Abstract
Angular limb deformity (ALD) affects many species of livestock and companion animals. The mechanisms of ALD development are not well understood, but previous research suggests the involvement of genetic risk factors. A case-control genome-wide association study (GWAS) was conducted with 40 ALD-affected and 302 unaffected Rambouillet rams and 40,945 single nucleotide polymorphisms (SNPs). Forelimbs of 6 ALD-affected rams were examined and diagnosed with osteochondrosis. Genome-wide or chromosome-wide significant SNPs were positioned exonic, intronic or within the 3′UTR of genes TSPAN18, NRG3 and NOVA2, respectively. These genes have previously described roles related to angiogenesis and osteoblast, osteoclast and chondrocyte proliferation and differentiation, which suggests the possibility for their involvement in the pathogenesis of osteochondrosis. Functional consequences of SNPs were evaluated through transcription factor binding site analysis, which predicted binding sites for transcription factors of known importance to bone growth, including SOX6, SOX9 and RUNX2. The identification of genetic risk factors for ALD may help to improve animal welfare and production in Rambouillet, a breed known to be at risk for ALD development. This study proposes genes TSPAN18, NRG3 and NOVA2 as targets for further research towards understanding the etiology of ALD in Rambouillet sheep.
Similar content being viewed by others
Introduction
Angular limb deformity (ALD) is a welfare and longevity concern for livestock and domestic animals. The term ALD describes deviations in the frontal plane of the limb that can be medial or lateral, unilateral or bilateral and congenital or acquired1. To the best of our knowledge there is no single disease pathway responsible for ALD. Congenital ALD has been associated with perinatal factors such as maternal nutrition, teratogen exposure, premature birth, twin pregnancy or placentitis1,2. The causes of acquired ALD are likely multifactorial and may involve the interaction of trauma, genetics, rapid growth rate, endocrinological and/or nutritional factors3,4. Dietary mineral imbalance, such as of vitamin D, calcium, phosphorus, copper, zinc or magnesium, have also been known to contribute to ALD3,5. Evidence in multiple species strongly suggests a genetic component to ALD etiology, which may function in conjunction with other predisposing environmental influences2,6,7.
Osteochondrosis is frequently described as the underlying pathology of acquired ALD in sheep, cattle, horses, pigs, dog and poultry8,9. Osteochondrosis is a non-infectious disorder characterized by focal disturbance of endochondral ossification that may affect the articular cartilage (osteochondrosis dissecans) or the physes (physeal osteochondrosis) of the long bones of animals and humans10. These disturbances may be caused by hypertrophic chondrocytes that fail to progress or a lack of chondrocyte organization in the proliferative zone, which ultimately lead to areas of retained cartilage promoting lameness and leg weakness11,12. The development of osteochondrosis is strongly associated with high-energy diets and rapid growth, and it is thought to involve the interaction of genetic, nutritional and mechanical factors, with no one factor being entirely responsible for disease progression9,12. Previous studies have reported genetic regions associated with osteochondrosis in horses and pigs, although specific genetic mechanisms remain undetermined13,14.
Incidence of acquired ALD have been observed in Rambouillet rams during central performance ram testing at North Dakota State University (NDSU) and University of Wyoming (UWY). In the present study, a subset of affected rams were examined with gross and microscopic pathology to better understand the cause of ALD in these animals. The relatively low rates of disease development suggest the contribution of genetic risk factors; therefore, genotype data were collected in order to test for SNPs associated with ALD. Genome-wide association study (GWAS) was conducted with 40 ALD-affected and 302 unaffected Rambouillet rams who were enrolled in NDSU and UWY ram tests over a three-year period. Three different models were evaluated for comparison and interpretation of GWAS results, and the significant GWAS markers and corresponding genes were investigated for known biological roles related to bone growth and development.
Results
Central performance ram test
Rams were 7 ± 3 months of age at the time of enrollment in the NDSU and UWY central performance ram tests. Of the 342 Rambouillet rams collected from ram tests in 2019, 2020 and 2021, 40 rams developed ALD and 302 rams remained unaffected by conclusion of the 140-day testing period. Rams were considered to be affected by ALD upon presentation of abnormal conformation (valgus or varus deviations) of one or both forelimbs, as determined by trained UWY and NDSU sheep center staff who had daily interactions with the animals (Fig. 1). Disease incidence varied from 2.70 to 21.67% by year and location, with an average of 14.04% of NDSU and 9.83% of UWY rams developing ALD, and an overall study incidence rate of 11.70% (Supplementary Table S1). A consistent nutritional profile was provided in each year of the study.
Visual presentation of ALD in a Rambouillet ram from UWY. The ram to the left demonstrates normal structure of the forelimbs and the ram to the right displays carpal valgus, indicative of ALD.
Gross pathology and histopathology examination
The examined forelimbs of six ALD-affected rams displayed varying degrees of carpal valgus with occasional slight deviation of the fetlock. The humerus, olecranon, radius, metacarpal and phalanges were sectioned and examined. Gross changes consisting of mild to moderate irregular physeal thickening primarily affected the distal radial physes, although some mild changes were observed in the olecranon physis of two rams and the distal metacarpal physis of one ram.
The histologic changes were similar amongst the different animals and varied only in the degree of severity. Changes were primarily characterized by irregular thickening of the physis (growth plate) consisting of both the proliferating and hypertrophic zones that extend down into the metaphysis (Fig. 2). Occasional foci of retained cartilage were present in the metaphysis and epiphysis. Chondrocyte maturation appeared normal with trabecular cartilage cores lacking appreciable chondrocytes. Evidence of increased fragility such as microfractures were not observed. The histological changes of all ALD-affected forelimbs were consistent with physeal osteochondrosis which reflected an underlying failure of endochondral ossification.
