|
|
Received: 25 February 2018 Revised: 1 September 2018 Accepted: 3 September 2018 DOI: 10.1111/jpn.12999
ORIGINAL ARTICLE
Effect of zinc oxide sources and dosages on gut microbiota and integrity of weaned piglets Wei Wang1 | Noémie Van Noten1 | Jeroen Degroote1 | Agathe Romeo2 | Pieter Vermeir3 | Joris Michiels1 1 Department of Animal Sciences and Aquatic Ecology, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium 2
ANIMINE, Sillingy, France
3 Department of Green Chemistry and Technology, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium
Correspondence Joris Michiels, Department of Animal Sciences and Aquatic Ecology, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium. Email:
[email protected] Present Address Wei Wang, Beijing Advanced Innovation Center for Food Science, China Agricultural University, Beijing, China Funding information Animine
Abstract Zinc oxide (ZnO) supplied at pharmacological dosage in diets of weaned piglets im‐ proves growth performance. However, it causes environmental contamination and induces bacterial antibiotic resistance, yet this practice is debated. The effects on gut microbiota and integrity in weaned piglets of conventional ZnO at nutritional and pharmacological dosage (110 and 2,400 mg/kg Zn, respectively) were compared to an alternative ZnO source at 110 and 220 mg/kg Zn. Each of the four treatments was applied to four pens (two piglets/pen; weaning age, 20 days) for 15 days, and piglets were sampled on day 15 to determine indices of gut integrity. Feeding conventional ZnO at 2,400 mg/kg Zn reduced coliforms and Escherichia coli in distal small intestine as compared to conventional ZnO at 110 mg/kg (−1.7 and −1.4 log10 cfu/g, respec‐ tively), whereas the alternative ZnO reduced only coliforms, irrespective of dosage (−1.6 to −1.7 log10 cfu/g). Transepithelial electrical resistance of distal small intestinal mucosa was higher for pigs fed the alternative ZnO source as compared with groups fed 110 mg/kg Zn of conventional ZnO, in line with a trend for higher gene expres‐ sion of claudin‐1 and zona occludens‐1. Interestingly, the alternative ZnO source at 110 and 220 mg/kg Zn increased intestinal alkaline phosphatase gene transcript as compared to conventional ZnO at 110 mg/kg Zn, whereas the alternative ZnO source at 110 mg/kg Zn exhibited higher Zn concentrations in mucosa (2,520 μg/g) as com‐ pared to conventional ZnO at 110 mg/kg Zn (1,211 μg/g). However, assessing alka‐ line phosphatase activity, no significant effects were found. In conclusion, the alternative ZnO reduced digesta Enterobacteriaceae numbers and improved gut in‐ tegrity, albeit similar or better, depending on the dosage, to the effects of pharmaco‐ logical dosage of conventional ZnO. KEYWORDS
escherichia coli, intestinal alkaline phosphatase, tight junctions, transepithelial electrical resistance, zinc solubility
1 | I NTRO D U C TI O N
(2012), the dietary requirement of Zn for piglets between 5 and 11 kg is 100 mg/kg. However, when added as zinc oxide (ZnO)
Zinc (Zn) is an essential trace element for animals, owing to its key
at pharmacological dosage (2,000–3,000 mg/kg Zn) in the feed
roles as a structural, catalytic and signalling component (Kambe,
of weaning piglets, it exhibits improved performance, endorsed
Tsuji, Hashimoto, & Itsumura, 2015). As recommended by NRC
by the alleviation of post‐weaning diarrhoea incidence. Although
J Anim Physiol Anim Nutr. 2018;1–11.
wileyonlinelibrary.com/journal/jpn © 2018 Blackwell Verlag GmbH | 1
|
WANG et al.
