrometer. Separations were performed using a C18 column (HxSil C18 Reversed phase, 161 250×4.6 mm, 5 μm particle size, Hamilton Company, Reno, Nevada, USA), with an injection 162 volume of 40 μL and elution flow rate of 1 mL/min. The mobile phase composition, the gradient 163 and MS operating conditions are the same as reported in Martini et al. (2017). MS experiments were 164 performed in ESI negative ion mode. Identification of phenolic compounds and metabolites in all 165 samples was carried out using full scan, data-dependent MS 2 scanning from m/z 100 to 1700. 166 167

Statistic 168
All data are presented as mean ± SD for three replicates for each prepared sample. One-way 169 analysis of variance (one-way ANOVA) with Tukey's post-hoc test was applied using Graph Pad 170 prism 6.0 (GraphPad software, San Diego, CA, U.S.A.). The differences were considered 171 significant with P<0.05. 172

In vitro bioaccessibility of phenolic compounds in different types of dark chocolate 174
In our previous work we identified and quantified by high-resolution mass spectrometry 141, 155 175 and 142 phenolic compounds in dark chocolate (DC), dark chocolate enriched with Sakura green 176 tea (GTDC) and dark chocolate enriched with turmeric powder (TDC), respectively (Martini et al., 177 2018). In this work, we present data regarding the release and bioaccessibility of dark chocolate 178 total and individual phenolic compounds following in vitro gastro-intestinal digestion. Figure 1  179 shows the impact of in vitro gastro-intestinal digestion on total phenolic compounds. The chemical 180 extract of GTDC showed a significant higher amount (P<0.05) of total polyphenols ( phenolic compounds were released from the food matrices in DC, GTDC and TDC,respectively. 185 The amount of bioaccessible total phenolic compounds increased by 31%,26.5% and 20.1% in DC,186 GTDC and TDC, respectively, after two hours of gastric digestion (Figure 1). The incubation with 187 pancreatic solution further increased the bioaccessibility of total compounds in the different samples 188 but to a different extent (Figure 1) showed the lowest amount (P<0.05) of total bioaccessible phenolic compounds (7172.14 ± 512.02 195 μmol gallic acid equivalent/100 g of dark chocolate). These results are in agreement with previously 196 reported data showing that the gastro-intestinal tract behaved as an extractor promoting the release 197 of phenolic compounds from solid food matrices (Blancas-Benitez, Pérez-Jiménez, Montalvo-198 González, González-Aguilar, & Sáyago-Ayerdi, 2018;Tagliazucchi et al., 2010;Tagliazucchi et al., 199 2012). However, other studies found a decrease in bioaccessible total phenolic compounds during 200 the intestinal digestion (Bouayed et al., 2012;Lingua, Wunderlin, & Baroni, 2018). The different 201 results can be related to the higher stability of dark chocolate phenolic compounds to the intestinal 202 conditions respect to the other foods tested or to a different food matrix effect. However, it should 203 be taken into account that the Folin-Ciocalteau assay is strongly subject to interferences, especially 204 from sugars and vitamin C (Singleton et al., 1999). On the other hand, dark chocolate is rich in 205 Maillard reaction products that react in a concentration-dependent manner with the Folin-Ciocalteau 206 reagent, possibly resulting in an overestimation of bioaccessible total phenolic compounds 207 (Verzelloni, Tagliazucchi, & Conte, 2007). 208 intestinal digestion of DC, GTDC and TDC, respectively. This means that 45%, 21% and 39% of 213 individual phenolic compounds were not bioaccessible in DC, GTDC and TDC, respectively. A 214 significant lower amount of phenolic compounds was observed in all the samples after simulated 215 gastro-intestinal digestion respect to the chemical extracts. In TDC, only 17.6% of the total amount 216 of phenolic compounds was released from the food matrix or not degraded during digestion. In DC 217 and GTDC, the amount of bioaccessible total phenolic compounds at the end of the digestion was 218 23.0% and 23.2%, respectively ( Figure 2H and Table 6). 