Whole Milk Increases Intestinal ANGPTL4 ... - ACS Publications

Subscriber access provided by Fudan University

Article

Whole milk increases intestinal ANGPTL4 expression and excretion of fatty acids through feces and urine Søren Drud Nielsen, Bashar Amer, Karoline Blaabjerg, Trine Kastrup Dalsgaard, Randi Jessen, Bjørn Petrat-Melin, Martin Krøyer Rasmussen, Hanne Damgaard Poulsen, and Jette Feveile Young J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.6b04135 • Publication Date (Web): 22 Dec 2016 Downloaded from http://pubs.acs.org on January 2, 2017

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 free 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 accessible to all readers and 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.

Journal of Agricultural and Food Chemistry 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 38

Journal of Agricultural and Food Chemistry

Whole milk increases intestinal ANGPTL4 expression and excretion of fatty acids through feces and urine Søren Drud Nielsen1*, Bashar Amer1*, Karoline Blaabjerg2, Trine K. Dalsgaard1, Randi Jessen1, Bjørn Petrat-Melin1, Martin Krøyer Rasmussen1, Hanne D. Poulsen2, Jette F. Young1 1

Department of Food Science, Aarhus University, Denmark

2

Department of Animal Science, Aarhus University, Denmark

*these authors contributed equally to the work.

Corresponding authors: Tel: +4587158051; Fax: +4587154891; E-mail: [email protected] Tel: +4587157998; Fax: +4587154891; E-mail: [email protected]

1 ACS Paragon Plus Environment


1

ABSTRACT

2

The angiopoietin-like 4 (ANGPLT4) protein is involved in lipid metabolism and is known to

3

inhibit lipoprotein lipase in the blood stream. We investigated the effect of milk on intestinal

4

ANGPTL4 and metabolic profile of growing pigs and the effect of free fatty acids (FFA) on

5

ANGPTL4 in ex vivo and in vitro assays. Feeding pigs whole milk increased intestinal

6

ANGPTL4 mRNA and increased fecal excretion of long chain FFA compared to the control

7

group fed soybean oil (n=9). Furthermore, FFA (C4-C8) induced ANGPTL4 gene expression in

8

porcine intestinal tissue mounted in Ussing chambers and ANGPTL4 protein secretion to both

9

the apical and basolateral side of intestinal Caco-2 cells on permeable membranes. Altogether,

Page 2 of 38

10

these results support an ANGPTL4 induced secretion of fecal FFAs. Urinary levels of FFA (C4-

11

C12), 3-hydroxyadipic acid and suberic acid were also increased by milk consumption indicating

12

higher energy expenditure compared to the control group.

13


Page 3 of 38


14

INTRODUCTION

15

Obesity is a negative consequence of an energy dense diet and is linked to increased risk of

16

developing cardiovascular diseases.1 Nutritional guidelines recommend limiting intake of

17

saturated fatty acids (FA), such as animal fat including milk fat.2, 3 However, in several studies

18

milk consumption has been associated with decreased risk of becoming obese, though

19

conclusions are not equivocal.4, 5 Some milk intervention studies found no positive correlation

20

between milk intake and obesity,4 but rather showed an inverse relationship between high-fat

21

dairy consumption and adiposity.6

22

Milk fat contains a high level of saturated FA with various chain length (C4-C18) and

23

esterification.7, 8 Exceptionally, milk fat contains high amounts of the short chain FA, butyrate,

24

and it is one of the few food sources of medium chain FA.7, 9 Studies on rodents suggest that

25

butyrate can protect against diet induced obesity,10, 11 while medium chain FA have been shown

26

to decrease obesity and positively affect lipid metabolism due to increased thermogenesis and fat

27

oxidation.12

28

Recently, we showed the potential of milk fat and intact casein to up-regulate angiopoietin-like 4

29

(ANGPTL4) gene expression when added to human intestinal cells in vitro.13 The free fatty acids