2
TA B L E 1 Ingredient and nutrient composition of experimental diets fed to weaners (d0‐15 post‐weaning)
ZnO source
Conventional
Alternative
Supplementation (mg/kg)
110
110
220
Ingredient composition (%, unless otherwise)
mation, improved intestinal histomorphology and mucin compo‐ sition, increased mucosal insulin‐like growth factor‐I expression, enhanced tight junction integrity and the adjustment of gut mi‐ crobiota (Carlson, Poulsen, & Sehested, 2004; Hu, Song, Li, Luan,
Barley
25.00
& Zhu, 2013; Li et al., 2006; Liu et al., 2014; Mavromichalis, Peter,
Corn
17.31
Parr, Ganessunker, & Baker, 2000). However, the long‐term use
Wheat
12.00
of ZnO at pharmacological dosage can produce toxic symptoms
Soybean meal CP49
12.36
characterized by depressed feed intake and efficiency of gain
Toasted soybeans
12.00
(Brink, Becker, Terrill, & Jensen, 1959). Further, feeding of high
Oat flakes
8.000
ZnO dosages may stimulate the occurrence of resistance to Zn
Sweet whey powder
4.000
in the pig gut bacteria and may play a role in the co‐selection of
Lactose
2.000
Soybean oil
1.177
Premix min & vit a
1.000
Potato protein
1.000
Sugar beet pulp
1.000
Monocalcium phosphate
0.961
Limestone
0.932
L‐lysine HCl
0.452
Salt
0.200
DL‐methionine
0.234
approaches to reduce post‐weaning scouring (Michiels, VanNoten,
L‐threonine
0.191
& Degroote, 2017), deserves full attention. HiZox® (Animine,
L‐valine
0.132
Sillingy, France) is ZnO and is prepared by a patented manufactur‐
L‐tryptophan
0.053
ing technology providing unique physico‐chemical features, such
Conventional ZnO, mg/ kg
147
antibiotic resistance (Cavaco, Hasman, & Aarestrup, 2011; EFSA, 2014). High dietary ZnO leads to excessive faecal Zn excretion (Buff, Bollinger, Ellersieck, Brommelsiek, & Veum, 2005; Case & Carlson, 2002), and hence the accumulation of Zn in soils after in‐ tensive pig farming is concerned as a serious environmental issue (Romeo, Vacchina, Legros, & Doelsch, 2014). In 2017, in the EU, the Commission voted in favour of phasing out pharmacological ZnO, ultimately by 2022 (Commission Implementing Decision of 26.6.2017, C(2017) 4,529 Final). Therefore, new Zn formulations applied at lower dosages that might provide equal efficacy (Cho, Upadhaya, & Kim, 2015; Han, Ma, Lv, Wu, & Qian, 2014), or other
FeSO 4.H2O, mg/kg
as high purity standards, high specific surface area from elevated
3,200
porosity and unique particle size and shape, according to supplier
Alternative ZnO, mg/kg 1,521
1,521
147
294
information. Cardoso, Chevalier, and Romeo (2017) employed
1,521
1,521
low‐angle laser light scattering, electron microscopy and BET nitrogen adsorption isotherms to differentiate feed‐grade zinc
Calculated and analysed nutrient composition b
89.0
90.0
89.4
89.4
4.8
5.1
4.6
5.2
Crude proteinb
17.7
18.2
18.1
18.3
Dig. Lysinec
1.12
1.12
1.12
1.12
Ether extractb
5.7
5.8
5.7
5.6
Zn, mg/kgb
114
1,891
94
273
Net energy, MJ/kgc
9.91
9.91
9.91
9.91
Dry matter Ash
a
moting effect of ZnO likely benefits from the improvement of gut health, including reduced secretory responses, reduced inflam‐
Treatment
2,400
the specific mechanisms are not fully clarified, the growth‐pro‐
b
Providing per kg of diet: vit A (retinyl acetate), 15,000 IU; vit D3 (chole‐ calciferol), 2,000 IU; vit E (all‐rac‐alfa‐tocopheryl acetate), 50.0 mg; vit K3 (menadione), 4.0 mg; vit B1 (thiamine mononitrate), 3.1 mg; vit B2 (ri‐ boflavine), 8.0 mg; vit B3 (calcium‐D‐pantothenate), 20 mg; vit B6 (pyri‐ doxine hydrochloride), 6.0 mg; vit B12 (cyanocobalamine), 50.0 µg; vit B3 (niacinamide), 40.0 mg; folic acid, 2.0 mg; biotin, 0.3 mg; betaine an‐ hydrate, 285 mg; endo‐1,4‐beta‐glucanase E3.2.1.4, 250 TGU; endo‐1,4‐ beta‐xylanase E3.2.1.8, 560 TXU; Cu (copper(II)sulphate pentahydrate), 15.0 mg; Mn (manganese(II)oxide), 48.0 mg; I (calcium iodate anhydrate), 1.9 mg; Se (sodium selenite), 200 µg; Se (selenomethionine produced by Saccharomyces cerevisae NCYC‐R397), 100 µg; E306, extract of vegeta‐ ble oils rich in tocopherols, tocopherols; 228 mg; clinoptilolite, 1.64 g; aromatic compounds, 72 mg. bAnalysed. cCalculated.
