Effect of divergence in residual feed intake on expression of lipid metabolism-related genes in the liver of meat-type ducks

Abstract

Lipid metabolism is considered one of the important factors affecting residual feed intake (RFI). However, the relationship between RFI and expression of lipid metabolism-related genes is unknown in meat-type ducks. To address this issue, a total of 1,000 male meat-type ducks with similar body weight were randomly selected to measure body weight gain and feed intake from 21 to 42 d of age to estimate RFI. The 8 greatest- (high RFI [HRFI]) and lowest- (low RFI [LRFI]) ranking birds were then selected for the present study. Relative expressions of key genes, namely sirtuin 1 (Sirt1), forkhead box O1 (Foxo1), peroxisome proliferator-activated receptor gamma (PPARγ), sterol regulatory element-binding transcription factor 1c (SREBP-1c), fas cell surface death receptor (FAS), acetyl-CoA carboxylase alpha (ACC), carnitine palmitoyltransferase 1A (CPT1A), and acyl-CoA oxidase 1 (ACOX1), were then determined in the HRFI and LRFI ducks by quantitative PCR. The results showed that RFI, feed conversion ratio (FCR), and average daily feed intake (ADFI) were significantly lower (P < 0.05) in LRFI ducks than in HRFI ducks. In addition, expression of Sirt1, Foxo1, CPT1A, and ACOX1 were significantly higher in LRFI ducks than in HRFI ducks (P < 0.05), whereas PPARγ and FAS expression levels were significantly lower in LRFI ducks than in HRFI ducks (P < 0.01). Correlation analysis showed that Sirt1, CPT1A, and ACOX1 expressions were significantly negatively correlated with FCR (r = −0.81 to −0.93; P < 0.01), whereas PPARγ and FAS expressions were significantly positively correlated with FCR (r = 0.74 to 0.87; P < 0.01). PPARγ expression was significantly positively correlated with RFI (r = 0.83; P < 0.01), whereas CPT1A and ACOX1 expressions were significantly negatively correlated with RFI (r = −0.84 to −0.89; P < 0.01). Sirt1 mRNA expression was positively correlated with Foxo1, CPT1A, and ACOX1 mRNA expression (r = 0.78 to 0.92; P < 0.01). Association of Foxo1 with CPT1A and ACOX1 was positive (r = 0.88 to 0.96; P < 0.01). These results suggest that genes related to fatty acid oxidation are upregulated in the liver of ducks with high feed efficiency, while genes associated with lipid synthesis are downregulated. Furthermore, the inclusion of lipid metabolism-related genes in future breeding programs might be beneficial for selecting ducks with greater feed efficiency phenotype.

INTRODUCTION

The demand for meat rich in protein is increasing concomitant with the ever-increasing global population. Currently, poultry meat production is one of the most intensive types of meat production (Sell-Kubiak et al., 2017). However, feed cost represents the major cost of modern poultry production (Liu et al., 2017). Therefore, one of the main objectives in poultry breeding is improving feed efficiency to decrease the feed cost for growth and production (Connor et al., 2013). Residual feed intake (RFI), first proposed by Koch et al. (1963), was widely used to measure feed efficiency in animal production (Aggrey et al., 2010). It was initially presented as an efficiency variable and defined as the difference between the actual feed intake and expected feed intake based on weight gain and average animal weight (Fonseca et al., 2015). Moreover, RFI is a moderate heritability trait in meat-type chickens (Aggrey et al., 2010) and Peking ducks (Zhang et al., 2017b), and it is considered an important indicator for measuring the efficiency of feed utilization (Smith et al., 2010).

RFI is affected by several key biological events, such as adipose tissue growth, extracellular matrix formation, inflammatory and immune responses, and lipid metabolism (Horodyska et al., 2019). Of these, lipid metabolism is an important biological event involved in the regulation of feed efficiency (Lee et al., 2015). A recent whole genome resequencing showed that RFI-related genes in Beijing-You chickens were enriched in lipid and carbohydrate metabolism (Liu et al., 2018). Furthermore, it was reported that genes related to jejunal lipid uptake, muscular lipid utilization, and intestinal bile salt transport were potential mediators driving feed efficiency in broiler chickens (Reyer et al., 2018). At the transcriptome level, fatty acid metabolism was found to be a functional category; pigs with high feed efficiency demonstrated the downregulation of genes of this pathway in adipose tissue (Lkhagvadorj et al., 2010). Chickens with high feed efficiency show higher expression of triglyceride hydrolysis and cholesterol transport genes and lower expression of lipid synthesis genes (Zhuo et al., 2015). Therefore, our hypothesis was that genes related to lipid metabolism might affect the regulation of feed efficiency traits in meat-type ducks.

