Effect of Wild Blueberry Metabolites on Biomarkers of Gastrointestinal and Immune Health In Vitro

: Wild blueberries ( Vaccinium angustifolium Aiton.) are a rich source of dietary ﬁber and (poly)phenols with gastrointestinal and immune health-promoting properties, however, their mechanisms of action on the intestinal epithelial cells and transient tissue macrophages remain to be elucidated. In this study, we evaluated the individual effects of anthocyanins, short-chain fatty acids (metabolites derived from ﬁber), and a series of hydroxycinnamic and hydroxybenzoic acid metabolites common to anthocyanins and other polyphenols on epithelial gut homeostasis in human colon epithelial CCD-18 cells and murine RAW 264.7 macrophages. Gastrointestinal cell migration was enhanced in response to anthocyanin glucosides with the maximum effect observed for malvidin-3-glucoside, and a structural subset of hydroxybenzoic acids, especially 2-hydroxybenzoic acid. Enhanced staining for ZO-1 protein in the junctional complexes was observed in CCD-18 cells treated with malvidin and butyrate, as well as several phenolic metabolites, including hydroxybenzoic and hydroxycinnamic acids. Nitric oxide production and pro-inﬂammatory gene expression proﬁles in the LPS-stimulated macrophages were mostly affected by treatments with 3-caffeoylquinic (chlorogenic) and 3,4-dihydroxycinnamic (caffeic) acids, as well as 2-hydroxybenzoic acid. This study lays the foundation for future investigations evaluating the effects of dietary interventions on managing gastrointestinal and inﬂammatory outcomes. short-chain ﬁber propionate, butyrate) naturally in wild blueberry digests, were screened for their ability to promote CCD-18 gut epithelial cell migration. Cell migration was enhanced relative to the vehicle control in response to all three anthocyanin glucosides tested, with the maximum effect observed for malvidin-3-glucoside (169.5 ± 48.4%, p < 0.05) and no effects for either chlorogenic short-chain fatty acid metabolites


Introduction
Wild lowbush blueberries (Vaccinium angustifolium Aiton.) have a complex and dense phytonutrient profile enriched with dietary fiber (3-4 g/100 g fresh weight) [1] and bioactive polyphenols, including anthocyanins, proanthocyanidins, and hydroxycinnamic acids (300-600 mg/100 g FW) [2]. Their anthocyanidin profile is dominated by delphinidin and malvidin glycosides [3] and is in contrast to many other anthocyanin-rich fruits and berries that predominantly accumulate cyanidins [4]. The blueberry anthocyanin glycosides can be further acylated with organic acids [5], and this modification generally improves the stability of anthocyanins at wider pH ranges [6] and digestive conditions [7] but hinders their bioavailability [8]. Following absorption in the gut, low amounts of precursor polyphenolic compounds can be found in the systemic circulation in the form of intact or phase II-derived (glucuronide, methyl, or sulfate) conjugates, while significantly larger amounts of the polyphenolic compounds undergo colonic fermentation by the gut microbiome and enter the human body as low molecular weight phenolic microbial metabolites or catabolites [9]. Disruption of the intact microbiome with antibiotic treatment in Immuno 2022, 2 294 rodents results in reduced absorption of microbial metabolites and predominant excretion of polyphenols in their unmodified form [10]. While different levels of phenolic intake affect the magnitude of biological effects [11], not all plasma phenolic metabolites show a linear response in an intake-dependent manner [12,13]. The lack of a linear relationship can be explained in part by the different affinities of individual phenolic acids towards microbial species, transporters, carrier proteins, and metabolic enzymes that are responsible for their phase-II metabolism.
In recent years, consumption of blueberries was increasingly associated with positive metabolic and immune health outcomes, in part by observing improvements in insulin sensitivity [14], vascular and endothelial function [15], inflammation [16], lipid status [17], and brain perfusion [18]. Blueberries, like any other phytochemical or nutrient-rich food, can influence the human state of health by modifying the integrity of the gastrointestinal system, conditioning the composition of gut microbiota, or directly shifting host metabolism and energy balance in the body.
The gastrointestinal wall, in the form of a single or double mucus layer, a single layer of the gut epithelial cells, and the vascular endothelium serves as the primary impervious barrier to all dietary components that are not hydrophobic or do not have a specific system of active transport [19]. This is achieved by the formation of tight junctions at the apical membrane, and adherence junctions at the basal membrane of the gut epithelial cells to control the trafficking of macromolecules, food tolerance, and immune responses in the gut [20]. Loss of barrier function secondary to release of zonulin, a prehaptoglobin-2 protein that disassembles tight junctions through EGF activation [21], generally leads to altered intestinal permeability and cytokine-mediated dysfunction that affects the intestinal mucosa and other tissues in the context of sustained chronic inflammation [22]. Zonulin expression is regulated in part by IL-6 responsive elements found in its promoter region [23] and correlates with the circulating IL-6 in subjects with obesity-associated insulin resistance [24]. At the same time, care should be taken to extrapolate zonulin findings directly [25]. While the beneficial effects of blueberry polyphenols and fiber on microbiota have been shown in clinical studies [26][27][28], little is known about the physiological effects and molecular targets of precursor polyphenol molecules and their metabolites on the activity of gastrointestinal epithelial cells and tissue-resident macrophages that directly sample luminal content [29].
In this study, we examined the ability of major metabolites of wild blueberry anthocyanins (hydroxycinnamic and hydroxybenzoic acids), hydroxycinnamic acid esters (chlorogenic acid), triterpenoid ursolic acid, and fiber (acetate, propionate, and butyrate short-chain fatty acids) to modulate biomarkers of the intestinal barrier and inflammation in vitro, in order to access their potential to support gastrointestinal and immune health outcomes.

