Evaluation of antioxidant capacity and aroma quality of breast milk

Nutrition 25 (2009) 105–114 www.elsevier.com/locate/nut

Basic nutritional investigation

Evaluation of antioxidant capacity and aroma quality of breast milk Wende Li, Ph.D.a, Farah S. Hosseinian, Ph.D.a, Apollinaire Tsopmo, Ph.D.b, James K. Friel, Ph.D.b, and Trust Beta, Ph.D.a,* b

a Department of Food Science, University of Manitoba, Winnipeg, Manitoba, Canada Department of Human Nutritional Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

Manuscript received February 1, 2008; accepted July 14, 2008.

Abstract

Objective: It is important to understand the difference and similarity in antioxidant capacity and aroma quality between formula and breast milk for purposes of modifying infant formulas. We evaluated the antioxidant properties and aroma quality of infant formula and breast milk. Methods: Six breast milk samples and four infant formulas were used. Antioxidant properties were measured using the following methods: 2,2-diphenyl-1-picryhydrazyl free radical scavenging capacity, oxygen radical absorbance capacity, total phenolic content, and phenolic composition. Aroma quality was determined using the electronic nose. Results: The 2,2-diphenyl-1-picryhydrazyl free radical scavenging activity for formula and breast milk ranged from 45.3% to 61.8% and from 52.8% to 61.2%, respectively. Oxygen radical absorbance capacity ranged from 28.8 to 31.9 g/kg for formula and from 25.5 to 39.2 g/kg for breast milk. Total phenolic content ranged from 422 to 751 mg/kg and from 329to 797 mg/kg for formula and milk, respectively. p-Hydroxybenzoic acid, p-coumaric acid, and ferulic acid were detected with values ranging from 614 to 635, 1391 to 1444, and 1425 to 1490 ␮g/kg in breast milk and from 783 to 3594, 1449 to 1510, and 1447 to 1561 ␮g/kg in formulas. Electronic nose results indicated that the aroma quality of formula controls 2, 3, and 4 was similar to that of breast milk. Conclusion: Differences and similarities in antioxidant properties and aroma quality were found among some of the formulas and breast milk. The contribution of phenolic acids to total antioxidant capacity was limited. © 2009 Elsevier Inc. All rights reserved.

Keywords:

Breast milk; Infant formula; Antioxidant capacity; Total phenolic content; Oxygen radical absorbance capacity; Phenolic acids; Aroma quality

Introduction Breast milk (BM) is an ideal and natural food for newborn babies [1]. It contains all the nutrients necessary for newborn infants [2] and supplies a number of defense factors for growing infants [3]. Strong evidence indicates that reactive oxygen species and other free radicals play an important role in many degenerative diseases such as cancer, atherosclerosis, and diabetes [4]. To prevent or reduce oxidative damage to various body tissues, breast-feeding is important for supplementing food with antioxidants [5]. This work was funded by the Canada Foundation for Innovation New Opportunities Fund, Advanced Food Materials Network, and Canada Research Chairs Program. * Corresponding author. Tel.: ⫹204-474-8214; fax: ⫹204-474-7630. E-mail address: [email protected] (T. Beta). 0899-9007/09/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.nut.2008.07.017

Endogenous antioxidants have been reported in BM. These antioxidants are divided into enzyme antioxidants and nonenzyme antioxidants. Enzyme antioxidants include glutathione peroxidase, catalase, superoxide dismutase [6], and coenzyme Q10 [7–9]. Non-enzyme antioxidants are vitamin E [10,11], ␥-tocopherol [9], retinol [12], vitamin C [13,14], ␤-carotene [10], isoflavones [15], selenium [16], protein antioxidants such as thioredoxin [17] and caseins [18], and protein hydrolysate peptides [19 –21]. Coenzyme Q10 is one example of a lipophilic antioxidant [8]. Levels of coenzyme Q10 and ␣- and ␥-tocopherols in human milk directly correlate with the antioxidant capacity of the milk. Thioredoxin is a redox-regulating protein and high levels of thioredoxin in newborns may provide a unique protective mechanism against oxidative stress during the fetal–neonatal transition [17]. Milk caseins may inhibit lipid peroxidation,

106

W. Li et al. / Nutrition 25 (2009) 105–114

possibly through the mechanism of increasing iron autoxidation [18]. Peptides from casein protein hydrolysate have been shown to possess superoxide anion scavenging activity [19]. Vitamin C in BM helps in the maintenance of a natural barrier against infection, stimulates leukocytes for their phagocytic and antimicrobial activities, augments antibody production, and enhances synthesis of interferon [22]. The antioxidant potential of isoflavones, daidzein, and genistein in BM has been shown in the prevention of some cancers, possibly through multiple effects connected with the inhibition of carcinogenesis [23]. Other phenolics in BM have not been studied. Although it was reported that breast-fed children may have higher lipid peroxidation levels than do formula-fed children in the presence of similar antioxidant capacity [24], in general, reports have indicated that BM provides better antioxidant protection than do infant formulas [25,26]. Because of recent modifications to infant formulas, such as adding long-chain polyunsaturated fatty acids and nucleotides and fortifying with selenium, there are formulas available with compositions and functions closer to those of human milk [16]. The aim of this study was to evaluate antioxidant properties of BM and formula extracts using 2,2-diphenyl-1-picryhydrazyl free radical (DPPH●), oxygen radical absorbance capacity (ORAC), and total phenolic content (TPC) methods and to analyze phenolic acid compositions and aroma quality of BM and formulas using high-performance liquid chromatography and electronic nose (e-nose), respectively, to provide potentially useful data for developing new types of formulas.

