Bioavailability is defined by the FDA as “rate and extent to which the active ingredient or moiety is absorbed and becomes available at the site of action” [
21]. Methods for determination of bioavailability of anthocyanins include human (
in vivo) or simulated experiments performed in a laboratory (
in vitro).
In vivo methods provide direct data of bioavailability and have been used for a large variety of nutrients. On the other hand,
in vitro methods have the advantage of being more rapid, less expensive, less labor intensive, and do not have ethical restrictions.
In vitro methods simulate gastrointestinal digestion under controlled conditions using commercial digestive enzymes, whereas the final absorption process is commonly assessed using Caco-2 cell cultures [
22]. Here, we will focus on anthocyanin transport and metabolism through Caco-2 cells. For a more comprehensive overview on bioavailability aspects of anthocyanins, please refer to these other reviews [
7,
16,
19].
3.1. Caco-2 Cell Growth and Differentiation
The Caco-2 cell line has been established by Fogh and co-workers in 1977 from a human colon adenocarcinoma, and originally used for the screening of cytotoxic effects of anti-tumor drugs and for the study of drug resistance mechanisms [
23]. During the past few decades, this cell line has been extensively used for cellular permeability studies of polyphenols [
24,
25,
26,
27]. It has been well established that Caco-2 cells can undergo spontaneous differentiation in culture conditions and exhibit the characteristics of mature enterocytes. The cell surface facing the top medium develops a brush border that resembles the luminal membrane of the intestinal epithelium. The cell surface attaching to the permeable membrane and facing the bottom medium develops into the basolateral membrane [
28,
29]. Despite their colonic origin, Caco-2 cells express the morphological and functional characteristics of small intestinal cells. The Caco-2 monolayer houses multiple transporters, receptors and metabolic enzymes such as cytochrome P450 1A (CYP1A), sulfotransferases (SULTs), UDP-glucuronosyltransferases (UGTs), and glutathione S-transferases (GSTs) [
30].
Transport experiments are generally carried out using filter-based inserts, where cells are seeded and allowed to grow and differentiate to confluent monolayers for approximately 21 days post seeding. Before performing the transport experiment, the integrity of the Caco-2 monolayer is controlled by measuring the transepithelial electrical resistance (TEER), or, more reliably, by examining the permeability of paracellular markers such as mannitol, inulin, Dextran, PEG 4000, and lucifer yellow [
30]. TEER is a non-invasive technique, which measures the impedance between the lumen and basolateral tissue. TEER measurements use a constant direct current applied by two electrodes, one connected with the lumen side and the other one with the basolateral side. By applying Ohm’s law it is possible to measure the related cells resistance [
31]. It was reported in the literature that an acceptable TEER value for Caco-2 cell monolayers should be from 200 to 1000 ohm per cm
2 [
32]. Since its original isolation, the Caco-2 cell line has been propagated in several laboratories around the world, producing Caco-2 cells of different “age”, or number of passages in culture. The expressions of typical differentiation markers of intestinal enterocytes were shown to increase from early to late passages. Accordingly, the TEER value has also been demonstrated to increase in later passages of cell monolayer [
33].
In order to reduce the heterogeneity of the Caco-2 parental cell line and to improve the performance and the stability of this cellular model, some clonal cell lines have been obtained from Caco-2. Among them, TC-7 is often used to simulate polyphenol transport [
34,
35,
36]. The TC-7 clone exhibited similar cell morphology to Caco-2 cells, displaying the presence of brush-border membrane and microvilli, and the formation of tight junctions. Similarly, on the basis of biochemical attributes and permeability characteristics, the TC-7 subclone appears to be similar to Caco-2 cells and presents a suitable alternative to parental cells for intestinal permeability studies [
37].
The HT-29 cell line is another cell line from colorectal origin with epithelial morphology, and has been used as a model for absorption, secretion and transport by intestinal cells. Under standard culture conditions, these cells grow as a non-polarized, undifferentiated monolayer. However, altering culture conditions or treating the cells with different inducers results in a differentiated and polarized morphology, characterized by a redistribution of membrane antigens and development of an apical brush-border membrane [
38]. Other human intestinal cell lines are less popular for the simulation of the human intestinal epithelium, such as the HCT-116 and SW480 cell lines, which are mainly used in unraveling cancer-related mechanisms [
39], and the HuTu-80 cell line, a model for duodenal cells [
40].
