How low can you go? Broiler breast muscle possesses a mitochondrial content of just 2%

ABSTRACT In 1977 the pectoralis major (breast) muscle mitochondrial content of domestic broiler chickens, Gallus gallus domesticus, was estimated at 4.1%. However, in the subsequent ∼45 years broilers have been further modified by sustained intensive genetic selection on feed conversion efficiency and breast muscle mass. With a view to understanding the metabolic implications of this historic selection – and also laying a foundation to better interpret current commercial performance – we quantitated the cellular bioenergetic structure of modern broilers. The breast muscle of 4 64 d old Ross308 broilers was subject to morphometric analysis based on transverse Transmission Electron Microscopy (TEM) images. This approach determined an average mitochondrial content, diameter and circularity ratio of 2.1%, 0.42 µm and 0.72, respectively. Broiler breast mitochondrial content has thus approximately halved in the last 45 years and represents one of the lowest contents recorded for the muscle of any eukaryotic species.


Introduction
Meat-producing chickens (broilers) are one of the globes primary sources of animal protein, possessing a dramatically enlarged pectoralis major (breast) muscle that comprises a quarter of the birds' total body mass (Joiner, et al., 2014).Moreover, modern birds have a reputation for extremely high feed conversion efficiencies, with some producers reporting in elite performers a wet weight gain of 1 kg for every 1.38 kg of dry feed consumed.This makes domestic chickens compelling models of animal performance.However, interest in the species is not restricted to the exceptional commercial capabilities of industrial birds.The chick is also considered a leading comparative model of the vertebrate muscle development programme, particularly with respect to cell lineage tracking and patterns of morphogenesis (Scaal & Marcelle, 2018).
Along with other avian meat-producing species (such as turkeys, pheasants and quail) chickens are members of the Phasianidae taxonomic clade, the largest branch of the Galliformes (Bottje et al., 2017).The members of this group tend to be sedentary, resident ground-dwelling birds that use short, burst flights to escape predators (Bottje et al., 2017).This behaviour explains the functionally unusual breast muscle metabolism of the ancestral birds such as the Red Junglefowl (Gallus gallus) progenitor of modern domestic chickens.This ancestrytaken together with the impacts of domestication and subsequent selectionaccounts for a muscle structure in modern birds that is dominated by explosive, fast twitch contractile isoforms, large diameter type IIB fibres and a very low oxidative capacity (Bottje et al., 2017).
Recent work across a population of 80 industrial Cobb broilers has indicated variation in breast muscle mitochondrial content (as estimated by a proxy, mtDNA copy number) is significantly associated with growth performance phenotypes.For example, birds with a relatively low breast mtDNA copy number tend to be heavier, more muscular and have higher abdominal fat (Reverter et al., 2017).It has also been argued that selection for higher feed efficiency may reduce tissue mitochondrial content through considerations of economic design (Hudson, 2009).After all, it would be energetically wasteful to pay for unnecessary spare physiological capacity.Given that intensively reared domestic birds do not face the demands (e.g.escaping the threat of predation) of wild animals there is considerable scope for economising on physiological function.However, the only bona fide record of an absolute measurement of mitochondrial content for broilers we could find was from the late 1970s (Kiessling, 1976;Kiessling, 1977), where the breast muscle of 2 Hydro Compact broilers was estimated to be 4.1% using planimetric analysis of Transmission Electron Microscope (TEM) images.
This raises a fundamental question.What has happened to the metabolism of broiler breast muscle in the intervening 45 years given numerous subsequent generations of intensive genetic selection on both muscle mass and feed efficiency?To answer this, we phenotyped the breast muscle of modern Ross308 Aviagen broilers at 64 d age using both (a) traditional planimetric estimation of mitochondrial content from transverse TEM images and (b) mtDNA copy number estimations based on a duplex taqman qRT-PCR assay using total DNA as the nucleic acid template.We hypothesised that the more muscular, more feed efficient modern broilers would have a lower breast muscle mitochondrial content when compared to the historic 1970s birds.In so doing we provide an updated benchmark for modern broilers and interpret our findings in the twin contexts of comparative physiology and animal production.