Materials and methods Fresh pooled BM samples were provided by six mothers. The information regarding BM nature, such as baby birth date, BM type, and time and place of collection of BM, is presented in Table 1. The samples (BM1 to BM6) were stored at ⫺80°C (Forma ⫺86°C, ULT Freezer, Thermo Electron Corporation, Calgary, Alberta, Canada) until analysis. Four infant formulas used as controls (FC) included FC1 from Nestle for 6- to 18-mo-old infants, FC2 from Enfalac (Mead Johnson) for 6- to 18-mo-old infants, FC3 from Enfalac for newborn infants, and FC4 from Enfalac for

0- to 12-mo-old infants. DPPH●, 2,2=-azobis (2-methylpropionamide) dihydrochloride (AAPH), and 16 phenolic acid standards (gallic, gentisic, p-coumaric, m-coumaric, caffeic, sinapic, ferulic, syringic, o-coumaric, vanillic, protocatechuic, chlorogenic, isoferulic, ellagic, trans-cinnamic, and p-hydroxybenzoic acids) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Trolox and fluorescein were purchased from Fisher Acros Organics (Morris Plains, NJ, USA) for use with the ORAC assay. All other chemicals and solvents were of the highest commercial grade and used without further purification. The DPPH● and ORAC methods were selected for evaluating the antioxidant capacity of the FC and BM samples. Their total phenolic content and phenolic acid composition were determined by Folin and high-performance liquid chromatographic (HPLC) methods, respectively. Their aroma quality was analyzed using the e-nose. Sample preparation The BM samples from individual mothers were collected at different times and combined to form a pooled sample that was mixed and then freeze-dried before use. All infant formulas were powdered products. Formula and freezedried BM samples were preserved at ⫺20°C for further use. Extraction of antioxidative compounds in milk and formula was carried out according to methods previously described [27], but with some modifications. One normal solution of HC1 (1N)/95% ethanol (v/v, 15/85) was used for extraction. The extraction procedure involved addition of 20 mL of solvent to 3.0 g of powder each in 50-mL brown bottles and shaking the powder for 1 h at 30°C in a rotary shaker (MaxQ 5000, BI Barnstead/Lab-Line, Dubuque, IA, USA) set at 300 rpm. The mixture of solvent and powder was then centrifuged at 7800 ⫻ g (SS-34 Rotors, RC5C Sorvall Instruments, DuPont, Wilmington, DE, USA) at 5°C for 15 min. The supernatant fluids were kept at ⫺20°C in the dark until further analysis for DPPH● scavenging activity, ORAC, and TPC. Extraction was carried out in duplicate. Alkaline hydrolysates of BM and formula were prepared according to the methods of Adom and Liu [28], Nardini and Ghiselli [29], and Li et al. [30], but with some modifications. Briefly, samples (2 g) were hydrolyzed using 2 mol/L of NaOH (60 mL) containing ascorbic acid (1% w/v)

Table 1 Relative nature of BM Name

Baby birth date

BM collected place

BM collected time

BM type

BM1 BM2 BM3 BM4 BM5 BM6

August 7, 2005 January 17, 2006 May 20, 2006 NA May 16, 2006 October 8, 2006

Guelph, Ontario Oshawa, Ontario Winnipeg, Manitoba Winnipeg, Manitoba Toronto, Ontario Winnipeg, Manitoba

NA NA May 16–25, 2006 May 24–29, 2006 September 11–October 19, 2006 October 23–30, 2006

ND ND Colostrum ND Mature milk Transition

BM, breast milk; NA, not available; ND, not determined

W. Li et al. / Nutrition 25 (2009) 105–114

and ethylene-diaminetetra-acetic acid (10 mmol/L) for 2 h at 25°C in a rotary shaker (MaxQ 5000, BI Barnstead/LabLine, Dubuque, IA, USA) set at 280 rpm. The hydrolyzed mixture was adjusted to pH 1.5–2.5 using 6 mol/L of icecold HCl and then centrifuged at 7800 ⫻ g (RC5C, Sorvall Instruments, DuPont, Wilmington, DE, USA) at 5°C for 20 min. The supernatant was extracted two to three times with hexane to remove lipids. The final solution was extracted three times with ethyl acetate. The combined ethyl acetate fraction was first dehydrated by adding 2 g of anhydrous Na2SO4 and then evaporated to dryness at 35°C with a ¨ CHI rotary vacuum evaporator (RotaVapor R-205, BU Labortechnik AG, Postfach, Flawil, Switzerland). The residue was dissolved in 5 mL of 50% methanol, and then the solution was re-dried with nitrogen in a 10-mL tube. The residue was redissolved in 1 mL of 50% methanol and filtered through a 0.45-␮m nylon filter. The filtrate (alkaline hydrolysate) was stored in the dark at ⫺20°C and subsequently analyzed by high-performance liquid chromatography to obtain the phenolic acid composition of the samples. Hydrolysis of each sample was carried out in duplicate. DPPH● scavenging activity analysis The DPPH method was used according to Brand-Williams et al. [31] and Li el al. [32], with some modification. Briefly, a 60-␮mol/L DPPH● solution was freshly made in 95% ethanol solution. The milk extract (200 ␮L) was reacted with 3.8 mL of the DPPH● solution for 60 min. The absorbance (A) at 515 nm was measured using a spectrophotometer (Gary 50 BIO UV-Visible Spectrophotometer, Varian, Australia) against a blank of 95% ethanol at t ⫽ 0, 5, 10, 20, 30, 40, 50, and 60 min. The chemical kinetics of antioxidant activity of BM and formula was also recorded. Antioxidant activity was calculated as follows: % DPPH ● scavenging activity ⫽ (1 ⫺ [Asample t ⁄ Acontrol t⫽0]) ⫻ 100 All DPPH tests were carried out in duplicate. ORAC assay The ORAC assay, first developed by Cao et al. [33], was used in this study according to Huang et al. [34], Dávalos et al. [35], and Li et al. [32]. An FLx800 microplate fluorescence reader (Bio-Tek Instruments, Inc., Winooski, VT, USA) equipped with KC4 3.0 software (version 29) was used with fluorescence filters for an excitation wavelength of 485/20 nm and an emission wavelength of 528/20 nm. Trolox concentrations ranging from 6.25 to 50 ␮mol/L were used for the calibration curve. Fluorescein (0.082 ␮mol/L) and AAPH (0.15 mol/L) also were prepared immediately before automatic transferring. All reagents and samples were prepared in 75 mmol/L of potassium phosphate buffer, pH 7.4. Rutin (20 ␮mol/L) was used as a control. Three