3.2. Anthocyanin Transport through Caco-2 Cells
Studies investigating anthocyanin absorption by Caco-2 cells are presented in
Table 1 [
11,
28,
29,
41,
42,
43,
44,
45,
46,
47,
48,
49]. The majority of these studies suggest that unlike other flavonoids, anthocyanins could be transported through Caco-2 monolayers in intact glycone forms, with the exceptions of black currant and some grape anthocyanins. Steinert
et al. [
41] demonstrated that anthocyanins from black currant, namely Dp-3-Glu, Dp-3-Rut, Cy-3-Glu and Cy-3-Rut, were not detected in any serosal solution. However, the authors showed that anthocyanins disappeared from the luminal side, not due to the anthocyanin degradation process but rather due to physiological actions of the cells. Similarly, Dp-3-Glu from grape extract was not transported [
48]. Diglucosylated Mv-3,5-DGlu and Pn-3,5-DGlu from grape/blueberry extract were also not transported in quantifiable concentrations [
49]. On the other hand, studies that observed anthocyanin transport, reported very low transport efficiencies. The transport efficiency of anthocyanins from blueberry extracts averaged
ca. 3%–4% (<1% in Dp-3-Glu) [
28]. Similarly, only about 1% of the red grape skin anthocyanins passed through a Caco-2 cell monolayer and reached the basolateral side [
42]. The percentage of transported monomeric anthocyanin glycosides from açaí fruit ranged from 0.5% to 4.9% [
43], whereas according to Cardona
et al. [
46] the transport rate of açaí anthocyanins was 1.2%. Transport efficiencies of Mv-3-Glu and Cy-3-Glu standards were found to be 4% and 0.8%–2.4%, respectively [
44,
45]. Moreover, Cy-3-Glu-Rut recovery from sour cherry fruit and nectar was
ca. 0.5%–4% [
11]. Trace amount of Pg-3-Glu, the predominant anthocyanin from strawberry extract, was found on the basolateral side of the epithelium [
47]. Transport efficiency of the major grape anthocyanin (Mv-3-Glu) was 0.35% [
48], while the absorption rates of Mv-3-Glu, Pn-3-Glu, Pt-3-Glu, Dp-3-Glu and Cy-3-Glu from grape/blueberry extract were 0.005%–0.06% [
49]. These results are in line with
in vivo studies showing a very low bioavailability of anthocyanins, with <1% of the ingested amount reaching the plasma or being excreted in the urine [
50,
51,
52,
53,
54].
Table 1.
Studies investigating anthocyanin absorption by Caco-2 cells.
Table 1.
Studies investigating anthocyanin absorption by Caco-2 cells.
Sample | Pre-Treatment | Anthocyanins | Anthocyanin Concentration | Cell Origin | Cell Differentiation | Incubation Time | Key Findings | Reference |
---|
Blueberry | Chemical extraction | Dp-3-Glu, Cy-3-Gal, Cy-3-Glu, Pt-3-Glu, Pn-3-Gal, Pn-3-Glu, Mv-3-Glu | 50 μg/mL | ATCC | 20–26 days | 0–120 min | Transport efficiency of ACNs averaged ca. 3%–4% (<1% in Dp-3-Glu); Glucose-based ACNs had higher bioavailability than galactose-based ACNs | [28] |
Black currant extract | - | Dp-3-Glu, Dp-3-Rut, Cy-3-Glu, Cy-3-Rut | 180 μM | DSWZ | 19–21 days | 0–80 min | ACNs were not detected in any serosal solution | [41] |
Red grape skin | Chemical extraction | Dp-3-Glu, Cy-3-Glu, Pt-3-Glu, Pn-3-Glu, Mv-3-Glu | 200 μg/mL | ATCC | 25 days | 4 days of pre-treatment + 6 min | Only ca. 1% of ACNs are transported; ACN transport significantly increased in the presence of ethanol; Cells pre-treated with ACNs showed ca. 50% increased transport; GLUT2 may be responsible for ACN transport | [42] |