Frontal sections of distal radius from two ALD-affected animals. Images show mild (a) to moderate (b) irregular physeal thickening primarily affecting the distal radial physes (arrows). The irregular physeal thickening consists of both the proliferating and hypertrophic zones which extend down into the metaphysis.
Genome-wide association study
Population structure and its relationship to phenotype, test location and test year were evaluated through principal component analysis (PCA). The first principal component (PC1) had an eigenvalue of 9.62 and explained 11.64% of the total genetic variance and the PC2 had an eigenvalue 7.56 and explained 9.15% of the total genetic variance. There did not appear to be any specific grouping of ALD-affected animals apart from control animals by either PC1 or PC2 (Fig. 3a). Similarly, rams from NDSU and UWY appeared to be evenly represented across PC1 and PC2 (Fig. 3b), as were rams from study years 2019, 2020, and 2021 (Fig. 3c).
Principal component analysis (PCA) of Rambouillet rams. In each panel, the PC1 is plotted on the x-axis and PC2 is plotted on the y-axis. (a) PCA with rams coded by phenotype, given by: red, ALD-affected rams; light blue, control rams; (b) PCA with rams coded by testing location, given by: yellow, NDSU; gray, UWY; (c) PCA with rams coded by test year, given by: violet, rams from test year 2019; black, 2020; green, 2021.
Genome-wide association study models were investigated using fastGWA linear regression in GCTA software, the Scalable and Accurate Implementation of GEneralized mixed model (SAIGE) R package, and a 1 degree of freedom (df) chi-square allelic test in plink software. In all models, the same three SNPs were significant by Bonferroni chromosome-wide (P-values < 1.08e−05) and/or permutation genome-wide (P-values < 2.00e−06) significance thresholds: rs160736723 located on chromosome 15, rs427563170 on chromosome 25, and rs416810983 on chromosome 14 (Fig. 4a). The estimated odds ratio (OR) of significant SNPs were 4.77, 4.03 and 3.50, respectively, and all OR were supported by 95% confidence intervals (Table 1, Supplementary Fig. S1). The SNP minor allele frequencies (MAF) of ALD-affected rams were 20.00%, 25.00% and 27.50%, compared to 4.98%, 7.64% and 9.77% for unaffected rams (Table 1). The genomic inflation factor (λGC) for the fastGWA, SAIGE, and chi-square models were 1.05, 1.02, and 1.06, respectively (Fig. 4b), indicating a lack of appreciable inflation of GWAS P-values15.
Multi-Manhattan and QQ plots displaying genome-wide association study results for incidence of ALD in Rambouillet rams. (a) Multi-Manhattan plot for the chi-square, fastGWA, and SAIGE models. Genome-wide significance is represented by the black line and defined by permutation testing (P-values < 2.00e−06); chromosome-wide significance is represented by the red line and defined by Bonferroni correction of the largest chromosome (P-values < 1.08e−05). (b) Quantile–quantile (QQ) plot for the chi-square, fastGWA, and SAIGE models displaying the expected versus observed P-values for each model.
Genomic context of significant SNPs
Genome-wide significant SNPs were located within genes tetraspanin 18 (TSPAN18; rs160736723, synonymous variant) and neuregulin 3 (NRG3; rs427563170, intronic variant). Additionally, the gene tumor protein p53 inducible protein 11 (TP53I11) and the uncharacterized ncRNA LOC121816715 were located within the region defined by ± 100 kb of SNP rs160736723. The chromosome-wide significant marker rs416810983 was located in a gene-rich region containing 10 genes within ± 100 kb (Supplementary Table S2) and was positioned 1,961 bp downstream in the unannotated 3′ UTR of the gene, NOVA alternative splicing regulator 2 (NOVA2). Although the 3′ UTR of NOVA2 is annotated as 352 bp in length in the sheep reference genome (ARS-UI_Ramb_v2.0), the assemblies of human and mouse contain much longer annotations; the Mus musculus NOVA2 3′ UTR is 6,701 bp (GRCm39 C57BL/6J genome) and the Homo sapiens NOVA2 3′ UTR annotation is 6,129 bp (GRCh38.p14 genome). Reference RNA tracks imply that the NOVA2 3′ UTR is a similar length in sheep, as RNA signal extends from the gene to roughly 5552 bp downstream according to the “RNA-seq exon coverage” track in the NCBI Genome Data Viewer. Therefore, it is likely that SNP rs416810983 is located within the Ovis aries NOVA2 3′ UTR.
Functional context of significant SNPs
The SNPs and corresponding genes implicated through GWAS were further investigated to better understand their potential functional context. Significant markers were located either exonic, intronic or in the 3′ UTR, suggesting the possibility for SNP alleles to be located within transcription factor binding site (TFBS) motifs16. The SNPs rs160736723, rs427563170 and rs416810983 were queried through the search tool FABIAN to compare against known TFBS motifs within the JASPAR 2022 database. Seventeen TFBS were predicted with score differences greater than ± 0.1 between reference and alternate SNP alleles, indicating either loss or gain of motif specificity (Table 2). The predicted TFBS included transcription factors with known relevance to bone growth, including RUNX2, NFE2L1, TBX5 at rs160736723, CREB1, CREM and SOX9 at rs427563170 and SOX6 at rs416810983.
