Professional Documents
Culture Documents
Short communication
PII: S0308-8146(17)31664-3
DOI: https://doi.org/10.1016/j.foodchem.2017.10.035
Reference: FOCH 21858
Please cite this article as: Kellett, M.E., Greenspan, P., Pegg, R.B., Modification of the Cellular Antioxidant Activity
(CAA) Assay to Study Phenolic Antioxidants in a Caco-2 Cell Line, Food Chemistry (2017), doi: https://doi.org/
10.1016/j.foodchem.2017.10.035
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Modification of the Cellular Antioxidant Activity (CAA) Assay
a
Department of Food Science & Technology, College of Agricultural and Environmental
Sciences, The University of Georgia, 100 Cedar Street, Athens, GA, 30602, United States
b
Department of Pharmaceutical & Biomedical Sciences, College of Pharmacy, The University of
Running Title: Phenolic Antioxidants and the Cellular Antioxidant Activity Assay
*Corresponding author: Dr. Ronald B. Pegg. Tel: (706) 542-1099. Fax: (706) 542-1050. E-mail:
rpegg@uga.edu
1
Abstract
In vitro assays are widely used to analyze the antioxidant potential of compounds, but they
cannot accurately predict antioxidant behavior in living systems. Cell-based assays, like the
cellular antioxidant activity (CAA) assay, are gaining importance as they provide a biological
perspective. When the CAA assay was employed to study phenolic antioxidants using
hepatocarcinoma (HepG2) cells, quercetin showed antioxidant activity in HepG2 cells; 25 and
Catechin, a phenolic antioxidant present in many foods, bestowed virtually no CAA in HepG2
cells. When Caco-2 cells were employed, more robust antioxidant activity was observed; 50 M
Based on these results, likely due to differences in active membrane transport between the cell
types, the Caco-2-based CAA assay appears to be a more appropriate method for the study of
Caco-2 Cells
2
1 1. Introduction
2 Multiple in vitro methods of analyzing dietary antioxidants are widely used today, but their
3 overall usefulness has been called into question (Schaich, Tian, & Xie, 2015). This is because the
4 results from in vitro chemical tests under non-physiological conditions simply cannot be
5 extrapolated to predict the performance of the tested sample in biological systems (Lpez-
6 Alarcn, & Denicola, 2013). Most in vitro tests analyze a compounds capability to quench
7 radicals via one of two mechanisms: the hydrogen atom transfer (HAT) method or the single
8 electron transfer (SET) method, which is primarily due to a compounds redox properties (Craft,
9 Kerrihard, Amarowicz, & Pegg, 2012). This oversimplifies the antioxidant action, as in many
10 cases a single system (food or otherwise), or even a single compound, can use both mechanisms
11 concurrently (Migliavacca, Carrupt, & Testa, 1997; Prior, Wu, & Schaich, 2005). Moreover,
12 radical quenching is not the whole role of antioxidant defenses: to protect cells from oxidative
13 stress, antioxidant compounds may up-regulate antioxidant enzymes, modulate gene expression,
14 or change cell signaling (Jones, 2006; Lpez-Alarcn, & Denicola, 2013). Test tube assays,
15 such as oxygen radical absorbance capacity (ORACFL) and ferric reducing antioxidant potential
16 (FRAP), are incapable of examining these additional factors, and for that reason, new methods of
18 Cell-based assays are gaining traction as a more biologically relevant technique that serves
19 as a middle ground between in vitro assays and animal feeding studies or human clinical trials.
20 One such assay is the quantitative Cellular Antioxidant Activity (CAA) assay, popularized in the
21 last decade by Wolfe and Liu (2007) although originally conceived and reported by Wang and
22 Joseph (1999). The CAA assay is useful for examining the bioavailability of food antioxidants,
23 because it can take several additional factors into consideration including cellular uptake, and
24 metabolism. Originally, the assay used PC12 cells from the adrenal gland of rats; today HepG2
25 cells, derived from human hepatocellular carcinoma, are reported frequently in the literature
26 (Wolfe & Liu, 2007; Wolfe, Kang, He, Dong, Zhang, & Liu, 2008; Song, Derito, Liu, He, Dong,
27 & Liu, 2010). Though the HepG2 cell line is of human origin, and thus an improvement over
28 PC12 cells, liver cells cannot be the ideal cell for measuring the effectiveness of dietary
29 antioxidants. For this reason, there is a more logical move towards the employment of human
30 colorectal adenocarcinoma (Caco-2) cells; they are of similar morphology, marker enzyme,
31 microvillar structure, tight junction and permeability as small intestinal epithelial cells (Wang &
32 Joseph, 1999). These human colonocytes retain their function in culture; hence, Caco-2 cells are
33 frequently used for modeling the intestinal barrier (Sergent et al., 2005; Wan, Dong, Yu, Sun, &