The liver is the most important target organ for investigating lipid metabolism in birds (Honda et al., 2016), and it plays key roles in the utilization of absorbed nutrients, including the storage, de novo synthesis, and recycling of nutritional components (Reyer et al., 2017). Moreover, the liver of avian species accounts for 95% of de novo fatty acid synthesis (Huang et al., 2017). Nevertheless, the understanding of the molecular regulation of key genes associated with lipid metabolism in the livers of meat-type ducks with divergent feed efficiency is limited. Selection programs based on combinations of gene markers associated with feed efficiency traits and current traditional methods might provide an important method for genetic improvement in feed efficiency breeding (Lee et al., 2015). Therefore, the objective of the present study was to investigate the associations between the expression of key genes related to lipid metabolism in the liver and feed efficiency traits in meat-type ducks. Our results may provide important gene markers for feed efficiency traits and contribute to elucidating the relationships between genes related to lipid metabolism and feed efficiency traits in meat-type ducks.

MATERIALS AND METHODS

All the experimental procedures were conducted according to the Regulations and Guidelines for the Administration of Affairs Concerning Experimental Animals established by the Ministry of Science and Technology (Beijing, China, revised in 2004) and were approved by the Institutional Animal Care and Use Committee of Anhui Agricultural University (permit number: SYXK 2016-007)

Experimental Population and Animal Husbandry

All meat-type ducks used in the present study were obtained from a random mating population, provided by Huangshan Qiangying Duck Breeding Co., Ltd. (Anhui, China). The ducks were derived from genetically unrelated sources and selected as closed populations according to the feather color, growth, feed efficiency, and carcass traits within each generation. The experimental ducks were of the sixth generation and were pedigreed by mating 200 male with 1,000 female ducks in 2 hatches. All ducks were weighed individually, sexed, wing-banded on the day of hatching, and raised indoors. At the age of 1 d, ducks with similar body weight were selected and kept together for feeding until the age of 21 d. For the first 3 d after hatching, the ducks were subjected to continuous illumination (24 L: 0 D); then a lighting program of 20 L: 4D was used for them from 4 to 42 d of age. The ducks were raised on the floor for the first 3 wk. All ducks were reared in the same house with similar environmental conditions and management procedure in accordance with the company’s management instructions. At 21 d of age, a total of 1,000 male ducks with similar body weight were finally selected and transferred to individual cages (55 × 50 × 40 cm) with particular troughs used for recording individual feed intake. Feed and water were provided ad libitum during the entire experimental period. All ducks were reared in the same house and fed the same basal diet. The feed compositions and calculated nutrients are provided in Table 1.

Table 1.

Ingredients and calculated nutrient levels of the experimental diet

Item	1–14 d	15–42 d
Composition (%)
Corn	56.60	61.00
Wheat bran	11.00	9.40
Soybean meal	28.00	22.70
Soybean oil	0.00	2.70
Limestone	1.60	1.40
Calcium hydrogen phosphate	1.30	1.50
Sodium chloride	0.40	0.32
Lysine	0.50	0.45
Methionine	0.20	0.13
Threonine	0.10	0.10
Premix1	0.30	0.30
Total	100.0	100.0
Nutrient levels
Metabolic energy (MJ/kg)	12.00	12.13
Crude protein (%)	19.5	17.5
Calcium (%)	0.95	0.85
Total phosphorus (%)	0.60	0.60
Lysine (%)	1.28	1.10
Methionine (%)	0.48	0.40
Cysteine (%)	0.30	0.25

¹Premix supplied per kilogram of supplement (15–42 d): vitamin A 8,000 IU; vitamin D₃ 2,800 IU; vitamin E 20 IU; vitamin K₃ 2.0 mg; vitamin B₁ 1.5 mg; vitamin B₂ 12 mg; vitamin B₆ 2.5 mg; vitamin B₁₂ 0.02 mg; D-pantothenic acid 12.0 mg; biotin 0.15 mg; nicotinic acid 45.0 mg; folic acid 0.8 mg; choline chloride 800 mg; Fe 60.0 mg; Cu 10.0 mg; Zn 55.0 mg; Se 0.35 mg; I 0.15 mg; Cr 0.15 mg; Mn 80.0 mg.

