Subscriber access provided by University of Winnipeg Library
Food and Beverage Chemistry/Biochemistry
Cholesterol content in human milk during lactation. A comparative study of enzymatic and chromatographic methods ISLAM J.A. HAMDAN, Luis Manuel Sanchez-Siles, Esther Matencio, Guadalupe Garcia Llatas, and María Jesús Lagarda J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b02205 • Publication Date (Web): 01 Jun 2018 Downloaded from http://pubs.acs.org on June 1, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 34
Journal of Agricultural and Food Chemistry
Cholesterol content in human milk during lactation. A comparative study of enzymatic and chromatographic methods
Islam J.A. Hamdan1, Luis Manuel Sanchez-Siles2, Esther Matencio2, Guadalupe Garcia-Llatas1*, María Jesús Lagarda1 1
Nutrition and Food Science Area, Faculty of Pharmacy, University of Valencia,
Avda. Vicente Andrés Estellés s/n, 46100 – Burjassot (Valencia), Spain. 2
R&D Department, Institute of Infant Nutrition, Hero Group. Avda. Murcia 1, 30820
– Alcantarilla (Murcia), Spain. * To whom correspondence should be addressed (Telephone +34-963543766; Fax +34-963544954; E-mail:
[email protected])
Authors’ contribution: Islam J.A. Hamdan contributed with the analytical assays and with data acquisition and interpretation. Guadalupe Garcia-Llatas and María Jesús Lagarda contributed with the conception and design of the study. Guadalupe Garcia-Llatas, María Jesús Lagarda, Luis Manuel Sánchez-Siles and Esther Matencio interpreted the data. All authors collaborated in the drafting and revision of the article, and approved the final version of the manuscript for submission.
1 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
1
Page 2 of 34
Abstract
2
This study validates a gas chromatography (GC) method for determining the sterol
3
profile of human milk (HM) and compares it with an enzymatic-spectrophotometric (E-
4
S) method. Good linearity (r > 0.97) and low limits of detection and quantification were
5
obtained with the GC method (< 1.8 and < 6 µg/100 g of HM, respectively). Suitable
6
intra- and inter-assay precisions (all < 18%) and satisfactory recovery percentages (80-
7
109%) were obtained for both methods. In addition, both methodologies were used to
8
assess cholesterol evolution in HM during lactation, showing a 50% decrease at 6
9
months versus colostrum. The E-S method overestimated cholesterol content by < 20%
10
versus the GC method. The results indicate that both methods may be used by the
11
industry and in research to better understand the differences between the sterol profiles
12
of infant formulas and HM.
13
Keywords
14
Breast
15
milk;
cholesterol;
colostrum;
enzymatic-spectrophotometry;
chromatography; sterols.
2 ACS Paragon Plus Environment
gas
Page 3 of 34
Journal of Agricultural and Food Chemistry
16
INTRODUCTION
17
Human milk (HM) is a very complex biological fluid with a composition that
18
changes with the lactation stage, the time of day, within feedings (more carbohydrates
19
in foremilk and more lipids in hindmilk), delivery or gestation length (preterm or term
20
HM), maternal parity, body mass index (BMI), age, diet and residential area, among
21
other factors1-4. These changes allow HM to adjust to the needs of the infant, ensuring
22
correct growth and development, but also make HM very difficult to mimic. For the
23
aforementioned reasons, HM is considered the gold standard for feeding infants up to 6
24
months of age, and exclusive breastfeeding is recommended for infants up to that age5.
25
In HM, lipids are the most variable macronutrient, being dispersed as fat globules
26
formed by a core (rich in triglycerides) surrounded by a membrane known as the milk
27
fat globule membrane (MFGM). The latter in turn is composed of an inner membrane
28
(rich in phosphatidylethanolamine, phosphatidylserine, phosphatidylinositol and
29
cholesterol) and an outer membrane (rich in phosphatidylcholine, sphingomyelin,
30
cholesterol, peptides, glycoproteins, cerebrosides and gangliosides, among other
31
compounds)6. The MFGM of HM is a good source of cholesterol, which is crucial to
32
infant growth, since it is essential for the synthesis of bile acids, hormones, vitamin D
33
and lipoproteins7. Furthermore, cholesterol is needed as a structural component of all
34
cell membranes, and especially as a constituent of myelin during the rapid brain growth
35
period8.
