Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Physiology

Residual feed intake studies in Angus-sired cattle reveal a potential role for hypothalamic gene expression in regulating feed efficiency1,2

 

This article in JAS

  1. Vol. 92 No. 2, p. 549-560
     
    Received: Aug 13, 2013
    Accepted: Dec 16, 2013
    Published: November 24, 2014


    3 Corresponding author(s): tdb0006@auburn.edu
 View
 Download
 Share

doi:10.2527/jas.2013-7019
  1. S. D. Perkins*,
  2. C. N. Key*,
  3. C. F. Garrett*,
  4. C. D. Foradori,
  5. C. L. Bratcher*,
  6. L. A. Kriese-Anderson* and
  7. T. D. Brandebourg 3
  1. Department of Animal Sciences, Auburn University, Auburn, AL 36849
    Department of Anatomy, Physiology and Pharmacology, Auburn University, Auburn, AL 36849

Abstract

Mechanisms underlying variation in residual feed intake (RFI), a heritable feed efficiency measure, are poorly understood while the relationship between RFI and meat quality is uncertain. To address these issues, 2 divergent cohorts consisting of High (HRFI) and Low (LRFI) RFI individuals were created by assessing RFI in 48 Angus-sired steers during a 70 d feeding trial to identify steers with divergent RFI. The association of RFI with indices of meat quality and expression of genes within hypothalamic and adipose tissue was then determined in LRFI and HRFI steers. While on test, feed intake was recorded daily with BW and hip heights recorded at 14 d intervals. Ultrasound measurements of rib eye area (REA) and backfat (BF) were recorded initially and before harvest. Carcass and growth data were analyzed using a mixed model with RFI level (LRFI, HRFI) as the independent variable. The least-square means (lsmeans) for RFI were –1.25 and 1.51 for the LRFI and HRFI cohorts (P < .0001). Dry matter intake was higher for the HRFI individuals versus the LRFI steers (P < .0001) while on test BW gain was not different between the 2 groups (P < 0.73). There were no differences detected in marbling score (P < 0.93), BF (P < 0.61), REA (P < 0.15), yield grade (P < 0.85) or objective Hunter color measures between LRFI and HRFI steers indicating that there was no relationship between RFI and meat quality. Neuropeptide-Y (NPY), relaxin-3 (RLN3), melanocortin 4 receptor (MC4R), and GnRH mRNA expression was 64%, 59%, 58%, 86% lower (P < 0.05), respectively, while gonadotropin inhibiting hormone (GnIH) and pro-opiomelanocortin (POMC) mRNA expression was 198% and 350% higher (P < 0.01) in the arcuate nucleus of LRFI steers. Expression of agouti-related protein (AGRP), relaxin/insulin-like family peptide receptor 1 (RXFP1), and melanocortin 3 receptor mRNA was similar between LRFI and HRFI animals. Pituitary expression of FSHβ (P < 0.03) and LHβ (P < 0.01) was correlated to hypothalamic GnRH levels suggesting that changes in gene expression within the arcuate nucleus had functional consequences. Leptin mRNA expression was 245% higher in the adipose tissue of LRFI steers consistent with lower levels of NPY and higher expression of POMC in their hypothalami. These data support the hypothesis that differences in hypothalamic neuropeptide gene expression underlie variation in feed efficiency in steers while the gonadotropin axis may also influence feed efficiency.



INTRODUCTION

Production efficiency needs to improve in beef cattle to meet the projected food needs of a growing world population and to strengthen a production chain that forms the critical economic foundation for many communities (Archer et al., 2004). However, such efforts are hampered by a lack of knowledge concerning the molecular mechanisms regulating feed efficiency in cattle. Residual feed intake (RFI) is a heritable feed efficiency measure that allows cattle to be ranked based on individual variation in feed intake that is independent of growth rate and other production traits (Herd and Arthur, 2009). Thus, RFI can be used experimentally to study the regulation of feed efficiency.

Several hypothalamic neuropeptides control feed intake and influence metabolic rate in mammals and thus represent attractive candidates that regulate feed efficiency (Berthoud, 2002; Elmquist et al., 2005). Neuropeptide-Y (NPY) and agouti-related protein (AGRP) stimulate feeding while pro-opiomelanocortin (POMC)-derived α-melanocortin stimulating hormone (α-MSH) inhibits feeding behavior (Cowley, 2003; Cota et al., 2006). Circulating hormones such as insulin, adipose tissue-derived leptin and gut-derived ghrelin reciprocally regulate NPY/AGRP and α-MSH neurons with leptin acting to decrease feed intake and ghrelin stimulating feed intake (Arnold et al., 2006).

Neuropeptides regulating the reproductive axis potentially regulate feed efficiency as well. For instance recent studies suggest that GnRH and gonadotropin-inhibitory hormone (GnIH) represent a molecular switch between reproduction, energy balance, and feeding (Johnson et al., 2007; Qi et al., 2009; Clarke et al., 2012).

The objective of the current study was to determine the association of RFI with indices of carcass merit and the expression of genes within the hypothalamus and adipose tissue that have been implicated in regulating feed intake and energy balance.


MATERIALS AND METHODS

All procedures involving animals were approved by the Auburn University Animal Care and Use Committee.