107

hundred microliters of buffer solution (blank), diluted sample solutions, rutin, and standards were manually transferred in duplicate to a 96-well flat bottom polystyrene microplate (Corning Incorporated, Corning, NY, USA). A full automation of plate-to-plate liquid transfer was programmed using a Precision 2000 automated microplate pipetting system (Bio-Tek Instruments, Inc., Winooski, VT, USA) as previously described [32]. Peroxyl radical was generated by AAPH during measurement, and fluorescein was used as the substrate [36]. All reaction mixtures were measured in duplicate, and at least three independent assays were performed for each sample. Data were processed according to Cao and Prior [37] and Huang et al. [34]. The net area under the curve was obtained by subtracting the area under the curve of the blank from that of the sample. ORAC values were expressed as Trolox equivalents using the standard curve. Determination of TPC The TPCs of extracts of BM and formula were determined using the Folin-Ciocalteau method [38,39], with some modifications. A sample (200 ␮L) was added to 1.9 mL of freshly diluted 10-fold Folin-Ciocalteau reagent (BDH Inc., Toronto, ON, Canada). Sodium carbonate solution (1.9 mL, or 60 g/L) was then added to the mixture. After 120 min of reaction at ambient temperature, the absorbance of the mixture was measured at 725 nm against a blank of distilled water. Ferulic acid was used as a standard and results expressed as ferulic acid equivalents. All analyses were performed in duplicate. Determination of phenolic acid composition Phenolic acid composition of milk hydrolysates was determined using an HPLC method [40], with some modification. HPLC analysis was performed on a Waters Alliance 2695 model chromatograph instrument (Waters, Mississauga, ON, Canada) equipped with a Waters 2996 photodiode array detector. Phenolic acids were separated on a reverse-phase Phenomenex C18 column (150 ⫻ 4.6 mm) with a gradient of solvent A (water containing 1% [v/v] formic acid) and solvent B (methanol containing 0.1% [v/v] formic acid) for 72 min at a flow rate of 0.7 mL/min. The solvent gradient was programmed as follows: at 0 min 15% B, 7 min 20% B, 8 to 20 min 15% B, 21 to 33 min 24% B, 34 to 36 min 13% B, 37 to 45 min 20% B, 46 to 62 min 42% B, 63 to 68 min 100% B, and 69 to 72 min 15% B. Phenolic acids in the eluants were monitored at 270 and 325 nm synchronously. Identification of the phenolic acids was accomplished by comparing the retention times of peaks in samples with those of phenolic acid standards. Highperformance liquid chromatograms of the 16 standards are shown in Figure 2A. The HPLC analyses were carried out in duplicate.

DPPH radical scavenging (%)

108

W. Li et al. / Nutrition 25 (2009) 105–114

60

soft independent model class analogy (SIMCA). PCA is used in qualitative analysis for assessing the discrimination and similarities between samples and groups. SIMCA is used for comparison to ensure good similarity of a new product to a gold standard [44]. The software package for PCA and SIMCA was from Alpha M.O.S. Eleven of the 16 sensors, LY2/LG, LY2/G, LY2/AA, LY2/Gh, LY2/gCT1, LY2/gCT, T30/1, P10/1, P40/1, T70/2, and PA2, were selected for data treatment by PCA and SIMCA done to obtain the best aroma discrimination among samples. Four replicate measurements were made for each sample.

FC1 FC2 FC3 FC4 BM1 BM2 BM3 BM4 BM5 BM6

50 40 30 20 10 0 0

10

20

30

40

50

60

70

Reaction time (min) Fig. 1. Antioxidant activity kinetics of FC and BM extracts using DPPH●. BM, breast milk; DPPH●, 2,2-diphenyl-1-picryhydrazyl free radical; FC, formula control.

Aroma analysis of BM and formula using e-nose The alpha-Fox 3000 SAS e-nose used in this study comprises an array of metal oxide or conducting polymer sensors with a broad and partly overlapping selectivity for the measurement of odor molecules [41]. It is one part of ␣-Prometheus equipment (Alpha M.O.S., Toulouse, France), which combines the SAS module coming from the ␣-Fox instrument and an FMS module coming from ␣-Kronos instrument, equipped with a headspace autosampler (HS100) to the ␣-Prometheus system. The measurement procedure and parameter option of the SAS e-nose were according to Bleibaum et al [42] and Supriyadi et al. [43], with some modifications. Briefly, the sample (1 g) was placed into a 10-mL glass vial and sealed with septa crimped onto the top. The vial was heated for 8 min at 80°C to produce an equilibrium headspace. The vial sample was agitated at 500 rpm during heating. The headspace gas (1000 ␮L) was automatically taken and injected into the e-nose at 1000 ␮L/s using a syringe preheated to 90°C. The equipment was continuously purged with dry air set at 150 mL/min. The acquisition time was 300 s. Data from the SAS e-nose were analyzed using principal component analysis (PCA) and

Statistical analysis Data were subjected to one-way analysis of variance for comparison of means, and significant differences were calculated according to Duncan’s multiple range test at the 5% level. Data were reported as means ⫾ standard deviations. Differences at P ⬍ 0.05 were considered statistically significant. PCA and SIMCA maps acquired from the e-nose were statistical results of all sensor data. Quantitative results were generally expressed on a dry-weight basis for BM and formula.