36
The cholesterol content in HM decreases in the course of the lactation stages, and in
37
any case is higher than that found in infant formulas (IFs). Thus, different qualitative
38
and quantitative profiles of cholesterol and other sterols (cholesterol precursors and
39
plant sterols) can be found in HM and IFs9-18. This fact causes breast-fed infants to
40
consume up to 5 times more cholesterol than formula-fed infants (18.2 versus 3.4 3 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
41
mg/kg/day at 20 weeks of age and with a mean body weight of 6.8 kg)19.
42
Claumarchirant et al.16 reported similar cholesterol intakes (3.7 mg/kg/day) from IFs in
43
infants of the same age and weight.
44
This increased intake of cholesterol by breast-fed infants has been postulated as one
45
possible reason why serum cholesterol levels are higher in breast-fed infants than in
46
formula-fed infants, while the opposite occurs in relation to plant sterol levels20-21.
47
Several studies have suggested that higher serum cholesterol levels in breast-fed infants
48
could prevent cardiovascular diseases in adult life20,22-27 by downregulating hepatic
49
hydroxymethyl glutaryl coenzyme A reductase via epigenetic modifications28. However,
50
it has been mentioned that further studies are needed to better understand programming
51
in the neonatal period and the role of the other sterols present in HM8,29.
52
Despite the importance these bioactive compounds may have for the infant, few
53
studies have examined the sterol profile (focused mainly on cholesterol and
54
desmosterol) in HM (Table 1), and only three publications involve validated
55
methodology15,30-31. Chromatographic methods have been the most widely used option
56
for the identification and quantification of sterols in HM, particularly gas
57
chromatography (GC), given its advantage of being able to establish the complete sterol
58
profile. However, such methods are time consuming and are not suitable for large
59
numbers of samples. Enzymatic-spectrophotometric (E-S) methods are faster and less
60
expensive, and can also be used for sterol determination - though they are not as specific
61
as GC, since the cholesterol content is overestimated. In this regard, cholesterol oxidase,
62
used in E-S, oxidizes the 3β-hydroxyl group present in cholesterol and other sterols
63
(desmosterol, lathosterol, campesterol, stigmasterol and β-sitosterol). 32-33
64
On the other hand, few studies have compared analytical methods for the
65
determination of cholesterol in HM. Harzer et al.12 and Haug & Harzer33 compared the
4 ACS Paragon Plus Environment
Page 4 of 34
Page 5 of 34
Journal of Agricultural and Food Chemistry
66
cholesterol contents
obtained
by
E-S assay and
high performance
liquid
67
chromatography (HPLC), and described a 20% lower content with the latter technique -
68
this difference being attributed to the non-specificity of E-S. Kamelska et al.34 in turn
69
reported variable contents obtained by attenuated total reflectance-Fourier transform
70
infrared spectroscopy (ATR-FTIR) or GC. Furthermore, Kamelska et al.31 obtained
71
higher values with E-S than with GC when direct saponification was applied to HM.
72
Due to the abovementioned reasons, the two main objectives of the present study
73
were: (a) to validate a GC method for sterol (cholesterol, desmosterol, lathosterol,
74
campesterol, stigmasterol, lanosterol, β-sitosterol) determination in HM and compare it
75
with a proposed E-S method; and (b) to assess the evolution of cholesterol content in
76
HM samples from different lactation stages using both techniques. Results obtained in
77
this study would facilitate selection of the analytical technique best suited for
78
cholesterol determination in HM.
79
MATERIALS AND METHODS
80
Reagents
81
Epicoprostanol (5β-cholestan-3α-ol, 96.9% purity) as internal standard (IS),
82
cholesterol (5-cholesten-3β-ol, 99% purity), stigmasterol (5,22-stigmastadien-3β-ol,
83
95% purity), lanosterol (8,24-lanostadien-3β-ol, 99%), lathosterol (5α-cholest-7-en-3β-
84
ol, 99.5%), β-sitosterol (5-stigmasten-3β-ol, 98% purity), N,O-bis-(trimethylsilyl)
85
trifluoroacetamide with 1% trimethylchlorosilane (BSTFA+1% TMCS) and potassium
86
hydroxide (KOH) were purchased from Sigma–Aldrich Chemical (St. Louis, MO,
87
USA).
