Antimicrobial peptides (AMP) are evolutionarily conserved components of the innate immune response, found among all classes of life ranging from prokaryotes to humans, and a promising family of natural alternatives to antibiotic growth promoters (AGP) with vast diversity in terms of sources, structures, functionalities, antimicrobial spectra and modes of action(Reference Hancock and Lehrer1–Reference Keymanesh, Soltani and Sardari3). Currently, 1711 AMP have been identified from a wide range of organisms (http://aps.unmc.edu/AP/main.php; accessed February 2011). Recent research demonstrated that AMP could improve the growth performance of piglets(Reference Hong, Kim and Hwang4, Reference Tang, Yin and Zhang5), growing–finishing pigs(Reference Cromwell, Davis and Morrow6), chickens(Reference Bao, She and Liu7) and fishes(Reference Zhou, Wang and Li8), and the integrity of intestinal surfaces and intestinal mucosal response of chickens(Reference Bao, She and Liu7, Reference Liu, She and Wang9, Reference Wang, Ma and She10). It seems that AMP are suited to be alternatives to AGP, but the use of AMP from natural sources is not at present cost-effective. Those AMP consisting of a single polypeptide chain of common amino acids are well suited for large-scale production by recombinant expression methods using micro-organisms that are resistant to the produced AMP.
Cecropins A (K W K L F K K I E K V G Q N I R D G I I K A G P A V A V V G Q A T Q I A K) and D (W N P F K E L E K V G Q R V R D A V I S A G P A V A T V A Q A T A L A K) belong to the cecropin family of AMP. They were originally isolated from the insect Hyalophora cecropia, and exerted strong antibiotic activity against Gram-positive and Gram-negative bacteria at micromolar concentrations(Reference Hultmark, Engström and Bennich11). The chimeric peptide, cecropin A(1-11)-D(12-37) (K W K L F K K I E K V – G Q R V R D A V I S A G P A V A T V A Q A T A L A K), which has the first eleven residues from the N terminus of cecropin A and the last twenty-six residues from the C terminus of cecropin D, had greater antimicrobial activities than the parent peptides(Reference Christensen, Fink and Merrifield12, Reference Fink, Memifield and Baoman13). Because cecropins with an amidated C terminus have broad-spectrum properties(Reference Li, Merrifield and Boman14), cecropin A(1-11)-D(12-37) was amidated by adding Asn to its C-terminus to produce a more potent AMP, cecropin A(1-11)-D(12-37)-Asn (K W K L F K K I E K V – G Q R V R D A V I S A G P A V A T V A Q A T A L A K – N) (CADN), through the application of the recombinant DNA expressed in Pichia pastoris, using plasmid pPlCZα-A (Invitrogen) as the expression vector(Reference Huang, Huang and Wen15).
The purpose of the present study was to verify the in vivo antimicrobial activity of CADN and its effects on the growth performance, nutrient utilisation and intestine villus morphology of broilers, to evaluate the feasibility of CADN serving as a possible alternative to AGP in poultry diets.