30

(FFA) of the milk were identified as the major inducers14 and confirm previous studies showing

31

that FFAs induce ANGPTL4 gene expression in several different intestinal cell lines.15, 16

32

ANGPTL4 inhibit lipoprotein lipase in the blood stream17, 18 and is suggested to act as a signal

33

protein released from adipocytes and other tissues causing a decreased lipid uptake and increased

34

fat metabolism.19 Hence, decreased levels of ANGPTL4 in the blood is associated with increased

35

bodyweight, increased circulating FFA, and increased waist-to-hip ratio, strongly suggesting a

36

role of ANGPTL4 in fat storage.20 Furthermore, studies have suggested that an increased



Page 4 of 38

37

intestinal ANGPTL4 may result in a reduction of adipocyte-associated lipoprotein lipase activity,

38

thus limiting uptake of FFAs by the adipose tissue.17 Recent results further indicate that

39

ANGPTL4 inhibits the pancreatic lipase in the intestinal lumen as ANGPTL4 (-/-) knock-out

40

mice show higher levels of triglycerides (TG) in the gut than normal mice.18 Furthermore, milk

41

inhibited the uptake of dietary fat, which was ascribed to the high calcium content. However,

42

dairy calcium seemed more efficient than non-dairy calcium indicating that other inducers of fat

43

excretion may be present in milk.21, 22 Thus, the excretion of FA after milk intake may be a

44

combined calcium- and ANGPTL4 effect and short and medium chain FA in milk seemed to be

45

essential as they are more potent inducers of ANGPLT4 gene expression compared to longer

46

chain FA.14

47

The present study investigates the effects of whole milk in a dietary intervention compared to a

48

control diet containing soybean oil using pigs as a model for humans. ANGPTL4 gene expression

49

of intestinal tissue and metabolite profile of the blood, urine and feces were investigated as well

50

as FFA of feces and urine. For specific investigation of short and medium chain FFA on

51

ANGPTL4 gene expression and mode of action, dissected pig intestinal tissue23 and the intestinal

52

Caco-2 cell line were used as a models.

53

MATERIALS AND METHODS

54

Materials. Methanol and heptane were purchased from Rathburn Chemicals Ltd., Walkerburn

55

Scotland, UK, and methoxamine hydrochloride from Alfa Aesar GmbH, Karlsruhe, Germany.

56

Pyridine, anhydrous ethanol, ethyl chloroformate (ECF), butyric acid, linoleic acid and linolenic

57

acid standards, hexanoic acid-d11 (98%) internal standard, phosphate buffered saline (PBS) and

58

trimethylsilyl cyanide from Sigma-Aldrich CHEMIE GmbH, Steinheim, Germany. Caproic acid,

59

caprylic acid, capric acid, lauric acid, myristic acid, palmitic acid, stearic acid and oleic acid 4 ACS Paragon Plus Environment

Page 5 of 38


60

standards came from Fluka Chemie GmbH (Buchs, Switzerland). Myo-inositol-d6 (98%), citric-

61

2,2,4,4-d4 acid (98%), butyric acid-d7 (98%), octanoic acid-d15 (98%), decanoic acid-d19

62

(98%), lauric acid-d23 (98%), myristic acid-d27 (98%), palmitic acid-d31 (98%) and stearic

63

acid-d35 (98%) internal standards were obtained from Cambridge Isotope Laboratories, Inc.

64

Andover, MA, USA.

65

Chloroform from Chem Solute (Renningen, Germany), and sodium hydroxide from J. T. Baker

66

(Deventer, The Netherlands). Ultra-pure water came from a Milli-Q system, Millipore SAS

67

(Molsheim, France) and anhydrous sodium sulphate was from Merck (Darmstadt, Germany).

68

Approvals. The Danish Animal Experiments Inspectorate, the Danish Ministry of Justice,

69

Copenhagen, Denmark, approved the experimental protocols for the three experiments with pigs.