88
hydroxycholesterol (5-cholestene-3β,7β-diol, 97% purity) and desmosterol (5,24-
89
cholestadien-3β-ol, 98% purity) were purchased from Steraloids Inc. (Newport, RI,
7α-hydroxycholesterol
(5-cholestene-3β,7α-diol,
5 ACS Paragon Plus Environment
98%
purity),
7β-
Journal of Agricultural and Food Chemistry
90
USA). Campesterol (24α-methyl-5-cholesten-3β-ol, 96.2% purity) was obtained from
91
Chengdu Biopurify Phytochemicals Ltd. (Sichuan, China).
92
Hexane and 2-propanol were acquired from Scharlau Chemie, S.A. (Barcelona,
93
Spain). Ethanol absolute grade was purchased from Panreac Química SLU (Barcelona,
94
Spain) and pyridine extra dry (99.5% purity) from Acros Organics (Geel, Belgium). The
95
E-S kit for cholesterol determination was obtained from Boehringer Mannheim / R-
96
Biopharm (R-Biopharm AG, Darmstadt, Germany). Hydrogen (purity >99%, NMH2-
97
500; LNI Schmidlin S.A., Geneva, Switzerland) and nitrogen (N2) (purity >99%, Zefiro
98
35 LC-MS, F-DGS SAS, Evry, France) generators were used.
99
All standards used were of GC grade, and all reagents were of analytical grade.
100
Millipore MilliQ deionized water (Millipore Ibérica S.A., Barcelona, Spain) was used.
101
Samples
102
For the determination of analytical parameters, a pool of mature HM was prepared
103
from milk samples obtained from four healthy Spanish donors aged between 25-35
104
years. The donors were non-smokers, well-nourished and without caloric restrictions.
105
The volunteer mothers had gestations > 37 weeks and infants with normal weight at
106
birth. The study of the evolution of cholesterol content throughout the lactation period
107
was carried out with pools of HM samples obtained from 65 volunteer mothers of term
108
infants between 18-40 years of age, as described by Claumarchirant et al.35 These pools
109
were obtained from different stages of lactation (colostrum, transition, 1, 3 and 6
110
months) and three cities in Spain (Madrid, Murcia and Valencia). The study protocol
111
was approved by the Clinical Research Ethics Committee and Breastfeeding Committee
112
of Puerta de Hierro-Majadahonda University Hospital (Madrid, Spain). All subjects
113
were informed and gave written consent before inclusion in the study.
114
Determination of sterols by GC
6 ACS Paragon Plus Environment
Page 6 of 34
Page 7 of 34
Journal of Agricultural and Food Chemistry
115
Methodology
116
Direct saponification and extraction of the unsaponifiable fraction were carried out
117
for sterol determination. Two µg of epicoprostanol (IS) and 3 ml of an ethanol (90%)
118
solution of 2 N KOH were added to 0.5 ml (~ 0.54 g) of HM and placed in a 10 ml
119
Pyrex glass tube, followed by heating of the mixture at 80ºC during 45 min in a shaking
120
water bath (~100 rpm, Stuart SBS30)36. Then, once the sample had cooled, MilliQ
121
water and hexane were added (1 ml each), and vortexing (1 min) and centrifugation at
122
3200 rpm, 4°C, 10 min (Eppendorf 5810R) were performed. The hexane phase was
123
transferred to a new tube, and extraction with hexane was repeated twice more,
124
collecting all the organic phases. The unsaponifiable fraction was then dried with N2 and
125
subjected to derivatization at 65ºC during 1 h in a thermoblock (Stuart SBH200D) with
126
200 µl of BSTFA+1% TMCS: pyridine (10:3, v/v)17. The trimethylsilyl ether (TMSE)
127
derivatives obtained were dissolved in hexane, filtrated (using a syringe-driven Millex-
128
FH filter unit, 3 ml, 0.45 µm Millipore, Milford, MA, USA), and the solvent was
129
evaporated with N2. Each sample was dissolved in 800 µl of hexane for cholesterol and
130
in 40 µl for the determination of other sterols. From each solution, one microliter was
131
injected into a gas chromatography–mass spectrometry system (GC-MS, Trace GC-
132
Ultra ITQ ion trap 900, Thermo Scientific, Waltham, MA, USA) for identification, and
133
into a GC-flame ionization detector (GC–FID, Autosystem XL, Perkin-Elmer,
134
Nortwalk, CT, USA) for quantification purposes. Both instruments were equipped with
135
CP-Sil 8 low bleed/MS (50 m × 0.25 mm × 0.25 µm film thickness) capillary columns
136
(Chrompack-Varian, Middelburg, The Netherlands).