Cell Culture
The normal colonic epithelial CCD-18 Co cells (ATCC CRL-1459, obtained from American Type Culture Collection; Livingstone, MT, USA) and the mouse macrophage cell line RAW 264.7 (ATCC TIB-71) were maintained in Dulbecco's modified Eagle's medium (DMEM, Life Technologies, Carlsbad, CA, USA), supplemented with 10% fetal bovine serum (Life Technologies), 100 IU/mL of penicillin, and 100 µg/mL of streptomycin (Fisher Scientific, Pittsburg, PA, USA) at a density not exceeding 5 × 10 5 cells/mL. All cell lines were grown in 57-cm 2 Nunc cell culture dishes (Nalge Nunc International, Rochester, NY, USA) maintained at 37 • C in a humidified 5% CO 2 Thermo Forma Series II incubator (Fisher Scientific), and routinely passaged every 3-4 days. Cell viability and dose range determination studies did not show any detrimental effects at all levels up to the maximum concentration tested as described earlier [30][31][32], with the exception of ursolic acid which was cytotoxic at 15 µM and therefore excluded from the subsequent experiments (data not shown).

Cell Migration and Formation of Junctional Complexes
CCD-18 cells were seeded in the 96-well Oris Cell Migration plates (Platypus Technologies, Madison, WI, USA) with a stopper positioned in the center of a well to create a cell-free detection zone after 24 h of incubation. Following the removal of the stoppers, cells were allowed to migrate (with and without treatments as indicated, using 0.1% DMSO as a vehicle control) into the clear field for 48 h, stained with NucBlue Hoechst 33342 Live ReadyProbes Reagent (Fisher Scientific, Waltham, MA, USA), and central fluorescence was measured at a 360 nm excitation/460 nm emission using a BioTek SynergyH1 microplate reader (Agilent, Santa Clara, CA, USA), both pre-and post-migration. Elevated levels of growth factors found in 20% FBS were used as a positive control.
In a separate set of experiments, CCD-18 cells were allowed to form monolayers for 5 days (with and without treatments as indicated, using 0.1% DMSO as a vehicle control), fixed with cold 4% formaldehyde for 15 min, permeabilized with 0.3% Triton X-100 for 5 min, and blocked with 5% BSA for 30 min at 37 • C. Zonulin accumulation in junction complexes was visualized using ZO-1/TJP1 antibody Alexa Fluor 488 (Fisher) (1:200 in 0.1% BSA for 3 h) and quantified using mean fluorescent intensity at a 490 nm excitation/525 nm emission.