Animal resource
Twenty Ross308 birds were purchased from 9Dorf Farms, a pasture raised poultry farm at Lilydale, QLD, at 50 d and grown to 64 d at the Poultry Unit on the Gatton campus of The University of Queensland.Twenty was deemed enough to clearly differentiate individuals on the phenotype of muscle weight, based on a power calculation.Prior to purchase these birds received mixed feed and had access to grass, comparable to free range birds.This husbandry practice slowed their growth in comparison to standard, intensive industrial methods.After purchase they were fed an industry relevant broiler starter diet (see Table 1 for nutritional details).All birds were weighed at 64 d and euthanized by cervical dislocation.They were grown out to 64 d (which is longer than typical for broiler production systems) as part of an experimental strategy to promote 'white striping' (which did not readily materialise).
Skinless and deboned breast was mechanically separated from the carcass, using a knife to cut around the furcula alongside the keel bone.The entire dissected right breast muscle was then weighed.Pectoralis major necropsy samples were consistently taken from the same portion of the right pectoral muscle as previously described (Reverter et al., 2017), placed into heavy duty aluminium foil, snap frozen in liquid nitrogen then stored at −80°C.We selected 10 birds for qPCR estimation of pectoralis mitochondrial content after scoring on extent of white striping, with 5 scored as none and 5 as moderate.We also selected 4 birds for fixation and subsequent planimetric analysis of pectoralis mitochondrial content, with 2 scored as none and 2 as moderate.

TEM estimation of pectoralis major mitochondrial content
Breast tissue samples (1 mm × 1 mm cubes) from 4 individual birds were fixed overnight (for 24 h total) in 2.5% glutaraldehyde, washed with PBS then stored under PBS at 4°C until ready for mounting and sectioning.They were mounted such that transverse cross sections could be taken with respect to the orientation of the muscle fibres and the associated mitochondrial network.The sections were visualised using a Jeol JEM-1011 electron microscope operated at 80 kV.Images were captured with an Olympus Morada camera using the Olympus AnalySIS software within the Centre of Microscopy and Microanalysis at The University of Queensland.Planimetric analysis (mitochondrial content, mitochondrial diameter and circularity ratio) was performed after importing the images into ImageJ software.In brief, 15 images were assessed for each of the 4 individual birds.Planimetric analysis was used to determine mitochondrial content by quantitating the total area covered by the tubules of the mitochondria with respect to the total area of the image under investigation.For the quantitation of mitochondrial diameter, the longest and shortest diameters of each of 10 transverse tubules of the mitochondrial network were estimated per image.For the estimation of circularity, the relationship between the shortest and longest diameter of each tubule under consideration was expressed as a ratio.

qRT-PCR estimation of pectoralis major mtDNA copy number
The mtDNA copy number is an established proxy for mitochondrial content.Total DNA contains the compositional information given that the number of copies of haploid mtDNA can be compared to the number of copies of a single copy gene localised to the diploid nuclear genome.Total DNA was extracted from the pectoralis of 10 birds using Qiagen DNeasy Blood and tissue kits following the manufacturers' instructions.
The development and performance of the duplex taqman assay is described in another paper (Reverter et al., 2017).In brief mtDNA copies were estimated by selective amplification of the mitochondrially encoded gene ND2.These values were compared to selective amplification of a single copy nuclear gene MSTN, with both reactions proceeding simultaneously in the same reaction vessel.We used the following formula to calculate mtDNA copies for every 2 nuclear genes based on the PCR cycles at which the fluorescence threshold was crossed (Reverter et al., 2017).
Where nDNA Ct and mtDNA Ct are the PCR cycles at which the fluorescence threshold was crossed for the nuclear (MSTN) and mitochondrial (ND2) DNA segments, respectively.
Given chickens have a diploid nuclear genome this formula equates to a mtDNA copy number estimate expressed for every 2 copies of the nuclear gene.This is, in effect, expressing the phenotype on a 'per nucleus' basis.For most chicken tissues composed of mononucleate cells this is equivalent to a 'per cell' basis.For the particular assessment of multinucleate muscle fibres it is best seen as representing a correction 'per unit volume of cytoplasm.' We ran all assays on a Qiagen Rotorgene system using 20 μl total reaction volumes, Thermofisher's Fast Advanced mastermix and PCR cycling parameters (95°C denaturation for 20 s followed by 40 cycles of 95°C for 1 s and 60°C annealing for 20 s) recommended by Thermofisher.A given 20 μl reaction contained 10 μl 2x Fast Advanced mastermix, 8 μl ultrapure molecular grade water and 1 μl each of the mtDNA and nDNA primer / probe combinations.The fluorescence threshold was set manually at 0.05 and fixed across runs.No template controls were included for each run.Prior to quantitation all gDNA was standardised to a concentration of 1 ng / μl.