Trait Measurement

The body weight of each duck was recorded at the beginning (21 d old) and the end (42 d old) of the experimental period. The average daily weight gain (ADG), average daily feed intake (ADFI), metabolic body weight (MBW^0.75), and feed conversion ratio (FCR) were measured throughout the experimental period. RFI (Aggrey et al., 2010) was estimated as follows using SAS version 9.4 (SAS Institute, Cary, NC):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{R}\mathrm{F}\mathrm{I}=\mathrm{A}\mathrm{D}\mathrm{F}\mathrm{I}-(b_0+b_1\times \mathrm{M}\mathrm{B}{\mathrm{W}}^{0.75}+b_2\times \mathrm{A}\mathrm{D}\mathrm{G})}\end{array}}$$

where b₀ is the intercept, and b₁ and b₂ are the partial regression coefficients of RFI on MBW^0.75 and ADG, respectively. At the end of the trial, 7 ducks were removed because of their incomplete phenotype or pedigree information. The RFIs for each of the remaining 993 were calculated.

Tissue Collection and Total RNA Extraction

In the present study, the 30 greatest RFI (inefficient; HRFI) and 30 lowest RFI (efficient; LRFI) ducks were chosen to prioritize samples because the average RFI rank was subjected to outlier or extreme values. Finally, 8 ducks with the LRFI and 8 ducks with the HRFI were randomly selected from each group to represent 2 distinct RFI group. After exsanguination, liver samples were immediately collected and stored in RNAlater (ThermoFisher Scientific, Waltham, MA) at 4 °C and subsequently transferred to the laboratory and stored at −80 °C. Total RNA was extracted using the Bizol RNA kit (Biomiga, San Diego, CA) and purified with RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Total RNA was quantified using a NanoDrop 2000 spectrophotometer (ThermoFisher Scientific) with all samples exhibiting a ratio of OD260/OD280 between 1.8 and 2.0 and OD260/230 between 1.8 and 2.2. RNA integrity was assessed by 1.2% agarose gel electrophoresis stained with ethidium bromide and an Agilent 2100 Bioanalyzer system (Agilent Technologies, Santa Clara, CA). RNA samples were selected for subsequent analyses according to OD260/OD280 ratio (1.9–2.1) and RNA integrity number (RIN ≥ 8.0).

Primers and Complementary DNA Synthesis

In this study, key genes related to lipid metabolism included sirtuin 1 (Sirt1), forkhead box O1 (Foxo1), peroxisome proliferator-activated receptor gamma (PPARγ), sterol regulatory element-binding transcription factor 1c (SREBP-1c), fas cell surface death receptor (FAS), acetyl-CoA carboxylase alpha (ACC), carnitine palmitoyltransferase 1A (CPT1A), and acyl-CoA oxidase 1 (ACOX1). The primers used in this study were designed based on mRNA sequences obtained from the GenBank database (https://www.ncbi.nlm.nih.gov). Specific primer pairs were designed for each target gene using Primer Premier 5.0 (Primer Premier, Palo Alto, CA) and UCSC In-Silico PCR to detect the sensitivity and specificity (http://genome.ucsc.edu/). All PCR reactions were conducted using intron-spanning primers with amplified efficiencies ranging from 90% to 100%, and R-square ranging from 0.992 to 0.998 as validated with standard curves. All primers were synthesized by Beijing Tsingke Biological Technology (Beijing, China). The primer sequences are presented in Table 2. First-strand complementary DNA (cDNA) was synthesized using the Primer Script RT Reagent kit (TaKaRa Bio, Otsu, Japan) according to the manufacturer’s instructions. Total RNA (1.0 μg) from each sample was reverse transcribed into cDNA. The converted cDNA was diluted to 60.0 ng/μL working aliquots and stored at −20 °C until analyzed.