70

Dietary intervention with whole milk (Exp. 1). In total, 18 crossbred pigs

71

(Landrace*Yorkshire*Duroc) comprising nine litters of two pigs each were used. The pigs were

72

reared at a commercial farm, weaned at day 28 of age and transported to the research facilities at

73

Aarhus University, Foulum, where they were fed a standard pig diet based on barley, wheat,

74

soybean meal and rape cake for 11 weeks prior to the experimental period. During the

75

experimental period, the pigs with an initial body weight of 46.9 ± 5.6 kg were individually

76

housed in pens with visual and physical contact to neighboring pigs and ad libitum access to

77

water. The littermates were fed one of the two experimental diets twice a day for three weeks

78

after which the pigs were deprived of feed 18 h before slaughter. The composition of the two

79

diets is shown in Table S1. The control diet was optimized according to the Danish

80

recommendations for pigs between 65 and 105 kg for all nutrients.24 The milk-based diet was

81

similar to control diet apart from replacing soybean oil with full fat milk (Lærkevang, 3.5% fat,

82

Arla, Denmark) mixed with the solid part of the diet at feeding. The pigs were fed restrictively 5 ACS Paragon Plus Environment


Page 6 of 38

83

and the feeding level in the first week (control diet: 1.76 kg feed/day; milk diet: 1.71 kg feed/day

84

+ 0.66 kg milk/day) was increased by 17 and 33% in the second and third week, respectively.

85

One pig fed the control diet was omitted from the experiment due to hernia. All of the remaining

86

pigs stayed healthy and none of them had feed refusals.

87

Upon termination, samples of adipose tissue, muscle, liver, and pancreas were taken and

88

immediately frozen in liquid nitrogen. Blood serum, feces and urine were collected the day prior

89

to the dietary intervention, and on the day the experiment was terminated. The blood samples

90

were allowed to coagulate for 30 min at room temperature, and thereafter centrifuged at room

91

temperature at 1700 x g for 10 min. Urine, serum and fecal samples were homogenized and

92

stored at -80 °C until analysis. Stripped live epithelium from the proximal part of the ileum were

93

washed with ice cold PBS six times, and transferred to a 15 mL tube with 10 mL cell recovery

94

solution (BD Bioscience, Albertslund, Denmark) and incubated on ice for 1 h. Subsequently,

95

each tissue piece was transferred to a petri dish and the upper cell layer was carefully scraped

96

with a flexible inoculation loop. The cell recovery solution containing the free floating cells from

97

the tissue was transferred to a 15 mL tube on ice and incubated for 30 min. Afterwards PBS was

98

added to a total volume of 15 mL, and centrifuged at 350 x g for 5 min at 4 °C. Cells were

99

washed three times in PBS and then lysed in 350 µL lysis buffer from the RNeasy minikit

100

(Qiagen, Copenhagen, Denmark). The cell lysate was vortexed for 30 sec and stored at -80 °C

101

until RNA purification. Serum concentrations of total cholesterol (TC), HDL-cholesterol (HDL-

102

C), LDL-C and TG were measured on a Pentra 400 analyser (HORIBA ABX, Montpellier,

103

France) as previously described.25

104

Ussing chamber experiment (Exp. 2). Prior to the Ussing chamber experiment, we

105

investigated the constitutive expression of ANGPTL4 along the pig small intestine. Three female 6 ACS Paragon Plus Environment

Page 7 of 38


106

pigs weighing 49 ± 2 kg were terminated as described in Exp. 1. The pigs were crossbreeds and

107

reared at the same commercial unit as the pigs in Exp. 1. Six samples were taken from the small

108

intestine by removing a 10 cm piece of the intestine. The length of the small intestine was 15.1 ±

109

0.1 m and samples were taken as follows: sample 1 at 20 cm posterior to the pylorus and samples

110

2, 3, 4, 5 and 6 every 3 m so that sample 6 was taken 20 cm anterior to the ileal-caecal junction.