The STRING database was used to identify known and predicted protein-protein interactions of the genes implicated by GWAS results. The Homo sapiens orthologs of TSPAN18, NRG3 and NOVA2 were queried and results revealed five protein interactions with TSPAN18 and eight protein interactions each with NRG3 and NOVA2 (Table 3). Interactant proteins with notable biological roles relevant to bone health included ADAM10 with TSAPN18 and HNRNPK and MBNL1 with NOVA2. For NRG3, notable interactions included EGFR, ERBB3, ERBB4, HRAS, KRAS and HSP90AA1.
Discussion
Osteochondrosis has previously been described as a cause of ALD in sheep, horses, pigs and other species8,9. Physeal osteochondrosis is characterized by lesions of retained cartilage in the bone resulting from the focal failure of endochondral ossification. Disease development is thought to require several contributing factors, including high-energy diet, rapid growth rate and genetics12, and early lesions may heal and never progress to clinical disease41. Although a number of biological mechanisms have been proposed, failure of blood delivery to the growth cartilage is thought to be the initiating event in disease pathogenesis8. Chondrocyte-driven lengthening of long bones is an intricate process that involves the coordinated control of transcription factors, growth factors and cytokines, that culminates in endochondral ossification, the replacement of avascular cartilage with highly vascularized bone42. Immature chondrocytes direct chondrocyte differentiation through expression of SOX5, SOX6 and SOX9 transcription factors. These cells also produce the extracellular matrix components collagen type II α1 and aggrecan43. Hypertrophy is initiated by downregulation of SOX5, SOX6, SOX9 and collagen type II α1, accompanied by the simultaneous upregulation of RUNX2, which regulates both chondrocyte maturation and osteoblast differentiation12,42. Hypertrophic chondrocytes produce type X α1 collagen to create the matrix template for later calcification42,43,44. In later stage hypertrophy, chondrocytes produce vascular endothelial growth factor (VEGF), matrix metalloproteinase 13 (MMP13) and osteopontin to direct angiogenesis and the invasion of blood cells into the cartilage matrix. These vessels facilitate the delivery of osteoclasts, osteoblasts, and bone marrow cells to the ossification front, which orchestrate cartilage resorption and the deposition of mineralized bone42,43,45,46. Vascular invasion of the cartilage-bone interface is critical for matrix calcification and successful endochondral ossification12.
The biological relevance of the results of this study are supported by the current literature. The genes and transcription factors implicated by this GWAS have previously described roles in angiogenesis and the differentiation and proliferation of cartilage and bone cells, indicating pertinence to endochondral ossification and suggesting potential roles in osteochondrosis development. The most significant GWAS result is SNP rs160736723, a synonymous variant within TSPAN18 (Fig. 5a). The gene TSPAN18 has been found to be highly expressed in endothelial cells and promotes the cell surface expression of the store-operated Ca2+ entry (SOCE) channel pore-subunit ORAI147,48. Experiments in mice have suggested that Orai1 has a critical role in bone homeostasis through the regulation of osteoblasts and osteoclasts47. The lack of SOCE in MC3T3-E1 pre-osteoblastic cells had an inhibitive effect on mineralized nodule formation and Orai1−/− stromal cells showed impaired mineral deposition in in vitro experiments. Similarly, osteoclastogenesis was impaired in Orai1−/− due to defects in the fusion of pre-osteoclasts47. In mouse and zebrafish, Tspan18 expression was noted in the vasculature of developing embryos and in the adult during wound healing. Deficiency of Tspan18 was found to result in the downregulation of VEGFR2, suggesting Tspan18 as either a direct or indirect regulator of the VEGF pathway49. Additionally, TSPAN18 has an experimentally determined interaction with ADAM10 (Table 3), a protease found to affect the long bone longitudinal growth in mice through the CXCL12/CXCR4 signaling axis31.
Summary of biological implications of GWAS results. (a) Deficiency of TSPAN18 gene expression may have negative repercussions on endochondral ossification through the downregulation of VEGFR2 leading to inhibition of the VEGF pathway and angiogenesis, or through decreased cell-surface expression of ORAI1 resulting in reduced Ca2+ signaling affecting osteoclast and osteoblast cell maturation. Predicted transcription factor binding site (TFBS) motifs, denoted by the dashed line, for NFE2L1, RUNX2 and TBX5 show loss of specificity with the rs160736723 SNP allele, potentially affecting TSPAN18 transcription; (b) Changes to NRG3 gene expression may impact chondrocyte growth and proliferation through the NRG3/ERBB4 pathway. Binding site motif specificities for transcription factors CREM, CREB1 and SOX9 are predicted to increase with the SNP allele at rs427563170; (c) Decreased gene expression of NOVA2 may decrease vascularization necessary for endochondral ossification. The SNP allele at rs416810983 results in decreased TFBS motif specificity of SOX6.
The significant SNP rs427563170 is an intronic variant within gene NRG3 (Fig. 5b). The extracellular epidermal growth factor (EGF)-like domain of NRG3 is a ligand that binds to the extracellular domain of ERBB4 (HER4), a member of ERBB family of receptor tyrosine kinases50. ErbB4 has been shown to be expressed in developing mouse cartilage and bone cells51 and may have roles in chondrocytic growth and differentiation52. The STRING database indicates that NRG interacts with EGFR, a receptor protein critical for bone formation (Table 3), as well as with proteins ERBB3, HRAS, HSP90AA1 and KRAS that promote angiogenesis and neovascularization32,34,36.