Table 2.

Specific primers used for quantitative PCR

Gene1	Accession no.	Primer2 (5′ to 3′)	Efficiency (%)	R 2	Length (bp)
Sirt1	XM_027463311.1	F: CTTGTGGGAGTAGTAGCGAAAG R: TGGAGCATCCTCATCCTCTAT	100	0.997	110
Foxo1	XM_005014856.4	F: Acacagtgaaccccatgtca R: ttaacaggggcatacgggtt	98	0.993	124
PPARγ	NM_001310398.1	F: GAATGTCACACAATGCCATCAG R: AGATCAGCAGATTCAGGGTTTAG	99	0.996	112
SREBP-1c	JQ080310.1	F: CGAGTACATCCGCTTCCTG R: TCTTCTGCACGGCCATCCT	98	0.995	92
FAS	XM_027459847.1	F: tggaaccctactaagcagcc R: aagattgtccgccttcctga	97	0.992	110
ACC	XM_013093070.2	F: CGCTAAACCCCTGGAGATGA R: CAGACACAGGCGACAAAGAC	100	0.994	140
CPT1A	XM_027457811.1	F: TGGACACTGCAAAGGAGATAC R: CTCAGAGACCTCTCAATCACTTC	98	0.995	102
ACOX1	XM_027471372.1	F: tgcagagttctaaggcagct R: gtcacatccaagtcacgagc	99	0.996	123
GAPDH	XM_027449739.1	F: ccatcacagccacacagaag R: ggatgacttttcccacagcc	100	0.998	126

¹Sirtuin 1 (Sirt1), forkhead box O1 (Foxo1), peroxisome proliferator-activated receptor gamma (PPARγ), sterol regulatory element-binding transcription factor 1c (SREBP-1c), fas cell surface death receptor (FAS), acetyl-CoA carboxylase alpha (ACC), carnitine palmitoyltransferase 1A (CPT1A), acyl-CoA oxidase 1 (ACOX1), glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

²F = forward primer, R = reverse primer.

Quantitative PCR

Quantification of mRNAs was carried out by quantitative PCR using the ABI 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA). To minimize the variation, each sample was run in triplicate and no-template control was included on each plate. The reaction mix had a final volume of 20.0 μL and contained 1.0 μL cDNA, 1.0 μL gene-specific primer, 8.0 μL distilled/deionized H₂O, and 10.0 μL 2 × SYBR Green Master Mix (Applied Biosystems). Thermal cycling parameters were as follows: 95 °C for 5 min; 40 cycles of 95 °C for 15 s and 60 °C for 1 min; and a melting curve program of 95 °C for 15 s, 60 °C for 1 min, 95 °C for 15 s, and 60 °C for 15 s. Gene expression results were normalized using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an endogenous reference gene and calculated using the 2^−ΔΔCt method. We first tested the specificity and efficiency using β-actin and GAPDH as housekeeping genes and found that GAPDH was more stable and appropriate than β-actin for this study. Data are shown as fold change relative to baseline and determined as previously described by (Livak and Schmittgen, 2001).

Statistical Analysis

The phenotypic data of each duck, including feed efficiency and relevant traits, were tested by linear regression of body weight and ADFI against the day of experiment using the general linear model of SAS version 9.4 (SAS Institute, Cary, NC). For all parameters tested, bird served as the experimental unit. All data were examined for normality and homogeneity of variance by histograms, q-q plots, and formal statistical tests as part of the UNIVARIATE procedure of SAS version 9.4. Mean values were separated by least squares means using a Bonferroni adjustment with a significant level of α = 0.05. Differences in phenotypic data and gene expression between HRFI and LRFI ducks were analyzed using Student’s t-test of SAS version 9.4. Pearson’s correlation coefficients among feed efficiency traits, growth traits, and gene expression values were determined using PROC CORR of SAS version 9.4. All data are expressed as least squares means ± SE. The differences were considered statistically significant at P <0.05.