Materials and methods
Preparation of cecropin A(1-11)-D(12-37)-Asn liquid sample
In brief, one colony of genetically engineered P. pastoris strain GS-CAD (Chinese patent no.: CN01107584·8) was selected from yeast-growing plate (1 % yeast extract+2 % peptone+2 % dextrose+2 % agar, pH 6·3–6·5) containing 500 μg zeocin/ml (stored at 4°C; Invitrogen), grown in 100 ml of sterile yeast-growing broth (1 % yeast extract+2 % peptone+2 % dextrose, pH 6·3–6·5) in 500 ml flasks and shaken (30°C, 200 rpm) for 24 h to obtain 100 ml of inoculum. The inoculum was added to 35 litres of sterile yeast fermentation broth (YPCT, 1 % yeast extract+2 % peptone+2 % maize steep powder+0·2 % trisodium phosphate, pH 6·3–6·5) in a bioreactor (BIOF-50L/S; Shanghai Gaoji Biology Engineering Company Limited) with 50 litres of working volume. The temperature was maintained at 30°C, vessel pressure to be 0·08–0·1 MPa and impeller speed set to 300 rpm. Medical grade oxygen was fed to the bioreactor to maintain dissolved oxygen concentration at 6–7 mg/l. The pH was adjusted to the desired point during fermentation by automatic feeding of 30 % (w/w) ammonium hydroxide. About 0·35 litres of 95 % methanol were fed in the logarithmic growth phase to induce the expression of the CADN gene for 72 h. When the fermentation was completed, saturated steam was passed through the fermentation vessel for 10 min to kill the yeast cells, and YPCT was centrifuged at 6000 g for 10 min and the supernatant fractions were harvested for the experimental use. Determination of in vitro antibacterial activity of the CADN liquid sample (CADNL) was performed as described previously(Reference Hultmark, Steiner and Rasmuson16). One unit of antibacterial activity was defined as the amount of AMP giving 50 % reduction of the absorbance at 570 nm compared with the control. The number of units was estimated by the following formula: 
$U = \sqrt {( A _{0} - A )/ A } $, where A is the absorbance in the sample and A 0 is the absorbance in the control. The in vitro antibacterial activity against Escherichia coli K12D31 of the CADNL used in the experiment was about 6000 U/ml.
Preparation of yeast fermentation broth free of cecropin A(1-11)-D(12-37)-Asn
In brief, one colony of wild-type (untransformed) P. pastoris strain GS 1151 was grown in YPCT, and YPCT was harvested according to the above procedures.
Diets
For the experiment, two basal diets meeting or exceeding the National Research Council(17) nutrient requirements for grower and finisher broilers were made, respectively. The chemical composition of the two basal diets is shown in Table 1. Broilers were fed with five experimental grower or finisher diets which were prepared by supplementing their basal diet with CADNL or YPCT free of CADN, and then produced in pellet form. The five grower diets were as follows: (1) basal diet+YPCT at 8 ml/kg, (2) basal diet+YPCT at 6 ml/kg+CADNL at 2 ml/kg, (3) basal diet+YPCT at 4 ml/kg+CADNL at 4 ml/kg, (4) basal diet+YPCT at 2 ml/kg+CADNL at 6 ml/kg, (5) basal diet+CADNL at 8 ml/kg. Similarly, the five finisher diets were prepared using the finisher basal diet and also produced in pellet form.
Table 1 Ingredients and chemical composition of the experimental basal diets*
CP, crude protein; AME, apparent metabolisable energy.
* Estimated from the data of the Chinese Animal Nutrition Association(39).
† Premix contains the following: vitamin premix 1000 mg; mineral premix 1000 mg; antioxidant butylhydroxyanisol 250 mg; maduramycin 5 mg; mould inhibitors 500 mg; wheat bran 245 mg. The mineral premix provided (per kg of diet): 80 mg Fe; 8 mg Cu; 40 mg Zn; 60 mg Mn; 0·35 mg I; 0·07 mg Co; 0·15 mg Se. The vitamin premix provided (per kg of diet): 4968 μg vitamin A palmitate; 45 μg cholecalciferol; 30 mg dl-α-tocopheryl acetate; 1 mg menadione; 1 mg thiamin; 10 mg riboflavin; 4 mg pyridoxine; 0·02 mg cyanocobalamin; 30 mg niacin; 12 mg pantothenic acid; 0·5 mg folic acid; 0·2 mg biotin. Vitamin supplement was over-adjusted for each vitamin so that after the feed pelleting process, vitamins would still be available.