Animals and Design

Forty-eight back-grounded Angus-sired steer calves averaging 315 ± 16 d of age were purchased from the Auburn University E.V. Smith Experiment Station Fall 2009 calf herd and were transported to the Auburn University Beef Cattle Evaluation Center where they remained for the duration of the project between September 1, 2010 and February 1, 2011. Weights and hip heights were recorded on each steer on arrival. Steers were stratified based on contemporary group, hip height, and BW into 4 pens (12 hd/pen) to minimize social dominance. Each pen of cattle had indoor and outdoor access with a capacity of 12 cattle per pen. Pens were 6.1 × 9.1 m inside and 18.3 × 92.7 m outside. The outside portion of each pen was 18.6 m at the widest point by 92.7 m long and divided into three 6.2 m strips. Steers were allowed access to a different strip weekly to maintain groundcover (common bermudagrass: Cynodon dactyloncl) which served to minimize erosion and promote hoof health. Each pen contained 12 Calan Gates (American Calan, Northwood, NH) to measure individual feed intake and cattle had continuous access to automatic water troughs.

To determine the association of RFI with indices of carcass merit and the expression of genes within the hypothalamus and adipose tissue that have been implicated in regulating feed intake and energy balance, calves were trained to the Calan Gate system during a 21-d acclimation period during which they were fed a corn-based diet at 2% of their BW and hay (bermudagrass and bahiagrass mix). Steers were fed twice daily. After the initial 21 d warm up period, hay access was removed and steers were hand fed ad libitum amounts of the corn-based diet. Steers were weighed and measured for hip height on 2 consecutive days to establish on-test BW and hip heights and were subsequently weighed every 2 wk throughout the test period. An initial real-time ultrasound was performed on all steers to determine on test ultrasound 12th rib and rump fat thickness, percentage intramuscular fat, and longissimus dorsi area with subsequent ultrasound measurements taken monthly until a final set of ultrasound data were recorded 24 h before harvest. All ultrasound data were collected by the same Ultrasound Guidelines Council certified technician using an Aloka 500 (Aloka America, Wallingford, CT) with a 17-cm transducer using CUP Lab image capture software.

Once trained to the Calan Gate system, steers underwent a 70 d feeding trial conducted using Beef Improvement Federation Guidelines for measuring RFI in cattle (BIF, 2010) to assess daily feed intake and RFI. The diet consisted of sorghum-sudan silage with a concentrate top dress (concentrate and minerals). The diet was fed ad libitum for the duration of the project and chemical composition was assessed by proximate analysis. At the end of the 70 d test period, RFI was calculated for each steer. The best 8 and worst 6 steers based on RFI were retained until reaching slaughter BW with the intermediary 32 steers being marketed. At the end of the feeding period, retained cattle were transported to Auburn University Lambert Powell Meats Laboratory where they were humanely harvested under USDA-FSIS inspection. Adipose hypothalamic tissues were collected to facilitate gene expression studies as detailed below. Carcasses were evaluated for quality and yield grade as well as color and pH.

Chemical Analysis of Forage

One kg silage and top dress samples were randomly sampled and placed in collection bags and evaluated in duplicate. Samples were then dried for 48 h at 55°C in a convection oven (model 420, NAPCO, Winchester, PA) and weighed before and after drying (equilibrated to room temperature for 24 h) to determine DM. Dry samples were ground in a Thomas-Willey mill (model 4, Thomas Scientific) to pass through a 1-mm mesh screen, labeled, and placed in sealed plastic containers. Chemical analyses were conducted on ground samples to determine DM, ash, NDF, ADF, and CP. Methods described by AOAC (1998) were used to determine DM, ash, and CP. Nitrogen was determined using a Leco TruSpec (Leco Corporation, St. Joseph, MI) and multiplied by 6.25 to estimate CP. Forage NDF and ADF were determined on samples according to Van Soest et al. (1991) using an Ankom200 fiber analyzer and Ankom F57 filter bags (Ankom Technology Corp., Fairport, NY). Chemical composition of the diet is shown in Table 1.


View Full Table | Close Full ViewTable 1.

Ingredients and chemical composition of diet

 
Diet ingredient, % (as-fed basis) Content
Sorghum-Sudan baylage 50.0
Concentrate 50.0
Composition baylage (DM basis)
    DM, % 61.1
    CP, % 11.7
    NDF, % 22.4
    TDN, %1 14.0
    ME, Mcal/kg2 2.13
Composition concentrate (DM basis)
    DM, % 87.2
    CP, % 13.2
    NDF, % 9.5
    TDN, %1 79.3
    ME, Mcal/kg2 3.10
1TDN = (105.2- 0.667 * NDF) * 0.88
2ME = (TDN * 1.01 *.04409)- 0.45

Carcass Analyses

Hot carcass weight was recorded after harvest, and carcasses were chilled at 2 ± 1°C for 24 h. At 24 h postmortem, carcass pH was recorded in the left side round using a pH Spear probe (Oakton Instruments, Vernon Hills, IL) and each carcass was split between the 12th and 13th ribs for evaluation of back fat (BF); ribeye area (REA); KPH; skeletal maturity; lean maturity; average maturity; and marbling. Longissimus muscles were evaluated for objective color measurements at the 12th and 13th rib interface using a Hunter Miniscan XE Plus (Hunter Lab, Reston, VA) for Hunter L*, a*, and b* values. The Miniscan was calibrated according to manufacturer’s recommendations and utilized a D65 light source, a 10° viewing angle, and a 35 mm viewing area.