34 Li, 2015).
36 to monitor the inhibition of peroxyl radical-induced oxidation inside the cell. The ester form of
37 the dye (DCFH-DA) is employed, because it is nonionic and non-polar and thus can rapidly
38 transport across the cell membrane (LeBel, Ischiropoulos, & Bondy, 1992; Wang & Joseph,
39 1999). Once inside the cell, the DCFH-DA is deacetylated by endogenous cellular esterases and
40 left in the more oxidizable DCFH form (Figures 1 and 2). A free-radical generator, namely 2,2-
41 azobis (2-amidinopropane) dihydrochloride (ABAP), is added to the system and begins to form
42 peroxyl radicals. Other reactive oxygen species (ROS) are generated endogenously through
43 cellular function (Sies, 1993). If an antioxidant is introduced and able to enter the cell, it can
44 compete with radicals and quench them in a variety of ways, keeping the DCFH from being
45 oxidized to form the fluorescent 2,7-dichlorofluorescein (DCF). The principle of the CAA
46 assay, along with possible pathways of radical quenching can be found in Figure 1. The
47 effectiveness of the sum total of these pathways is evaluated by monitoring the intracellular
48 fluorescence resulting from DCFH over a time course: reduced fluorescence compared to a
49 control (i.e., wells void of antioxidant compounds) indicates a sample has substantial antioxidant
51 The CAA assay is quantitated by the CAA unit (i.e., the relative reduction in fluorescence,
52 expressed in percent), which takes into account relative areas under the curves of sample wells
53 and control wells. The higher the CAA unit, the more effective the antioxidant is in a cellular
54 system. The assay has been used with HepG2 cells for a variety of pure phytochemicals found in
55 food, such as quercetin, gallic acid, and caffeic acid, as well as several common fruit and
56 vegetable extracts literature (Wolfe & Liu, 2007; Wolfe et al., 2008; Song et al., 2010).
57 Commercial products, such as the OxiSelect TM Cellular Antioxidant Assay Kit, are available
58 (Cell Biolabs, 2013). Recently, the assay was modified for analysis in the Caco-2 cell line. Wan
59 et al. (2015) changed cell plating density, probe concentration, and the choice of peroxyl radical
61 As the assay is still relatively new and is thus continually being modified for a variety of
62 applications, standardization has not yet occurred. In fact, there is no universally accepted
63 method for the measurement of antioxidants (Niki, 2010). For these reasons, discrepancies
65 Previously, CAA assays have been performed for fruit and vegetable extracts using HepG2
66 cells, all with quercetin as the standard (Wolfe et al., 2008; Song et al., 2010; Wan et al., 2015).
67 Other foodstuffs, such as tree nuts, are relatively poor in quercetin, but rich in flavan-3-ols like
69 because the intestinal barrier and bioavailability are the main factors affecting the function of
70 antioxidants in vivo, the Caco-2 cell line should serve as a superior model to screen
71 phytochemical-rich foods for potential antioxidant activity. Most food extracts being tested now
72 employ the HepG2 cell line and the quercetin standard curve, which has limitations. In this
73 study, the CAA assay was performed with both HepG2 and Caco-2 cells using both quercetin
74 and (+)-catechin. Thus, the objectives were to assess the CAA assay for the performance of
75 Caco-2 cells in initial stages of antioxidant research and to describe an assay suitable for
76 measuring phenolic antioxidants, such as the flavan-3-ol species abundant in tree nuts.