Results DPPH radical scavenging activity Free radical scavenging activities of BM and FC extracts are presented in Table 2. The DPPH● scavenging activities of FC ranged from 45.3% to 61.8% at 60 min. The range for BM was 52.8% to 61.2% at 60 min. Significant differences in DPPH● scavenging activity were found among some of the FCs, BMs, and between some FCs and BMs, respectively, such as between FC1 and FC2, between BM3 and BM5, and between FC1 and all six BM samples. DPPH● scavenging activities for FC1 (45.3%) and FC4 (61.8%) were the lowest and the highest, respectively among the four

Table 2 DPPH● scavenging activity (at 60 min), ORAC and TPC of BM and FC extracts* Name

DPPH● scavenging (%)

ORAC (equivalent of Trolox, g/kg)

TPC (equivalent of ferulic acid, mg/kg)

FC1 FC2 FC3 FC4 BM1 BM2 BM3 BM4 BM5 BM6

45.3 ⫾ 1.9 51.4 ⫾ 1.1c 60.8 ⫾ 5.3a 61.8 ⫾ 2.0a 57.2 ⫾ 0.2ba 55.5 ⫾ 0.8bc 52.8 ⫾ 1.9bc 57.6 ⫾ 0.8ba 61.2 ⫾ 0.7a 60.7 ⫾ 0.4a

30.1 ⫾ 0.5 31.9 ⫾ 0.5bcd 30.8 ⫾ 0.9cd 28.8 ⫾ 1.1ed 39.2 ⫾ 4.9a 34.5 ⫾ 1.6bc 32.5 ⫾ 1.3bcd 31.4 ⫾ 0.7cd 35.8 ⫾ 0.3ba 25.5 ⫾ 1.4e

422 ⫾ 52.5c 567 ⫾ 12.8b 739 ⫾ 58.5a 751 ⫾ 63.5a 553 ⫾ 18.9b 412 ⫾ 14.6c 797 ⫾ 50.7a 341 ⫾ 17.8c 329 ⫾ 10.2c 331 ⫾ 18.7c

d

cd

BM, breast milk; DPPH●, 2,2-diphenyl-1-picryhydrazyl free radical; FC, formula control; ORAC, oxygen radical absorbance capacity; TPC, total phenolic content * Values are means ⫾ SDs. Values with the same letter are not statistically different at the 5% level (Duncan’s multiple range test).

W. Li et al. / Nutrition 25 (2009) 105–114

FC samples. BM3 (52.8%) and BM5 (61.2%), respectively, had the lowest and highest DPPH● scavenging activities among the six BM samples. DPPH● scavenging activities of all BM samples were slightly higher than that of FC2 (51.4%) and slightly lower than that of FC4 (61.8%). Reaction kinetics of FC and BM antioxidants with DPPH radicals are shown in Figure 1. The similarities and differences among samples are clearly indicated in Figure 1. For instance, DPPH● scavenging activities of BM5 and BM6 were 27.6% and 31.8% at 5 min and 50.3% and 50.9% at 30 min, respectively. Oxygen radical absorbance capacity The ORAC values of FC and BM extracts are listed in Table 2. The ORAC values of FC ranged from 28.8 (FC4) to 31.9 g/kg (FC2) and no significant differences were found among the FC samples. The ORAC values of the six BM samples ranged from 25.5 (BM6) to 39.2 g/kg (BM1). There were significant differences in ORAC values among some of the BM samples, such as between BM6 and all other BM samples and between BM1 and BM4. A significant difference in ORAC values was also found among some of the FC and BM samples, such as between BM1 and all FC samples. Total phenolic content The TPCs of FC and BM extracts are presented in Table 2. The TPC of FC ranged from 422 (FC1) to 751 mg/kg (FC4). Significant differences in TPC were found among

109

FC samples, such as between FC1 and FC2 (567 mg/kg), or FC3 (739 mg/kg), or FC4 and between FC2 and FC3 or FC4, respectively. The TPCs of BM ranged from 329 (BM5) to 797 mg/kg (BM3). Significant differences in TPC were found among the six BM samples, such as between BM1 and all other BM samples and between BM3 and BM2, or BM4, or BM5, or BM6, respectively. Significant differences in TPC were also observed among some of FC and BM samples, such as between FC2 and BM3 and between FC1 and BM1 or BM3, respectively. The four FC samples were all within the range (329 –797 mg/kg) of the lowest and highest TPCs of BM samples. The TPC (797 mg/kg) of BM3 was the highest among all FC and BM samples. Phenolic acid composition Phenolic acid composition of FC and BM hydrolysates is shown in Table 3. Three types of phenolic acids were detected in FC (Fig. 2B) and BM (Fig. 2C) samples. The phenolic acids included p-hydroxybenzoic acid (p-HA), pcoumaric acid (p-CA), and ferulic acid (FA). The levels of p-HA, p-CA, and FA ranged from 783 to 3549 ␮g/kg, 1449 to 1652 ␮g/kg, and 1447 to 1561 ␮g/kg in FC samples and from 614 to 635 ␮g/kg, 1391 to 1444 ␮g/kg, and 1425 to 1490 ␮g/kg in BM samples, respectively. p-HA (3549 ␮g/ kg) in FC1 was significantly high when compared with that of other FC and all the BM samples (Table 3). Significant differences in p-HA and p-CA were found among some FC samples. There were also significant differences in p-HA

Fig. 2. High-performance liquid chromatographic profile of standards and sample hydrolysate. (A) High-performance liquid chromatographic profiles for the 16 standard phenolic acids, (B) hydrolysate of formula control 2, and (C) the hydrolysate of breast milk 1. (A–C) Numbers indicate the following standard chemicals: 1, gallic acid; 2, protocatechuic acid; 3, p-hydroxybenzoic acid; 4, gentisic acid; 5, chlorogenic acid; 6, vanillic acid; 7, caffeic acid; 8, syringic acid; 9, p-coumaric acid; 10, ferulic acid; 11, m-coumaric acid; 12, sinapic acid; 13, isoferulic acid; 14, o-coumaric acid; 15, ellagic acid; 16, trans-cinnamic acid. (B, C) Isoferulic acid was used as an internal standard. AU, arbitrary unit.