Gene Expression Analysis

Upon exsanguination, subcutaneous adipose tissues were immediately collected, snap frozen in liquid nitrogen, and stored at –80°C until mRNA analysis. To facilitate this, a plug of subcutaneous adipose tissue and longissimus muscle was removed by cutting through the hide immediately on verifying that the animal had expired. Hypothalamic and pituitary tissues were collected from the severed head. Tissue samples were then processed immediately in a wet lab adjacent to the harvest floor. Processed samples were thoroughly snap-frozen by emersion in liquid nitrogen and buried in dry ice within 5–10 min of exsanguination. A large slice of medial basal hypothalamus containing the arcuate nucleus was collected and frozen intact. Arcuate nucleus tissue was later harvested from the frozen hypothalamic sample using a punch. Total RNA was extracted from frozen tissue using a 2-step purification protocol with total RNA first being extracted from whole tissue using RNAzol® RT (MRC, Inc, Cincinnati, OH) followed by purification using RNAeasy spin columns (QIAGEN, Inc., Valencia, CA) including a DNase treatment according to the manufacturers’ recommendations. Total RNA was quantified using a BioTek Synergy 4 plate reader utilizing the Take3 system (BioTek U.S., Winooski, VT) with all samples exhibiting an OD 260/280 between 1.8 and 2.0 and an OD 260/230 value between 1.8 and 2.2. Spectral scans ranging from 200 to 400 nm further verified sample purity as all RNA samples produced smooth curves exhibiting 1 peak at 260 nm. Total RNA integrity was assessed both visually by resolving 2 μg of RNA on a denaturing formaldehyde gel containing ethidium bromide and by determining an RNA Integrity Number (RIN) using an Agilent 2100 bioanalyzer (Agilent Technologies, Inc., Clara, CA). All samples demonstrated sharp ribosomal bands with a 28S to 18S ratio greater than 1 and RIN values greater than 7.0 and were thus judged intact and non-degraded. Total RNA was then reverse transcribed using 160 units of Superscript II reverse transcriptase (Promega Inc, Madison, WI) and 0.5 μg Oligo(dT)15 primers in a reaction volume of 20 μL also containing 1 μg RNA/reaction, 6 mM MgCl2, 0.5 mM each of dNTP, and 20 units RNasin with the reaction being performed in a single cycle with the following steps: heating for 5m at 65°C, annealing for 5m at 25°C, elongation for 50m at 42°C and heating for 15m at 70°C. The cDNA was subsequently stored at –80°C until used in gene expression assays. Real-time PCR was performed on the resultant cDNA using a Roche Lightcycler 480 Real-time PCR machine and LightCycler 480 SYBR Green I Master Mix (Roche Applied Science, Indianapolis, IN) according to manufacturer’s directions in a reaction volume of 20 μL consisting of 10 μL of master mix and 10 μL of H2O containing 50 ng of cDNA and.5 μM each of the forward and reverse primers under the following conditions: 1 pre-incubation step of 5m at 95°C followed by 45 cycles with each cycle consisting of a melt step of 10s at 95°C, and annealing step of 10s at 57°C, and an elongation step of 10s at 72°C. Each sample was run in 3 separate PCR runs with resulting Cp values averaged across values obtained from 3 separate plates. All PCR reactions were performed using intron-spanning primers under optimized conditions with primer efficiencies ranging between 90 and 100% (Table 2) as verified with standard curves. All primers were designed to have an optimal annealing temperature of 57°C. Product purity was assessed by melting curve analysis and expected amplicon sizes were verified on a 2% agarose gel stained with ethidium bromide. Values were normalized to eukaryotic translation initiation factor 3 subunit K (EIF3K) mRNA expression. The EIF3K mRNA levels represent an appropriate control as the efficiency of the EIF3K primers was 100% (Table 2) and EIF3K mRNA expression was not different between any groups tested (P < 0.93). Data are expressed as fold change relative to baseline and calculated according to Pfaffl (2010).


View Full Table | Close Full ViewTable 2.

Oligonucleotide polymerase chain reaction primers

 
Gene1 Accession no. Primer sequence (5’3’) Orientation Efficiency
NPY NM_001014845.3 TAG CGG AGC GTG ATT GCC CG Forward 100%
GGG GGT GTC CGG AGC AGG TT Reverse
AGRP EU374213.1 GGG CAC CCC TCT TGT AGA GCC Forward 95%
GGC CCA CAC GTG ACT GCT TCC Reverse
POMC NM_174151.1 GCC GCT GAA CAT CCT CGC CC Forward 100%
CTC CAG GCA CCA ACC ACG CA Reverse
RLN3 XM_002688750.1 GCA GTG GCC TCC AGC GAG TG Forward 96%
CCC TGG GGT TCC CTG CCA GT Reverse
MC3R XM_002701108.1 CGT GCT GTC TGC ACT CGG CT Forward 93%
TTG CCG TTC CTG ACC ACG GC Reverse
MC4R FJ430565.1 GCA AGC GGC AGG CTC GACG Forward 96%
GGC AGC CTC AGC ACC TTT CTC C Reverse
RXFP1 XM_610789.5 TGT GTC TGC AGT TAC GTG CTT TGG Forward 93%
GCC AGG GAC CCC AGA AGC TG Reverse
GnRH NM_177514.2 TCA CCT TCA GCT GCC TCT TCA TCA Forward 100%
TCA GCC GAG CTC GTG GTA TAT TGT Reverse
GnIH NM_001127268.1 CTG GAA AGG CTG TCC TCT CT Forward 92%
CCA ACC TGC CAC TGA GAT TT Reverse
FSHβ NM_174060.1 ACG TAC CCA GTA GCC ACT GAA TGT Forward 97%
AGC CTA CGC ACA TGT ACA CAC AGA Reverse
LHβ NM_173930.1 GCC CTG TCT GTA TCA CTT TCA CCA Forward 95%
TTA GAG GAA GAG GAT GTC TGG GAG Reverse
Leptin NM_173928.2 TCG TGA CCT TCT TTG GGA TTT Forward 91%
CAC ACT GGA ATA CTT CCC TCT C Reverse
EIF3K NM_001034489.2 TGA CAG ACA GCC AGC TAA AGG TGT Forward 100%
TCT TCT CCA CGA TGT TCT TGG GCT Reverse
1Neuropeptide-Y (NPY), agouti-related protein (AGRP), Pro-opiomelanocortin (POMC), relaxin-3 (RLN3), melanocortin 3 receptor (MC3R), melanocortin 4 receptor (MC4R), relaxin/insulin-like family peptide receptor 1 (RXFP1), gonadotropin releasing hormone (GnRH), gonadotropin inhibiting hormone (GnIH), and pituitary mRNA expression of follicle stimulating hormone beta polypeptide (FSHβ) and luteinizing hormone beta polypeptide (LHβ), and eukaryotic translation initiation factor 3 subunit K (EI3K).