77
82 (DMSO), (+)-catechin hydrate (98% purity by HPLC), and quercetin dihydrate (98% purity by
83 HPLC) were purchased from the Sigma-Aldrich Chemical Co. (St. Louis, MO, USA).
85 phosphate-buffered saline (PBS), Hanks Balanced Salt Solution (HBSS), fetal bovine serum
89
91 Human hepatocellular carcinoma (HepG2) cells were purchased from the American Type
92 Culture Collection (ATCC, Manassas, VA, USA) and cultured via the method described by
93 Wolfe and Liu (2007). Briefly, cells were cultured in sterile Williams Medium E (WME), along
95 hydrocortisone. Human colon adenocarcinoma (Caco-2) cells were also acquired from the
96 ATCC. Following the method of Xie, Kosiska, Xu, and Andlauer (2013) cells were cultured in
97 Advanced DMEM supplemented with 10% endotoxin-free, heat-inactivated fetal bovine serum
99
101 Stock solutions of quercetin dihydrate (98% purity by HPLC) and (+)-catechin hydrate
102 (98% purity by HPLC) were prepared in DMSO at a concentration of 50 mg/mL. On the day of
103 analysis, samples were diluted to final concentrations ranging from 5 to 1,000 M in serum-free
104 culture media; the final treatment solutions contained <2% DMSO. A 12.5 mM stock solution of
105 the fluorescent probe DCFH-DA in CH3OH was prepared monthly. Before each experiment,
107 ABAP stock solution in HBSS was prepared weekly and diluted to 600 M before experimental
108 use. All DCFH-DA and ABAP solutions were held at -20 C between uses, while samples in
110
112 Cellular antioxidant measurements were made following the method of Wolfe and Liu (2007)
113 with modifications. After HepG2 or Caco-2 cells reached confluence in the Corning 25-cm2
114 culture flasks, cells were twice washed with sterile PBS (pH 7.4), and then dissociated from the
115 surface using 0.05% trypsin-EDTA. Cells were plated (6.0 104) in 100-L cell culture
116 media/well in Corning Costar 96-well, black, flat bottom tissue culture-treated dishes (VWR
117 International, Suwanee, GA, USA) and incubated until confluent (24 to 48 hours). Confluence
118 was confirmed using the microscope. Wells on the perimeter were left empty to reduce any
119 variation due to plate location. Growth medium was removed after confluence was achieved, and
120 the cells were then washed with PBS to remove any non-adherent and dead cells.
121 Next, 50 L of 25 M DCFH-DA working solution were added to each well, followed by 50
122 L of the antioxidant treatments (in triplicate wells). For a control, 50 L of DCFH-DA and 50
123 L of serum-free culture media (no antioxidant included) were applied to triplicate wells. Once
124 the DCFH-DA and antioxidant treatments were added, the cells were placed in the incubator for
125 1 h at 37 C.
126 After this period, the cells were washed 3 with PBS to ensure that any antioxidant effect
127 observed later in the assay was due only to compounds internalized by the cells. One hundred L
128 of the free radical generator, ABAP (600 M), were then added. The cells were immediately
129 placed in a BMG FLUOstar Omega microplate reader (BMG LABTECH Inc., Cary, NC, USA),
130 where real-time fluorescence was read initially and every five minutes then after for 1 h (i.e., 13
131 readings were read in total). Fluorescence was measured at an excitation wavelength of 485 nm
133
135 To test the viability of both HepG2 and Caco-2 cell lines after catechin and quercetin
136 treatments were applied (i.e., at concentrations ranging between 2.5 and 250 M), the
138 employed (Gentile, Allegra, Angileri, Pintaudi, Livrea, & Tesoriere, 2012). DCFH-DA and the
139 phenolics were applied to cells in the same fashion as described for the CAA assay and allowed
140 to incubate for 1 h. Then, the cells were thoroughly washed with PBS to remove all phenolic
141 residues and traces of media. Control wells contained MTT and media, but no phenolics. One
142 hundred microliters of serum-free culture media were applied to the cells along with 10 L of 12
143 mM MTT. After 4 h, 85 L of media were removed and 50 L of DMSO were added to each
144 well to dissolve the generated cellular formazan chromagen. The samples were incubated at 37
145 C once more for 10 min and the absorbance of the purple colored formazan was measured at
146 540 nm. The assay was quantitated by comparing the absorbance of samples against that of the
147 control; treatments resulting in significantly (p < 0.05) lower absorbance readings are considered
149