110

W. Li et al. / Nutrition 25 (2009) 105–114

and p-CA between some FC and BM samples. p-HA levels in BM samples were lower than in FC samples. Aroma quality A PCA map of FC and BM samples from the analysis of the ␣-Fox 3000 SAS e-nose is shown in Figure 3. The discrimination index was 89 and the principal components, C1 on the x axis and C2 on y axis, were 99.69% and 0.25%, respectively, when LY2/LG, LY2/G, LY2/AA, LY2/Gh, LY2/gCT1, LY2/gCT, T30/1, P10/1, P40/1, T70/2, and PA2 sensors were selected. Based on the PCA map, all FC and BM samples were clearly divided into two groups on the x axis, with one group including only FC1 and another group including all other samples (Fig. 3). Different groups on the x axis exhibited differences in aroma among the samples, whereas those in the same group on the x axis indicated similarity in aroma quality. Slight differences in aroma among samples in the same group are indicated through different distances on the y axis. Wide distances listed in Table 4 indicated large differences in aroma between samples. The similarity in distances between samples was divided into three groups for all samples (Table 4). The distance of the first group for FC1 to all other samples ranged from 1.602 to 1.789. It was wider in comparison with the second group (range 0.102– 0.200) and the third group (range 0.011– 0.098). The difference in aroma between FC1 and all other samples was confirmed through these distances. SIMCA maps of FC and BM samples are shown in Figures 4A and 4B, respectively. When FC2, FC3, and FC4 were used as “gold standards” for predicting good or bad scores of other samples, the results indicated that the scores of the six BM samples and FC1 were good and bad, respectively (Fig. 4A). The SIMCA map further confirmed the similarity in aroma among the FC2, FC3, FC4, and BM samples. However, when BM samples were used as gold standards for predicting good or bad scores of FC samples, the results indicated that the scores of the FC samples were all bad (Fig. 4B). Table 3 Phenolic acid composition of FC and BM* Name

p-HA (␮g/kg)

p-CA (␮g/kg)

FA (␮g/kg)

FC1 FC2 FC3 FC4 BM1 BM2 BM3 BM4 BM5 BM6

3594 ⫾ 263 1114 ⫾ 15b 890 ⫾ 84c 783 ⫾ 86dc 615⫾18d 615 ⫾ 25d 627 ⫾ 11d 614 ⫾ 10d 617 ⫾ 23d 635 ⫾ 13d

1510 ⫾ 69 1652 ⫾ 87a 1485 ⫾ 50cb 1449 ⫾ 63cb 1429 ⫾ 14cb 1444 ⫾ 18cb 1417 ⫾ 23cb 1391 ⫾ 15c 1423 ⫾ 20cb 1422 ⫾ 18cb

1532 ⫾ 42a 1561 ⫾ 53a 1546 ⫾ 60a 1447 ⫾ 64a 1464 ⫾ 57a 1490 ⫾ 57a 1451 ⫾ 67a 1425 ⫾ 32a 1438 ⫾ 39a 1455 ⫾ 74a

a

b

BM, breast milk; FA, ferulic acid; FC, formula control; p-CA, p-coumaric acid; p-HA, p-hydroxybenzoic acid * Values are means ⫾ SDs. Values with the same letter are not statistically different at the 5% level (Duncan’s multiple range test).

Fig. 3. PCA map of BM and FC samples using the electronic nose. The x axis separates different quality of samples, and the y axis shows differences between samples of similar quality using PCA. BM, breast milk; FC, formula control; PCA, principal component analysis.

Discussion Reactive oxygen species produced during metabolic and physiologic processes and harmful oxidative reaction possibly occurring in organisms could be removed or restrained through enzymatic and non-enzymatic antioxidative mechanisms [45]. It was reported that oxidative stress was significantly lower in breast-fed infants than in formula-fed infants [26], and human milk had better antioxidant protection than did formula [25]. However, results from Granot et al. [24] indicated that BM-fed infants exhibited increased peroxidative injury in the presence of a similar antioxidant capacity in comparison with formula-fed infants, likely due to differences in amount and composition of dietary fatty acids between human milk and cow milk’s–modified formula. Significant differences and similarities in antioxidant properties were found among FC samples, among BM samples, and between FC and BM samples in our in vitro antioxidative capacity evaluation (Table 2). The differences in antioxidant properties appearing between and BM should be considered if the main objective is to determine antioxidants responsible for the action. For instance, human milk contains more long-chain polyunsaturated fatty acids (C20 –22), which are highly susceptible to lipid peroxidation and associated with enhanced oxidative injury when compared with formula. If retardation of lipid peroxidation is the main objective, antioxidants may protect long-chain polyunsaturated fatty acids in human milk against oxidation. The nutrient and antioxidant compositions of BM are affected by many factors, such as different geographic areas for the same ethnic group and different ethnic groups in the same area [46], smoking [13], passive smoking [26], and cold storage condition [47]. It was found that human milk from Nigerian women was significantly higher in antioxidant content than that from Nepal women, and among the Nigerian populations,

W. Li et al. / Nutrition 25 (2009) 105–114

111

Table 4 Similarity distance from the electronic nose* Name–name

S-distance

Name–name

S-distance

Name–name

S-distance

BM5–FC1 BM4–FC1 BM6–FC1 BM3–FC1 BM2–FC1 FC4–FC1 BM1–FC1 FC2–FC1 FC3–FC1

1.789 1.783 1.769 1.753 1.735 1.704 1.689 1.608 1.602

BM5–FC3 BM5–FC2 BM4–FC3 BM4–FC2 BM6–FC3 BM6–FC2 BM3–FC3 BM3–FC2 BM2–FC3 BM2–FC2 BM1–BM5 BM1–BM4 FC3–FC4 BM5–FC4 FC2–FC4

0.200 0.198 0.193 0.191 0.176 0.174 0.160 0.157 0.140 0.137 0.114 0.105 0.105 0.103 0.102

BM4–FC4 BM1–FC3 BM1–FC2 BM1–BM6 BM6–FC4 BM1–BM3 BM3–FC4 BM2–BM5 BM2–BM4 BM2–FC4 BM1–FC4 BM1–BM2 BM3–BM5 BM2–BM6 BM3–BM4 BM5–BM6 BM2–BM3 BM4–BM6 BM3–BM6 FC2–FC3 BM4–BM5

0.098 0.095 0.092 0.088 0.081 0.071 0.070 0.065 0.056 0.054 0.053 0.050 0.044 0.038 0.035 0.028 0.021 0.019 0.018 0.012 0.011