Statistical Analysis

Growth and carcass traits were analyzed as a completely randomized design using a mixed linear model of SAS v9.2 (SAS Inst. Inc., Cary, NC). Animal served as the experimental unit. The RFI classification was used as the fixed independent variable in the model using the following model: y = μ + Xβ + Zu + e. Means were separated using least squares means with a Bonferroni adjustment using a significance level of α = 0.05. Individual animal RFI was determined using the following regression model: DMI = β0 + β1 (ADG) + β2(WT) +RFI where DMI is average daily feed intake, β0 is regression intercept, β1 is partial regression coefficient of DMI on ADG, β2 is the partial regression coefficient of DMI on BW, while ADG and middle test weight (MidWt0.75), calculated as off-test weight + on-test weight/2, were used as regressors on daily DMI (Koch et al., 1963). Changes in gene expression were calculated from the cycle threshold, after correction using EIF3K expression and analyzed using the Pair Wise Fixed Reallocation Randomization Test of REST-MCS v2.0 (http://rest.gene-quantification.info/). Correlations between gene expression levels in subcutaneous adipose tissue samples were also determined using the Proc Corr procedures of SAS (v9.2, SAS Inst. Inc.).


RESULTS

Growth Performance

Upon completion of the feeding trial, RFI was calculated for each individual steer and animals were classified as low (LRFI), average, or high (HRFI) based on their RFI values. Low, average and high group RFI means were –1.27, 0.03, and 1.51, respectively, with low and high RFI group means separated by greater than 2 SD (P < 0.0001) and r2 = 0.58 for DMI regressed on ADG and mid weight. As anticipated, there were no differences in initial BW (P < 0.91), initial hip height (P < 0.87), final BW (P < 0.94), total gain (P < 0.84), or ADG (P < 0.83) between LRFI, average or HRFI groups (Table 3). Importantly however, DMI differed significantly across groups with DMI lowest in LRFI steers and highest in HRFI steers (P < 0.0005) indicating HRFI steers consumed significantly more feed to achieve similar gain as LRFI steers (Table 3).


View Full Table | Close Full ViewTable 3.

Growth and performance traits of steers by residual feed intake (RFI) group1,2

 
Variable Low Average High P-value
RFI –1.27 ± 0.16c 0.03 ± 0.08b 1.51 ± 0.18a 0.0001
Initial BW, kg 351 ± 12 345 ± 6 345 ± 14 0.91
Initial hip height, cm 127 ± 0.8 127 ± 1.3 125 ± 1.9 0.87
Dry matter intake, kg 438 ± 16b 483 ± 7.9a 544 ± 19a 0.0005
Final BW, kg 466 ± 15 455 ± 7 457 ± 17 0.94
Gain, kg 115 ± 6.2 111 ± 3.1 113 ± 7.2 0.84
ADG, kg·d–1 1.64 ± 0.09 1.58 ± 0.04 1.67 ± 0.10 0.83
1Values are least square means (lsmeans) ± SEM
2n = 8 (low), 34 (average), 6 (high)

Carcass Merit and Meat Quality

To assess a potential relationship between RFI status and carcass merit, indices of meat quality were assessed on carcasses from LRFI and HRFI steers. Consistent with a lack of differences in gain across groups, there were no differences in yield grade (P < 0.85), REA (P < 0.15), back fat thickness (P < 0.61) or marbling score (P < 0.93) between LRFI and HRFI steers (Table 4). Likewise, objective Hunter color score was unaffected by RFI classification as L* values indicating degree of lightness (P < 0.68), a* values indicating degree of redness (P < 0.46), and b* values indicating degree of yellowness (P < 0.66) not different between LRFI and HRFI steers.


View Full Table | Close Full ViewTable 4.