137
In GC-MS, the carrier gas was hydrogen operating at a constant flow of 1 ml/min.
138
The temperature of the injector port was 280ºC, and splitless injection mode was
139
applied. The oven was initially programmed at 150 °C maintained during 3 min, heated
7 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
140
to 280 °C at a rate of 30 °C/min, and kept during 28 min, then raised to 295 °C at a rate
141
of 10 °C/min. Finally, this temperature was maintained for 10 min. The transfer line
142
temperature was 310 ºC, the ion source temperature was 250 ºC, the scan range was 50-
143
650 m/z, and the filament emission current was -70 eV. Identification of sterols was
144
carried out by comparing their retention times and MS spectra with those of standards
145
and the NIST v.2 library. In GC-FID, hydrogen was also the carrier gas at 0.91 ml/min,
146
the detector port was set at 300 ºC, and the oven and injector conditions were identical
147
to those of GC-MS. Both instrument conditions were slightly modified from Garcia-
148
Llatas et al.37 and Hamdan et al.17
149
Validation
150
Validation was made by studying the analytical parameters of selectivity, linearity,
151
limit of detection (LOD), limit of quantification (LOQ), precision and accuracy,
152
according to the guidelines of the International Conference on Harmonization (ICH) and
153
the Association of Official Analytical Chemists guidelines for the validation of
154
analytical methods38-39.
155
Selectivity and linearity
156
Sterol calibration curves were performed with 0.25 ml of HM and without addition
157
of matrix to check any interference from other compounds present in the HM
158
(selectivity assay). For establishing linearity, a minimum of 5 calibration points were
159
prepared with a constant amount of IS (2 µg), with derivatization using the
160
aforementioned method and analysis by GC-FID. Calibration curves were plotted for
161
each sterol, representing the sterol area/IS area ratio versus their respective quantities
162
(µg sterol/µg IS) (see Table 2).
163
Limit of detection (LOD) and limit of quantification (LOQ)
8 ACS Paragon Plus Environment
Page 8 of 34
Page 9 of 34
Journal of Agricultural and Food Chemistry
164
The LOD and LOQ of the method were determined from the background response of
165
6 method blanks using 0.5 ml of water. They were calculated as three and 10 times (for
166
LOD and LOQ, respectively) the standard deviation (SD) of the response against the
167
slope of the calibration curves for each sterol.
168
Precision
169
Intra- and inter-assay precisions were evaluated by calculating the relative standard
170
deviation (RSD) of the analysis of the sample on two different days.
171
Four aliquots of the sample were analyzed on the same day (intra-assay) and another
172
four aliquots on a different day (inter-assay).
173
Accuracy
174
The accuracy of the method for all sterols was established from recovery assays. The
175
sterol contents of four aliquots of the sample (0.25 ml of HM), and another four spiked
176
aliquots (see Table 3) were determined. The spiked amounts of sterols were also
177
verified in quadruplicate. Recovery (as a percentage) was calculated as: [(µg sterol in
178
spiked aliquots - µg sterol in non-spiked aliquots) x 100] / [µg spiked sterol], and
179
expressed as the mean ± SD.
180
Determination of sterols by E-S
181
Methodology
182
An aqueous solution of 8.02 N KOH (0.8 ml) and 1.2 ml of ethanol were added to 1
183
ml of HM (~1.08 g) in a 10 ml Pyrex glass tube. The mixture was heated at 70ºC during
184
30 min in a shaking water bath (Stuart SBS30)15. After cooling the sample at room
185
temperature, extraction of the unsaponifiable fraction was performed as previously
186
described for the chromatographic method, with different centrifugation conditions
187
(4000 rpm, 4 °C, 5 min). The collected hexane phases were evaporated and dissolved
188
with 1 ml of 2-propanol. This dissolution, referred to as the “sample solution”, was
9 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
189
processed with the enzymatic kit according to the instructions of the manufacturer. The
190
kit is based on the oxidation of cholesterol to cholestenone, producing hydrogen
191
peroxide. In the presence of catalase, the hydrogen peroxide produced oxidizes
192
methanol to formaldehyde, which in turn reacts with acetyl acetone, forming a yellow
193
lutidine dye in the presence of ammonium ions. The lutidine dye formed (measured at
194
405 nm) is stoichiometric to the concentration of cholesterol (Spectrometer Lambda 2
195
UV/Vis, Perkin Elmer®).