Birds and management
The feeding experiment was carried out in the Poultry Research and Teaching Farm of South China Agricultural University, with all procedures approved by the Animal Care and Use Committee of South China Agricultural University, according to the Regulations for the Administration of Affairs Concerning Experimental Animals (approved by the State Council of the People's Republic of China on 31 October 1988 and promulgated by Decree no. 2 of the State Science and Technology Commission of the People's Republic of China on 14 November 1988). A total of 1500 fourteen-day-old Lingnan Yellow male chickens, 222 (sd 13) g of body weight, from the Guangdong Academy of Agricultural Sciences (Guangzhou, Guangdong Province, China), in a completely randomised design, were randomly allocated to five groups with five replicate cages of sixty birds each, fed ad libitum the five grower (15–28 d of age) diets, respectively, and subsequently fed the five finisher (29–42 d of age) diets, respectively. During the feeding period, the chicks were provided water ad libitum and raised in illumination (24 h/d) by overhead fluorescent lighting, the daily feed intake (F) per cage of birds and daily feed intake per bird were recorded, and mortality in each cage was recorded if there were birds dead.
At the end of each feeding period, average daily feed intake (ADF, g/bird per d), average daily weight gain (ADG, g/bird per d), ADF:ADG ratio (F:G, g/g) and average mortality (%) for each cage were calculated, respectively.
Metabolic experiment
During the feeding experiments, about 100 g excreta from each cage of chicks was collected daily by laying a stainless-steel plate (100 cm × 150 cm) directly beneath each cage from 08.00 to 10.00 hours, and immediately stored in a zip-lock plastic bag at − 20°C. At the end of each feeding period, the daily excreta collections were pooled for each cage of birds, oven-dried at 65°C to constant weight, weighed and ground to a consistent particle size through a 40 mesh screen before analysis. Samples of the diets and excreta were assayed for gross energy with a bomb calorimeter (HWR-15C; Shanli Detecting Instrument Factory under the Shanghai Testing Technology Institute), nitrogen contents with a Kjeltec auto-analyzer (Kjeltec 2300; Foss Tecator), diethyl ether extract contents with a Soxtec Avanti fat extraction system (Soxtec 2050; Foss Tecator), and acid-insoluble ash contents using the procedure described earlier(Reference Van Keulen and Young18) with a spectrophotometer (UV-2550/2450; Shimadzu).
The apparent metabolisable energy (AME), nitrogen retention (NR), apparent digestibility of diethyl ether extract (ADEE) for the diet consumed by each cage of chicks were calculated as follows:
$$\begin{eqnarray} AME\,(MJ/kg) = (GE_{d} - GE_{e}\times AIA_{d}/AIA_{e})/1000;NR\,(\%) = 100\times (N_{d} - N_{e}\times AIA_{d}/AIA_{e})/N_{d};ADEE\,(\%) = 100\times (EE_{d} - EE_{e}\times AIA_{d}/AIA_{e})/EEd, \end{eqnarray}$$
where GEd, is the gross energy of the diet (J/g); GEe, is the gross energy of the excreta (J/g); AIAd, is acid-insoluble ash contents (g/kg) of the diet; AIAe, is acid-insoluble ash contents (g/kg) of the excreta; Nd, is the nitrogen content of the diet (%); Ne, is the nitrogen content of the excreta (%); EEd, is the diethyl ether extract content of the diet (%); EEe, is the diethyl ether extract content of the excreta (%).