Carcass and meat quality traits by residual feed intake (RFI) group1,2

 
Variable Low High P-value
Yield grade3 2.6 ± 0.3 2.7 ± 0.4 0.85
REA, cm2 68.4 ± 5.6 63.9 ± 4.8 0.15
Back fat, mm 7.9 ± 0.7 7.4 ± 0.8 0.61
Marbling score4 474 ± 29 470 ± 33 0.93
L*, lightness 35.5 ± 1.0 34.9 ± 1.1 0.68
a*, redness 23.7 ± 1.0 22.5 ± 1.2 0.46
b*, yellowness 23.2 ± 1.0 22.5 ± 1.1 0.66
1Values are lsmeans ± SEM
2n = 8 (low), 6 (high)
3USDA Yield Grade: 1 to 5 scale, where 1 = leanest and 5 = fattest
4Marbling Score: 300 = Traces; 400 = Slight; 500 = Small; etc.
5Average group were sold at auction pre-harvest and therefore carcass data for these steers are not available

Gene Expression

Neuropeptides in the arcuate nucleus of the hypothalamus are known to regulate feeding behavior in mammals. Figure 1 illustrates the association between RFI status and the expression of several genes in the arcuate nucleus of LRFI and HRFI steers that have been implicated in regulating satiety. The mRNA expression of neuropeptide-Y (NPY) and relaxin-3 (RLN3) was decreased 64% and 59%, respectively (P < 0.05), while Pro-opiomelanocortin (POMC) mRNA expression was 350% higher in the arcuate nucleus of LRFI steers versus HRFI steers (P < 0.01). The mRNA expression of agouti-related protein (AGRP) was unchanged between efficient (LRFI) and inefficient (HRFI) groups (P < 0.64).

Figure 1.
Figure 1.

The mRNA expression of neuropeptides, (A) neuropeptide-Y (NPY), (B) agouti-related protein (AGRP), (C) Pro-opiomelanocortin (POMC), and (D) relaxin-3 (RLN3) in the arcuate nucleus of steers identified as low Residual Feed Intake (RFI) or high RFI. Expression was determined by real-time RT-PCR. Values were normalized to eukaryotic translation initiation factor 3 subunit K (EIF3K) expression. Data are expressed as fold change relative to High RFI steers and calculated according to Pfaffl (2010). Bars denoted by * differ (P < 0.05), Low, n = 8, High, n = 6.

 

Figure 2 illustrates the association between RFI status and the expression of several genes in the arcuate nucleus of LRFI and HRFI steers that encode the cognate receptors for neuropeptides relaxin-3 and a cleavage product of POMC, melanocyte-stimulating hormone (α-MSH). The mRNA expression of melanocortin 4 receptor (MC4R) was 58% lower in LRFI steers compared to their HRFI counterparts (P < 0.05) while the expression of melanocortin 3 receptor (MC3R; P < 0.78) and relaxin/insulin-like family peptide receptor 1 (RXFP1; P < 0.63) mRNA was not different between LRFI and HRFI steers.

Figure 2.
Figure 2.

The mRNA expression of neuropeptide receptors, (A) melanocortin 3 receptor (MC3R), (B) melanocortin 4 receptor (MC4R), and (C) relaxin/insulin-like family peptide receptor 1 (RXFP1) in the arcuate nucleus of steers identified as either low Residual Feed Intake (RFI) or high RFI. Expression was determined by real-time RT-PCR. Values were normalized to eukaryotic translation initiation factor 3 subunit K (EIF3K) expression. Data are expressed as fold change relative to High RFI steers and calculated according to Pfaffl (2010). Bars denoted by * differ (P < 0.05), Low, n = 8, High, n = 6.

 

Figure 3 illustrates the association between RFI status and the expression of neuronal genes involved in the gonadotropin axis. The mRNA expression of GnRH was 86% lower (P < 0.01) while the expression of gonadotropin inhibiting hormone (GnIH) mRNA was 198% higher (P < 0.01) in the hypothalamus of LRFI versus HRFI steers. Consistent with the expression of gonadotropin regulatory hormone genes, the pituitary expression of follicle stimulating hormone β polypeptide (FSHβ) and luteinizing hormone β polypeptide (LHβ) mRNA was decreased in LRFI steers compared to HRFI steers (P < 0.05). Pearson correlation coefficients and associated P values for hypothalamic GnRH and pituitary gonadotropin mRNA expression are reported in Table 5. The expression of GnRH mRNA correlated with the abundance of FSHβ (P < 0.03) and LHβ (P < 0.01) mRNA in the pituitary indicating that when GnRH gene expression was lower, so too was expression of gonadotropin genes.

Figure 3.
Figure 3.

The hypothalamic expression of (A) gonadotropin releasing hormone (GnRH) and (B) gonadotropin inhibiting hormone (GnIH) mRNA and pituitary expression of (C) follicle stimulating hormone β polypeptide (FSHβ) and (D) luteinizing hormone β polypeptide (LHβ) mRNA in steers identified as either low Residual Feed Intake (RFI) or high RFI. Expression was determined by real-time RT-PCR. Values were normalized to eukaryotic translation initiation factor 3 subunit K (EIF3K) expression. Data are expressed as fold change relative to High RFI steers and calculated according to Pfaffl (2010). Bars denoted by * differ (P < 0.05), Low, n = 8, High, n = 6.

 

View Full Table | Close Full ViewTable 5.

Correlation of hypothalamic GnRH gene expression with pituitary gonadotropin gene expression1

 
FSHβ LHβ GnRH
FSHβ 0.953 0.589
P < 0.0001 P < 0.03
LHβ 0.953 0.783
P < 0.0001 P < 0.01
GnRH 0.589 0.783
P < 0.03 P < 0.01
1Pearson correlation coefficients and associated P values for arcuate nucleus mRNA expression of gonadotropin releasing hormone (GnRH) and pituitary mRNA expression of follicle stimulating hormone beta polypeptide (FSHβ) and luteinizing hormone beta polypeptide (LHβ). n = 8 (low), 6 (high).