Nitric Oxide Production and Gene Expression in Macrophages
RAW 264.7 cells were seeded in 96-well plates in triplicate at the concentration of 5 × 10 4 cells/well in a 200 µL culture medium and allowed to adhere for 24 h. The cells were then pre-treated with the indicated lower dose ranges of wild blueberry metabolites established in the gastrointestinal cell assays and elicited with 1 µg/mL of LPS for an additional 6 h. Nitric oxide released from the stimulated macrophages was indirectly quantified by measuring nitrite accumulation in the medium using the Greiss reagent system (Promega, Madison, WI) and a SynergyH1 microplate reader (BioTek) at 530 nm. For gene expression studies, the cells were seeded in 24-well plates at the concentration of 5 × 10 5 cells/well in a 1 mL culture medium and treated as indicated.

RNA Extraction, Purification, and cDNA Synthesis
The total RNA was isolated from RAW macrophages using TRIzol reagent (Life Technologies) following the manufacturer's instructions. RNA was quantified using the SynergyH1/Take 3 plate (BioTek). The cDNAs were synthesized using 2 µg of RNA for each sample using a high-capacity cDNA Reverse Transcription kit following the manufacturer's protocol on an ABI GeneAMP 9700 (Life Technologies).

Quantitative PCR Analysis
The resulting cDNA was amplified by real-time quantitative PCR using SYBR green PCR master mix (Life Technologies). To avoid interference due to genomic DNA contamination, only intron-overlapping primers were selected using Primer Express software,

Statistical Analysis
Statistical analyses were performed using Prism 8.0 (GraphPad Software, San Diego, CA, USA). Data were analyzed by one-way ANOVA with treatment as a factor. Post hoc analyses of differences between individual experimental groups were made using Dunnett's multiple comparison tests. The significance was set at p < 0.05. Values are reported as means ± SEM.

Physiological Dose Ranges of Gastrointestinal Metabolites
Certain dietary factors, such as glutamine [33], histidine [34], phenolic acids [35], and short-chain fatty acids derived from carbohydrate metabolism by microbiota [36] have the potential to directly affect intestinal mucosa and immunity. Gastrointestinal cells are typically exposed to higher levels of these metabolites than those found in the systemic circulation, as both intestinal tissues and fecal samples of healthy individuals were reported to contain around 50-150 mM of short-chain fatty acids [37] and 2-16 µmol of individual phenolic compounds [38]. These ranges were used to establish physiologically relevant concentrations of the gastrointestinal metabolites in the subsequent cell assays (up to 200 µM for short-chain fatty acids, and up to 60 µM for phenolic acid metabolites, respectively). The nomenclature of phenolic acid metabolites is listed in Table 1.

Gastrointestinal Epithelial Cell Migration
Active gastrointestinal cell migration is critical for steady-state epithelial turnover and healing to adequately maintain a selective permeability barrier separating luminal content from the underlying gastrointestinal tissues [39]. The effects of polyphenols, namely anthocyanins (cyanidin-, delphinidin-, and malvidin-3-glucosides), their major phenolic metabolites (Figure 1), and short-chain fatty acids derived from fiber (acetate, propionate, and butyrate) that naturally present in wild blueberry digests, were screened for their ability to promote CCD-18 gut epithelial cell migration. Cell migration was enhanced relative to the vehicle control in response to all three anthocyanin glucosides tested, with the maximum effect observed for malvidin-3-glucoside (169.5 ± 48.4%, p < 0.05) and no effects for either chlorogenic acid (Figure 2a) or short-chain fatty acid metabolites (Figure 2b).