RESULTS

Basic Statistics of Phenotypic Traits

The differences in RFI and relevant traits between HRFI and LRFI are presented in Table 3. The ADFIs of HRFI and LRFI ducks were 273.21 and 262.03 g, respectively. The FCRs of HRFI and RFI groups were 2.08 and 1.89, respectively. The RFIs of HRFI and RFI ducks were 11.56 and -12.43, respectively. The HRFI ducks consumed 4.2% more feed than LRFI ducks did (P < 0.05). The FCR of HRFI group was 10% higher than that of LRFI group (P < 0.05). The RFI, FCR, and ADFI were significantly lower (P < 0.05) for LRFI ducks than for HRFI ducks. Nevertheless, no significant difference was observed in ADG or MBW^0.75 (P > 0.05) between the 2 groups.

Table 3.

Difference in residual feed intake and relevant traits between HRFI and LRFI ducks^1,2

Trait3	HRFI	LRFI
ADFI, g/d	273.21 ± 16.58a	262.03 ± 17.42b
ADG, g/d	131.42 ± 6.43	138.45 ± 8.52
MBW0.75, g/d	358.34 ± 15.12	346.55 ± 10.32
FCR, g/g	2.08 ± 0.03a	1.89 ± 0.02b
RFI, g/d	11.56 ± 5.66a	−12.43 ± 6.43b

¹Values are least squares means ± SEM.

² n = 8 (HRFI); n = 8 (LRFI).

³ADFI = average daily feed intake; ADG = average daily weight gain; MBW^0.75 = metabolic body weight; FCR = feed conversion ratio; RFI = residual feed intake.

^a,bMeans with different superscripts in the same row indicate a significant difference (P < 0.05).

Gene Expression

The relative gene expression of key genes related to lipid metabolism in the liver of HRFI and LRFI groups is shown in Fig. 1. The expressions of Sirt1, Foxo1, CPT1A, and ACOX1 were significantly higher in LRFI ducks than in HRFI ducks (P < 0.05), whereas PPARγ and FAS expression levels were significantly lower in LRFI ducks than in HRFI ducks (P < 0.01). SREBP-1c and ACC expression levels were not different between the 2 RFI groups (P > 0.05).

Figure 1.

Effect of residual feed intake phenotype on relative expression of lipid metabolism-related genes in the liver of meat-type ducks. (A) sirtuin 1 (Sirt1), (B) forkhead box O1 (Foxo1), (C) peroxisome proliferator-activated receptor gamma (PPARγ), (D) sterol regulatory element-binding transcription factor 1 (SREBP-1c), (E) fas cell surface death receptor (FAS), (F) acetyl-CoA carboxylase alpha (ACC), (G) carnitine palmitoyltransferase 1 (CPT1A), (H) acyl-CoA oxidase 1 (ACOX1).

*P < 0.05, **P < 0.01. Low RFI, n = 8; high RFI, n = 8.

Association of Gene Expression With Feed Efficiency Traits

The correlation coefficients of the expression of key genes and phenotypic traits are presented in Table 4. The expression levels of Sirt1, CPT1A, and ACOX1 were significantly negatively correlated with FCR (r = −0.81 to −0.93; P < 0.01), and CPT1A and ACOX1 expression levels were significantly negatively correlated with RFI (r = −0.84 to −0.89; P < 0.01). The expression levels of PPARγ and FAS were significantly positively correlated with FCR (r = 0.74 to 0.87; P < 0.01), and PPARγ was significantly positively correlated with RFI (r = 0.83; P < 0.01).

Table 4.

Associations of expression of genes related to lipid metabolism in the liver with phenotypic traits in meat-type ducks^1,2

Gene3	ADFI	ADG	MBW0.75	FCR	RFI
Sirt1	−0.39	0.62	−0.25	−0.81**	−0.72
Foxo1	−0.31	0.51	0.06	−0.57	−0.53
PPARγ	−0.76	−0.32	0.48	0.87**	0.83**
SREBP−1c	0.66	0.20	0.76	0.37	0.33
FAS	0.49	−0.44	−0.50	0.74**	0.65
ACC	0.06	0.18	0.07	−0.09	−0.04
CPT1A	−0.64	0.31	−0.51	−0.93**	−0.89**
ACOX1	−0.57	0.51	−0.45	−0.90**	−0.84**

¹ADFI = average daily feed intake; ADG = average daily weight gain; MBW^0.75 = metabolic body weight; FCR = feed conversion ratio; RFI = residual feed intake.