The chromosome-wide significant SNP, rs416810983, is positioned within the 3′ UTR of NOVA2 (Fig. 5c). NOVA2 is an alternative splicing regulator that is expressed in vascular endothelial cells and is essential for angiogenesis and vascular morphogenesis53. There are known STRING interactions and/or protein coexpressions of NOVA2 with MBNL1 and hnRNPK (Table 3). Expression of the lncRNA MBNL1 has been found to be upregulated in vascular smooth muscle cells during differentiation40. Transcription of osteocalcin, a factor that stimulates matrix mineralization, is repressed by hnRNPK. Additionally, hnRNPK has regulatory roles in osteogenesis through interactions with lncRNA‐OG in mesenchymal stem cells39.
The expression of TSPAN18, NRG3, and NOVA2 may be regulated in part by transcription factors with known importance to bone growth. The alternate SNP allele at rs160736723 within TSPAN18 was predicted to result in a loss of TFBS motif specificity for RUNX2, NFE2L1 and TBX5 transcription factors (Table 2). Runx2 is a major transcription factor in bone growth with regulatory roles in osteogenesis, chondrogenesis, chondrocyte maturation and osteoblast differentiation23. Importantly, chondrocyte hypertrophy, which initiates endochondral ossification, is controlled by upregulation of RUNX246. Depletion of NFE2L1 (also known as NRF1) has been shown to impair osteoblast differentiation and osteoclastogenesis22. The gene TBX5 was identified as a candidate for osteochondrosis development in pigs through GWAS and differential transcript abundance of healthy versus defect cartilage13. Haploinsufficiency of TBX5 is known to cause Holt-Oram syndrome in human patients, which is characterized in part by skeletal malformations54.
Predicted TFBS for SOX9, CREB1 and CREM showed gain of specificity with the alternate SNP allele at rs427563170 within NRG3 (Table 2). Expression of SOX9 must be timed appropriately, as it is essential for chondrogenesis but prevents cartilage vascularization and successful endochondral ossification if not downregulated in the hypertrophic zone of the growth plate28. The CREB family are involved in regulating chondrocyte proliferation important for endochondral ossification26 and a member of the CREM family, ICER, is thought to regulate the activity and/or the expression of ATF/CREB transcription factors involved in osteoclastogenesis and bone homeostasis27. The predicted SOX6 TFBS showed loss of specificity with the rs416810983 alternate allele within the NOVA2 3′ UTR (Table 2). SOX6 is a member of the ‘SOX-Trio’ of transcription factors, together with SOX5 and SOX9, that regulate chondrogenesis28.
Many previous studies support a multifactorial etiology of osteochondrosis4,10, and a high plane of nutrition leading to rapid growth is often cited as a contributing factor of disease12. The high-energy diets fed during central performance ram tests are intended to provide balanced nutrition and support uninhibited growth to facilitate performance comparisons between animals55. We hypothesize that a combination of rapid growth and the presence of genetic risk factors likely lead to the development of clinical disease in the ALD-affected rams of this study. In view of the multifactorial nature of osteochondrosis, it is not surprising that the SNPs described in this study do not explain all of the disease variation. Unaffected rams with variant alleles may have had some level of subclinical osteochondrosis that was not observable or that resolved before progressing to clinical disease41 or may have not experienced enough environmental stress to facilitate lesion development.
This study suggests that the genes TSPAN18, NRG3 and NOVA2 may have biological functions in ALD disease progression. Despite the limited sample size, this study benefited from the environmental uniformity experienced by all affected and unaffected rams. Additionally, significant SNPs were identified through three different GWAS models, which lends confidence to these associations. Further research with a larger sample size is necessary to validate this study as well as to continue to elucidate the complex biological mechanisms that likely underpin ALD development.
In this study, Rambouillet rams with acquired ALD had gross and microscopic lesions consistent with physeal osteochondrosis. Rams with one or more alternate alleles at rs160736723, rs427563170 or rs416810983 were found to be 3.5 to 4.8 times more likely to develop ALD than rams with reference alleles. The genes implicated by GWAS results have previously described roles in angiogenesis, osteogenic differentiation and chondrocyte development and proliferation, suggesting their relevance to osteochondrosis. Further, the TFBS with predicted loss or gain of motif specificity at significant SNPs have known roles in the regulation and timing of endochondral ossification. Further work is necessary to validate GWAS findings and to elucidate the potential roles of TSPAN18, NRG3 and NOVA2 in the pathogenesis of physeal osteochondrosis. This study proposes three SNPs as genetic risk factors for ALD in Rambouillet rams and indicates potential genetic mechanisms that may contribute to disease.
Methods
Central performance ram test
Central performance ram testing was conducted as previously described (Becker et al. 2023). In each study year, ram lambs of 7 ± 3 months of age from more than 30 regional (WY, ND, SD, MT, CO) seedstock producers were brought to the University of Wyoming-Laramie Research and Extension Center (Laramie, WY; 41° 17′ N, − 105°40′ W) or North Dakota State University-Hettinger Research and Extension Center (Hettinger, ND; 46° 01′ N, − 102°65′ W). Rams were managed as a single cohort in a dry-lot management system at each testing location and were provided fully balanced ad libitum textured diets that met requirements for all phases of ram growth, containing 15% to 17% crude protein and 68% to 76% total digestible nutrients on a dry matter basis. Each ram test began in the fall and concluded in the spring and lasted for a total duration of 140 days. Ram test protocols were approved by North Dakota State University Institute for Animal Care and Use Committee (IACUC). All methods were carried out in accordance with relevant IACUC guidelines and regulations. The methods of this study have been reported in accordance with the ARRIVE guidelines.