$$\begin{eqnarray} AME\,intake\,(kJ/bird\,per\,d) = ADF\times AME.\quad Dietary\,AME:apparent\,metabolisable\,protein\,ratio\,(MEMP)\quad \,(MJ/kg) = 10\ 000\hairsp AME/(NR\times crude\,protein). \end{eqnarray}$$
Bacteriological examinations
At the end of the feeding experiment, one chick from each replicate cage, with its body weight close to the mean, was chosen and killed by intravenous injection of sodium pentobarbital at 100 mg/kg, and then the jejunal and caecal digesta were collected for the measurement of aerobic bacteria count according to the procedures of the Health Protection Agency(19). Approximately 1 g of each digesta sample was mixed with 9 ml of sterile Peptone Saline Diluent (0·1 % peptone+0·85 % NaCl, pH 7·0), and homogenised for 3 min. From the initial 10− 1 dilution, ten-fold serial dilutions were subsequently made in sterile Peptone Saline Diluent. For each dilution, 0·1 ml was inoculated onto the centre of a sterile and dried Plate Count Agar (0·25 % yeast extract+0·5 % tryptone+0·1 % glucose+1·2 % agar, pH 7·0) plate. A sterile spreader was used to spread the inoculum over the surface of each plate as soon as possible. The plates were left on the bench for approximately 15 min to allow absorption of the inoculum into the agar, then inverted and placed in an incubator at 30°C for 48 h. The colonies on each plate were counted as soon as the plates were removed from the incubator. The quantity of bacteria was expressed as log10 (colony-forming units/g wet digesta) for each digesta sample.
Scanning electron microscopic examination
After the chick from the control or CADNL8 group was killed, the midpoint of the duodenum or jejunum was cut open longitudinally. From each intestine segment, three pieces of intestine inner surface tissue (about 1·0 cm × 1·0 cm) were removed, rinsed by gently flushing the debris from the tissue surface using cold 0·1 m-phosphate buffer (PB; pH 7·2) until visible feed particles and the mucus had been removed, and then immediately fixed in 4 % glutaraldehyde (in PB) overnight. After being washed in PB for 5 min, these specimens were post-fixed in 1 % osmium tetroxide (in PB) for 2 h, washed in PB for 5 min again, and dehydrated, respectively, in graded ethanol solutions (50, 70, 80, 90, 100 and 100 %) and amyl acetate ester for 15 min. Afterwards, the specimens were vacuum-dried in a freeze-drying apparatus (CHRIST ALPHA 1-4) until all observable surface solvent had been removed, mounted on aluminium stubs, coated with gold palladium for 30 min, and finally examined with a scanning electron microscope (Philips XL-30 ESEM) using the secondary electron mode at an accelerating voltage of 10·0 kV and magnifications 200 × . The clearest villus images were selected for comparison of morphological changes in villi between the two treatment groups.
Light microscopic examination
After the chicks from each cage were killed, the duodenum from the gizzard to the pancreatic and bile ducts and the ileum from the Meckel's diverticulum to the ileum–caecal junction were collected, and flushed with saline (0·90 % NaCl, w/v). A 2 cm length of each intestinal midpoint was fixed in 10 % buffered formalin (pH 7·0) and embedded with paraffin wax. From each intestinal segment, ten transverse sections were cut at a thickness of 5 μm and fixed on each slide. After the ten transverse sections were stained with haematoxylin–eosin (Guangzhou Xiuwei Trading Company Limited), two villi per section were randomly selected to determine villus height (from the tip to the base, excluding the intestinal crypt) and crypt depth, using an image processing and analysis system (version 1; Leica Imaging Systems Limited).
From the ten stained sections in each intestinal segment per bird, the five most clearly stained sections were selected for the calculation of villus morphometric parameters; therefore, the averages of ten villus height, crypt depth and villus height:crypt depth values were expressed as the mean values, respectively, for each bird. Thereafter, one most clearly stained duodenum section of the control or CADNL8 group was photographed using an Olympus Vanox-S Light Microscope (Olympus).
Statistical analysis
The SPSS statistical software package (version 17.0 for Windows; SPSS China) was used for all statistical analyses. Multiple comparisons among means were determined by one-way ANOVA followed by Duncan's test (when equal variances were assumed) or Tamhane's T 2 test (when equal variances were not assumed). P values < 0·05 were considered significant. The correlate-bivariate and regression-curve estimation models were used to examine the linear and quadratic correlation between two sets of parameters. When a correlation coefficient (r) was up to 0·70, the correlation was considered high; and when the coefficient of determination (R 2) was up to 0·8, the goodness of fit of the observed data to the regression equation was considered acceptable.