Açaí pulp | Chemical extraction | Cy-3-Rut, Cy-3-Glu | 50–500 μg/mL | ATCC | 21 days | 30–120 min | Transport efficiency of ACNs was 0.5%–4.9%; Presence of polymeric ACNs decreased transport of monomeric ACN glycosides (up to 40.3%) | [43] |
Standard | - | Cat-Mv-3-Glu, Mv-3-Glu | 100 μM | n/a | 21 days | 30–120 min | Transport efficiency of Mv-3-Glu was 4%; Absorption efficiency of Cat-Mv-3-Glu was lower than Mv-3-Glu (ca. 3%) | [44] |
Sour cherry fruit and nectar | Chemical extraction | Cy-3-Glu-Rut | 55 μM | ATCC | 23–24 days | 360 min | Cy-3-Glu-Rut recovery was ca. 0.5%–4%; Cy-3-Glu-Rut transported 3 times more efficiently from nectar than fruit; Sucrose and citric acid enhanced the transport of Cy-3-Glu-Rut (ca. 5-fold); SPE reduced the transport efficiency of Cy-3-Glu-Rut by 5–10-fold | [11] |
Standard | Encapsulation | Cy-3-Glu | 37.5 μM | n/a | 20–26 days | 60 min | Nano-encapsulated Cy-3-Glu with apoferritin was more efficiently transported compared to free Cy-3-Glu | [29] |
Standard | - | Cy-3-Glu | 10–40 μM | ATCC | 13 days | 30–120 min | Transport efficiency of Cy-3-Glu was 0.8%–2.4%; Phloridzin and phloretin inhibited the absorption of Cy-3-Glu; SGLT1 and GLUT2 are probably involved in the absorption of Cy-3-Glu | [45] |
Açaí concentrate | Chemical extraction | Cy-3-Glu, Cy-3-Rut | 500 μg/mL | ATCC | 18–21 days | 0–120 min | Transport rate of ACNs was 1.22%; Phospholipids from soy lecithin and terpenes from cold pressed citrus oil increased the transport of ACNs | [46] |
Strawberry | Chemical extraction + in vitro digestion | Pg-3-Glu, Pg-3-Mal-Glu, Cy-3-Glu | 16.3 mg/100 g | ATCC | 21 days | 120 min | Trace amount of Pg-3-Glu was transported | [47] |
Grape | Chemical extraction | Mv-3-Glu, Pn-3-Glu, Pt-3-Glu, Cy-3-Glu, Dp-3-Glu | 1766.1 μg/mL | ATCC | 21 days | 30–240 min | Mv-3-Glu, Pn-3-Glu, Pt-3-Glu and Cy-3-Glu were transported, whereas Dp-3-Glu was not transported; Transport efficiency of major anthocyanin (Mv-3-Glu) was 0.35% | [48] |
Grape/blueberry extract | - | Mv-3-Glu, Pn-3-Glu, Pt-3-Glu, Dp-3-Glu, Cy-3-Glu, Mv-3,5-DGlu, Pn-3,5-DGlu | 2613 μM | ATCC | 21 days | 0–90 min | Absorption rates of Mv-3-Glu, Pn-3-Glu, Pt-3-Glu, Dp-3-Glu and Cy-3-Glu were 0.005%–0.06%; Mv-3,5-DGlu and Pn-3,5-DGlu were not transported in quantifiable concentrations | [49] |
Few studies compared the transport efficiency of anthocyanins across Caco-2 cells with other polyphenols. The transport of both Mv-3-Glu and catechin through Caco-2 cells was found to be time dependent and reached approximately to the same value (4%) after 120 min of incubation [
44]. Similarly, the recovery of epicatechin in the basolateral side (1%–4%) was also about the same with Cy-3-Glu-Rut (0.5%–4%) [
11]. Reported transport of some other flavonoids through Caco-2 cells was 30% for quercetin, 17% for genistein and 6% for epicatechin [
55].
The aglycone structure of anthocyanins is one of the many factors influencing their transport. For instance, Dp-3-Glu from blueberry extract showed lower transport efficiency compared to Mv-3-Glu and Pn-3-Glu. This may be a result of the higher number of hydroxyl groups in Dp or the greater hydrophobic structure of Mv that facilitated an increased portioning into cells and tissues. In addition, Dp has no OCH
3 group, while Pn has one and Mv has two OCH
3 groups (
Figure 1), indicating that hydrophilic and hydrophobic groups affect the absorption of anthocyanins [
28]. Similarly, for black currant anthocyanins the loss of delphinidins was significantly higher than cyanidins. Thus, the structural features might be crucial for anthocyanin stability [
41].