Leptin is an adipose tissue-derived hormone known to signal satiety by acting on neurons within the arcuate nucleus. Figure 4 illustrates the association between RFI status and the expression of leptin gene expression in subcutaneous adipose tissue. Consistent with the depressed expression of NPY and POMC in the arcuate nucleus of LRFI steers, leptin mRNA expression was 245% higher in LRFI steers relative to their HRFI steer counterparts (P < 0.05).

Figure 4.
Figure 4.

The adipose tissue expression of leptin mRNA in steers identified as either low Residual Feed Intake (RFI) or high RFI. Expression was determined by real-time RT-PCR. Values were normalized to eukaryotic translation initiation factor 3 subunit K (EIF3K) expression. Data are expressed as fold change relative to High RFI steers and calculated according to Pfaffl (2010). Bars denoted by * differ (P < 0.05), Low, n = 8, High, n = 6.

 


DISCUSSION

Feed intake is positively correlated with RFI and differences in DMI between low and high RFI cattle are associated with differences in feeding behavior exhibited by the 2 groups (Golden et al., 2008; Lancaster et al., 2009; Kelly et al., 2010; Montanholi et al., 2010; Durunna et al., 2011; Hafla et al., 2013). The addition of feeding behavior traits such as frequency or duration of feeding bouts to RFI models has been reported to explain up to 35% of the variation in DMI not explained by ADG and MidWt0.75 in base RFI models (Lancaster et al., 2009; Kelly et al., 2010; Kayser and Hill, 2013). Therefore, regions in the brain such as the hypothalamus that function to integrate metabolic, neural, and endocrine signals to coordinate feeding behavior, energy balance, and developmental trajectory in animals represent attractive candidate tissues for identifying mechanisms that might contribute to or serve as markers for differences in feed intake and thus variations in RFI (Baile and McLaughlin, 1987; Sartin et al., 2011). To address this hypothesis, the present study was conducted to examine potential associations between the expression of hypothalamic genes implicated in regulating satiety and RFI status in growing steers.

Classic studies where lesions were experimentally induced within nuclei of the hypothalamus established a role for the arcuate nucleus (ARC) as a key center for appetite regulation in livestock and it is now clear that 2 neuronal populations residing within the ARC are crucial for controlling feed intake and regulating BW (Baile and McLaughlin, 1987; Berthoud, 2002; Hillebrand et al., 2002; Ellacott and Cone, 2004; Sartin et al., 2011). Orexigenic neurons express the neuropeptides, NPY and AGRP, which stimulate feeding (Elmquist et al., 2005; Heisler et al., 2006). The opposing anorexigenic neurons express neuropeptides such as POMC-derived α-MSH that signal satiety by inhibiting feeding behavior (Cota et al., 2006). The peptides α-MSH and AGRP have opposing actions on the melanocortin receptors (MCR) as α-MSH activates MC3R and MC4R to signal satiety while AGRP acts as an MCR antagonist (Cowley, 2003). The specialized blood-brain barrier of the ARC allows circulating hormones such as insulin, adipose tissue-derived leptin and gut-derived ghrelin to serve as peripheral signals which reciprocally regulate the NPY/AGRP and α-MSH neurons with leptin acting to decrease feed intake and ghrelin stimulating neurons that increase feed intake (Arnold et al., 2006). Ancillary neuropeptides such as relaxin-3 acting though its cognate receptor, relaxin/insulin-like family peptide receptor 1 (RXFP1), can modulate these circuits to stimulate feeding behavior (Hossain et al., 2013). Additionally, as yet unappreciated satiety factors may exist. These observations suggest a mechanism whereby differences in feeding behavior traits between animals could be associated with differences in the expression of genes within the hypothalamus and adipose tissue that have been implicated in regulating energy balance and feeding behavior.

The experimental approach utilized in the current study involved ranking steers by RFI to create divergent groups consisting of low and high RFI steers. This approach has been used successfully by others and the spread in RFI between LRFI and HRFI groups in the current study was larger than previous studies (Baker et al., 2006; Kolath et al., 2006; Nkrumah et al., 2006). As expected, cattle exhibited similar growth performance across RFI classes over the course of the 70 d feeding trial (Archer et al., 1999). Mean values of DMI, ADG, and total gain in this experiment were similar to values reported for other studies conducted under comparable conditions (Baker et al., 2006; Kolath et al., 2006; Kelly et al., 2010). Consistent with previous studies, RFI was not associated with ADG while there was a positive correlation between RFI and DMI (Herd and Bishop, 2000; Arthur et al., 2001; Herd et al., 2003; Baker et al., 2006; Kolath et al., 2006; Kelly et al., 2010). These data indicate that divergent LRFI and HRFI groups were successfully created in the current study.