111

Intestinal epithelium cell lysate was obtained as described in Exp. 1 and stored at -80 °C until

112

RNA purification.

113

For the Ussing chamber experiment, four female pigs weighing 47 ± 1 kg were fed a standard

114

diet optimized according to the Danish recommendations for all nutrients. Pigs were deprived of

115

feed 19 h before slaughtering. Immediately after slaughtering, the small intestine (17 ± 3 m) was

116

removed and 150 cm of the distal part (300 cm anterior to the ileal-caecal junction) was collected

117

and placed in an oxygenated and phosphate-buffered Ringer solution (25 mM NaHCO3, 120 mM

118

NaCl, 1.0 mM MgSO4, 6.3 mM KCl, 2.0 mM CaCl, 0.32 mM phosphate buffer (pH 7.4) and 16

119

mM glucose) preheated to 38 °C. Within 15 min, the epithelium was stripped of the muscle

120

layers and mounted onto eight Ussing chambers with an opening area of 1.995 cm2 (WPI,

121

Sarasota, FL, USA). Fifteen mL aerated (95% O2 and 5% CO2) Ringer solution was used as

122

bathing medium on the mucosal and serosal sides.26 The Ringer solution at the mucosal side

123

contained equal amounts of sodium salts of butyrate, caproate and caprylate to a total

124

concentration of 0, 3 or 9 mM (chamber concentration) while the Ringer solution at the serosal

125

side contained the same amount of mannitol (chamber concentration), respectively, to keep

126

osmolality equal on the two sides. Two chambers were controls without FFA whereas three

127

chambers each contained 3 or 9 mM FFA. Thirty min. after mounting of the epithelium, the

128

Ringer solution was replaced by an identical solution. Hence, the tissue was exposed to FFA for



Page 8 of 38

129

3 h, however, the transport of FFA over the tissue was measured in the Ringer solution incubated

130

with the tissue for 2.5 h. The difference in the electrical potential (Pd, mV) over the epithelium

131

was measured in open circuit at 1, 2 and 2.5 h after mounting. The remaining time the Pd was

132

clamped to 0 mV by an external current called short-circuit current (Isc, µA) and the epithelial

133

conductance (G), was calculated from Isc and Pd using Ohm’s law. To test the viability of the

134

tissue, 2.27 mM theophylline (chamber concentration), a phosphodiesterase inhibitor, was added

135

bilaterally to all chambers to measure the cAMP-dependent Cl- secretion of the intestine after 3

136

h. The increase in Isc due to theophylline addition (∆Isctheo) was calculated by subtracting the

137

basal Isc measured 5 min before stimulation from the peak Isc 15 min after stimulation. At 3 h of

138

incubation, the Ringer solution from both sides was sampled for FFA analyses. The epithelium

139

was gently removed from the chambers and a heterogenic population of epithelial cells was

140

scraped off as described in the previous section.

141

Caco-2 trans-well experiment (Exp. 3). Culture conditions and trans-well studies on the human

142

intestinal cell line Caco-2 (ATCC, HTB-37) were conducted as described previously.27 Briefly,

143

cells were seeded in 24 well plates at a density of 6.5 × 104 cells cm-2 for the ANGPTL4 gene

144

expression assays. For the protein secretion assays combined with ANGPLT4 gene expression,

145

Caco-2 cells were grown on permeable membranes (0.4 µm PCF, Millicell, Hellerup, Denmark)

146

placed in a 24 well plate, with 600 µL culture medium on the basolateral side and 400 µL on the

147

apical side. Cells where left in culture for 21 days to differentiate with change of culture medium

148

every 2-3 days.28 Transepithelial resistance and lucifer yellow passage were monitored as

149

evidence of monolayer integrity.27 After 21 days, culture medium was changed on the apical side

150

to treatment sample and incubated for 24 h at 37 °C. After 24 h the culture medium from the