² n = 8 (HRFI); n = 8 (LRFI).

³Sirtuin 1 (Sirt1), forkhead box O1 (Foxo1), peroxisome proliferator-activated receptor gamma (PPARγ), sterol regulatory element-binding transcription factor 1c (SREBP-1c), fas cell surface death receptor (FAS), acetyl-CoA carboxylase alpha (ACC), carnitine palmitoyltransferase 1A (CPT1A), acyl-CoA oxidase 1 (ACOX1).

*P < 0.05, **P < 0.01.

Correlation of Expression of Genes Related to Lipid Metabolism

The correlation coefficients of the expression of key genes in the liver of meat-type ducks are presented in Table 5. The expression level of Sirt1 was significantly positively correlated with that of Foxo1, CPT1A, and ACOX1 (r = 0.78 to 0.92; P < 0.01), whereas it was significantly negatively correlated with that of FAS (r = −0.76; P < 0.05). Foxo1 expression was significantly positively correlated with CPT1A and ACOX1 expression (r = 0.88 to 0.96; P < 0.01), whereas it was significantly negatively correlated with SREBP-1c and FAS expression (r = −0.79 to −0.80; P < 0.05). PPARγ expression was significantly positively correlated with FAS expression (r = 0.83; P < 0.05), whereas it was significantly negatively correlated with ACC expression (r = −0.61; P < 0.05). CPT1A expression was significantly negatively correlated with SREBP-1c and FAS expression (r = −0.82 to −0.97; P < 0.01), whereas it was significantly positively correlated with ACC and ACOX1 expression (r = 0.68 to 0.94; P < 0.01).

Table 5.

Correlation of expression of genes related to lipid metabolism in the liver of meat-type ducks¹

Gene2	Sirt1	Foxo1	PPARγ	SREBP-1c	FAS	ACC	CPT1A	ACOX1
Sirt1	1	0.83**	0.27	−0.63	−0.76*	0.39	0.92**	0.78**
Foxo1		1	0.42	−0.80*	−0.79*	−0.06	0.96**	0.88**
PPARγ			1	0.53	0.83**	−0.61*	0.32	0.28
SREBP-1c				1	0.84	−0.54	−0.82**	−0.71
FAS					1	0.36	−0.97**	0.20
ACC						1	0.68**	0.42
CPT1A							1	0.94**
ACOX1								1

¹ n = 8 (LRFI); n = 8 (HRFI).

²Sirtuin 1 (Sirt1), forkhead box O1 (Foxo1), peroxisome proliferator-activated receptor gamma (PPARγ), sterol regulatory element-binding transcription factor 1c (SREBP-1c), fas cell surface death receptor (FAS), acetyl-CoA carboxylase alpha (ACC), carnitine palmitoyltransferase 1A (CPT1A), acyl-CoA oxidase 1 (ACOX1).

*P < 0.05, **P < 0.01.

DISCUSSION

Improving feed efficiency can potentially reduce the production cost and decrease nitrogen and phosphorus emissions, which contribute to environmental pollution (Saintilan et al., 2013). In the present study, HRFI ducks consumed 4.2% more feed than did LRFI ducks, and the FCR and RFI of HRFI ducks were significantly higher than those of LRFI ducks. The results are consistent with previous studies in chickens (Zhuo et al., 2015) and mule ducks (Drouilhet et al., 2016). These data imply that animals with low feed efficiency (HRFI) have a greater cost of production. In addition, molecular breeding can markedly shorten the breeding period and increase breeding efficiency; therefore, improving feed efficiency using gene markers in future breeding programs will significantly improve the profitability of meat-type ducks (Ye et al., 2017).