Gross pathology and histopathology examination
Nine forelimbs from six ALD-affected rams (three rams each from NDSU and UWY) were evaluated for gross and histological pathology. Each limb was evaluated grossly and the humerus, olecranon, radius, metacarpal and phalanges were sectioned. Decalcified, 5 µm thick, hematoxylin and eosin stained frontal section of the distal radius and the sagittal section of the olecranon of each submitted forelimb and the frontal section of the distal metacarpus of two forelimbs were evaluated. Images were taken on a BX 43 microscope, 2 × 0.08 NA plan apo lens on a DP 23 camera at 3088 × 2076 resolution.
DNA genotyping and quality control
Whole blood or tissue samples were collected from 342 Rambouillet rams over a three-year period. Rams were sampled from NDSU in 2019 (n = 44), 2020 (n = 60), 2021 (n = 55), and from UWY in 2019 (n = 74), 2020 (n = 43) and 2021 (n = 66). The genotype data of 313 of these rams have been analyzed in previously published work56. DNA was isolated from blood at the University of Idaho using the phenol–chloroform method57 and tissue samples were provided to AgResearch for DNA extraction. The DNA samples were genotyped with the Applied Biosystems™ Axiom™ Ovine Genotyping Array (50K) consisting of 51,572 single nucleotide polymorphism (SNP) markers (Thermo Fisher Scientific, catalog number 550898) or the AgResearch Sheep Genomics 60K SNP chip consisting of 68,848 SNP markers (GenomNZ, AgResearch, New Zealand). Duplicate markers within a panel were filtered to retain the marker with the highest call rate (CR). Compatible markers between panels were matched by marker name and genome position and plink v1.9 was used to merge genotype array data58 (http://pngu.mgh.harvard.edu/purcell/plink/). Markers were filtered with plink v1.9 to remove non-autosomal markers and those with CR < 0.90, minor allele frequency < 0.01, and Hardy-Weinberg Equilibrium P-value < 1e−06. A total of 40,945 SNP markers were carried through to final analysis. All rams had a CR of 95% or greater.
Genome-wide association study
Principal component analysis (PCA) was conducted to visualize the population structure and phenotypic distribution among ALD-affected and unaffected rams. Analysis was conducted with plink v1.9 and visualized with the packages ggplot2 and patchwork in R version 4.2.258,59,60,61. Each plot was constructed with the first principal component (PC1) on the x-axis and the second principal component (PC2) on the y-axis. Rams were color-coded by phenotype (ALD-affected or unaffected), test location (NDSU or UWY) and test year (2019, 2020 or 2021). The proportion of variance explained by each PC were calculated by dividing the PC eigenvalue by the sum of eigenvalues.
In total 342 Rambouillet rams, including 40 ALD-affected and 302 unaffected, were examined in three genome-wide association study (GWAS) models. The case-to-control ratio of the study was 1-to-7.55, which approached but did not exceed the threshold of 1-to-10 generally used to define an imbalanced GWAS62. For this reason, three models with the ability to handle imbalance between study groups were chosen for analysis and comparison. The data were analyzed through a generalized linear mixed model (GLMM)-based fastGWA-GLMM in GCTA version 1.93.2 software63, a Scalable and Accurate Implementation of GEneralized (SAIGE) mixed model v0.45 available as a R package62, and a 1 degree of freedom (df) chi-square allelic test in plink v1.9 software58,64. In the fastGWA-GLMM, the genetic variance statistic (Vg) was not significant, likely due to the lack of highly related individuals within the dataset. Therefore, the program implemented the linear regression (fastGWA) model instead of GLMM65. The fastGWA model used a sparse genomic relationship matrix (GRM) to account for population stratification and was run with default parameters. The SAIGE model was run using the R scripts “step1_fitNULLGLMM.R” and “step2_SPAtests.R” with default minor allele count and frequency thresholds, and the leave one chromosome out (LOCO) method was used to construct a different GRM for each chromosome. In plink v1.9, the flags ‘–assoc’, ‘–adjust’, and ‘–ci 0.95’ were used to run the chi-square test and to report 95% confidence intervals for estimated odds ratio (OR) statistics. The GWAS results were visualized with the package CMplot and OR confidence intervals were visualized with gridextra and ggplot2 in R60,66,67.
To avoid over interpretation of results and to limit the possibility of Type I or Type II error, the genome-wide significance threshold for this study was determined by adaptive permutation68. Permutation testing was carried out in plink v1.9 (flags ‘–aperm’, ‘–assoc perm’). Each SNP underwent a minimum of 500 permutations and up to 1,000,000 permutations under default conditions for the pruning alpha threshold, pruning beta, initial pruning interval and the interval increase rate. The most significant P-value achieved by chance after 1,000,000 permutations was used to define the threshold for genome-wide significance. Chromosome-wide significance was calculated using the Bonferroni adjustment for the largest chromosome (chromosome 1; 4622 markers).
Genomic context of significant SNPs
The database Ensembl69 was used to identify the variant type of significant SNPs. In order to understand the genomic contexts of SNPs, the reference genome ARS-UI_Ramb_v2.070 was searched using the GenomeBrowser tool in NCBI71. Genes or noncoding elements located within ± 100 kb of significant SNPs were recorded for further investigation.