151

apical and basolateral sides were collected into individual tubes and stored at -80 °C until 8 ACS Paragon Plus Environment

Page 9 of 38


152

analysis. The cells were lysed with 350 µL lysis buffer (RNeasy Mini Kit, Qiagen) and stored at

153

-80 °C until RNA-purification. Monolayers were exposed for 24 h to 9 mM of either the single

154

FFA butyrate, caproate or caprylate or a FFA-mix of these in culture medium. Differentiated

155

Caco-2 cells on permeable membranes were added 3 or 9 mM butyrate or caprylate to determine

156

ANGPTL4 gene induction and protein secretion to either the apical or basolateral side of the cell

157

monolayer. A cell viability assay based on the activity of a cellular oxidoreductase (wst-1) was

158

used to verify no significant reduction in the viability of cells.

159

ANGPTL4 expression (real-time PCR). RNA purification, reverse transcription and

160

quantification of the relative amount of cDNA from all three experiments were conducted

161

according to Nielsen et al. (2014). The sequences of forward primers, reverse primers and

162

hydrolysis probes were as follows; Porcine ANGPTL4 5’TCTCTGGTGGTTGGTGGTTTG’3,

163

5’GCTGCCGAGGGATGGAAT’3 and 5’CCACTCCAACCTCAATGGCCAGTACTTC’3,

164

Human ANGPTL4; 5’GACCCGGCTCACAATGTCA’3,

165

5’ATCTTGCAGTTCACCAAAAATGG’3 and 5’TGCACCGGCTGCCCAGGG’3, β-actin; 5’

166

ACCCAGATCATGTTCGAGACCTT’3, 5’TCACCGGAGTCCATCACGAT’3,

167

5’CTGTATGCCTCTGGCCGCACCA3.

168

ANGPTL4 protein. ANGPTL4 protein levels in the culture medium from Caco-2 cells grown

169

on permeable membranes from Exp. 3 were determined using a sandwich enzyme-linked

170

immunoabsorbent assay as described previously. 14, 29

171

Metabolomics. Sample preparation. Frozen fecal, serum and urine collected from Exp. 1 were

172

thawed in an ice bath, 80% methanol extraction was performed on samples followed by

173

trimethylsilyl derivatization as described in supplementary material.



Page 10 of 38

174

GC-ToF MS analysis. The samples (1 µL aliquot) were injected in a 1:40 split mode into an

175

Agilent Technologies 7890B gas chromatography system coupled to an Agilent Technologies

176

7200 Accurate-Mass Q-ToF mass spectrometer (Agilent Technologies, Germany). Separation of

177

the metabolites was performed on a HP-5MS capillary column coated with polyimide (20 m,

178

0.180 mm i.d., 0.18 µm film thickness; Agilent Technologies). The initial temperature of the

179

oven was held at 60 ºC for 2 min, ramped to 320 ºC at a rate of 30 ºC min-1 and then held at 320

180

ºC for 2 min. Helium was used as carrier gas at a constant flow rate of 1 mL min-1 through the

181

column. The temperatures of the ion source and injector were 230 and 300 ºC, respectively. The

182

mass spectral analysis was performed in scan mode with a quadruple temperature of 150 ºC and

183

a fragmentation voltage of 70 eV with a solvent delay of 2.00 min. Mass detector was switched

184

off not to detect the urea peak at 5.1-5.6 min. Samples from the two diet groups and from before

185

and after intervention were randomly analyzed.