Results
The growth performance and nutrient utilisation of growers are shown in Table 2. There were no differences in initial body weight (P = 0·598), mortality (P = 0·294), AME intake (P = 0·149) or MEMP (P = 0·724) among the groups, but there were significant differences in ADG (P = 0·011), ADF (P < 0·001), F:G (P < 0·001), ADEE (P < 0·001), NR (P < 0·001) or AME (P < 0·001). CADNL had a quadratic effect on ADG (ADG = − 0·045 CADNL2+0·302 CADNL+21·303; r 0·933, R 2 0·870; with the optimum CADNL dose being 3·4 ml/kg), negative linear effects on ADF (r − 0·933) and F:G (r − 0·885), and positive linear effects on ADEE (r 0·999), NR (r 0·980) and AME (r 0·985). AME had a positive linear correlation with ADEE (r 0·980) or NR (r 0·994), whereas ADF had a negative linear correlation with dietary AME (r − 0·980) but no correlations with MEMP.
Table 2 Growth performance and nutrient utilisation parameters for broilers (14–49 d)(Mean values with their pooled standard errors (PSE), n 5 replicates per treatment)
CADNL, cecropin A(1-11)-D(12-37)-Asn liquid sample; EE, diethyl ether extract; AME, apparent metabolisable energy of the diet; AMEI, AME intake of chicks; MEMP, dietary AME:apparent metabolisable protein ratio.
a,b,c,d,e Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
* Included in the diets at different supplement doses of 0, 2, 4, 6 and 8 ml/kg, respectively
The growth performance and nutrient utilisation of finishers are also shown in Table 2. There were no significant differences in mortality (P = 0·221) and F:G (P = 0·464), but significant differences in terminal body weight (P = 0·005), ADG (P = 0·012), ADF (P = 0·028), dietary ADEE (P < 0·001), NR (P < 0·001), AME (P < 0·001), AME intake (P = 0·001) and MEMP (P < 0·001). CADNL had quadratic effects on terminal body weight (r 0·894), ADG (ADG = − 0·245 CADNL2+1·802 CADNL+70·383; r 0·886, R 2 0·785; with the optimum CADNL dose being 3·7 ml/kg), ADF (r 0·838) and F:G (r 0·902), and had positive linear effects on dietary ADEE (r 0·974), NR (r 0·979) and AME (r 0·987). AME had a positive linear correlation with ADEE (r 0·903) or NR (r 0·997), whereas ADF had no correlations with AME but a quadratic correlation with MEMP (r − 0·967).
The effects of CADN on the jejunal and caecal aerobic bacterial counts of broilers at 42 d of age are shown in Table 3. There were significant differences (P>0·05) in aerobic bacterial counts in the digesta in the jejunum or caecum among the groups, and CADNL had a negative linear correlation with aerobic bacterial counts in the jejunal (r − 0·928, P = 0·023) or caecal digesta (r − 0·971, P = 0·006).
Table 3 Effect of cecropin A(1-11)-D(12-37)-Asn (CADN) on the jejunal and caecal aerobic bacterial counts (log10 colony-forming units/g wet digesta) of broilers at 42 d of age (Mean values with their pooled standard errors (PSE), n 5 replicate chicks per treatment)
CADNL, CADN liquid sample.
a,b,c,d,e Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
* Included in the diets at different supplement doses of 0, 2, 4, 6 and 8 ml/kg, respectively.
Intestinal villus height, crypt depth and villus height:crypt depth ratio of each group are shown in Table 4. There were significant differences (P < 0·001) in villus height, crypt depth and villus height:crypt depth ratio of the duodenum or ileum among the groups. CADN had a positive linear effect on villus height (r 0·958, P = 0·010) or villus height:crypt depth ratio (r 0·904, P = 0·035), but a negative linear effect on crypt depth (r − 0·865, P = 0·058) of the duodenum; similarly, CADN had a positive linear effect on villus height (r 0·884, P = 0·046) or villus height:crypt depth ratio (r 0·916, P = 0·029), but a negative linear effect on crypt depth (r − 0·875, P = 0·052) of the ileum.