Functional context of significant SNPs
Markers located within a gene or within the 3′ UTR were evaluated for predicted transcription factor binding site (TFBS) matrix score differences between reference and alternate SNP alleles. Reference genome sequences comprised of 11 bp total, with the SNP allele (reference or alternate) in the 6th position, were queried through the online software FABIAN72. Sequences were tested against detailed transcription factor flexible models (TFFM) compiled within the JASPAR 2022 database73. The proteins of implicated genes were queried through the STRING database (using Homo sapiens orthologs) for known and predicted protein-protein interactions74. Interacting proteins and predicted transcription factors were investigated through literature review for potential relevance to bone health and growth.
Approval for animal experiments
All ram test protocols were approved by North Dakota State University Institute for Animal Care and Use Committee (# 20210012). Informed consent was collected from all animal owners upon enrollment of their rams in the central performance ram test.
Data availability
All genotype data analyzed in this study have been made available from the EVA repository. Data accession details are as follows: project PRJEB58836, analyses ERZ15609617, available at https://www.ebi.ac.uk/eva/?eva-study=PRJEB58836.
References
García-López, J. M. Angular limb deformities: Growth augmentation. Vet. Clin. Equine Pract. 33, 343–351. https://doi.org/10.1016/j.cveq.2017.03.005 (2017).
Dittmer, K. E. & Thompson, K. G. Approach to investigating congenital skeletal abnormalities in livestock. Vet. Pathol. 52, 851–861. https://doi.org/10.1177/0300985815579999 (2015).
Orsini, J. A. & Kreuder, C. Musculoskeletal disorders of the neonate. Vet. Clin. North Am. Equine Pract. 10, 137–166. https://doi.org/10.1016/S0749-0739(17)30372-3 (1994).
van Oldruitenborgh-Oosterbaan, M. M. S., Mol, J. A. & Barneveld, A. Hormones, growth factors and other plasma variables in relation to osteochondrosis. Equine Vet. J. 31, 45–54. https://doi.org/10.1111/j.2042-3306.1999.tb05313.x (1999).
Dittmer, K. E. & Thompson, K. G. Vitamin D metabolism and rickets in domestic animals: A review. Vet. Pathol. 48, 389–407. https://doi.org/10.1177/0300985810375240 (2011).
Bradshaw, R. H., Kirkden, R. D. & Broom, D. M. A review of the aetiology and pathology of leg weakness in broilers in relation to welfare. Avian Poult. Biol. Rev. 13, 45–104. https://doi.org/10.3184/147020602783698421 (2002).
Lappalainen, A. K. et al. Breed-typical front limb angular deformity is associated with clinical findings in three chondrodysplastic dog breeds. Front. Vet. Sci. 9, 1099903. https://doi.org/10.3389/fvets.2022.1099903 (2023).
Olstad, K., Ekman, S. & Carlson, C. S. An update on the pathogenesis of osteochondrosis. Vet. Pathol. 52, 785–802. https://doi.org/10.1177/0300985815588778 (2015).
Craig, L. E., Dittmer, K. E., & Thompson, K. G. Bones and Joints. Jubb, Kennedy & Palmer’s Pathology of Domestic Animals: Volume 1, 16–163.e1; https://doi.org/10.1016/b978-0-7020-5317-7.00002-3 (2016).
Jeffcott, L. B. Osteochondrosis - An international problem for the horse industry. J. Equine Vet. Sci. 16, 32–37. https://doi.org/10.1016/s0737-0806(96)80063-3 (1996).
Wardale, R. J. & Duance, V. C. Characterisations of articular cartilage and growth plate cartilage collagens in porcine osteochondrosis. J. Cell Sci. 107, 47–59. https://doi.org/10.1242/jcs.107.1.47 (1994).
Ytrehus, B., Carlson, C. S. & Ekman, S. Etiology and pathogenesis of osteochondrosis. Vet. Pathol. 44, 429–448. https://doi.org/10.1354/vp.44-4-429 (2007).
Rangkasenee, N. et al. Genome-wide association identifies TBX5 as candidate gene for osteochondrosis providing a functional link to cartilage perfusion as initial factor. Front. Genet. https://doi.org/10.3389/fgene.2013.00078 (2013).
McCoy, A. M. et al. Identification and validation of risk loci for osteochondrosis in standardbreds. BMC Genom. 17, 1–11. https://doi.org/10.1186/s12864-016-2385-z (2016).
Williams, C. J. et al. Genome wide association study of response to interval and continuous exercise training: The Predict-HIIT study. J. Biomed. Sci. https://doi.org/10.1186/s12929-021-00733-7 (2021).
Khan, A. H., Lin, A. & Smith, D. J. Discovery and characterization of human exonic transcriptional regulatory elements. PLoS One. 7, e46098. https://doi.org/10.1371/journal.pone.0046098 (2012).
Satoda, M. et al. Mutations in TFAP2B cause Char syndrome, a familial form of patent ductus arteriosus. Nat. Genet. 25, 42–46. https://doi.org/10.1038/75578 (2000).
Abousoliman, I. et al. Genome-wide analysis for early growth-related traits of the locally adapted Egyptian Barki sheep. Genes 12, 1243. https://doi.org/10.3390/genes12081243 (2021).
Rahman, M., Akhtar, N., Jamil, H., Banik, R. & Asaduzzaman, S. TGF-β/BMP signaling and other molecular events: Regulation of osteoblastogenesis and bone formation. Bone Res. 3, 15005. https://doi.org/10.1038/boneres.2015.5 (2015).