Sugar moieties and polymeric structures may also have an influence on anthocyanin absorption by Caco-2 cells. For blueberry extracts, glucose-based anthocyanins had higher bioavailability than galactose-based anthocyanins. [
28]. On the other hand, for black currant anthocyanins no differences are shown between the respective glucose and rutinose sugar moieties indicating that sugar conjugates may have a minor effect on anthocyanin stability [
41]. The presence of polymeric anthocyanins in açaí fruit decreased the transport of monomeric anthocyanins glycosides in a dose-dependent manner by up to 40.3% [
43]. Similarly, the absorption efficiency of flavanol-anthocyanin dimer Catechin-Mv-3-Glu, an anthocyanin derivative reported in grape skins and red wine, was lower than Mv-3-Glu (
ca. 3%) [
44].
The presence of other food components has been shown to have a major impact on anthocyanin transport. Solid phase extraction (SPE) of sour cherry extracts reduced the transport efficiency of Cy-3-Glu-Rut by 5–10-fold [
11]. Ethanol, one of the main constituents of red wine, improved anthocyanin transport through Caco-2 cells [
42]. However, this hypothesis is open to debate as there are some
in vivo reports [
56,
57] claiming that ethanol has no influence on anthocyanin absorption. The ethanol concentration used in the cell culture study (1%), which was non-toxic to Caco-2 cells [
42], is much lower than the actual ethanol concentration in red wine. Therefore the impact of ethanol on anthocyanin absorption and bioavailability may depend on the models as well as the doses used [
42]. Citric acid also enhanced anthocyanin transfer to the basolateral side of Caco-2 cells [
11]. This may be linked to the fact that anthocyanins are more stable at low pH values [
16,
58] (
Figure 2). This effect of pH on transport across Caco-2 cells may have some physiological relevance. Although the cellular interstices and blood have a pH of around 7.4, the pH in the upper gastrointestinal tract under fasting conditions ranges from 5.0 to 6.5. In addition, the pH of the acidic microclimate just above the epithelial cell layer has been reported to be between 5.8 and 6.3 [
11,
28]. Furthermore, phospholipids from soy lecithin and terpenes from cold pressed citrus oil increased the transport of açaí anthocyanins in an
in vitro cell monolayer model with Caco-2 cells, and a combination of phospholipids and terpenes was found to be the most effective [
46].
In addition to the factors mentioned above, the physiological pH and temperature conditions (pH 7 and 37 °C) used in Caco-2 cell culture experiments may have a great influence on the stability of anthocyanins. In fact, a study on Cy-3-Glu [
59] showed that there was no significant difference between the Caco-2 cell and cell free incubations in terms of the losses of Cy-3-Glu and the appearance of degradation products. These findings suggest that the loss of anthocyanins may be the result of spontaneous chemical breakdown rather than Caco-2 cell induced enzymatic deglycosylation followed by chemical degradation.
Although the exact mechanism of anthocyanin absorption in the small intestine is still unclear, it has been proposed that anthocyanins could interfere with the transporters responsible for their own transport. The candidates for anthocyanin transporters were the glucose transporters, since anthocyanins possess a sugar moiety, in particular a glucose residue. SGLT1 and GLUT2 are the main hexose transporters described in Caco-2 cells. SGLT1 is an energy-dependent and sodium-dependent cotransporter, whereas GLUT2 is a facilitated transporter. SGLT1 is only present on the apical membrane and until a few years ago, GLUT2 was described to be present only in the basolateral membrane and in some pathologies on the apical membrane. Recently, it has been described and accepted that GLUT2 is present on the apical side and can be gathered to the membrane in the presence of a large amount of glucose, therefore becoming the main transporter responsible for glucose uptake [
42,
45]. It was found that GLUT2 expression assessed by RT-PCR was increased in Caco-2 cells pretreated with red grape skin anthocyanins, by comparison with controls, indicating that chronic consumption of anthocyanins could be favorable for their own bioavailability. In addition, the tested red grape skin anthocyanins interfered with glucose uptake resulting in an inhibitory effect (about 60% decrease) [
42]. Similarly, Pn-3-Glu from strawberry extract was able to influence glucose uptake into the cells and transport to the basolateral side by inhibiting activities of the glucose transporters [
27]. Another study also confirmed that exposure to anthocyanin rich berry extract derived from blueberry, bilberry, cranberry, elderberry, raspberry seeds and strawberry significantly reduce SGLT1 and GLUT2 expressions [
60]. Inhibition studies conducted using the pharmacological agents, phloridzin, an inhibitor of SGLT1, or phloretin, an inhibitor of GLUT2, revealed that the absorption of Cy-3-Glu was significantly inhibited in the presence of these agents [
45]. These data suggest that anthocyanins may prevent hyperglycemia by decreasing glucose transporter expressions.