151 Effectiveness of antioxidant treatments for both cell lines was quantitated by examining the
152 percent reduction in fluorescence. Briefly, curves were generated by the 13 fluorescence
153 response readings of each treatment over the course of the 1 h assay. The area under each curve
154 was calculated via integration with the MARS Data Analysis Software (Version 3.20 R2, BMG
155 LABTECH). As expected, control wells exhibited the maximum fluorescence, as there was no
156 inhibition of the ABAP and DCFH-DA reaction. The percent reduction (or the CAA unit) was
158 CAA units were determined for quercetin and (+)-catechin gathered from triplicate
160
161 3. Results and discussion
163 The CAA assay was performed using HepG2 cells with concentrations of quercetin ranging
164 from 25 to 250 M. The CAA unit, the index of antioxidant activity, increased in a
166 pattern as the effect tapered off at higher antioxidant concentrations. Fluorescence was reduced
167 by up to 58.61.0% when in the presence of 250 M quercetin. Our results are in agreement with
168 the methodology and literature recommended in the OxiSelectTM Cellular Antioxidant Activity
169 Kit (Cell Biolabs, Inc., San Diego, CA, USA); the kit recommends using concentrations of
170 quercetin ranging from 31.3 to 2000 M to develop a standard curve. Hofer, Jrgensen, and
171 Olsen (2014) also reported a half-maximal inhibitory concentration (IC50) value for quercetin as
172 71.238.6 M, which again is similar to the values we report as well as those determined using
173 the OxiSelect kit. These findings, however, were quite different from the work of Wolfe and
174 Liu (2007), who popularized the HepG2 version of the assay; they cited CAA values of up to
175 70.0 units with quercetin concentrations as low as 10 M! For easy comparison, these results
178 foods, (+)-catechin was also analyzed using the traditional CAA assay in HepG2 cells (Figure 4).
179 In contrast to that observed with quercetin, the CAA of catechin was surprisingly rather weak.
180 While 250 M quercetin reduced fluorescence considerably for a CAA value of 58.61.0, the
181 same concentration of catechin resulted in a CAA value of only 16.50.9. Wolfe and Liu
182 examined the structureactivity relationships of pure compounds and their effectiveness in
183 HepG2-based CAA measurements and found that catechin had virtually no antioxidant activity,
184 while quercetin was extremely active (Wolfe & Liu, 2008). Structural characteristics of catechin,
185 such as the ortho-dihydroxy moiety on the B-ring and the C-3 hydroxy group, both increase
187 Michel, & Saran, 1990). As monomeric and polymeric catechin/epicatechin and their derivatives
188 are prevalent in tree nuts, it was determined that the HepG2-based CAA assay would not be
189 useful for gaining biologically relevant information on the antioxidant potential of tree nuts such
190 as U.S. pecans. For this reason further experiments were performed using the Caco-2 cell line.
191 Results for both (+)-catechin and quercetin were quite different in the Caco-2 cell line. As
192 evident in Figure 5, CAA values for catechin were superior in Caco-2 cells: catechin imparted a
193 greater reduction in fluorescence than previously observed in HepG2 cells. Overall in Caco-2
194 cells, fluorescence was reduced by 25.5 to 61.0% across the varying concentrations of catechin
195 tested; whereas in HepG2 cells, the maximum reduction in fluorescence was much less, at only
196 16.5%. Rodriguez-Mateos et al. (2014) reported that the cellular accumulation of ()-epicatechin
197 after a 1 h incubation was more pronounced in Caco-2 monolayer cells (239 pmol/mg protein)
199 The cellular effects of flavonoids will ultimately depend on the extent to which they associate
200 with the cell type in question, either by interactions with membrane or uptake into the cytosol
201 (Rodriguez-Mateos et al., 2014). One possible explanation for the difference in the CAA
202 between the two cell lines is their membrane transport systems (Lodish et al., 2003). As passive
203 diffusion through membranes is similar for all cells, differences in active membrane transport is
204 likely responsible for the observed differences in the CAA. Unfortunately, the role of
205 transporters in both cellular accumulation and cellular efflux has only recently been appreciated
206 and our knowledge for individual phenolic compounds is at a basic level.