196
Validation
197
Precision and accuracy were assessed according to the guidelines of the ICH38 and
198
AOAC39, whereas the analytical parameters of linearity, LOD and sensitivity were
199
provided by the manufacturer.
200 201 202 203 204
Precision The intra- and inter-assay precisions were established as described above in section 2.3.2. Validation (precision). Accuracy Accuracy was determined as described above in section 2.3.2. Validation (accuracy)
205
using 1 ml of HM.
206
Statistical analysis
207
Normal data distribution and homogeneity of variance were checked using the
208
Shapiro-Wilk and Levene tests, respectively. A two-tailed Student t-test (n-2 degrees of
209
freedom) was performed to evaluate the linearity of the calibration curves, and
210
ANCOVA or Kruskal-Wallis testing for normal and non-normal distribution of data,
211
respectively, was used to compare the slopes of the calibration curves (with and without
212
addition of matrix). To assess significant differences, analysis of variance (ANOVA,
213
with Tukey’s post hoc test) was applied to chromatographic and E-S cholesterol
10 ACS Paragon Plus Environment
Page 10 of 34
Page 11 of 34
Journal of Agricultural and Food Chemistry
214
contents in the course of lactation, and a Student t-test was used to compare cholesterol
215
contents between the two methods40 . For all cases, α = 0.05 was applied.
216
RESULTS AND DISCUSSION
217
Validation of the GC method
218
Selectivity and linearity
219
The total analytical run time was 28 minutes, as can be seen from the chromatogram
220
of the pool of mature HM (Figure 1). The identified sterols (relative retention time with
221
respect to the IS, 18.3 min) were: 7α-hydroxycholesterol (1.08), cholesterol (1.13),
222
desmosterol (1.18), lathosterol (1.21), 7β-hydroxycholesterol (1.27), campesterol (1.28),
223
stigmasterol (1.36), lanosterol (1.44) and β-sitosterol (1.45). Their GC-MS spectra are
224
shown in Figure 2.
225
Table 2 reports the tested amount ranges, the adjusted linear equations, the statistical
226
tests to check the goodness of fit of the calibration curves (without and with matrix),
227
and the matrix interference assay. All correlation coefficients were ≥ 0.98, and the
228
calculated t values were higher than the critical t value (3.18) for all sterols, which
229
reflected good linearity (Table 2).
230
Since a matrix interference was detected for cholesterol, desmosterol, lathosterol and
231
β-sitosterol (p < 0.05) (Table 2), we decided to use calibration curves added with matrix
232
for all sterol quantifications. Few publications have described the use of calibration
233
curves to quantify sterols (only cholesterol and desmosterol) in HM. Ramalho et al.15
234
reported HPLC calibration curves for cholesterol and desmosterol, with r > 0.999 for
235
both compounds. However, these authors did not report any matrix interference assay,
236
though they performed direct saponification of the sample. Kamelska et al.41 developed
237
a calibration curve for cholesterol quantification in IF by GC which was further applied
11 ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
238
in HM samples31,42. The slope of the curve was 5-fold lower (y=0.1861x+7.78, r = 0.96)
239
than our own.
240
Limit of detection (LOD) and limit of quantification (LOQ)
241
The LOD and LOQ (as ng in assay and µg/100 g of HM) obtained are shown in
242
Table 3. This method allowed the detection and quantification of all studied sterols at
243
concentrations < 2 and < 6 µg/100 g of HM for LOD and LOQ, respectively. The
244
highest LOD and LOQ corresponded to campesterol (1.80 and 5.99 µg/100 g of HM,
245
respectively) and the lowest to stigmasterol (0.26 and 0.87 µg/100 g of HM,
246
respectively). Cholesterol was quantifiable above 1.9 µg/100 g HM. Our LOQs were
247
20-fold lower for cholesterol and 6-fold lower for desmosterol than those reported by
248
Ramalho et al.15 (40 and 20 µg/100 ml, respectively).
249
Precision
250
Intra- and inter-assay precisions (% RSD) are reported in Table 3. Intra-assay
251
precision was lower than 13% for all the sterols analyzed, and ranged from 0.8%
252
(desmosterol) to 12.5% (stigmasterol), while inter-assay precision was lower than 18%
253
for all the sterols analyzed (from 3% for cholesterol to 17.6% for β-sitosterol). The intra
254
and inter-assay precision values comply with those indicated by the AOAC39 for a
255
single laboratory: