It is widely predicted that the world population will increase to 9 billion by 2050( 1 , Reference Godfray, Beddington and Crute 2 ). At the same time, economic improvements in developing countries around the world are predicted to result in an increased demand for meat, milk and other animal products, as those societies become more ‘westernised’. Even though there are calls for people in developed countries to reduce meat consumption for health reasons, particularly processed red meat, the demand for meat is predicted to continue to increase at a similar rate to that seen in the previous 10+ years. Over the past 50 years, tremendous advances in animal genetics and animal nutrition have been made to meet the increasing demand, particularly in pigs and poultry, but this has mainly been achieved using high-quality feed ingredients such as wheat, maize and soya. Over recent years these ingredients have become increasingly more expensive, due to a combination of increased demand from the biofuels industry, as well as for animal and human nutrition, along with shortages due to crop failures in some parts of the world. It has been estimated that for many agricultural commodities the rate of production has already reached a peak( Reference Seppelt, Manceur and Liu 3 ). Hence, if we are to continue to meet the demand for animal products, we cannot simply feed more animals the same feed ingredients, as that would require more crops, land and water( 1 , Reference Godfray, Beddington and Crute 2 ).
Feed ingredients account for a large proportion of the overall costs of animal production, particularly in non-ruminant species( Reference Patience, Rossoni-Serao and Gutierrez 4 ). Continuing to rely on the same ingredients, in competition with human nutrition and biofuels, mean prices will increase and therefore the cost of meat and animal products will also increase. Therefore the aim of the present research is to improve the efficiency with which animals utilise their feeds, giving more products for the same amount of feed or the same amount of product for less feed. This is referred to as feed efficiency (FE), which is simply calculated as the change in body weight divided by the change in feed intake (kg gain/kg feed). Hence, increased efficiency would be greater gain per unit feed. Another term used is feed conversion ratio (FCR), which is kg feed/kg gain, with improved efficiency associated with a lower FCR value (less feed per unit gain). More recently animal scientists refer to residual feed intake (RFI), which compares the feed intake for each individual animal to the average for the herd/group at the same rate of growth( Reference Patience, Rossoni-Serao and Gutierrez 4 ). Hence, an animal with a low RFI (often a negative value) would be eating less for the same growth rate and therefore be more efficient than an animal with a high RFI (a positive value), which would be eating more.
There is no doubt that selective breeding and improved diet formulations over the past 20–30 years have improved the FE of pigs( Reference Patience, Rossoni-Serao and Gutierrez 4 ) and chickens( Reference Siegel 5 ), with FCR values of 2·0 or less currently achievable (i.e. >50 % efficiency). Indeed it is predicted that FCR values of 1·5 and less will be seen relatively soon for both pigs and chickens (note that the lowest value theoretically possible would be 1·0, meaning 100 % efficiency). In contrast, ruminants are a lot less efficient( Reference Berry and Crowley 6 ), with FCR values of 5·0 or more being normal (i.e. <20 % efficiency). However, we must remember that ruminants can utilise ingredients not used for human consumption (e.g. grass and silage) and are therefore not competing with human subjects, non-ruminants and biofuels for the high-quality ingredients. FE can be improved in ruminants by feeding higher quality ingredients as concentrates( Reference Martin, Morgavi and Doreau 7 ), but that is not the solution for the future. What we need is to maintain or improve the efficiency of livestock, while at the same time maintaining or improving the quality of the animal products, but using alternative (human inedible) feed ingredients as much as possible. In that way, we will be converting human inedible ingredients into high-quality, human-edible foods. This review will highlight a few ways in which this is being achieved or might be achieved in the future.
Use of enzymes as feed additives
A number of enzymes are already used commercially as feed additives, particularly in non-ruminant (pig and poultry) feeds, to increase the digestion and subsequent absorption of nutrients( Reference Bedford and Schulze 8 – Reference Humer, Schwarz and Schedle 10 ). They are mainly used to improve the digestion of feed components that the animals cannot normally digest or are only able to digest fairly poorly, such as complex carbohydrates and phytate. By increasing the digestibility of the feed, more nutrients enter the body and less pass through in the faeces, resulting in increased growth for the same level of feed intake, hence improving FE.
A number of enzyme feed additives are commercially available to improve the digestibility of cereal carbohydrates, particularly targeting xylans and arabinoxylans present in the cell walls( Reference Masey O'Neill, Smith and Bedford 9 ). By digesting these important structural carbohydrates in the cell wall, that then allows the animals’ own carbohydrate-digesting enzymes (e.g. α-amylase) better access to the main starch stores within the plant cells. Secondly, the digestion reduces the viscosity problems associated with arabinoxylans and β-glucans( Reference Masey O'Neill, Smith and Bedford 9 ). A number of studies have shown improved FE and/or FCR of pigs and chickens when these enzymes are added to the feed. For example, xylanase supplementation of feed was shown to improve FCR (1·41 v. 1·56 in controls) in broiler chickens by increasing weight gain, but not affecting feed intake( Reference Amerah, Mathis and Hofacre 11 ). As well as increasing the digestibility of the carbohydrate component of the feed and reducing the viscosity, there are suggestions that these carbohydrate-degrading enzymes might have prebiotic actions on the gut microflora via the oligosaccharides they produce( Reference Masey O'Neill, Smith and Bedford 9 ). This could be another potential mechanism for their effects on FE. The absorption of nutrients across the gut is also known to affect production of gut peptides, which can subsequently alter gut motility and feed intake. Indeed xylanase supplementation of feed has been shown to increase plasma peptide YY levels in broiler chickens( Reference Singh, Masey O'Neill and Ghosh 12 ) and we have recent data showing effects of xylanase supplementation on plasma peptide YY, gastric inhibitory polypeptide and glucagon-like peptide-1 concentrations in young pigs( Reference May, O'Sullivan and Brameld 13 ). Hence, the regulation of gut peptides and their subsequent effects on gut motility, feed intake and/or nutrient utilisation might be additional, alternative mechanisms for the effects of these carbohydrate-degrading enzymes on FE.
Phytase is another enzyme used commercially in non-ruminant (pig and poultry) feeds( Reference Humer, Schwarz and Schedle 10 ). Phytase digests phytate (also called phytic acid or inositol hexakisphosphate), the main storage form for phosphorus in plants. Phytate (hexakisphosphate) is inositol with six phosphate groups attached and phytase is able to cleave individual phosphate groups, thereby releasing them for absorption and use by the animal. Phytase supplementation results in greater absorption of phosphorus and calcium from the feed in broiler chickens and pigs( Reference Simons, Versteegh and Jongbloed 14 ), resulting in increased growth and reduced FCR. However, the increased growth may not simply be due to increased absorption of these important micronutrients. Chicken studies( Reference Liu, Ru and Li 15 ) have shown that high levels of phytate in the diet inhibit pepsin and trypsin activities and therefore inhibit protein digestion and amino acid absorption, resulting in increased FCR. Inclusion of phytase as well as high phytate in the diet reduced the inhibitory effect on proteolysis, resulting in improved (reduced) FCR( Reference Liu, Ru and Li 15 ).
Both of these feed additive enzymes have positive effects on FE in pigs and chickens fed cereal-based diets. They do so by different mechanisms, meaning their benefits are likely to be additive, but importantly they may allow the use of poorer quality (i.e. human inedible) feed ingredients, an important consideration for future sustainability and food security. These and other enzymes are also being investigated for use in ruminants( Reference Masey O'Neill, Bedford, Walker, Garnsworthy and Wiseman 16 ).
Use of growth promoters/metabolic modifiers/anabolic agents
There are three main classes of growth promoters( Reference Sillence 17 ) : β-adrenergic agonists (BA), anabolic steroids and growth hormone (GH, also called somatotropin). They all improve FE in livestock to some extent and this is associated with increased lean mass (particularly skeletal muscle) and reduced fat mass( Reference Sillence 17 ). Indeed they have all been in the news at different times in relation to their illegal use as performance enhancing drugs in sportsmen and women. Their effects on muscle and fat mass were first discovered in the 1950s (anabolic steroids) or 1980s (BA and GH) and a number of commercial products are currently licenced for livestock production around the world( Reference Sillence 17 ), although they are all banned in the European Union (EU). For example, ractopamine and zilpaterol (both BA) are licenced for use in pigs and/or cattle in North and South America, South Africa, India and Australia, but not China. Similarly, the anabolic steroid mix of trenbolone acetate and oestradiol is licenced for use in beef cattle in North and South America, South Africa, India, Australia and China and GH (either bovine or porcine somatotropin) is licenced for use in dairy cattle or pigs in the same areas. We were unable to find information for other parts of the world (e.g. Northern Africa and other parts of Asia); so to our knowledge only the EU has a total ban on the use of these agents in livestock production. This is despite much of the early research work being carried out in the EU, especially the UK, and the original scientific reports suggesting their use was safe( Reference Lamming, Ballarini and Baulieu 18 ), as long as appropriate guidelines were followed (e.g. a withdrawal period prior to slaughter).
At the University of Nottingham, we have been comparing the molecular modes of action of BA and GH in both sheep( Reference Al-Doski, Parr and Hemmings 19 – Reference Parr, Al-Doski and Hemmings 21 ) and pigs( Reference Brameld, Atkinson and Saunders 22 – Reference Sensky, Jewell and Ryan 24 ) combining transcriptomic and metabolomics technologies in a systems biology approach to identify novel mechanisms to achieve the same effects. Ultimately the aim is to identify novel target genes/proteins to develop more acceptable drugs or for targeted breeding or nutritional manipulations. We have made good progress and have identified up-regulation of the serine biosynthesis pathway( Reference Al-Doski, Parr and Hemmings 19 , Reference Parr, Al-Doski and Hemmings 21 , Reference Brameld, Ryan and Williams 23 ) and a number of other novel changes in response to BA and/or GH treatments. We are currently performing proof-of-principle studies to determine whether the novel genes we have identified really do regulate growth, body composition and/or FE. If successful, the next stage will be to use this information to develop breeding strategies, new dietary regimens or drugs that result in improved FE in livestock.