oxide sources in terms of particle size, morphology and specific area. HiZox® represented different characteristics than conven‐ tional ZnO sources. It showed aggregates having a sponge internal structure, whereas other sources were dense materials. Also, it presented large agglomerates and small aggregates, while other samples did not differ from aggregates and agglomerates. The po‐ tentiated ZnO showed a porous high specific area, whereas others had very low specific area and were not porous. For example, the potentiated ZnO had a specific surface area of 42 m2/g versus 0.5– 2.4 m2/g for other ZnO sources. This potentiated ZnO source con‐ tains minimum 75% Zn. Ex vivo, in chyme from donor piglets that was supplemented with ZnO sources, it was shown that this ZnO source showed higher solubility and enhanced bacterial growth reduction more drastic and more rapid as compared to analytical ZnO at the same concentration (Vahjen, Zentek, & Durosoy, 2012). When added to the starter diet of piglets (from 14 to 35 days post‐ weaning), the alternative ZnO source supplemented at 110 mg/kg Zn improved growth and feed:gain as compared to conventional ZnO at 2,400 mg/kg Zn (Morales, Cordero, Piñeiro, & Durosoy,
|
3
WANG et al.
2012). Cho et al. (2015) demonstrated that this alternative ZnO
piglets. Scoring was done daily for all piglets. The number of diar‐
source added at 220 mg/kg Zn could equally improve growth in the
rhoea days accounts for all piglets on each day having a faeces
prestarter phase as pharmacological ZnO, whereas Raquipo et al.
score 3.
(2017) demonstrated that the alternative ZnO source at 220 mg/ kg Zn was advantageous over a coated ZnO source. The objective of the current study was to test the effects on gut microbiota and
2.2 | Sampling
integrity in weaned piglets (prestarter phase) of conventional ZnO
At d15, piglets were euthanized by electrical stunning followed by
at nutritional and pharmacological dosage (110 and 2,400 mg/kg
exsanguination. Subsequently, the entire gastrointestinal tract was
of Zn, respectively) compared to the alternative ZnO source at 110
removed. The small intestine (defined as the part of the gastroin‐
and 220 mg/kg of Zn, dosages compliant to local legislation.
testinal tract between the pyloric sphincter and the ileo‐caecal valve; SI) was obtained, and its length was measured. A segment of
2 | M ATE R I A L S A N D M E TH O DS 2.1 | Animals and diets
20 cm posterior to 75% of small intestinal length, referred as distal SI, was taken and rinsed with saline and used for Ussing chamber experiments. A mucosal sample at the same site was taken by plac‐ ing another segment of 20 cm on an ice‐cold surface after rinsing
The study was conducted in accordance with the European rec‐
with saline. The tissue was slit along its length, and the mucosal side
ommendations for the protection of animals used for agricultural
was softly cleaned with a paper towel to remove residual fluid. The
research (EU Directive 91/630/EEG and 98/58/EG). Thirty‐
mucosa was obtained by gently scraping with a glass slide. Mucosa
two newly weaned piglets (Topigs × Piétrain, 20 days of age,
was snap‐frozen in liquid nitrogen and stored at −80°C. Next, the
6.04 ± 0.77 kg) were blocked by origin (litter), sex (equal number
contents of the stomach and segments 0%–25% (proximal SI) and
of females and castrated males) and weight and assigned to 16
75%–100% (distal SI) of small intestinal length were quantitatively
pens with two piglets per pen. It means that for every block of four
collected, pH of the contents was determined, and aliquots were
pens, piglets were originating from two litters in a way that each
taken to determine Escherichia coli and coliform counts by plating
pen had one piglet from each litter. Four dietary treatments were
technique and for qPCR quantification of bacterial groups. The re‐
applied to four pens each, following a randomized block design.