219 The apparent lowest bioaccessible value of phenolic compounds in TDC was ascribed to the poor 220 bioaccessibility (0.24%) of curcuminoids (Table 4). 221 Among the different phenolic classes, phenolic acids showed the highest bioaccessibility ( Figure  222 2B and 2G) with hydroxycinnamic acids displaying higher bioaccessibility (from 41.2% to 45.1%) 223 than hydroxybenzoic acids (from 28.1% to 43.5%). These compounds were efficiently released 224 from the food matrices and stable under gastro-intestinal conditions. When the effect of gastro-225 intestinal digestion in coffee and cardoon was studied, chlorogenic acids were proved to be quite 226 stable (Juániz et al., 2017;Monente et al., 2015). Similarly, Tagliazucchi et al. (2010) and Bouayed 227 et al. (2012) found that caffeic and coumaric acids were quite stable during in vitro gastro-intestinal 228 digestion. On the other hand, Bouayed et al. (2012) observed a bioaccessibility of 31.6%-56.5% of 229 hydroxycinnamic acids in selected apple varieties following in vitro gastro-intestinal digestion. 230 Hydroxycinnamic acid-aspartate derivatives were the most bioaccessible hydroxycinnamic acids in 231 the tested dark chocolates ( Table 2). Ferulic acid (the most abundant hydroxycinnamic acid in dark 232 chocolates) was detected in lower concentration in the intestinal environment respect to 233 hydroxycinnamic acid-aspartate derivatives ( Table 2) (Table 7). Similarly, other isomerization reactions might take place among galloylquinic 248 acid isomers as observed in GTDC subjected to in vitro digestion (Table 5). 249 Flavan-3-ols were the dominant class of phenolic compounds in the tested dark chocolates. 250 However, due to their low bioaccessibility, hydroxycinnamic acids dominated the phenolic profile 251 in in vitro digested chocolates, with the only exception of GTDC (Figure 2A and Table 1). While 252 the monomeric flavan-3-ols appeared to be in some way bioaccessible, the recovered amount of 253 procyanidins was extremely low and most of them were not found in the intestinal environment. 254 (Epi)gallocatechin isomers were only detected after in vitro digestion of GTDC probably because 255 they were present in higher concentration in GTDC respect to the other dark chocolate samples. The 256 high instability of catechins and procyanidins had been reported earlier (Bouayed et al. 2012). In a 257 previous study, procyanidin B2 was almost completely degraded into the monomeric epicatechin 258 during gastric digestion (Kahle et al., 2011). The degradation of procyanidin B2, epicatechin and 259 catechin into unknown degradation products in artificial intestinal conditions was also observed 260 (Kahle et al., 2011;Zhu et al., 2002;Bouayed et al. 2012). In another study, epigallocatechin and 261 epigallocatechin gallate were found to be sensitive to gastro-intestinal digestion with less than 10% 262 recovery after in vitro digestion of green tea (Green, Murphy, Schulz, Watkins, & Ferruzzi, 2007). 263 Some new compounds appearing in the intestinal environment may be indicative of catechin 264 monomers degradation (Table 7). For example, trihydroxybenzene may be originated from the B-265 ring of (epi)gallocatechin and epigallocatechin gallate. Indeed, after in vitro digestion of GTDC two 266 new compounds were detected and identified as (epi)gallocatechin homodimers (theasinensin 267 isomer and P2 analogue) (Neilson et al., 2007). Finally, the highest bioaccessibility of 268 (epi)gallocatechin isomers and the higher content of gallic acid observed after digestion of GTDC, 269 respect to the contents found in the chemical extract, could be explained as a consequence of 270 hydrolysis of epigallocatechin gallate. 271 272

Effect of in vitro digestion on antioxidant activities 273