BM, breast milk; FC, formula control; S-distance, similarity distance

the human milk from Fulani women contained significantly lower antioxidant capacity than that from the other ethnic groups of their country [46]. Women who smoke have lower vitamin C levels in transition milk and mature milk than those who do not [13]. Antioxidants are decreased in passive-smoker infants and their mothers when compared with those of non-smokers [26]. Antioxidant activity of human milk was significantly decreased during refrigeration and the freezing process, and refrigeration was better than freezing and thawing to preserve antioxidant activity. Infant health is closely related to the nutritional components, antioxidant level, and quality of the mother’s milk. Known antioxidants in BM are superoxide dismutase [6], coenzyme Q10 [7], vitamin E [11] and vitamin C [13], ␤-carotene [10], selenium [16], thioredoxin [17], isoflavones [15] etc. To our knowledge, phenolic acid composition in BM and formula has not been reported. Phenolic compounds are the main contribution to the in vitro total antioxidant capacity of grains, such as when using methanol extraction, with TPCs of 1108 mg/kg for Chinese black-grained wheat whole meal [30] and 910 mg/kg for high amylose corn whole meal [32]. The TPC range of BMs and formulas was 329 to 797 mg/kg, lower when compared with that of wheat and corn whole meal. Although three types of phenolic acids were found in the FC and BM samples (Table 3), low levels of p-HA, p-CA, and FA indicated that the contribution of phenolic acids to total antioxidant capacity was limited. Standard phenolic acids decreased in antioxidant capacity in the following order: protocatechuic acid ⬎ chlorogenic acid ⬎ caffeic acid ⬎ p-HA ⬎ gentisic acid ⬎ FA ⬎

vanillic acid ⬎ syringic acid ⬎ p-CA [48]. Before FC and BM alkaline hydrolysis, polyphenolic compounds should be the main form of phenolic compounds, similar to bound phenolic complexes that appear in plants [29]. We found a low recovery (61–74%) of internal standard isoferulic acid after FC and BM hydrolysis, probably due to the high content of proteins that easily form gel aggregates and to the application of hexane for lipid removal that led to formation of precipitates [23]. Although the protein in hydrolysate was mainly removed by centrifugation, the gel precipitates still appeared when lipid was being removed using hexane. The levels of phenolic acids from HPLC determination are likely lower than the actual values in the FC and BM samples because of their low recovery. The biological actions of dietary plant phenolic compounds are mainly in the prevention of DNA singlestrand breakage and cytotoxicity through the role of metal chelation [49], free radical scavenging through inhibiting reactive oxygen species [50] reducing ␣-tocopheroxyl radical activity [51], and modulation of enzymatic activity [52,53]. Phenolic compounds also have antitumor effects [54], anti-inflammatory activity through inhibition of transcription factor nuclear factor-␬B [55], and inhibitory effects on nuclear transcription factor’s activation [56] and extracellular signal-regulated protein kinase 1/2 (ERK1/2) activation [57]. Phenolic compounds are able to play a putative role in preventing some diseases, such as their anti– colon cancer potential [58]. Because of the importance of phenolic compounds in health promotion, consumption of a diet rich in phenolic compounds for preventing some diseases in newborn infants is likely an effec-

112

W. Li et al. / Nutrition 25 (2009) 105–114

infant formulas with similar aroma quality as BM and for monitoring their quality. The obvious differences in aroma quality between FC1 and other FC or BM samples had a possible relation with a high p-HA content in FC1. A possible effect of phenolic acids on aroma quality of infant formulas needs to be further confirmed. In conclusion, our study reported the antioxidant properties and aroma quality of FC and BM samples. Some differences and similarities in antioxidant properties and in aroma quality were found between FC and BM samples, among FC samples, and among BM samples. Evaluation regarding the antioxidant and aroma properties of FC and BM provided helpful data for developing novel infant formulas. Although three types of phenolic acids were found, their contribution to the total antioxidant capacity in the FC and BM samples was limited because of minor levels. Existing antioxidants in the BM samples would have undergone biotransformation after they had been absorbed from the diet and likely would possess high bioavailability for the infants. Although the phenolic levels found in this study were low, they are likely useful for preventing some infant diseases. Future work will include research on methods to further increase BM antioxidant capacity and improving BM quality through consumption of rich antioxidant diets. Fig. 4. Soft independent model class analogy map of BM and FC samples using the electronic nose. (A) FC2, FC3, and FC4 were used as good criteria (training scores) for predicting good or bad scores of FC1, BM1, BM2, BM3, BM4, BM5, and BM6. BM1, BM2, BM3, BM4, BM5, and BM6 were good (inside shadow section), and FC1 was poor (outside shadow section). (B) BM1, BM2, BM3, BM4, BM5, and BM6 were used as good criteria (training scores) for predicting good or bad scores of FC1, FC2, FC3, and FC4. FC1, FC2, FC3, and FC4 were poor (outside shadow section). BM, breast milk; FC, formula control.

tive way to increase the levels of phenolic compounds in BM. Aroma is an important quality for many food products in addition to their nutritional and antioxidant properties. Drastic changes in aroma properties of the same type of food will likely influence consumer acceptability of the product. The e-nose technology is rapid and simple for overall comparative analysis of volatile compounds [59]. It is important for quality control and quality assurance and can be used to track the quality desired by consumers [42]. Because of a sensor’s difference in sensitivity for volatile compounds in different samples, sensitive sensors were selected for the optimization of PCA and SIMCA analyses. Eleven sensors selected from a total of 16 sensors were sensitive to volatile compounds in the FC and BM samples. The PCA and SIMCA maps acquired using the e-nose were statistical results of the comparison of FC and BM aroma data obtained from the 11 sensors. After gold standards were set, the differences and similarities in aroma could be evaluated for the FC and BM samples through SIMCA analysis of e-nose technology. E-nose is a useful tool for developing new

Acknowledgments The authors are thankful for the technical assistance from Wan Yuin Ser of the Department of Food Science and Haifeng Yang of the Department of Human Nutritional Sciences at the University of Manitoba.