The liver is one of the most important metabolic organs, because it can transform lipid, glucose, and various proteins into ready-to-use energetic substrates. Moreover, the liver is an important organ for maintaining the balance of lipid metabolism, including lipid transport, synthesis, and catabolism (Li and Wu, 2018). In addition, similar to humans, birds use the liver rather than adipose tissue as the principal site of de novo lipid synthesis (Huang et al., 2017). This continuous turnover of lipid has a complex regulation. Various signaling pathways are involved in the hepatic lipid metabolism such as the Sirt1/Foxo1 pathway (Li and Wu, 2018) and the AMPK/ACC/CPT1 pathway (Cheng et al., 2018). Besides, numerous recent studies have indicated that transcription factors related to lipid metabolism play vital roles in the regulation of feed efficiency (Weber et al., 2016; Reyer et al., 2017; Asher et al., 2018). Thus, the current study examined the contribution of key genes related to lipid metabolism in the liver to the variation in feed efficiency in meat-type ducks.

A previous study has indicated that the genes of the Sirt1/Foxo1 pathway, including Sirt1, Foxo1, and SREBP-1c, are involved in lipid metabolism (Li and Wu, 2018). Sirt1 is a NAD⁺-dependent histone deacetylase that plays important roles in the regulation of energy metabolism and stress resistance (Rodgers et al., 2005). Recent studies have shown that hepatic Sirt1 plays a key role in triglyceride metabolism in the liver and controls hepatic lipid metabolism by mediating fatty acid β-oxidation (Rodgers and Puigserver, 2007; Zhang et al., 2017a). Likewise, Foxo1 plays a crucial role in regulating the level of gluconeogenesis in the liver. In the current study, the expression levels of Sirt1 and Foxo1 were significantly higher in LRFI ducks than in HRFI ducks, and Sirt1 was significantly negatively correlated with FCR. The activation of Sirt1 has been reported to promote lipolysis and the upregulation of fatty acid oxidation genes in liver (Yoo et al., 2017). Similarly, Foxo1 has been reported to promote lipolysis and inhibit adipocyte differentiation (Li et al., 2017). Moreover, Sirt1 controls lipolysis through the deacetylation of Foxo1, further mediating the expression of adipose triglyceride lipase (Chakrabarti et al., 2011). Taken together, the higher expression of Sirt1 and Foxo1 in the high feed efficiency (LRFI group) ducks reflects that the high feed efficiency trait is accompanied by increasing fatty acid oxidation and an upregulation of fatty acid oxidation genes in the liver. In addition, it was reported that the activation of Sirt1 induced the expression of Foxo1 (Li and Wu, 2018). A previous study indicated that Sirt1 could enable Foxo1 transcription in multiple systems, and a positive feedback mechanism was detected in Foxo1-dependent Sirt1 transcription (Xiong et al., 2011). These findings corresponded with our finding that Sirt1 expression was significantly positively correlated with Foxo1 expression.

Besides Sirt1 and Foxo1, CPT1A and ACOX1 also play crucial roles in the regulation of lipolysis in the liver. CPT1A is a rate-limiting enzyme in the transport of long-chain fatty acids for β-oxidation (Shi et al., 2016). It has been reported that fatty acid oxidation can be suppressed by downregulating the expression of CPT1A (Deminice et al., 2011). Likewise, ACOX1 is another key enzyme in fatty acid oxidation (Li et al., 2015). CPT1A and ACOX1 can promote fatty acid oxidation and reduce fat deposition (Lee et al., 2011). In the present study, we found that expression levels of CPT1A and ACOX1 were significantly higher in LRFI ducks than in HRFI ducks, and CPT1A and ACOX1 expressions were significantly negatively correlated with FCR and RFI. These data indicated that genes related to fatty acid oxidation might be upregulated in ducks with high feed efficiency. This result was consistent with those of a previous study that showed that genes involved in fatty acid oxidation were promoted in adipose tissues in LRFI pigs than in HRFI pigs (Gondret et al., 2017). Moreover, an early transcriptomic analysis in chickens showed that LRFI chickens had higher lipid oxidation compared to HRFI chickens through the upregulation of genes related to lipid transport (Lee et al., 2015).