In order to study how the antioxidant activity of the dark chocolate samples was modified 274 throughout the digestive process, antioxidant activity was determined by FRAP and ABTS assays at 275 each stage of the in vitro gastro-intestinal digestion and in the chemical extracts (Figure 3A and B). 276 The GTDC chemical extract was the sample with the highest activity for both the assays (

In vitro metabolism of digested dark chocolate phenolic-rich fraction in cell cultures 332
In order to verify the cell metabolism of dark chocolate phenolic compounds, Caco-2 and SW480 333 media were analysed by LC-MS ion trap after incubation (24 h) with in vitro digested dark 334 chocolate (at concentration corresponding to IC50). Different metabolic reactions, including 335 (de)hydroxylation, (de)hydrogenation, and conjugation with methyl, glucuronide, sulphate, and 336 glutathione moieties were monitored. Some parent compounds and newly formed metabolites were 337 detected in both cell types and reported in Table 8. 338 In addition to the parent compounds catechin and epicatechin, two newly formed metabolites were 339 tentatively identified. Methyl-(epi)catechin was found in the cell media of both the cell lines 340 whereas dimethyl-(epi)catechin was found only in Caco-2 medium. Previous studies identified 341 methyl-epicatechin and sulphate-epicatechin as the main metabolites in Caco-2 experiments with a 342 prevalence of methylation (Aragonès et al., 2017;Sanchez-Bridge et al., 2015). The lack of 343 identification of sulphate metabolites of (epi)catechin could be due to their low concentration in the 344 media (i.e. they could be formed but were below the limit of detection) or to the inhibition of the 345 specific enzymes as a consequence of the presence of other phenolic compounds. Sanchez-Bridge et 346 al. (2015) showed that the co-administration of epicatechin with flavonols, flavones and isoflavones 347 reduced the metabolism of epicatechin (especially sulphation) in Caco-2. Despite the appearance in 348 vivo of glucuronidated epicatechin metabolites, we did not find these substituted metabolites under 349 our experimental conditions. Previous studies suggested the absence of specific uridine 5'-350 diphospho-glucuronosyl-transferase isoforms able to form glucuronic acid conjugate of epicatechin 351 in Caco-2 cells (Actis- Goretta et al., 2013;Sanchez-Bridge et al., 2015). 352 The main hydroxycinnamic acid derivatives found after in vitro gastro-intestinal digestion of dark 353 chocolate were the conjugated forms with amino acids (such as aspartate and tyrosine, Table 2). 354 With the exception of trace amounts of feruloyl-aspartate (found in the media of both cell lines), we 355 were not able to identify these compounds after incubation with the two cell lines. Diversely, we 356 found ferulic and coumaric acids in the media of both cell lines and caffeic acid only after 357 incubation with SW480 cells. A sulphated form of coumaric acid and dihydro-ferulic acid were 358 tentatively identified as newly formed metabolites in the cell culture media after incubation with 359 SW480 and Caco-2, respectively. The concentration of coumaric acid increased from 5.10 ± 0.12 360 after in vitro digestion to 86.14 ± 3.19 and 96.80 ± 4.96 μmol/100 g of chocolate after 24 h of 361 incubation with Caco-2 and SW480, respectively ( Table 2 and Table 8 coumaroyl-aspartate and/or coumaroyl-tyrosine, catalysed by membrane-bound carboxypeptidases, 367 may be hypothesized. Indeed, coumaric acid has been found particularly stable when incubated with 368 Caco-2 or rat hepatic cells (Kahle et al., 2011;Kern et al., 2003). After 24 h of incubation with 369 Caco-2 an increased amount of ferulic acid respect to the concentration found at the end of the 370 digestion was detected ( Table 2 and Table 8). Methylation of caffeic acid by catechol-O-371 methyltransferase may account for the increase in ferulic acid concentration (Kern et al., 2003). 372 This conclusion is supported also by the evidence of the disappearance of caffeic acid from the 373 medium. Indeed, methylation of di-hydro-caffeic acid may account for the appearance of di-hydro-374 ferulic acid in the medium, as already suggested in Caco-2 cells by Poquet et al. (2008). 