207 We have studied the antioxidant activity of phenolic extracts from pecans in Caco-2 cells:
208 they behaved similarly to catechin at a 30 M concentration (unpublished data). As there have
209 been no previous CAA studies using phenolics derived from tree nuts, direct comparison to the
210 literature is not possible. Previous research performed by Wan et al. (2015) aimed to optimize
211 the CAA assay for Caco-2 cells. However, Wan et al. (2015) did not examine the same
212 compounds in multiple cell lines. These authors concluded that Caco-2 cells could be used very
213 effectively for such measurements of biological antioxidant activity, which has been further
215
217 For the MTT viability assay, quercetin and (+)-catechin standards were applied to both cell
218 lines at concentrations ranging between 2.5 and 250 M for 1 h; no statistically significant
219 effects (p > 0.05) between the sample and control cells were observed. From this, we can
220 conclude that the phenolics were not toxic to the cells at any of the concentrations examined in
221 the study and that reduced fluorescence was resultant from antioxidant defenses.
222
223 4. Conclusions
224 In summation, CAA results from Caco-2 cells (known to be good models of the intestinal
225 barrier) were directly compared to those from HepG2 cells. Caco-2 cells yielded more
226 antioxidant activity (shown as substantial reductions in fluorescence) for two common food
227 phytochemicals, as well as being able to be a potentially suitable cell line to determine the
228 antioxidant activity of tree nut extracts. As HepG2 cells were not suitable for flavan-3-ols, such
229 as catechin, the Caco-2 cell model would be more appropriate for assessing foods rich in these
230 phenolics.
231
232 Acknowledgments
233 The authors would like to acknowledge the USDA-NIFA-SCRI Award No. 2011-51181-30674
234 and the Georgia Agricultural Commodity Commission for Pecans (GACCP) for funding this
235 research.
236
239
240 References
241 Bors, W., Heller, W., Michel, C., & Saran, M. (1990). Flavonoids as antioxidants: Determination
243 Cell Biolabs, Inc. (2013). OxiSelectTM Cellular antioxidant activity assay kit (green
244 fluorescence).
245 Craft, B. D., Kerrihard, A. L., Amarowicz, R., & Pegg, R. B. (2012). Phenol-based antioxidants
246 and the in vitro methods used for their assessment. Comprehensive Reviews in Food Science
248 Gentile, C., Allegra, M., Angileri, F., Pintaudi, A. M., Livrea, M. A., & Tesoriere, L. (2012).
249 Polymeric proanthocyanidins from Sicilian pistachio (Pistacia vera L.) nut extract inhibit
253 chelators in two cellular free radical assays: Strong protection by luteolin. Journal of
255 Jones, D. P. (2006). Redefining oxidative stress. Antioxidants & Redox Signaling, 8, 18651879.
256 LeBel, C. P., Ischiropoulos, H., & Bondy, S. C. (1992). Evaluation of the probe 2,7-
257 dichlorofluorescin as an indicator of reactive oxygen species formation and oxidative stress.
259 Lodish, H., Berk, A., Matsudaira, P., Kaiser, C. A., Krieger, M., Scott, M. P., Zipursky, L., &
260 Darnell, J. (2003). Transport of ions and small molecules across cell membranes. In
261 Molecular cell biology, 5th edition (pp. 245-300). W.H. Freeman: New York.
262 Lpez-Alarcn, C., & Denicola, A. (2013). Evaluating the antioxidant capacity of natural
263 products: A review on chemical and cellular-based assays. Analytica Chimica Acta, 763, 1
264 10.
265 Migliavacca, E., Carrupt, P.-A., & Testa, B. (1997). Theoretical parameters to characterize
266 antioxidants. Part 1. The case of vitamin E and analogs. Helvetica Chimica Acta, 80, 1613
267 1626.