Formation of Epithelial Cell Monolayers
After establishing cell contacts, gastrointestinal epithelial CCD-18 cells developed junction complexes and established a monolayer resembling that of mature intestinal enterocytes [39]. The junction complexes are primarily composed of the membrane-anchored occludin and cytoplasmic zonula occludens-1 (ZO-1) proteins that regulate paracellular permeability [40]. Enhanced staining for ZO-1 protein was observed in CCD-18 cells treated with malvidin and butyrate ( Figure 3). Additionally, among the wild blueberry anthocyanin metabolites tested at the lower dose effective in promoting gastrointestinal cell migration, only methylated derivatives of hydroxybenzoic (vanillic acid 4H3MBA) and 3 hydroxycinnamic acid (3HCA) enhanced accumulation of the ZO-1 protein. The maximum effect was similar for ferulic, isoferulic, and caffeic acids in the range of 50-58% increase in the ZO-1 mean fluorescence signal (p < 0.05). Phenylacetic metabolites, such as 3,4-dihydroxyphenylacetic and 3-methoxy-4-hydroxyphenylacetic (homovanillic) acids had no biological activity in this assay (Figure 3).

Formation of Epithelial Cell Monolayers
After establishing cell contacts, gastrointestinal epithelial CCD-18 cell junction complexes and established a monolayer resembling that of mature enterocytes [39]. The junction complexes are primarily composed of the me chored occludin and cytoplasmic zonula occludens-1 (ZO-1) proteins that r cellular permeability [40]. Enhanced staining for ZO-1 protein was observe cells treated with malvidin and butyrate ( Figure 3). Additionally, among th berry anthocyanin metabolites tested at the lower dose effective in promoti testinal cell migration, only methylated derivatives of hydroxybenzoic (van 4H3MBA) and 3 hydroxycinnamic acid (3HCA) enhanced accumulation of tein. The maximum effect was similar for ferulic, isoferulic, and caffeic acid of 50-58% increase in the ZO-1 mean fluorescence signal (p < 0.05). Phenyla lites, such as 3,4-dihydroxyphenylacetic and 3-methoxy-4-hydroxyphenyla vanillic) acids had no biological activity in this assay (Figure 3).

Reduction in Inflammatory Response in Macrophages
Next, we investigated the ability of wild blueberry metabolites to mod acid production and gene expression profiles characteristic of acute and ch mation in the LPS-stimulated RAW 264.7 macrophages. Nitric oxide produ significantly affected by the parent anthocyanin glucosides when tested at centration (Figure 4). On the contrary, treatments with both 3-caffeoylquin (chlorogenic acid) and its major metabolite 3,4-dihydroxycinnamic acid (ca to strong reductions of nitric oxide release in the culture medium by 49.2% respectively. Figure 3. Effects of wild blueberry anthocyanins, short-chain fatty acids, and phenolic acid metabolites on the formation of junctional complexes in the CCD-18 gastrointestinal cells monolayers. Cells were allowed to form monolayers for 5 days with/without the indicated treatments and compared against vehicle control (0.1% DMSO). Zonulin accumulation in the junction complexes was visualized using ZO-1/TJP1 antibody Alexa Fluor 488, mean fluorescent intensity at 490 nm excitation/525 nm emission and reported as mean ± SEM (* p < 0.05, ** p < 0.01).