Importantly, the present study reveals associations between RFI and the expression of several hypothalamic and adipose-specific genes that are known to regulate feeding behavior while implicating the gonadotropin system as a novel endocrine axis that may influence feed efficiency. These data suggest a molecular signature of feed efficiency that, on further study, may allow the beef industry to better predict RFI phenotypes and ultimately speed the development of marker-assisted selection programs that decrease the need to evaluate RFI directly. Beef producers face significant economic pressure to improve feed efficiency to improve the profitability and sustainability of beef production. Costs associated with feed can represent up to 70% of non-fixed expenses associated with cattle production (Archer et al., 2004; Arthur et al., 2001, USDA-ERS, 2012). Reducing daily DMI by just .91 kg per head could reduce the cost of beef production by $1 billion annually within the United States and incorporating RFI into selection programs could improve profitability for beef producers by as much as 33% (Herd et al., 2003; Archer et al., 2004; Weaber, 2012). However, RFI is costly and labor intensive to measure which has limited its adoption while the development of alternative strategies is slowed by a poor understanding of molecular mechanisms underlying variation in feed efficiency and RFI in cattle (Herd et al., 2003; Crews 2005; Herd and Arthur, 2009). Any increase in understanding that facilitates reduced feed inputs without affecting growth performance or carcass merit has the potential to improve profitability of beef production systems and would be of great use to the beef industry.

While the present study focused on investigating a potential association of hypothalamic gene expression with RFI status in beef cattle, feed efficiency is likely influenced by complex interactions between organ systems, endocrine status and differences in cellular energy metabolism. Herd and Arthur (2009) have proposed that potential differences between multiple physiological processes underlie the significant variation in RFI that is often observed within a herd. These include differences in 1) gut-related parameters such as digestibility and fermentation; 2) factors affecting energy expenditure such as heat increment, basal metabolic rate, non-exercise induced thermogenesis, protein turnover, and differences in physical activity; and 3) differences in stress response as well as contributions by other processes. While the current study doesn’t allow either the determination of factors that induced the observed differences in target gene expression or evaluation of whether these differences caused the observed RFI phenotypes or rather were induced by RFI status, the novel associations between neuronal gene expression and RFI status observed in the present study support the hypothesis that differences in hypothalamic function between LRFI and HRFI steers represent another component that contributes to RFI status.

Clearly feed intake is regulated in mammals by the integration of peripheral signals acting on the hypothalamus (Baile and McLaughlin, 1987; Sartin et al., 2011). Regardless of whether neuronal activity within the ARC is driving RFI status or whether differences in cellular metabolism are inducing neuronal changes, the expression patterns of orexigenic and anorexigenic genes observed in the ARC of steers suggest that superior efficiency is associated with a reduced drive to eat in growing cattle. Both NPY and RLN3 mRNA expression was lower while mRNA expression for POMC was higher in the ARC of LRFI steers versus their HRFI counterparts. Consistent with the observation that NPY expression was lower in animals with lower DMI, Bahar and Sweeney (2008) have demonstrated that NPY potently stimulates feeding behavior in cattle. Furthermore, SNP in the NPY gene have been associated with ADG, BW, G:F and RFI status pointing to the potential importance of this gene in regulating energy balance and feed efficiency in growing cattle (Sherman et al., 2008; Zhang et al., 2011; Alam et al., 2012; Trujillo et al., 2013). Likewise, the observation that RLN3 mRNA was reduced in LRFI steers is consistent with the emerging role for relaxin-3 in regulating satiety as RLN3 has been shown to acutely stimulate feeding behavior in mammals when infused centrally (Bathgate et al., 2002; McGowan et al., 2005; Hida et al., 2006; McGowan et al., 2006; Hossain et al., 2013). To date, no study has associated the expression of RLN3 with DMI in cattle. It is well established that α-MSH, a cleavage product of POMC potently signals satiety via activation of the MC3R and MC4R receptors in humans and rodents and this appears to also be the case for both sheep and pigs (Clarke et al., 2003; Archer et al., 2004; Barb et al., 2004). Thus an increase in POMC mRNA expression is consistent with decreased feed intake. Recently, SNP in the POMC gene were associated with carcass merit in cattle (Gill et al., 2010; Deobald and Buchanan, 2011; Liu et al., 2013). Finally, multiple studies in cattle indicate that low RFI animals exhibit 10 to 24% fewer daily feeding bouts and up to 11 to 26% shorter feeding bout durations than their high RFI counterparts though some variability exists in these associations across studies in the literature (Nkrumah et al., 2007; Golden et al., 2008; Lancaster et al., 2009; Kelly et al., 2010; Montanholi et al., 2010; Durunna et al., 2011; Schwartzkopf-Genswein et al., 2011; Hafla et al., 2013; Kayser and Hill, 2013). Though feeding behavior was not measured in the present study, the elevated POMC and lower NPY and RLN3 mRNA levels observed in the ARC of LRFI animals in this study are consistent with reduced levels of feeding behavior traits reported for cattle exhibiting superior feed efficiency in the literature.

The expression profiles of genes measured in the current study support a model whereby greater leptin expression in the adipose tissue of efficient steers drive changes in the arcuate nucleus that lead to reduced feed intake. Consistent with decreased expression of orexigenic genes and elevated expression of anorexigenic genes in the ARC, leptin mRNA was elevated in the subcutaneous adipose tissue of LRFI versus HRFI steers, a relationship that would be predicted by the well established role of circulating leptin to act on NPY and POMC neurons in the ARC of mammals to lower NPY and increase POMC expression (Arnold et al., 2006). The elevated leptin expression was observed despite there being no differences in subcutaneous fat thickness or marbling score between LRFI and HRFI steers. Several studies have reported the significant association between leptin and RFI (Nkrumah et al., 2004; Hoque et al., 2009; Kelly et al., 2009) and between leptin growth performance (Corva et al., 2009). Taken together, these data support a role for leptin in regulating feed efficiency in cattle.