268 Niki, E. (2010). Assessment of antioxidant capacity in vitro and in vivo. Free Radical Biology &
270 Prior, R. L., Wu, X., & Schaich, K. (2005). Standardized methods for the determination of
271 antioxidant capacity and phenolics in foods and dietary supplements. Journal of Agricultural
276 Schaich, K. M., Tian, X., & Xie, J. (2015). Hurdles and pitfalls in measuring antioxidant
277 efficacy: A critical evaluation of ABTS, DPPH, and ORAC assays. Journal of Functional
279 Sergent, T., Garsou, S., Schaut, A., De Saeger, S., Pussemier, L., Van Peteghem, C., Larondelle,
280 Y., & Schneider, Y.-J. (2005). Differential modulation of ochratoxin A absorption across
281 Caco-2 cells by dietary polyphenols, used at realistic intestinal concentrations. Toxicology
283 Sies, H. (1993). Strategies of antioxidant defense. European Journal of Biochemistry, 215, 213
284 219.
285 Song, W., Derito, C. M., Liu, M. K., He, X., Dong, M., & Liu, R. H. (2010). Cellular antioxidant
286 activity of common vegetables. Journal of Agricultural and Food Chemistry, 58, 66216629.
287 Wan, H., Dong, L., Yu, X., Sun, H., & Li, Y. (2015). A Caco-2 cell-based quantitative
288 antioxidant activity assay for antioxidants. Food Chemistry, 175, 601608.
289 Wang, H., & Joseph, J. A. (1999). Quantifying cellular oxidative stress by dichlorofluorescein
290 assay using microplate reader. Free Radical Biology & Medicine, 27, 612616.
291 Wardman, P. (2008). Use of the dichlorofluorescein assay to measure reactive oxygen species.
293 Wolfe, K. L., & Liu, R. H. (2008). Structureactivity relationships of flavonoids in the cellular
294 antioxidant activity assay. Journal of Agricultural and Food Chemistry, 56, 84048411.
295 Wolfe, K. L., & Liu, R. H. (2007). Cellular antioxidant activity (CAA) assay for assessing
296 antioxidants, foods, and dietary supplements. Journal of Agricultural and Food Chemistry,
298 Wolfe, K. L., Kang, X., He, X., Dong, M., Zhang, Q., & Liu, R. H. (2008). Cellular antioxidant
299 activity of common fruit. Journal of Agricultural and Food Chemistry, 56, 84188426.
300 Xie, Y., Kosiska, A., Xu, H., & Andlauer, W. (2013). Milk enhances intestinal absorption of
301 green tea catechins in in vitro digestion/Caco-2 cells mode. Food Research International, 53,
302 793800.
303
Figure captions
radicals before they can reach the 2,7-dichlorofluorescin (DCFH) dye; (2) react with
formation of radicals; (3) inhibit lipid peroxidation; (4) react with alkyl peroxy
radicals to stop propagation of other radicals; and (5) inhibit intracellular redox
pathways that can also oxidize the DCFH (Wolfe & Liu, 2007; Wardman, 2008;
DCF, 2,7-dichlorofluorescein.
DCFH-DA and various concentrations of quercetin for 1 h and then washed and
excitation/emission wavelength pair of 485 nm and 538 nm; CAA units represent
cellular antioxidant capacity. Error bars not visible signify values < the symbol size
used. Our results ( ) are compared against those of Wolfe and Liu ( ).10
HepG2 cells. CAA units, a value reflecting cellular antioxidant capacity, were
538 nm. The data represent meanstandard error based on triplicate determinations of
four separate experiments. Error bars not visible signify values < the symbol size
used.
( ) cell types. Error bars not visible signify values < the symbol size used.
Figure 1.
Figure 2.
60
CAA Unit
40
20
0
0 50 100 150 200 250
Quercetin (M)
Figure 3.
60
50
40
CAA Unit
30
20
10
0
0 50 100 150 200 250
Phenolic concentration (M)
Figure 4.
70
60
50
CAA Unit
40
30
20
10
0
0 50 100 150 200 250
(+)-Catechin/Quercetin Concentration (M)
Figure 5.
Highlights
Choosing an appropriate cell line to evaluate the potential of food antioxidants is crucial.
Most cellular antioxidant activity (CAA) assays of foods use a HepG2 cell model, which is
not ideal.
A Caco-2 cell line was found to be superior for the CAA assay when analyzing flavan-3-ols.
Comparing the CAA of quercetin and catechin with both cells revealed Caco-2 to be more
utilitarian.