Reduction in Inflammatory Response in Macrophages
Next, we investigated the ability of wild blueberry metabolites to modulate nitric acid production and gene expression profiles characteristic of acute and chronic inflammation in the LPS-stimulated RAW 264.7 macrophages. Nitric oxide production was not significantly affected by the parent anthocyanin glucosides when tested at a lower concentration (Figure 4). On the contrary, treatments with both 3-caffeoylquinic acid (chlorogenic acid) and its major metabolite 3,4-dihydroxycinnamic acid (caffeic acid) led to strong reductions of nitric oxide release in the culture medium by 49.2% and 34.7%, respectively. mmuno 2022, 2, FOR PEER REVIEW Figure 4. Effects of wild blueberry anthocyanins, short-chain fatty acids, and pheno lites on nitric oxide production in activated macrophages. Cells were pre-treated wi olites and inflammatory response was induced with 1 µg/mL LPS for 6 h. Changes in tration as an indirect measure of nitric oxide production were reported as mean ± the LPS controls (* p < 0.05, ** p < 0.01, **** p < 0.0001).
The remaining methylated hydroxycinnamic acid metabolites used were also effective at reducing nitric oxide production to a similar degree the range of 22-31%. Finally, there was a trend to decrease nitric oxide relea rophages were exposed to hydroxybenzoic acid metabolites, with benzoic most potent (a reduction of 54.6%, p < 0.05). Short-chain fatty acid, phenylac nylpropanoic acid metabolites were largely ineffective in this assay.
All groups of wild blueberry metabolites showed varying levels of m gene expression profiles associated with acute and chronic biomarkers of (Cox-2, iNOS, IL-1β, and IL-6). Cox-2 expression was most strongly a caffeoylquinic acid −1.72x (−72%) and 2-hydroxybenzoic acid −1.76x (−76% metabolites from the hydroxybenzoic and hydroxycinnamic acids in the ran and 38-61%, respectively, while short-chain fatty acids, phenylacetic, as nylpropionic acids used in this study, showed little to no effects. A wider a olites affected the expression of iNOS, albeit the overall magnitude of t smaller than that of Cox-2. Many of the metabolites of both fiber and phen fected the expression levels of the early inflammatory IL-1β gene with no cl towards any subclass used in this study. The strongest inhibition of IL-1β e achieved by treatment with 3-caffeoylquinic acid and its major metabo droxycinnamic acid (−1.73x and −1.81x, respectively), as well as 2-hydrox (−1.65x). The mRNA levels of IL-6 were least affected by bioactive wild blu olites, as significant IL-6 mRNA decreases were observed only when cells to parent anthocyanin molecules (22-59%) and their hydroxycinnamic aci were in the range of 24-55%, as observed for the 3,4-dihydroxycinnamic, Figure 4. Effects of wild blueberry anthocyanins, short-chain fatty acids, and phenolic acid metabolites on nitric oxide production in activated macrophages. Cells were pre-treated with target metabolites and inflammatory response was induced with 1 µg/mL LPS for 6 h. Changes in nitrite concentration as an indirect measure of nitric oxide production were reported as mean ± SEM relative to the LPS controls (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).
The remaining methylated hydroxycinnamic acid metabolites used in this study were also effective at reducing nitric oxide production to a similar degree of potency in the range of 22-31%. Finally, there was a trend to decrease nitric oxide release when macrophages were exposed to hydroxybenzoic acid metabolites, with benzoic acid being the most potent (a reduction of 54.6%, p < 0.05). Short-chain fatty acid, phenylacetic, and phenylpropanoic acid metabolites were largely ineffective in this assay.
All groups of wild blueberry metabolites showed varying levels of modulation of gene expression profiles associated with acute and chronic biomarkers of inflammation (Cox-2, iNOS, IL-1β, and IL-6). Cox-2 expression was most strongly affected by 3-caffeoylquinic acid −1.72x (−72%) and 2-hydroxybenzoic acid −1.76x (−76%), followed by metabolites from the hydroxybenzoic and hydroxycinnamic acids in the range of 22-56% and 38-61%, respectively, while short-chain fatty acids, phenylacetic, as well as phenylpropionic acids used in this study, showed little to no effects. A wider array of metabolites affected the expression of iNOS, albeit the overall magnitude of the effect was smaller than that of Cox-2. Many of the metabolites of both fiber and phenolic origin affected the expression levels of the early inflammatory IL-1β gene with no clear specificity towards any subclass used in this study. The strongest inhibition of IL-1β expression was achieved by treatment with 3-caffeoylquinic acid and its major metabolite 3,4-dihydroxycinnamic acid (−1.73x and −1.81x, respectively), as well as 2-hydroxybenzoic acid (−1.65x). The mRNA levels of IL-6 were least affected by bioactive wild blueberry metabolites, as significant IL-6 mRNA decreases were observed only when cells were exposed to parent anthocyanin molecules (22-59%) and their hydroxycinnamic acid metabolites were in the range of 24-55%, as observed for the 3,4-dihydroxycinnamic, 4-hydroxy-3-methoxycinnamic, and 3-hydroxy-4-methoxycinnamic acids ( Figure 5).