Statistical analysis
Summary statistics (means and standard deviations) of mitochondrial content (%), mitochondrial diameter and mitochondrial circularity ratio were calculated using the = average and = stdev functions in Microsoft Excel.Pairwise Pearson correlations between bird liveweight, breast muscle weight and breast muscle mitochondrial DNA copy number were calculated using the correlation matrix function in GraphPad Prism software version 9. Statistical significance was declared if the P value was ≤ 0.05 and a trend declared if 0.05 < P ≤ 0.10.Estimates of mitochondrial content were log2 transformed prior to statistical analysis to reduce skewness.

TEM estimation of pectoralis major mitochondrial content
The summary mitochondrial content, diameter and circularity data for the 4 broilers can be found in Table 2.A representative TEM image can be found in Figure 1.The average content, diameter and circularity for the 4 broilers were 2.1%, 0.42 µm and 0.72, respectively.

qRT-PCR estimation of pectoralis major mtDNA copy number
The mtDNA copy number values and associated bird phenotype data for 10 broilers can be found in Table 2. Bird weight and breast muscle weight were positively correlated and approached significance (r = 0.57, P = 0.088).Breast muscle weight comprised an average of 28% of overall bird mass.The established negative relationship between mtDNA copy number and bird weight was not detected as significant in this study (r = −0.31,P = 0.39).For one individual (bird 17) we have both TEM estimates of mitochondrial content and qRT-PCR estimates of mtDNA copy number (Table 3).