For proof-of-principle studies we often utilise transgenic animals (mainly mice) where the gene of interest is either overexpressed or knocked out/down (i.e. genetic manipulation (GM)), often in a tissue-specific manner. This is done to investigate whether manipulation of the specific gene results in the predicted changes in tissue growth and/or metabolism, as well as changes in FE or whole-body energy expenditure. Such studies cannot be performed in cultured cells, so must be done in animals. Although technically challenging, GM can now be achieved in livestock( Reference Niemann, Kuhla and Flachowsky 25 ), so that it will theoretically be possible to produce herds of transgenic livestock. Indeed the Chinese government is funding work using GM aimed at developing new breeds of livestock for agricultural use in the future, including research into their safety( 26 ). One of the main advantages of GM over conventional animal breeding is that GM speeds up the process and is more gene-specific; whereas conventional breeding, while very successful over the past 50 years, can result in unwanted side effects, both on animal welfare and also product quality. The halothane pig( Reference Rosenvold and Andersen 27 ) and Callipyge sheep( Reference Tellam, Cockett and Vuocolo 28 ) are prime examples of this. Both have increased growth rates, particularly muscle, but one (halothane) results in highly stressed pigs and both result in poorer meat quality.
Molecular studies of low residual feed intake animals
The concept of low and high RFI has progressed rapidly over recent years( Reference Herd, Oddy and Richardson 29 , Reference Sartin 30 ). Studies are being carried out around the world aimed at identifying specific genes (or markers) for improved FE in virtually all livestock species (cattle, pigs, sheep and poultry). The genetic approach has been to identify markers (quantitative trait loci or SNP) of low RFI for subsequent use in selective breeding programmes. For example, a Chinese group( Reference Luo, Sun and Ma 31 ) recently identified a SNP in a microRNA (miR-1596) gene in chickens that resulted in reduced expression of miR-1596 in livers and was associated with low RFI. Interestingly, they suggested that there were more than seventy target genes for miR-1596( Reference Luo, Sun and Ma 31 ), which were mainly involved in energy metabolism, apoptosis and immune responses, with some being important proteins for assembling mitochondria.
We collaborated with another Chinese group( Reference Jing, Hou and Wu 32 ), to investigate differential gene expression in skeletal muscle from pigs with low v. high RFI using a deep sequencing (RNAseq and miRNAseq) approach. A number of mRNA (IGF2, FABP3 and PGC1a) and miRNA (miR1, miR30, miR10b and miR145) were found to be differentially expressed, but importantly the majority of mitochondrial genes were down-regulated. The data suggested that low RFI was linked with changes in expression of mRNA and miRNA associated with increased muscle growth and reduced mitochondrial activity in skeletal muscle( Reference Jing, Hou and Wu 32 ).
Effects on mRNA or miRNA associated with mitochondria appear to be a recurring theme in the low RFI studies( Reference Bottje and Kong 33 , Reference Grubbs, Huff-Lonergan and Gabler 34 ) and this agrees with some of our growth promoter studies, where we also see down-regulation of a number of genes associated with mitochondria, including both tricarboxylic acid cycle and oxidative phosphorylation genes (JM Brameld, T Parr et al., unpublished results).
Once again the genes being identified in these various RFI studies could be potential targets for novel drugs, dietary regimens or GM in animals, as well as being used for conventional breeding strategies to improve FE in livestock.
There are tools already available to improve FE in meat production, including the use of enzyme feed additives and growth promoters. Recent molecular studies are starting to identify other mechanisms that might be utilised in the future, including manipulation of gut microflora or gut peptides and targeting of gene expression in skeletal muscle or other tissues using drugs or GM technologies. Whether the use of drugs or GM technologies will be acceptable to the EU general public in the future remains to be seen, but we cannot simply wait until food and meat availability becomes limited (or very expensive) before starting research on these more controversial topics. At present, food and meat are readily accessible and reasonably affordable throughout most of the EU, so the current ban on the use of growth promoters does not really affect the consumer. However, this might change if feed ingredients continue to increase in price and there are issues with crop failures around the world limiting their availability for animal feeds. The EU might then have to reconsider the ban or accept that meat and animal products will become more expensive and less accessible, as well as potentially limiting the countries we import meat from. We should emphasise that safety and quality of the products will always be a primary concern and must not be ignored in the drive to improve FE for meat production. Indeed we would suggest that research into the safety aspects must be carried out alongside the research into the manipulation of FE, as is currently happening in China. Finally, we suggest that greater emphasis is needed on the use of poorer quality ingredients in animal feeds in future, to reduce the competition with human nutrition and biofuels for the high-quality ingredients, such as wheat, maize and soya.
We would like to acknowledge the numerous PhD students and collaborators (both academic and industrial) that have contributed to the work included in this review, of which there are too many to name.
The work included has been funded by the Biotechnology and Biological Sciences Research Council (BBSRC), Zoetis (formerly Pfizer Animal Health) and AB Vista.
Conflicts of Interest
The studies we have done on feed enzymes have been funded by AB Vista and the recent growth promoter studies are funded by Zoetis/Pfizer Animal Health.
Both the authors contributed equally to the planning and writing of this manuscript.