Table 4 Villus height, crypt depth and villus height:crypt depth ratio of the duodenum and ileum of broilers at 42 d of age (Mean values with their pooled standard errors (PSE), n 5 replicate chicks per treatment)
CADNL, CADNL, cecropin A(1-11)-D(12-37)-Asn liquid sample.
a,b,c,d Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
* Included in the diets at different supplement doses of 0, 2, 4, 6 and 8 ml/kg, respectively.
The comparisons of scanning electron microscope microphotographs of the duodenum and jejunum villi of broilers at 42 d of age between the control (CADNL0) and CADNL8 groups are shown in Fig. 1. The duodenum villi of the control group (Fig. 1(a)) were tongue-like, while those of CADNL8 (Fig. 1(b)) were leaf-like. As to the jejunum villi, the villi of the two groups were all tongue-like, but the villi of the control group appeared to have severe inflammation.
Fig. 1 Scanning electron microphotographs of the intestinal villi of Lingnan Yellow broilers at 42 d of age (at magnification 200 × and 10·0 kV; scale bar = 100 μm). (a) Duodenum villi of the control group; (b) duodenum villi of the cecropin A(1-11)-D(12-37)-Asn liquid (CADNL8) group; (c) jejunum villi of the control group; (D) jejunum villi of the CADNL8 group.
The difference in the histomorphology of villi in the duodenum between the control and CADNL8 groups is shown in Fig. 2. The duodenum villi of CADNL8 (Fig. 2(b)) were longer than the villi of the control group (Fig. 2(a)); furthermore, the villi of CADNL8 (Fig. 2(b)) anastomosed to form a netted leaf-like structure, which was not found in the intestines of the control group.
Fig. 2 Comparison of the villus histomorphology of broilers at 42 d of age. (a) Duodenum villi of the control group (40 × ); (b) anastomosed duodenum villi of the cecropin A(1-11)-D(12-37)-Asn liquid (CADNL8) group (40 × ).
Discussion
In the present study, CADN had a quadratic effect on ADG for both grower and finisher broilers, similar to the effect of salinomycin in pigs(Reference Lindemann, Kornegay and Stahly20) and lasalocid in beef cattle(Reference Bretschneider, Elizalde and Pérez21). For the growers, MEMP was not significantly different among the groups, letting dietary energy be the limiting factor for feed intake and broilers decrease their feed intakes with dietary AME increasing to satisfy their requirements for a certain amount of energy for growth, in accordance with a previous result(Reference Veldkamp, Kwakkel and Ferket22). For the finishers, MEMP was significantly different among the groups, letting broilers adjust their feed intakes in a quadratic response to varying MEMP to satisfy their requirements for optimum nutrient balance at a certain MEMP, in accordance with a previous result(Reference Mbajiorgu, Ng'ambi and Norris23).
Because CADN is a macromolecular peptide, and the level of the macromolecular peptide transferred to the portal blood of the animal was extremely low(Reference Wakabayashi, Kuwata and Yamauchi24), CADN should act only in the gastrointestinal tract (GIT). There was a widespread concern that AMP could be inactivated inside the GIT, because the GIT is rich in acids and proteinases. The present study indicated that CADN was effective in inhibiting bacterial growth in a dose-dependent manner in the GIT of the chickens, i.e. CADN could resist, to some extent, if not all, to the hydrolysis in the GIT of the chickens.