References [1] Oddy WH. The impact of breast milk on infant and child health. Breastfeed Rev 2002;10:5–18. [2] Gossage CP, Deyhim M, Yamini S, Douglass LW, Moser-Veillon PB. Carotenoid composition of human milk during the first month postpartum and the response to b-carotene supplementation. Am J Clin Nutr 2002;76:193–7. [3] Granot E, Kohen R. Oxidative stress in childhood—in health and disease states. Clin Nutr 2004;23:3–11. [4] Beckman KB, Ames BN. The free radical theory of aging matures. Physiol Rev 1998;78:547– 81. [5] Oxman T, Shapira M, Diver A, Klein R, Avazov N, Rabinowitz B. A new method of long-term preventive cardioprotection using Lactobacillus. Am J Physiol Heart Circ Physiol 2000;278:H1717–24. [6] L’Abbe M, Friel JK. Superoxide dismutase and glutathione peroxidase content of human milk from mothers of premature and full term infants during the first 3 months of lactation. J Pediatr Gastroenterol Nutr 2000;31:270 – 4. [7] Compagnoni G, Giuffre B, Lista G, Mosca F, Marini A. CoQ10 plasmatic levels in breast-fed infants compared to formula-fed infants. Biol Neonate 2004;86:165–9. [8] Niklowitz P, Menke T, Giffei J, Andler W. Coenzyme Q10 in maternal plasma and milk throughout early lactation. Biofactors 2005; 25:67–72.

W. Li et al. / Nutrition 25 (2009) 105–114 [9] Quiles JL, Ochoa JJ, Ramirez-Tortosa MC, Linde J, Bompadre S, Battino M, et al. Coenzyme Q concentration and total antioxidant capacity of human milk at different stage of lactation in mothers of preterm and full-term infants. Free Radic Res 2006;40:199 –206. [10] Ostrea EMJR, Balun JE, Winkler R, Porter T. Influence of breastfeeding on the restoration of the low serum concentration of vitamin E and beta-carotene in the newborn infant. Am J Obstet Gynecol 1986;154:1014 –7. [11] Korchazhkina O, Jones E, Czauderna M, Spencer SA. Effects of exclusive formula or breast milk feeding on oxidative stress in healthy preterm infants. Arch Dis Child 2006;91:327–9. [12] Lindeman JHN, van Zoeren-Grobben D, Schrijver J. The total free radical trapping ability of cord blood plasma in preterm and term babies. Pediatr Res 1989;26:20 – 4. [13] Ortega RM, Lopes-Sobaler AM, Quintas ME, Martinez RM, Andres P. The influence of smoking on vitamin C status during the third trimester of pregnancy and on vitamin C levels in maternal milk. J Am Coll Nutr 1998;17:379 – 84. [14] Hoppu U, Rinne M, Salo-Vaananen P, Lampi AM, Piironen V, Isolauri E. Vitamin C in breast milk may reduce the risk of atopy in the infant. Eur J Clin Nutr 2005;59:123– 8. [15] Franke AA, Custer LJ, Tanaka Y. Isofavones in human breast milk and other biological fluids. Am J Clin Nutr 1998;68(suppl):1466S–73. [16] Carver JD. Advances in nutritional modifications of infant formulas. Am J Clin Nutr 2003;77:1550S– 4. [17] Todoroki Y, Tsukahara H, Ohshima Y, Shukunami KI, Nishijima K, Kotsuji F, et al. Concentration of thioredoxin, a redox-regulating protein, in umbilical cord and breast milk. Free Radic Res 2005;39: 291–7. [18] Cervato G, Cazzola R, Cestaro B. Studies on the antioxidant activity of milk caseins. Int J Food Sci Nutr 1999;50:291– 6. [19] Suetsusa K, Ukeda H, Ochi H. Isolation and characterization of free radical scavenging activities peptides derived from casein. J Nutr Biochem 2000;11:128 –31. [20] Peña-Ramos EA, Xiong YL. Antioxidative activity of whey protein hydrolysates in liposomal system. J Dairy Sci 2001;84:2577– 83. [21] Hernández-Ledesma B, Dávalos A, Bartolomé B, Amigo L. Preparation of antioxidant enzymatic hydrolysates from alpha-lactalbumin and beta-lactoglobulin. Identification of active peptides by HPLCMS/MS. J Agric Food Chem 2005;53:588 –93. [22] Romeu-Nadal M, Morera-Pons S, Castellote AI, López-Sabater MC. Rapid high-performance liquid chromatographic method for vitamin C determination in human milk versus an enzymatic method. J Chromatogr B 2006;830:41– 6. [23] Franke AA, Custer LJ. Daidzein and genistein concentrations in human milk after soy consumption. Clin Chem 1996;42:955– 64. [24] Granot E, Golan D, Rivkin L, Kohen R. Oxidative stress in healthy breast fed versus formula fed infants. Nutr Res 1999;19:869 –79. [25] Friel JK, Martin SM, Langdon M, Herzberg GR, Buettner GR. Milk from mothers of both premature and full-term infants provides better antioxidant protection than does infant formula. Pediatr Res 2002;51: 612– 8. [26] Aycicek A, Erel O, Kocyigit A, Selek S, Demirkol MR. Breast milk provides better antioxidant power than does formula. Nutrition 2006; 22:616 –9. [27] Li W, Pickard MD, Beta T. Effect of thermal processing on antioxidant properties of purple wheat bran. Food Chem 2007;104:1080 – 6. [28] Adom KK, Liu RH. Antioxidant activity of grain. J Agric Food Chem 2002;50:6182–7. [29] Nardini M, Ghiselli A. Determination of free and bound phenolic acids in beer. Food Chem 2004;84:137– 43. [30] Li W, Shan F, Sun S, Corke H, Beta T. Free radical scavenging properties and phenolic content of Chinese black-grained wheat. J Agric Food Chem 2005;53:8533– 6. [31] Brand-Williams W, Cuvelier ME, Berset C. Use of a free radical method to evaluate antioxidant activity. Lebensm Wiss Technol 1995;28:25–30.