Hyc, A., Osiecka-Iwan, A. & Moskalewski, S. Could BMPs therapy be improved if BMPs were used in composition acting during bone formation in endochondral ossification?. Int. J. Mol. Sci. 23, 10327. https://doi.org/10.3390/ijms231810327 (2022).
Stegen, S. & Carmeliet, G. Hypoxia, hypoxia-inducible transcription factors and oxygen-sensing prolyl hydroxylases in bone development and homeostasis. Curr. Opin. Nephrol. Hypertens. 28, 328–335. https://doi.org/10.1097/MNH.0000000000000508 (2019).
Kim, H. M., Han, J. W. & Chan, J. Y. Nuclear factor erythroid-2 like 1 (NFE2L1): Structure, function and regulation. Gene 584, 17–25. https://doi.org/10.1016/j.gene.2016.03.002 (2016).
Hallett, S. A., Ono, W. & Ono, N. The hypertrophic chondrocyte: To be or not to be. Histol. Histopathol. 36, 1021. https://doi.org/10.14670/HH-18-355 (2021).
Ariki, R. et al. Homeodomain transcription factor Meis1 is a critical regulator of adult bone marrow hematopoiesis. PloS One 9, e87646. https://doi.org/10.1371/journal.pone.0087646 (2014).
Chang, M. R. & Griffin, P. R. RORβ modulates a gene program that is protective against articular cartilage damage. Plos one. 17, e0268663. https://doi.org/10.1371/journal.pone.0268663 (2022).
Long, F., Schipani, E., Asahara, H., Kronenberg, H. & Montminy, M. The CREB family of activators is required for endochondral bone development. Development. 128, 541–550. https://doi.org/10.1242/dev.128.4.541 (2001).
Chandhoke, T. K. et al. Osteopenia in transgenic mice with osteoblast-targeted expression of the inducible cAMP early repressor. Bone 43, 101–109. https://doi.org/10.1016/j.bone.2008.03.012 (2008).
Hattori, T. et al. SOX9 is a major negative regulator of cartilage vascularization, bone marrow formation and endochondral ossification. Development 137(901–911), 2010. https://doi.org/10.1242/dev.045203 (2010).
Syed, M., Mushtaq, S., Loya, A. & Hassan, U. NKX3. 1 a useful marker for mesenchymal chondrosarcoma: An immunohistochemical study. Ann. of Diagn. Pathol. 50, 151660. https://doi.org/10.1016/j.anndiagpath.2020.151660 (2021).
Courel, M., Friesenhahn, L. & Lees, J. A. E2f6 and Bmi1 cooperate in axial skeletal development. Dev. Dyn. 237, 1232–1242. https://doi.org/10.1002/dvdy.21516 (2008).
Mizuno, S. et al. ADAM10 is indispensable for longitudinal bone growth in mice. Bone. 134, 115273. https://doi.org/10.1016/j.bone.2020.115273 (2020).
Chandra, A. et al. Epidermal growth factor receptor (EGFR) signaling promotes proliferation and survival in osteoprogenitors by increasing early growth response 2 (EGR2) expression. J. Biol. Chem. 288, 20488–20498. https://doi.org/10.1074/jbc.M112.447250 (2013).
Duan, B. et al. Overexpression of ERBB3 promotes proliferation, migration, and angiogenesis in nasopharyngeal carcinoma. Int. J. Clin. Exp. Pathol. 12, 2931–2940 (2019).
Liang, X. et al. Overexpression of ERBB4 rejuvenates aged mesenchymal stem cells and enhances angiogenesis via PI3K/AKT and MAPK/ERK pathways. FASEB J. 33, 4559–4570 (2019).
Serban, D., Leng, J. & Cheresh, D. H-ras regulates angiogenesis and vascular permeability by activation of distinct downstream effectors. Circ. Res. 102, 1350–1358. https://doi.org/10.1161/CIRCRESAHA.107.169664 (2008).
Karar, J. & Maity, A. PI3K/AKT/mTOR pathway in angiogenesis. Front. Mol. Neurosci. https://doi.org/10.3389/fnmol.2011.00051 (2011).
Xiang, X., You, X. M. & Li, L. Q. Expression of HSP90AA1/HSPA8 in hepatocellular carcinoma patients with depression. Onco. Targets Ther. 11, 3013–3023. https://doi.org/10.2147/OTT.S159432 (2018).
Figueras, A. et al. The impact of KRAS mutations on VEGF-A production and tumour vascular network. BMC Cancer 13, 1–11. https://doi.org/10.1186/1471-2407-13-125 (2013).
Wang, Z. et al. The emerging roles of hnRNPK. J. Cell. Physiol. 235, 1995–2008. https://doi.org/10.1002/jcp.29186 (2020).
Lim, Y. H., Ryu, J., Kook, H. & Kim, Y. K. Identification of long noncoding RNAs involved in differentiation and survival of vascular smooth muscle cells. Mol. Ther. Nucl. 22, 209–221. https://doi.org/10.1016/j.omtn.2020.08.032 (2020).
Laverty, S. & Girard, C. Pathogenesis of epiphyseal osteochondrosis. Vet. J. 197, 3–12. https://doi.org/10.1016/j.tvjl.2013.03.035 (2013).