Since the high instability of anthocyanins has a direct impact on their potential health benefits, food processing technologies such as encapsulation may be used to improve their bioavailability [
61,
62,
63]. Accordingly, the nano-encapsulated Cy-3-Glu with apoferritin was more efficiently transported through a Caco-2 cell monolayer compared to free Cy-3-Glu [
29]. In another study, processing of sour cherry fruit into nectar led to three times more efficient transport of Cy-3-Glu-Rut through a Caco-2 cell monolayer [
11].
3.3. Anthocyanin Metabolism by Caco-2 Cells
Polyphenols undergo Phases I and II transformations in the human body. Phase I transformations consist of oxidation, reduction and hydrolysis, but these transformations occur less frequently. Phase II biotransformations taking place in the liver and the intestine occur more intensively. These Phase II transformations consist of conjugation reactions where different are formed (methyl, glucuronic and sulfate derivatives) [
64]. Spontaneous transformation of anthocyanins to phenolic acids and aldehydes is reported to occur under biological conditions [
65]. Confirming that, under cell culture conditions, the main metabolites of Cy-3-Glu and Cy are detected as protocatechuic acid (PCA) and phloroglucinaldehyde (PGA), which are derived from the A and B rings of the parental compound (
Figure 3). With action of enzymes, these metabolites can be further degraded to glucuronide and sulfate conjugates [
59]. Another important Phase II reaction of anthocyanins is the methylation, which alters the number of hydroxyl and methoxyl groups in ring B in comparison with the native compound. Although not so intense as genuine anthocyanins, methylated metabolites of Cy-3-Glu, Dp-3-Glu and Pt-3-Glu displayed some antiproliferative activity for the Caco-2 cell line [
66]. On the other hand, some other anthocyanin metabolites including gallic acid, 3-
O-methylgallic acid, and PGA reduced cell proliferation in Caco-2 cells more effectively compared to parental anthocyanins [
67]. Therefore when assessing the health benefits of anthocyanins, potential effects of such metabolites should be taken into account.
Figure 3.
Metabolites (PGA: phloroglucinaldehyde; PCA: protocatechuic acid) of cyanidin-3-glucoside (Cy-3-Glu) and cyanidin (Cy).
Figure 3.
Metabolites (PGA: phloroglucinaldehyde; PCA: protocatechuic acid) of cyanidin-3-glucoside (Cy-3-Glu) and cyanidin (Cy).
3.4. Bioactive Properties of Anthocyanins on Caco-2 Cells
Although the bioavailability of anthocyanins is very low, anthocyanins from several different food sources have been shown to exert health-promoting effects on Caco-2 cells (
Table 2). These potential bioactive properties included antiproliferative [
68,
69,
70,
71,
72,
73,
74,
75,
76,
77,
78,
79], antioxidant [
79,
80,
81,
82,
83,
84,
85] and anti-inflammatory [
48,
86,
87] effects. MTT (3-(4,5-dimethylthiazolil-2-yl)-2,5-diphenyl-tetrazoliumbromide) is the most common assay that is applied to evaluate the antiproliferative effects of anthocyanins on Caco-2 cells [
68,
69,
73,
74,
76,
77,
78]. This assay is based on the conversion of the yellow tetrazolium salt MTT by mitochondrial dehydrogenase of live cells to the purple formazan. Trypan blue [
70,
79], thymidine incorporation [
71], CCK-8 (cell counting kit 8) [
72], LDH (lactate dehydrogenase) [
73] and SRB (sulphorhodamine B) [
74] are some other assays that are used to determine the cell viability after anthocyanin treatment. For the antioxidant activity, the formation of intracellular ROS (reactive oxygen species) is often determined using a fluorescent probe, DCFH-DA (2′,7′-dichloro-dihydro-fluorescein diacetate) [
79,
80,
81,
82,
84,
85]. Prior to anthocyanin treatment, the oxidation is initiated with different compounds including t-BHP (
tert-butyl hydroperoxide) [
79,
80,
85] and AAPH (2,2′-azobis (2-amidinopropane) dihydrochloride) [
81,
82]. Anti-inflammatory activities of anthocyanins included reduction of NF-κB (nuclear factor κB) activation [
86,
87], inhibition of NO (nitric oxide) secretion [
87], downregulating the expression of pro-inflammatory cytokines (particularly IL-8) and reducing the levels of adhesion molecules [
48].