Figure 5.
Heatmap of anti-inflammatory effects of wild blueberry anthocyanins, short-chain fatty acids, and phenolic acid metabolites based on qPCR gene expression profiles of key biomarkers of acute and chronic inflammation, including cyclooxygenase-2 (Cox-2), inducible nitric oxide synthase (iNOS), interleukin-1β (IL-1β), and interleukin-6 (IL-6). Macrophages were pre-treated with individual compounds as the specified and inflammatory response was induced with 1 µg/mL LPS for 6 h. Dexamethasone (Dex) at 10 µM was used as a positive control. Total RNAs were isolated from duplicate treatments and pooled for qPCR analysis. Fold changes in gene expression are reported as means relative to healthy controls.

Discussion
Although early humans relied on mixed diets of leafy greens, storage tubers, fruits (seeds, nuts), and opportunistic meats, diets based on modern crops show significant decreases in many health-promoting metabolites as we selected for new crop varieties with improved taste profiles and shelf life [40]. Wild blueberries are ideally positioned to start to close this gap, as they generally contain higher amounts of fibers and anthocyanins than their cultivated counterparts. Blueberries are considered a good source of fiber and healthpromoting polyphenols, specifically anthocyanins [41]. Low bioavailability of these high molecular weight polyphenols and dietary fibers, at least partially resistant to digestion, has been a matter of research and debate for several decades. We now know that chemical degradation and bacterial catabolism of these molecules into the smaller and more bioavailable catabolites, such as phenolic acids [9] and short-chain fatty acids [42] form a critical connection between xenobiotic metabolism, bioactivation, and host-gut biochemical interactions with these structures. Microbiome-derived fermentation products from foods naturally rich in polyphenols and fibers may also play an important role in the gastrointestinal mucoprotection, feeding of the intestinal epithelial cells, maintaining the intestinal barrier, and regulating the immune outcomes in the gut [43].
This study described the differential effects of short-chain fatty acids and phenolic metabolites on migration and formation of tight junctions by the gastrointestinal epithelial cells. A group of hydroxybenzoic acid metabolites significantly enhanced CCD-18 cell migration. This outcome was the net of cell proliferation and migration into the target area. Benzoic acids were common in the microbial degradation metabolite pools originating from dietary aromatic compounds naturally present in berries [44] and fermented dairy products [45]. In recent studies, benzoic acid supplementation increased the digestibility of total nitrogen, therefore improving the utilization of nutrients [46] and improving the mucosal-epithelial integrity in pigs [47]. A generation of benzoic acid metabolites via the intestinal microbial metabolism of dietary aromatic compounds modified urinary hippuric acid profiles and correlated with positive health outcomes in patients with gastrointestinal disorders, such as Crohn's disease [48]. Reduced synthesis of benzoic acid metabolites by the intestinal microbiota may therefore be implicated in the gut epithelial pathogenesis, loss of barrier function, and warrants further investigation. Benzoic acid Figure 5. Heatmap of anti-inflammatory effects of wild blueberry anthocyanins, short-chain fatty acids, and phenolic acid metabolites based on qPCR gene expression profiles of key biomarkers of acute and chronic inflammation, including cyclooxygenase-2 (Cox-2), inducible nitric oxide synthase (iNOS), interleukin-1β (IL-1β), and interleukin-6 (IL-6). Macrophages were pre-treated with individual compounds as the specified and inflammatory response was induced with 1 µg/mL LPS for 6 h. Dexamethasone (Dex) at 10 µM was used as a positive control. Total RNAs were isolated from duplicate treatments and pooled for qPCR analysis. Fold changes in gene expression are reported as means relative to healthy controls.