Discussion
Modern Ross308 broilers purchased in 2019 at 50 d age and raised on an industry relevant ad libitum diet to 64 d possess a breast muscle mitochondrial content of 2.05% (Table 2).This is half that of Hydro Compact broilers phenotyped at 4.1% using the same planimetric analysis of breast muscle electron micrographs in the late 1970s (Kiessling, 1976;Kiessling, 1977).While both the 1970s study and the present study both used the standard stereometric technique as originally described by (Weibel, 1969), unfortunately, the age, nutritional background and rearing conditions of the 1970s broilers were not reported.This does make direct comparisons of the two studies somewhat challenging, but at the same time our new data provides an up-to-date benchmark against which further studies can confidently build upon and cross-reference.
Nevertheless, and all else being taken as equal, 45 years of additional genetic selection on muscle mass and feed efficiency has dramatically (∼2-fold) reduced muscle aerobic capacity from what were already very low levels.While we accept that some of the observed reduction could be environmental in origin (e.g.thermal regimen or physical activity), a formal partitioning analysis based on reciprocally feeding '1957 equivalent' and 2001 broilers 1957 and 2001 diets concluded that numerous improvements in broiler performance are 85-90% attributable to genetic gain (Havenstein, et al., 2003;Reyer, et al., 2015).Based on this partitioning analysis we assert that differences in nutrition between our birds and those from the 1970s likely accounts for only a small fraction of the difference we observe.Moreover, given our birds were free-ranging prior to purchase at d 50 their muscle mitochondrial content would have, if anything, been stimulated by exercise compared to more intensively reared, more rapidly grown, heavier (∼4 kg) industry birds.As such the very low 2.05% estimate we have produced for these ∼3 kg birds can actually be considered a conservative estimate, with the true value of modern industry birds likely even lower.
The combination of data from both TEM and qRT-PCR technologies taken from similar tissue samples of agestandardised birds raised under identical environmental conditions now allows us to reassess the likely mitochondrial contents of other bird tissues for which we previously had only mtDNA copy numbers (Reverter et al., 2017).Building on our previous study on Cobb broilers (Reverter et al., 2017) a revised breast muscle mitochondrial content of 2.05% (rather than the previously assumed 4%) suggests that in modern broilers heart, drumstick and white fat have actual mitochondrial contents of 9.3%, 6.6% and 1.0%, respectively.The predicted heart mitochondrial content value of 9.3% is dramatically lower than 10 differently sized mammals whose heart content has been typically measured to be between 24% (for larger mammals like dogs) -37% (for smaller mammals like mice) (Barth, et al., 1992).
The present finding of a relative reduction in breast muscle mitochondrial content in the modern (more muscular and more feed efficient) birds is entirely consistent with similar observations made in extant hypermuscular, feed efficient breeds representing a number of mammalian production species.These include MSTN mutant cattle (Kambadur, et al., 1997;McPherron & Lee, 1997), MSTN mutant sheep (Clop et al., 2006), Callipyge sheep (Clop et al., 2006;Jackson, et al., 1997aJackson, et al., , 1997b)), Large White pigs (Jing et al., 2015) and Yorkshire pigs (Sales & Kotrba, 2013).In those muscular, feed efficient breeds a muscle fibre composition shift favouring the low mitochondrial content Type IIB fibres has been repeatedly observed.
Where does a skeletal muscle mitochondrial content of just 2.05% sit in the broader context of life on earth?In fact, it is among the lowest measurements we could find for any muscle in any species.Intriguingly, it is at lower end of the estimations of 2.0-3.9% made for limb muscles of two wild cheetah, Acinonyx jubatus, the ultimate sprint adapted mammal (Williams et al., 1997).It is a staggering 17 fold lower than the 35.5% documented in the breast muscle of the highly athletic hummingbird, Selaphorus rufus (Mathieu-Costello, et al., 1992).Indeed, the only lower estimates we could identify come from two very large ruminants: the semitendinosus muscle in 190 kg Zebu cattle, Bos indicus (1.1 -1.3%), and the longissimus dorsi muscle in 240 kg Eland, Taurotragus oryx (1.4%) (Mathieu et al., 1981).Given muscle tissue mitochondrial content scales negatively with body mass e.g.(Hudson, 2009) such that smaller animals have higher contents, the extremely low breast muscle content observed here in relatively small ∼3 kg animals is even more striking that when taken at face value.Moreover, given the breast muscle in these birds is 28% of total body mass, the implications for the birds' overall bioenergetics are far greater than for the ruminants.
One conclusion is certain.For a given muscle in every individual of every species, there must be a lower limit beyond which muscle mitochondrial content can no longer be further diminished.Beyond this metabolic 'point of no return,' sustainable aerobic ATP supply will be unable to satisfy basic cellular ATP demand and tissue damage will inevitably ensue.Even for low bioenergetic demand sedentary commercial broilers reared in highly controlled settings there are indications we may already be approaching that particular tipping point.Concerning muscle pathologies including 'wooden breast' (Mutryn, et al., 2015) and 'white striping' (Boerboom, et al., 2018) have recently emerged.These are particularly noticeable in rapidly grown larger birds which we know are the very individuals possessing the lowest mitochondrial contents (Reverter et al., 2017).
Taking this suite of comparative and production data together reinforces the notion that the modern broiler represents a remarkable and very extreme model of skeletal muscle function.We suggest that the broiler industry needs to be cognisant not to push the birds bioenergetic system too hard (i.e. its mitochondrial content too low) in the relentless quest for increasing feed efficiencies.

Figure 1 .
Figure 1.Representative TEM microgram taken at 5000x magnification of a transverse section through the mitochondrial network of broiler breast muscle.The mitochondrial network comprises individual tubules that are of very small diameter and sparse.The transverse cross section of an example mitochondrial tubule is identified by a red arrow.

Table 1 .
The nutritional composition of the broiler starter diet.

Table 2 .
The summary statistics for breast mitochondrial content, diameter and circularity ratio.
a Values represent the mean ± SD obtained from 15 images.b Values represent the mean ± SD obtained from 10 transverse tubules in each of 15 images.

Table 3 .
The phenotype and mtDNA copy number data from the breast muscle of ten broilers.