maining contents of stomach, proximal and distal SI were subjected
A wheat–barley–corn–soybean basal diet with no antibiotics, and
to freeze‐drying pending analysis for soluble and total Zn.
free of supplemental Zn and Fe, was prepared. Four experimental diets were made as follows (Table 1): basal diet +110 mg/kg Zn from conventional ZnO (feed grade, 75% Zn; sourced from a local
2.3 | Ex vivo measurements in Ussing chambers
commercial feed mill) (control), basal diet +2,400 mg/kg Zn from
Segments of distal SI were used to measure intestinal permeabil‐
conventional ZnO, basal diet +110 mg/kg Zn from the alternative
ity and ion transport with the Ussing chamber system according to
ZnO source (HiZox®, Animine, Sillingy, France; 75% Zn) and basal
Michiels, Missotten, Dierick, Fremaut, and De Smet (2010) and Wang
diet +220 mg/kg Zn from the alternative ZnO source. The basal
et al. (2016). In brief, for each piglet, the segment (20 cm) was flushed
diet contained a high level of supplemental Fe (500 mg/kg), ex‐
with saline, then cut along the mesenteric border, stripped of the
ceeding by multifold the requirements outlined by NRC (2012), to
muscle layers and mounted in modified Ussing chambers on a seg‐
exacerbate gastrointestinal disturbances and burden the antioxi‐
ment holder, as flat sheets with an exposed tissue area of 1.07 cm2.
dant system (Li, Hansen, Borst, Spears, & Moeser, 2016). Nutrient
Paracellular permeability of the tissue was assessed with the macro‐
analysis of diets showed similar amounts for proximate nutrients
molecular probe FD‐4, 4 kDa, in two chambers. Parameters of intes‐
across treatments (Table 1). Treatments that were supplemented
tinal ion transport and the tissue integrity were assessed in two other
at 110 mg Zn per kg diet on top of Zn from ingredients showed
chambers, including transepithelial electrical resistance (TEER), base‐
slightly differing Zn levels, whereas diets with pharmacological
line short‐circuit current (Isc) and Cl− secretion stimulated by agonists:
level of ZnO contained close to 2,000 mg/kg Zn, still considered
serosal 0.1 mM 5‐HT (serotonin, neural regulated secretagogue) and
as growth‐promoting. Pigs were housed with two pigs per pen
bilateral 5 mM theophylline (cAMP/cGMP‐mediated secretagogue).
(2.10 m2/pen) with full slatted floors, conventional ventilation scheme and with a starting ambient temperature at 30°C and a 24 L light schedule till d5 post‐weaning. From d6 till 15 post‐weaning,
2.4 | Bacteriological determinations
ambient temperature was linearly adjusted to 28°C with 18L:6D
Escherichia coli and coliform counts (viable counts; log10 CFU/g fresh
light schedule (dark form midnight till 6 a.m.). Animal performance
digesta) in digesta of stomach, proximal and distal SI were done
and health (BW, ADG, ADFI, F:G ratio, faeces score and diarrhoea
using the ring‐plate technique (Michiels, Missotten, Van Hoorick,
days) were monitored. Faeces score was as follows: score 1, nor‐
et al., 2010). Serial 10‐fold dilutions were made from 1 g aliquots
mal brown soft‐formed stool; score 2, yellow or dark black (bloody)
of fresh digesta, using a sterilized peptone solution (1 g peptone
soft‐formed sticky faeces; and score 3, watery, liquid, unformed
+0.4 g agar +8.5 g NaCl in 1 L distilled water) and plated onto se‐
stool, yellow or dark black (bloody) diarrhoea, wet backsides
lective media in duplicate. Selective media were used for counting
|
WANG et al.