186

Free fatty acids quantification. Quantification of FFA in blood, urine and feces from Exp. 1 as

187

well as from the Ringer solution of the Ussing chamber experiment (Exp. 2) was performed

188

using the ECF-FFA method described by Amer et al. (2013) with modification for feces as

189

described by Amer et al. (2015).30 Three out of six urine samples in each group were excluded

190

from the analysis due to contamination with feces either in samples drawn before or after the

191

intervention. Epithelial transport rates of FFA measured in the Ussing chambers were calculated

192

as follows:

193

Mucosal clearance (MD) = Cmuc(0) – Cmuc(2.5)

194

Cmuc(0) = [FFA] (mmol/L) on mucosa side, immediately after replacing the Ringer solution

195

Cmuc(2.5) = [FFA] (mmol/L) on mucosa side after 2.5 h

196

Serosal release (SR) = Cser(2.5) - Cser(0)


Page 11 of 38


197

Cser(0) = [FFA] (mmol/L) on serosa side, immediately after replacing the Ringer solution

198

Cser(2.5) = [FFA] (mmol/L) on serosa side after 2.5 h

199

The clearance of FFA at the mucosal side and was not recovered at the serosal side, is assumed

200

to constitute tissue content and metabolized loss (TM). TM = MD - SR

201

Statistical analysis. Statistical analysis of gene expressions, protein concentrations and blood

202

lipids was performed in R 2.14.0 software with linear mixed models adjusted for multiple

203

comparisons using Tukey’s HSD Post Hoc to compare the values between treatments.

204

Ussing chamber experiment. Measurements of the electrophysiological parameters (Pd, Isc, G

205

and ∆Isctheo) were analyzed by the MIXED procedure in SAS. Treatment differences were

206

separated by the PDIFF option. Data on Pd, Isc, and G were analyzed by a model containing the

207

fixed effects of FFAs concentration, incubation time, and their interaction. Pig and pig x

208

intestinal piece x chamber were considered as random factors. Incubation time within pig x

209

intestinal piece x chamber was used as repeated measure. The results on ∆Isctheo were analyzed

210

by a model containing the fixed effects of FFA concentration. Pig and pig x intestinal piece were

211

considered as random factors.

212

Intervention study. Animal performance data were analyzed by a model containing diet as fixed

213

effect (control or milk) and the initial weight as covariate. Litter was the random variable. FFA

214

changes for each pig was analyzed by analysis of variance.

215

GC-ToF MS metabolomics data. Raw GC-ToF MS data were deconvoluted using Mass Hunter

216

Unknown Analysis software (Quantitative Analysis, Version B.07.00/Build 7.0.457.0, Agilent

217

Technologies), detected components from all samples were transferred into Mass Profiler

218

Professional software (Version 12.6.4-Build 196252, Agilent Technologies) for retention time

219

alignment, peaks filtration and metabolites identification using NIST11 library. Data sets were



Page 12 of 38

220

normalized by dividing features over the sum of all features of each observation after removing

221

background peaks from column material, solvents and derivatization reagents.

222

After processing, integrated peak areas for the detected features, data were imported to SIMCA

223

14.0 32-bits (Umetrics, Sweden) for multivariate data analysis. Before statistical analysis, data

224

were visually examined for outliers (Hotelling’s T2 value > 95% confidence limit) in principal

225

component analysis (PCA). Both PCA and orthogonal partial least squares discriminant analysis

226

(OPLS-DA) were applied to develop compliance models based on the most discriminative

227

features detected in serum, urine and feces. Models were calculated for the whole data set and for

228

different subsets of the samples according to time point of sample collection or diet group. Data

229

was auto scaled (unit variance scaling), and the dietary patterns and/or time point were given as

230

class information in the OPLS-DA models. OPLS-DA models were cross-validated to evaluate

231

the performance of the models. The importance of the variables for the projection values, which

232

summarize the overall contribution of each X-variable to the PLS model, were used to select

233

metabolites for univariate data analysis when comparing different groups or different time

234

points. Univariate data analysis was done on metabolites selected from multivariate data analysis

235

using R software. Internal standards peaks were used to normalize GC-MS data before analysis

236

for higher accuracy.31

237

RESULTS

238

Dietary intervention with whole milk (Exp. 1). The control and milk diets were balanced for

239

energy and lipid content, whereas the protein content of the milk diet was higher resulting in a

240

higher daily protein intake of 8% for pigs fed the milk diet compared to the control diet.