Contrary to our initial hypothesis, mRNA expression of AGRP was not different between LRFI and HRFI steers. An orexigenic effect of AGRP has been well established in sheep as fasting induces AGRP gene expression in the ARC and infusion of AGRP into the hypothalamus increases feed intake in ewes (Henry et al., 2001; Adam et al., 2002; Archer et al., 2002; Wagner et al., 2004). However, injection of AGRP into the hypothalamus of pigs failed to illicit a feeding response indicating the effect of AGRP may differ across species (Barb et al., 2004). Little information exists in the literature concerning the role for AGRP in regulating satiety in cattle. The current study suggests that AGRP mRNA expression does not vary by RFI status but the observation that AGRP mRNA expression was not different between LRFI and HRFI steers does not preclude a role for AGRP in regulating feed intake in the bovine.

Given the effects of neuropeptides on satiety are mediated through interaction with their respective receptors, we measured expression of the RXFP1, MC3R and MC4R genes in the ARC of LRFI and HRFI animals. The expression of RXFP1 and MC3R mRNA levels were not different across RFI groups. However, surprisingly the expression of MC4R mRNA was lower in LRFI versus HRFI steers suggesting a potential blunted ability to respond to α-MSH in low RFI steers. However, α-MSH can activate both MC3R and MC4R receptors and both melanocortin receptor family members have been implicated in suppressing feed intake and preventing a positive energy balance (Tao, 2010). It is possible that functional redundancy mitigated effects of decreased MC4R expression. Also, because only MC4R mRNA abundance was measured in the present study, it is unknown whether receptor number or activity was altered. Finally, the role each receptor subtype plays in regulating satiety in cattle is not well characterized.

Interestingly, mRNA expression for GnRH was lower while expression of GnIH mRNA was elevated in the hypothalamus of LRFI vs. HRFI steers and the expression of GnRH was highly correlated with gonadotropin gene expression in the pituitary. The reproductive axis is controlled by hypothalamic GnRH which acts on the pituitary gland to stimulate gonadotropin secretion leading to sex steroid production by the testis or ovaries. This action is inhibited by steroidal feedback from the gonads and from hypothalamic expression of GnIH. It has recently become apparent that GnIH also stimulates feeding without reducing energy expenditure in rodents, primates, and ruminants (Johnson et al., 2007; Qi et al., 2009; Clarke et al., 2012). Interestingly, GnRH has been identified as a potential regulator of feed efficiency in cattle using a genome wide association study and intracerebroventricular infusions of GnRH have been shown to significantly alter food consumption in shrews and zebrafish (Kauffman and Rissman, 2004; Rolf et al., 2012). These data support a role for GnRH/GnIH in regulating energy balance and feeding. Consistent with this argument, a growing literature suggests that vaccinating against GnRH increases feed intake in pigs (reviewed by Millet et al., 2011; Batorek et al., 2012). Likewise the surgical removal of the gonads in pigs is associated with increases in food intake and a shift in body composition toward increased adiposity (Campbell et al., 1989; Oliver et al., 2003). As a whole these data suggest members of the gonadotropin axis may influence feed efficiency in growing livestock.

Given the significant economic pressure for producers to finish cattle to an endpoint that merits quality grades of USDA Prime or Choice, it is important that selection for efficiency does not decrease carcass merit (Archer et al., 1999; Anderson, 2012). Earlier studies have indicated that up to 5% of the variation in RFI can be attributed to differences in carcass composition with low RFI animals tending to produce leaner carcasses (Richardson et al., 2001; Richardson and Herd, 2004). Therefore, our first objective was to determine the association of RFI with indices of carcass merit. In the current study there were no differences between LRFI and HRFI steers in yield grade, ribeye area, subcutaneous adipose tissue mass as measured by backfat or intramuscular fat deposition as measured by marbling score. These results are consistent with several studies that did not find a significant correlation between RFI and these traits (Carstens et al., 2002; Schenkel et al., 2004; Baker et al., 2006; Nkrumah et al., 2007; Ahola et. al., 2011). Likewise, there were no differences in objective Hunter color scores for steaks harvested in the current study. Baker et al. (2006) did not see a difference in L* or a* across RFI groups but high RFI steers did exhibit higher b* values in that study. While measures of tenderness or palatability were not recorded in the current study, the few studies that have addressed these parameters indicated that RFI has little impact on Warner-Bratzler Shear Force and sensory panel outcomes consistent with the lack of evidence for an association between RFI and meat quality that was observed in the current study (McDonagh et. al., 2001; Ahola et al., 2011).

In conclusion, this study provides no evidence that RFI is associated with indices of meat quality suggesting that incorporation of RFI into selection programs would not negatively impact carcass merit. Overall, the expression profiles of genes measured in the current study support the hypothesis that greater leptin expression in the adipose tissue of efficient steers drive changes in the arcuate nucleus that would be expected to decrease feed intake. This is the first study to demonstrate that expression of orexigenic genes NPY and RLN3 is decreased while the expression of anorexigenic gene, POMC, is elevated in the arcuate nucleus of efficient steers relative to inefficient ones. Furthermore this is the first study in cattle to implicate GnRH and GnIH as potential mediators of feed efficiency. Currently, there is a great need to better understand the regulation of physiological processes driving variation in RFI so that improvements in feed efficiency can be hastened through either the development of new technologies or improved selection programs. This study identifies several novel targets that may regulate feed efficiency and underlie the variation in RFI observed in growing cattle.

 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.