113

[32] Li W, Wei CV, White PJ, Beta T. High-amylose corn exhibits better antioxidant activity than typical and waxy genotypes. J Agric Food Chem 2007;55:291– 8. [33] Cao GH, Alessio HM, Cutler RG. Oxygen-radical absorbance capacity assay for antioxidants. Free Radic Biol Med 1993;14:303–11. [34] Huang D, Ou B, Hampsch-Woodill M, Flanagan JA, Prior RL. Highthroughput assay of oxygen absorbance capacity (ORAC) using a multichannel liquid handling system coupled with a microplate fluorescence reader in 96-well format. J Agric Food Chem 2002;50: 4437– 44. [35] Dávalos A, Gómez-Cordovés C, Bartolomé B. Extending applicability of oxygen radical absorbance capacity (ORAC-fluorescein) assay. J Agric Food Chem 2004;52:48 –54. [36] Awika JM, Rooney LW, Wu X, Prior RL, Cisneros-Zevallos L. Screening methods to measure antioxidant activity of sorghum (Sorghum bicolor) and sorghum products. J Agric Food Chem 2003;51: 6657– 62. [37] Cao GH, Prior RL. Measurement of oxygen radical absorbance capacity in biological samples. Oxidants and antioxidants. Methods Enzymol 1999;299:50 – 62. [38] Singleton VL, Rossi JA. Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. Am J Enol Vitic 1965;16: 144 –58. [39] Beta T, Nam S, Dexter JE, Sapirstein HD. Phenolic content and antioxidant activity of pearled wheat and roller mill fractions. Cereal Chem 2005;82:390 –3. [40] Li W, Pickard MD, Beta T. Evaluation of antioxidant activity and electronic taste and aroma properties of antho-beers from purple wheat grain. J Agric Food Chem 2007;55:8958 – 66. [41] Harper WJ. The strengths and weakness of the electronic nose. Adv Exp Med Biol 2001;488:59 –71. [42] Bleibaum RN, Stone H, Tan T, Labreche S, Saint-Martine E, Isz S. Comparison of sensory and consumer results with electronic nose and tongue sensors for apple juices. Food Qual Prefer 2002;13:409 –22. [43] Supriyadi, Shimizu K, Suzuki M, Yoshida K, Muto T, Fujita A, et al. Maturity discrimination of snake fruit (Salacca edulis Reinw.) cv. Pondoh based on volatiles analysis using an electronic nose device equipped with a sensor array and fingerprint mass spectrometry. Flavour Fragr J 2004;19:44 –50. [44] Tan T, Schmitt V, Isz S. Electronic tongue: a new dimension in sensory analysis. Food Technol-Chicago 2001;55(10):44, 46, 48, 50. [45] Halliwell B. Free radicals, antioxidants, and human disease: curiosity, cause, or consequence? Lancet 1994;344:721– 4. [46] VanderJagt DJ, Okolo SN, Costanza A, Blackwell W, Glew RH. Antioxidant content of the milk of Nigerian women and the sera of their exclusively breast-fed infants. Nutr Res 2001;21:121– 8. [47] Hanna N, Ahmed K, Anwar M, Petrova A, Hiatt M, Hegyi T. Effect of storage on breast milk antioxidant activity. Arch Dis Child Fetal Neonatal Ed 2004;89:518 –20. [48] Onyeneho SN, Hettiarachchy NS. Antioxidant activity of durum wheat bran. J Agric Food Chem 1992;40:1496 –500. [49] Garcia-Alonso FJ, Guidarelli A, Periago MJ. Phenolic-rich juice prevents DNA single-strand breakage and cytotoxicity caused by tert-butyldroperoxide in U973 cells: the role of iron chelation. J Nutr Biochem 2007;18:457– 66. [50] Aboul-Enein HY, Kruk I, Kładna A, Lichszteld K, Michalska T. Scavenging effects of phenolic compounds on reactive oxygen species. Biopolymers 2007;86:222–30. [51] Pazos M, Andersen ML, Medina I, Skibsted LH. Efficiency of natural phenolic compounds regenerating ␣-tocopherol from ␣-tocopheroxyl radical. J Agric Food Chem 2007;55:3661– 6. [52] Koshihara Y, Neichi T, Murota S, Lao A, Fujimoto Y, Tatsumo T. Caffeic acid is a selective inhibitor for leukotriene biosynthesis. BBA-Lipid Lipid Met 1984;792:92–7. [53] Sud’ina GF, Mirzoeva OK, Pushkareva MA, Korshunova GA, Sumbatyan NV, Varfolomeev SD. Caffeic acid phenethyl ester as a

114

W. Li et al. / Nutrition 25 (2009) 105–114

lipoxygenase inhibitor with antioxidant properties. FEBS Lett 1993;329:21– 4. [54] Yoshioka K, Deng T, Cavigelli M, Karin M. Antitumor promotion by phenolic antioxidant: inhibition of AP-1 activity through induction of Fra expression. Proc Natl Acad Sci U S A 1995;92:4972– 6. [55] Karamenderes C, Konyalioglu S, Khan S, Khan IA. Total phenolic contents, free radical scavenging activity and inhibitory effects on the activation of NF-kappa B of eight centaurea L. species. Phytother Res 2007;21:488 –91. [56] Natarajan K, Singh S, Burke TR, Grunberger D, Aggarwal BB. Caffeic acid phenethyl ester is a potent and specific inhibitor of

activation of nuclear transcription factor NF-kB. Proc Natl Acad Sci U S A 1996;93:9090 –5. [57] Karhumäki P, Tiitinen SL, Turpeinen H, Parkkinen J. Inhibition of ERK1/2 activation by phenolic antioxidants protects kidney tubular cells during cold storage. Transplantation 2007;83:948 –53. [58] Shoeb M, Jaspars M, MacManns SM, Celik S, Nahar L, Kong-ThooLin P, Sarker SD. Anti-colon cancer potential of phenolic compounds from the aerial parts of Centaurea gigantea (Asteraceae). J Nat Med 2007;61:164 –9. [59] Xu XQ. Electronic nose for characterisation of flavour patterns of deep frying oils. Food Aust 2006;58:89 –91.