In the GIT, commensal bacteria may induce intestine inflammation by releasing metabolites, e.g. lipopolysaccharide from Gram-negative bacteria and lipoteichoic acid from Gram-positive bacteria(Reference Niewold25, Reference Sukhotnik, Yakirevich and Coran26). The metabolites may bind to a metabolite-binding protein and subsequently to CD14 to activate inflammatory responses through initiating Toll-like receptors 4 and 2 on macrophages(Reference Niewold25, Reference Scott, Rosenberger and Gold27). In the present study, inflammation was found to be evident in the jejunum villi of the control group, indicating that CADN had an anti-inflammatory effect. The anti-inflammatory effect of CADN might result indirectly from its antibacterial effect, but whether it would have the following effects, similar to other AMP, needs to be studied: (1) neutralising bacterial metabolites to prevent the formation of metabolite-binding protein complex(Reference Hancock and Scott28), (2) directly and selectively inhibiting the expression of genes encoding the pro-inflammatory molecules in macrophages(Reference Scott, Rosenberger and Gold27), (3) increasing the expression of programmed cell death-1 to reduce CD8+T cell-mediated enteric autoimmunity response and the subsequent inflammation(Reference Hancock and Scott28, Reference Reynoso, Elpek and Francisco29), (4) inducing caspase-independent apoptosis in the mucosal inflammatory cells(Reference Cerón, Contreras-Moreno and Puertollano30).
The commensal microbiota hydrolyses bile acids and their salts required for proper fat digestion and absorption, and competes with the host for the uptake of nutrients and energy, thus decreasing utilisation efficiency of fat, protein and energy(Reference Dibner and Richards31), so in the present study, while decreasing aerobic bacterial counts in the digesta in a dose-dependent manner, CADN increased ADEE, NR and AME of the diets in a dose-dependent manner. This was similar to the results of many previous studies. An increase in dietary potato antimicrobial protein reduced populations of total aerobic bacteria in the caecum (day 42) of broilers, and linearly improved retention of dry matter (day 20 to 21) and crude protein (day 20 to 21 and day 41 to 42) in the diet(Reference Ohh, Shinde and Jin32). Butyrate in the diet decreased E. coli numbers in the duodenum and improved the feed conversion of chickens(Reference Panda, Rama Rao and Raju33). CADN increasing the nutrient utilisation of the diets might also result from its anti-inflammatory effect, because the effect of AGP causes reduced accumulation of inflammatory cells in the mucosa, a thinner intestinal wall, thereby sparing energy for production(Reference Niewold25).
In the present study, while decreasing the bacterial counts in the digesta and improving the efficiency of nutrient utilisation, CADN improved intestinal villus structures in a dose-dependent manner. Because the commensal microbiota has adverse effects on intestinal villus structures by producing harmful metabolites(Reference Sukhotnik, Yakirevich and Coran26, Reference Reynoso, Elpek and Francisco29, Reference Dibner and Richards31, Reference Barszcz and Skomiał34) and increased available nutrients promote the growth of the intestinal villi(Reference Drozdowski and Thomson35–Reference Buwjoom, Yamauchi and Erikawa38), the mechanism for CADN improving villus structures might involve permitting villus growth by inhibiting bacterial growth and providing more available nutrients for villus growth. Whether CADN may directly promote villus growth should be studied with germ-free animals.
Conclusions
CADN inhibited gut bacterial growth, improved nutrient utilisation and intestine villus structures, and thus improved the growth of grower and finisher broilers with optimum CADNL doses being 3·4 and 3·7 ml/kg, respectively. CADN is therefore a possible alternative to AGP in broiler feeds, but its safety for consumers and the environment needs to be studied.
Acknowledgements
The authors would like to thank Guo-qing Huang from Shenzhen Yipeng Biological Engineering Company, Limited (Shenzhen, Guangdong Province, China) for supporting the preparation of CADNL and YPCT free of CADN. We also thank the referees for their constructive and helpful comments, especially suggestions concerning the interpretation of the data and discussion on the manuscript. This study received no specific grant from any funding agencies in the public, commercial or not-for-profit sectors. J.-G. H. is the guarantor of the article. L.-F. W. is the principal investigator of the study. Both authors participated in the design, execution and data interpretation of the study and contributed to the various drafts of the manuscript. The authors approved the final draft of the manuscript. Neither of the authors has any conflicts of interests.