4
the following bacterial groups: E. coli (Tryptone Bile X‐glucuronide Agar, CM0945, Oxoid; incubated for 24 hr at 44°C aerobically) and coliform bacteria (Eosin Methylene Blue Agar, CM0069, Oxoid;
2.6 | Alkaline phosphatase activity in distal small intestinal mucosa
incubated for 24 hr at 37°C aerobically). Denaturing gradient gel
Alkaline phosphatase activity was determined using an alkaline
electrophoresis (DGGE) was applied to separate PCR products of
phosphatase assay kit (Abcam; Cambridge, UK), with p‐nitrophe‐
the 16S rRNA genes of the overall microbial community composi‐
nyl phosphate, which turns yellow when dephosphorylated by al‐
tion. The general bacterial primers 338F‐GC and 518R were used
kaline phosphatase, used as a phosphatase substrate. Absorbance
for PCR. Gels had a denaturating gradient ranging from 45% to
values were measured at 405 nm. The activity was expressed as
60% and were run using a DCode™ Universal Mutation Detection
µmol min−1 g−1 mucosa.
System (Bio‐Rad, Nazareth, Belgium). Data analysis was carried out using GelCompar version 6.6 (Applied Maths, Sint‐Martens‐Latem, Belgium), enabling the calculation of richness and Shannon diver‐ sity indices. qPCR was applied for quantification of Lactobacilli, Clostridium cluster
XIVa
(C. coccoides–E. rectale group)
and
Enterobacteriaceae. The PCR mixture (total volume of 15 µl) con‐
2.7 | RT‐qPCR to determine transcript of tight junction proteins, pro‐inflammatory cytokines, Toll‐ like receptor 4 and intestinal alkaline phosphatase Relative gene expressions of intestinal tight junctions, occludin
tained 5 µl template (between 1 and 10 µg/µl DNA, corresponding
(OCLN), claudin‐1, claudin‐5, claudin‐7 (CLDN‐1,‐5,‐7) and zonula oc‐
to a 1:100 dilution of the original DNA extract) and 10 µl QPCR
cludens 1 (ZO‐1), and inflammatory cytokines, tumour necrosis factor
SYBR Green ROX Mix (Westburg), forward and reverse primer
alpha (TNF‐α), interferon gamma (IFN‐γ), interleukin‐1 beta (IL‐1β), as
(10 µM each). Negative controls for each batch of samples included
well as Toll‐like receptor 4 (TLR4) and intestinal alkaline phosphatase
template consisting of qPCR water. Samples were incubated in a
(IAP), were quantified using RT‐qPCR. Primers used in this study
StepOnePlus real‐time PCR device (Applied Biosystems) for 40
were designed using Primer 3 Plus using criteria described in Wang
cycles with the programmed profile as indicated in Supporting in‐
et al. (2016) and shown in Supporting information Table S2. The total
formation Table S1. Samples were checked for correct peaks in the
RNA of mucosal samples was extracted using the Bio‐Rad Aurum
melt curve. The standard curve in all of the different runs should
Total RNA Kit (Bio‐Rad Laboratories, Inc., Hercules, CA, USA), in‐
have an efficiency between 95% and 105%. If not, the run was re‐
cluding an on‐column DNase I treatment to remove genomic DNA
peated. Each sample was analysed in triplicate, and outliers (more
(gDNA). The concentration of extracted RNAs was measured by
than 1 C T difference) in triplicates were omitted. Resulting values
NanoDrop ND‐1000 (NanoDrop Technologies, Thermo Scientific,
were converted to copies/µl by multiplying with 102 , that is, by
Wilmington, DE, USA). RNA integrity was checked by analysing 1 μg
taking into account the dilution (there was a 1:100 dilution of DNA
RNA by 1% agarose gel electrophoresis (28S and 18S rRNA bands).
extract prior to the qPCR).
A minus‐RT control PCR was performed using primer for YWHAZ to verify the absence of gDNA. Subsequently, 1 μg of high‐quality
2.5 | Soluble and total Zn in intestinal contents and mucosa
DNA‐free RNA was reverse transcribed in the 20 μl reverse tran‐ scription reaction with the ImProm‐II cDNA Synthesis Kit (Promega, Madison, WI, USA), containing both oligo‐dT and random primers.