241

Consequently, pigs fed the milk diet tended to have a higher finishing weight and average daily


Page 13 of 38


242

gain (ADG) and a lower feed conversion ratio (FCR) compared with pigs fed the control diet

243

(Table 1).

244

The constitutive gene expression of ANGPTL4 in different tissues of the pigs fed the control diet

245

was investigated and revealed the highest expression in the adipose tissue, followed by the liver,

246

small intestine and muscle while ANGPTL4 mRNA could not be detected in pancreas [average

247

Ct values: 24.4, 25.1, 27.3 and 29.6, respectively]. The ANGPTL4 gene expression in different

248

tissues from the pigs fed the milk diet revealed no differences in muscle, liver or fat tissue, while

249

ANGPTL4 gene expression of the small intestine was significantly increased 2.5 fold by the milk

250

diet compared to the control.

251

A targeted approach was used for quantification of FFA content while a global metabolite profile

252

was performed on fecal, urine and serum samples. PCA was carried out on 100, 141 and 198

253

entries in serum, urine and feces, respectively, from samples collected before and after three

254

weeks of intervention. A stable baseline was obtained showing no separation i.e. no difference

255

between the two groups in any of the three matrices (data not shown). After intervention a clear

256

separation along PCA1 of the control and the milk diet was obtained in all three matrices

257

explaining 19% of the variation for serum, 17% for urine and 34% for feces samples,

258

respectively (Figure 1A, 1B and 1C). As PCA is an unsupervised method only directed by the

259

variation of the two datasets, this result confirms a different metabolite profile when milk was

260

included in the diet compared to the control diet.

261

OPLS-DA models were used to identify metabolites that separated the groups receiving the

262

control and milk diets (Figure S1 in supplementary material) and to identify which of the

263

detected metabolites were responsible for the observed effect (data not shown). All identified



Page 14 of 38

264

metabolites with changed level (selection criteria: variable importance for the projection (VIP)

265

value > 1.0) in urine and feces are summarized in Table 2.

266

Serum and feces levels of the amino acids: alanine, valine, leucine, proline, serine, methionine,

267

pyroglutamic acid, glutamic acid, phenylalanine and lysine was higher at the end of experimental

268

period for pigs fed milk diet. Only alanine and serine was higher in urine from pigs fed the milk

269

diet. In feces, also the amino acids sarcosine, glycine, methionine, aspartic acid and tyrosine

270

were higher in the milk group compared to the control. As the protein level was higher in the

271

milk diet this result was expected. Also the amino acid related metabolites, benzoic acid and 2-

272

hydroxyisocaproic acid were higher in feces from pigs fed the milk diet compared to the control.

273

In urinary the level of the medium chain FAs oxidation metabolites, 3-hydroxyadipoc acid and

274

suberic acid, was increased as a result of milk diet both as compared to before the intervention

275

(data not shown) and also compared to the control diet at the end of the intervention (Table 2),

276

whereas gluconic acid lactone was lower upon milk diet compared to the control.

277

The compounds m-xylene, o-xylene, ethylbenzene, 5-hydroxy-2-pentanone and D-cellobiose

278

were lower in feces from pigs fed milk diet compared to the control. We have not been able to

279

define specific roles for these metabolic differences.

280

Pigs given the milk diet excreted more of long-chain saturated FFAs (C14:0, C15:0, C16:0 and

281

C18:0) in feces compared to those on control diet, whereas the short and medium chain FFAs

282

(C4:0, C8:0, C10:0 and C12:0) was elevated in the urine (Figure 2).

283

No changes in TC, HDL-C, LDL-C, LDL-C:HDL-C, TC:HDL-C, TG concentration and their

284

delta values were observed between the two dietary groups (Table 3). However, the TC:HDL-C

285

ratio tended to be lower (P

Whole Milk Increases Intestinal ANGPTL4 ... - ACS Publications

Recommend Documents