375 The same conclusions can not be drawn for SW480. In the medium of this cell line we found some 376 residual caffeic acid and the amount of ferulic acid did not increase during incubation ( Table 2 and  377   Table 8). Indeed, we did not identify di-hydro-ferulic acid in the medium of SW480. This evidence 378 suggested that caffeic and di-hydro-caffeic acid were not substrates for the catechol-O-379 methyltransferase in SW480, despite its presence as indicated by the appearance of methylated 380 (epi)catechin as reported above. Therefore, hydroxycinnamic acids metabolism under our 381 experimental conditions resulted in the accumulation of coumaric and ferulic acids in cell media 382 with only minor phase II metabolism. Figure 5 reported the hypothetical pathways of 383 hydroxycinnamic acids metabolism leading to the accumulation of coumaric and ferulic acids. 384 Quercetin-hexoside and quercetin-pentoside were tentatively identified after 24 h of incubation with 385 SW480 cell line, despite their low concentration in the sample after in vitro gastro-intestinal 386 digestion. This is indicative of their relative stability in cell culture medium as already suggested by 387 Xiao, & Högger (2015). Instead, quercetin-hexoside was not identified after 24 h of incubation with 388 Caco-2. The aglycone quercetin, which was not present in dark chocolate after in vitro gastro-389 intestinal digestion, appeared after incubation with Caco-2, suggesting that this cell line was able to 390 de-glycosylate quercetin-hexoside releasing the corresponding aglycone. De-glycosylation of 391 flavonoid glycosides can be catalysed by the action of membrane-bound lactase phloridzin 392 hydrolase and/or cytosolic β-glucosidase (Németh et al., 2003). Some previous studies failed to 393 detect de-glycosylation of quercetin-glucoside by using Caco-2 cells (del Mar Contreras, Borrás-394 Linares, Herranz-López, Micol, & Segura-Carretero, 2015;Walgren, Walle, & Walle, 1998). 395 However, Caco-2 cells express both lactase phloridzin hydrolase and cytosolic β-glucosidase 396 (Németh et al., 2003). This discrepancy can be due to the shorter incubation time in the previous 397 studies (1-2 h vs 24 h in our study). Quercetin was not identified in SW480 cell culture medium, 398 suggesting that this cell line was not able to hydrolyse quercetin-hexoside. 399 Finally, one methylated derivative of ellagic acid was tentatively identified only in the SW480 cell 400 medium. 401 The addition of green tea leaves or turmeric powder in dark chocolate recipe lead to a modification 418 of dark chocolate healthy properties. Functionalization with green tea leaves resulted in a higher 419 amount of flavan-3-ols and flavonols after in vitro digestion than dark chocolate, achieving a more 420 efficient antioxidant activity. Similarly, the addition of turmeric powder may lead to an increased 421 anti-proliferative activity against adenocarcinoma cell lines respect to DC and GTDC. 422 In this way, the potential healthy effect of dark chocolate consumption could be maximized, 423 reducing the amount of energy and calories introduced with chocolate itself and resulting in a lower 424 intake to achieve the same biological effects. 425  Food Research International, 112, 1-16. Minekus, M., Alminger, M., Alvito, P., Ballance, S., Bohn, T., Bourlieu, C., et al. (2014). A standardised static in vitro digestion method suitable for food -an international consensus. Food & Function, 5, 1113−1124. Monente, C., Ludwig, I. A., Stalmach, A., de Peña, M. P., Cid, C, & Crozier, A. (2015. In vitro studies on the stability in the proximal gastrointestinal tract and bioaccessibility in Caco-2 cells of chlorogenic acids from spent coffee grounds. International Journal of Food Sciences and Nutrition, 66, 657-664. Neilson, A. P., Hopf, A. S., Cooper, B. R., Pereira, M. A., Bomser, J. A., & Ferruzzi, M. G. (2007).