Soluble and total Zn in intestinal contents were determined as modi‐
The obtained cDNA was diluted 10 times with molecular‐grade
fied from Schlegel, Nys, and Jondreville (2010). For total Zn meas‐
water, and a control PCR using 2 μl cDNA was performed to verify
urement, freeze‐dried contents were ground and ashed at 500°C
the reverse transcription reaction. The RT‐qPCR was performed
for 4 hr in a muffle furnace. Next, approximately 1 g of ashes was
on the CFX96 Touch Real‐Time PCR Detection System (Bio‐Rad
solubilized in 5 ml 16 N nitric acid and 5 ml Milli‐Q, and digested for
Laboratories, Inc.). Briefly, 2 μl cDNA template, 5 μl 2X KAPA SYBR
2 hr. The digests were further diluted to 50 ml with Milli‐Q and fil‐
FAST qPCR Kit Master Mix (Kapa Biosystems, Inc., Wilmington, MA,
trated (45 μm). To determine the soluble Zn content, samples were
USA), 2 μl molecular‐grade water, 0.5 μl forward primer and 0.5 μl
rehydrated (approximately 0.3 g freeze‐dried content and 10 ml
reverse primer (5 μM) were added to a total volume of 10 μl. The
Milli‐Q) and stirred constantly at 38°C during 2 hr. Supernatants
amplification conditions were as follows: (a) enzyme activation and
were obtained by centrifugation for 1 h at 20°C and 18,000 g. Next,
initial denaturation (95°C for 3 min); (b) denaturation (95°C for 20 s)
the supernatants were filtrated (45 μm). Mucosal samples of distal SI
and annealing/extension and data acquisition (annealing tempera‐
stored at −80°C were equilibrated to room temperature. Five mL of
ture depending on primer for 40 s) repeated 40 cycles; and (c) dis‐
digestion acid (3:1:1; v/v/v; HNO3, HClO 4, H2SO 4, all concentrated)
sociation (melt curve analysis from 70 to 90°C with 0.5°C increment
was added to approximately 1 g of wet tissue and heated to 110°C
every 5 s). In this study, PCR amplification efficiencies were consist‐
for 2 hr (Eller & Cassinelli, 1994). The hotplate temperature was in‐
ently between 90% and 110%, estimated as described by Bustin et
creased to 250°C until approximately 0.5 ml remained. After cool‐
al. (2009). Gene‐specific amplification was verified by agarose gel
ing, 5 ml Milli‐Q was added. Finally, one drop of 16 N HNO3 was
electrophoresis and melting curve analysis. Efficiency was used to
added to samples (digests, supernatants and tissue extracts) prior to
convert the Cq value into raw data with the highest expressed sam‐
quantification by ICP‐AES.
ples (lowest Cq value) as a calibrator for the normalization of raw
|
5
WANG et al.
TA B L E 2 Effect of experimental diets fed to weaners (d0–15 post‐weaning) on performance, faeces score and number of diarrhoea days (n = 4)*
Treatment ZnO source
Conventional
Alternative
Supplementation (mg/kg)
110
2,400
110
220
SEM
Initial BW (kg)
6.17
6.01
6.02
5.94
0.268
0.962
Final BW (kg)
6.12
5.86
5.58
5.75
0.285
0.669
p
Period d0‐5
ADG (g/d)
−51
−31
−88
−39
16.3
0.293
ADFI (g/d)
80
123
50
118
16.5
0.441
F:G (g/g)
−1.70
1.31
5.83
−1.00
1.890
0.564
Period d5‐15 Final BW (kg)
7.67
8.28
6.80
7.85
0.434
0.110
ADG (g/d)
105b
255a
130 b
222a
23.9
0.020
ADFI (g/d)
256
357
254
317
19.9
0.091
F:G (g/g)
1.33
1.40
1.72
1.44
0.237
0.344
ADG (g/d)
55b
156a
55b
131a
19.1
0.040
Total period d0‐15
*
ADFI (g/d)
193
275
184
248
16.6
0.089
F:G (g/g)
1.51
1.80
2.28
1.94
0.218
0.407
Faeces score
1.6
1.2
1.8
1.4
0.13
0.425
Number of diarrhoea days
4.0
3.0
5.3
1.3
0.88
0.365
Values with different superscripts within a row are significantly different at p