A previous study showed that the knockdown of Sirt1 can suppress the expression of CPT1A in isolated hepatocytes and in rat liver (Thakran et al., 2013). Another study reported that Sirt1 expression was positively correlated with ACOX1 expression in colorectal cancer cells and in xenografts (Sun et al., 2017). In agreement with these finding, Sirt1 was positivity associated with the mRNA expression of CPT1A and ACOX1 in the current study. However, a previous research demonstrated that the inhibition of Sirt1 decreased phosphorylation of Foxo1, which not only upregulated SREBP-1c gene and increased fatty acid synthesis, but also downregulated CPT1A genes and decreased fatty acid β-oxidation (Sun et al., 2013). This finding was consistent with the current result that Foxo1 was positively associated with the CPT1A and negatively associated with SREBP-1c.

Sterol regulatory element-binding proteins (SREBPs) belong to the family of transcription factors of the basic-helix-loop-helix leucine zipper (bHLHLZ), which mainly regulate the biosynthesis of fatty acids, triglycerides, and cholesterol (Horton et al., 2002). SREBP-1c is the major isoform expressed in the liver and is involved in cholesterol synthesis (Ponziani et al., 2015). The activity of SREBP-1c suppresses fatty acid oxidation and Foxo1 expression (Hu et al., 2012). Moreover, it is well documented that hyperinsulinemia increases SREBP-1c expression and further promotes hepatic de novo lipogenesis by inhibiting Foxo1 (Ido-Kitamura et al., 2012). These studies might explain our finding that SREBP-1c expression was negatively correlated with Foxo1 expression.

PPARγ is a master transcriptional regulator of adipogenesis. There is an increasing evidence that PPARγ enhances fatty acid storage by upregulating lipogenic genes, including lipid transporters and enzymes in fatty acid and triglyceride synthesis (Honda et al., 2016). In the current study, PPARγ expression was significantly lower in LRFI ducks than in HRFI ducks, and it was significantly positively correlated with FCR and RFI. Hepatic-specific PPARγ-knockout was found to improve glucose tolerance and decrease the upregulation of lipogenic, β-oxidation, and gluconeogenic genes in high-fat diet-induced mice (Morán-Salvador et al., 2011). Similarly, a previous study demonstrated the higher expression of triglyceride hydrolysis and cholesterol transport genes and lower expression of lipid synthesis genes were detected in chickens with high feed efficiency when compared with chickens having low feed efficiency (Zhuo et al., 2015). Our present results might be because low feed efficiency increased the expression of genes related to lipid synthesis in the liver. Moreover, it was reported that Sirt1 inhibited PPARγ-induced activity and expression of ACC gene (Qu et al., 2016). This agreed with our result, which showed negative correlation between PPARγ and ACC.

A recent study reported that FAS played a crucial role in the regulation of hepatic lipid metabolism, mitochondrial function, and fatty acid oxidation (Item et al., 2017). Our results showed that FAS expression was significantly lower in LRFI ducks than in HRFI ducks, and that it was significantly positively correlated with FCR. It is established that hepatic FAS overexpression promotes hepatic lipid accumulation and insulin resistance by compromising fatty acid oxidation, mitochondrial respiration, and the abundance of mitochondrial respiratory complexes in mice, while hepatocyte-specific ablation of FAS improves mitochondrial function and glucose tolerance (Item et al., 2017). In line with these findings, one study indicated that enhancing mitochondrial function contributed to improved feed efficiency in cattle (Connor et al., 2010). Mitochondrial function and mitochondrial respiratory complexes were found to be greater in high feed efficiency poultry than in low feed efficiency poultry (Iqbal et al., 2005). Therefore, the lower expression of FAS in LRFI ducks might be attributed to enhanced mitochondrial function or fatty acid oxidation in the livers of birds with high feed efficiency.

In summary, this is the first study to investigate the effect of divergence in RFI on the expression of lipid metabolism-related genes in the liver of meat-type ducks. Our findings suggest that genes related to fatty acid oxidation are upregulated in the liver of high feed efficiency animals, while genes associated with lipid synthesis are downregulated. In addition, this study identifies several novel targets that might regulate feed efficiency and underlie the variation in RFI observed in meat-type ducks. Furthermore, there is a great need to elucidate the molecular mechanisms and physiological processes of these novel targets on feed efficiency traits in meat-type ducks. After appropriate verification, it is possible that the inclusion of these novel targets in future breeding programs will be beneficial in the selection of ducks with superior feed efficiency phenotype.