Discussion
Although early humans relied on mixed diets of leafy greens, storage tubers, fruits (seeds, nuts), and opportunistic meats, diets based on modern crops show significant decreases in many health-promoting metabolites as we selected for new crop varieties with improved taste profiles and shelf life [40]. Wild blueberries are ideally positioned to start to close this gap, as they generally contain higher amounts of fibers and anthocyanins than their cultivated counterparts. Blueberries are considered a good source of fiber and health-promoting polyphenols, specifically anthocyanins [41]. Low bioavailability of these high molecular weight polyphenols and dietary fibers, at least partially resistant to digestion, has been a matter of research and debate for several decades. We now know that chemical degradation and bacterial catabolism of these molecules into the smaller and more bioavailable catabolites, such as phenolic acids [9] and short-chain fatty acids [42] form a critical connection between xenobiotic metabolism, bioactivation, and host-gut biochemical interactions with these structures. Microbiome-derived fermentation products from foods naturally rich in polyphenols and fibers may also play an important role in the gastrointestinal mucoprotection, feeding of the intestinal epithelial cells, maintaining the intestinal barrier, and regulating the immune outcomes in the gut [43].
This study described the differential effects of short-chain fatty acids and phenolic metabolites on migration and formation of tight junctions by the gastrointestinal epithelial cells. A group of hydroxybenzoic acid metabolites significantly enhanced CCD-18 cell migration. This outcome was the net of cell proliferation and migration into the target area. Benzoic acids were common in the microbial degradation metabolite pools originating from dietary aromatic compounds naturally present in berries [44] and fermented dairy products [45]. In recent studies, benzoic acid supplementation increased the digestibility of total nitrogen, therefore improving the utilization of nutrients [46] and improving the mucosal-epithelial integrity in pigs [47]. A generation of benzoic acid metabolites via the intestinal microbial metabolism of dietary aromatic compounds modified urinary hippuric acid profiles and correlated with positive health outcomes in patients with gastrointestinal disorders, such as Crohn's disease [48]. Reduced synthesis of benzoic acid metabolites by the intestinal microbiota may therefore be implicated in the gut epithelial pathogenesis, loss of barrier function, and warrants further investigation. Benzoic acid metabolites also showed a partial tendency towards downregulating the COX-2/PGE 2 signaling pathway at the level of gene expression ( Figure 5).
Several hydroxycinnamic acids, including 3,4-dihydroxycinnamic (caffeic acid), 4-hydroxy-3-methoxycinnamic (ferulic acid), and 3-hydroxy-4-methoxycinnamic (isoferulic acid) also improved gut epithelial cell migration. This data was in agreement with the earlier reported effects of ferulic acid on the LPS-induced epithelial barrier dysfunction [49]. At the same time, short-chain fatty acids derived from dietary fibers had little effect on the migration of the gastrointestinal cells, suggesting that this process does not directly depend on additional colonic energy supplies and trophic factors provided by these metabolites [50]. However, one of the short-chain fatty acids, butyrate, showed a significant effect on the formation of junctional complexes in CCD-18 cell cultures. This effect can be partially explained by the ability of butyrate, and to a lesser extent, propionate, to modulate gastrointestinal gene transcription via histone acetylation and induce permeability changes in tight junctions [51]. Among phenolic metabolites tested, an increase in hydrophobicity because of methylation was generally associated with stronger changes in junctional complexes, as measured by the accumulation of ZO-1 protein. It is possible that more hydrophobic phenolic metabolites had an increased capacity to improve the gastrointestinal barrier function and therefore impede the effects of luminal bacteria on the circulation and mesenteric lymph content [52]. On the other hand, we would expect more polar hydroxybenzoic acid metabolites to have stronger effects on the gastrointestinal tissues that face luminal content as previously reported in the turkey [53] and pig [54] models.
The integrity of intestinal mucosa is not only the major critical factor for nutrient digestion and absorption, but also plays a key role in the recognition and activation of the immune system in response to physiological stimuli and pathogen infection. Lipopolysaccharide induces an increase in intestinal permeability and subsequently activates residential tissue macrophages that promote differentiation and development of the pro-inflammatory response [55]. These effects are mediated by changes in the expression levels of cytokine biomarkers that drive activation of the classical pro-inflammatory M1 pathway, including the inducible nitric oxide synthase (iNOS), interleukin-1β (IL-1β), and interleukin-6 (IL-6) [56]. Cyclooxygenase-2 (Cox-2)/PGE 2 production in activated macrophages further drives distinct intracellular pathways that modulate inflammation and host immune responses [57].
In this study, caffeoylquinic, hydroxybenzoic, and hydroxycinnamic acids reduced nitric oxide release and associated gene expression levels of iNOS in macrophages, and this effect was largely absent when cells were exposed to hydroxyphenylacetic and hydroxyphenylpropionic acids (Figures 4 and 5). Among those, 3,4-dihydroxycinnamic (caffeic acid) and chlorogenic acid (that carries the caffeic acid moiety as a part of its structure) showed the strongest inhibition of pro-inflammatory gene expression in macrophages at the concentration tested. This was an interesting parallel to a previous study that showed the importance of the caffeic acid pharmacophore to metabolic health [31]. Methylated derivatives of caffeic acid showed stronger effects on COX-2 and IL-6 expression as compared to iNOS and IL-1β genes, suggesting different efficacies at various pathways of the inflammatory regulatory networks.
While all subclasses of phenolic metabolites are typically observed in human plasma or serum metabolome following the consumption of plant foods, only some metabolites achieve maximum observed concentrations of 10-25 µM, as reported for 3-phenylpropanoic, 2-hydroxybenzoic, and hippuric acids, while the individual concentrations of the remaining phenolic metabolites typically range between 10 nM and 1 µM [58]. The mean cumulative concentrations of these series of phenolic metabolites, however, often reach 30-40 µM in the systemic circulation and may be responsible for additive biological effects as previously shown for hydroxybenzoic acid metabolites in vascular smooth muscle cells [59]. Ranges of concentrations for major phenolic acid metabolites identified in plasma, urine, and fecal water support the notion that these can be achieved by dietary supplementation with various foods, including blueberries [41]. It, therefore, seems plausible that body tissues located in immediate proximity to the gastrointestinal luminal content (gut epithelial and smooth muscle cells) and blood vessels (endothelial and smooth muscle cells) are exposed to higher levels of polar microbial catabolites derived from dietary fiber and polyphenols, while methylated phenolic metabolites are better positioned to achieve biologically relevant outcomes in the more distant metabolically active tissues (adipose, liver, and tissue residential immune cells). The yet unexplored structure-activity relationships between different subclasses of phenolic metabolites with respect to their methylation and hydrophobicity, which facilitates passive membrane permeation and interaction with molecular and intercellular targets, may partially explain the various findings reported in adipocyte [32], liver [31,60], smooth muscle [59], and endothelial [61] cell culture models.

Conclusions
The observed activity of wild blueberry metabolites on key biomarker measures of gastrointestinal and immune cells was specific to the particular series and structures derived from digestion and fermentation of dietary fiber and polyphenolic compounds, and varied depending on the tissue analyzed. Further work is required to elucidate the underlying mechanisms and targets involved in the recognition of these signals. This work also raised an intriguing possibility that hydroxybenzoic acid metabolites play an important role in the modulation of the physiological outcomes in the gastrointestinal epithelial cells, similar to the short-chain fatty acids.
Overall, findings also pointed to the hydroxybenzoic and hydroxycinnamic phenolic acid metabolites as being more potent in regulating gene expression profiles that are central to sustaining acute and chronic inflammation in the gastrointestinal and immune cells. Thus, our results suggested that diets that incorporate wild blueberries hold a high potential to supplement humans with small molecular weight metabolites that may improve the management and outcomes of immune and metabolic risk factors linked to sustained inflammation.
Author Contributions: C.D.K. and S.K. conceived the study and outlined the scope of work; J.L. and D.E. established cell line assays; T.R. managed metabolite stocks and treatments; K.S. performed RNA extractions and qPCR analysis; S.K. performed assay quantifications and overall data analysis; T.R. and S.K. wrote the manuscript; C.D.K. and D.E. reviewed the manuscript. All authors have read and agreed to the published version of the manuscript. Data Availability Statement: Data may be obtained from a corresponding author upon a reasonable request.