Dai, J. & Rabie, A. B. M. VEGF: An essential mediator of both angiogenesis and endochondral ossification. J. Dent. Res. 86, 937–950. https://doi.org/10.1177/154405910708601006 (2007).
Li, J. & Dong, S. The signaling pathways involved in chondrocyte differentiation and hypertrophic differentiation. Stem Cells Int. https://doi.org/10.1155/2016/2470351 (2016).
Shen, G. The role of type X collagen in facilitating and regulating endochondral ossification of articular cartilage. Orthod. Craniofac. Res. 8, 11–17. https://doi.org/10.1111/j.1601-6343.2004.00308.x (2005).
Mackie, E., Ahmed, Y. A., Tatarczuch, L., Chen, K. S. & Mirams, M. Endochondral ossification: how cartilage is converted into bone in the developing skeleton. Int. J. Biochem. Cell Biol. 40, 46–62. https://doi.org/10.1016/j.biocel.2007.06.009 (2008).
Chan, W. C. W., Tan, Z., To, M. K. T. & Chan, D. Regulation and role of transcription factors in osteogenesis. Int. J. Mol. Sci. 22, 5445. https://doi.org/10.3390/ijms22115445 (2021).
Hwang, S. Y. et al. Deletion of Orai1 alters expression of multiple genes during osteoclast and osteoblast maturation. Cell Calcium. 52, 488–500. https://doi.org/10.1016/j.ceca.2012.10.001 (2012).
Gavin, R. L., Koo, C. Z. & Tomlinson, M. G. Tspan18 is a novel regulator of thrombo-inflammation. Med. Microbiol. Immunol. 209, 553–564. https://doi.org/10.1007/s00430-020-00678-y (2020).
Li, G. X. et al. Tetraspanin18 regulates angiogenesis through VEGFR2 and Notch pathways. Biol. Open. https://doi.org/10.1242/bio.050096 (2021).
Zhang, D. et al. Neuregulin-3 (NRG3): A novel neural tissue-enriched protein that binds and activates ErbB4. PNAS 94, 9562–9567. https://doi.org/10.1073/pnas.94.18.9562 (1997).
Fisher, M. C., Clinton, G. M., Maihle, N. J. & Dealy, C. N. Requirement for ErbB2/ErbB signaling in developing cartilage and bone. Dev. Growth Differ. 49, 503–513. https://doi.org/10.1111/j.1440-169X.2007.00941.x (2007).
Nawachi, K. et al. Tyrosine kinase-type receptor ErbB4 in chondrocytes: interaction with connective tissue growth factor and distribution in cartilage. FEBS Lett. 528, 109–113. https://doi.org/10.1016/S0014-5793(02)03263-5 (2002).
Meldolesi, J. Alternative splicing by NOVA factors: From gene expression to cell physiology and pathology. Int. J. Mol. Sci. 21, 3941. https://doi.org/10.3390/ijms21113941 (2020).
Pizard, A. et al. Connexin 40, a target of transcription factor Tbx5, patterns wrist, digits, and sternum. Mol. Cell. Biol. 25, 5073–5083. https://doi.org/10.1128/MCB.25.12.5073-5083.2005 (2005).
Burton, D. J., Ludden, P. A., Stobart, R. H. & Alexander, B. M. 50 years of the Wyoming ram test: How sheep have changed. JAS 93, 1327. https://doi.org/10.2527/jas.2014-8150 (2015).
Becker, G. M., Woods, J. L., Schauer, C. S., Stewart, W. C. & Murdoch, B. M. Genetic association of wool quality characteristics in United States Rambouillet sheep. Front. Genet. 13, 1081175. https://doi.org/10.3389/fgene.2022.1081175 (2023).
Sambrook, J., Fritsch, E. F., and Maniatis, T. Molecular Cloning: A Laboratory Manual. 2nd Edn. (Cold Spring Harbor Laboratory Press, 1989).
Purcell, S. et al. Plink: A tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81, 559–575. https://doi.org/10.1086/519795 (2007).
Pedersen, T. patchwork: The composer of plots. R package version 1.1.1 Available at: https://CRAN.R-project.org/package=patchwork (2020).
Kassambara, A. Ggpubr: 'ggplot2' based publication ready plots. R package version 0.4.0 Available at: https://CRAN.R-project.org/package=ggpubr (2020).
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. www.R-project.org/. (2021).
Zhou, W. et al. Efficiently controlling for case-control imbalance and sample relatedness in large-scale genetic association studies. Nat. Genet. 50, 1335–1341. https://doi.org/10.1038/s41588-018-0184-y (2018).
Jiang, L., Zheng, Z., Fang, H. & Yang, J. A generalized linear mixed model association tool for biobank-scale data. Nat. Genet. 53, 1616–1621. https://doi.org/10.1038/s41588-021-00954-4 (2021).
McHugh, M. L. The chi-square test of independence. Biochem. Med. 23, 143–149. https://doi.org/10.11613/bm.2013.018 (2013).
Jiang, L. et al. A resource-efficient tool for mixed model association analysis of large-scale data. Nat. Genet. 51, 1749–1755. https://doi.org/10.1038/s41588-019-0530-8 (2019).
Auguie, B. gridExtra: Miscellaneous Functions for "Grid" Graphics. R package version 2.0.0. http://CRAN.R-project.org/package=gridExtra (2015).
Yin, L. CMplot: Circle manhattan plot. R package version 4.1.0 Available at: https://CRAN.R-project.org/package=CMplot. (2022).
Che, R. et al. An adaptive permutation approach for genome-wide association study: Evaluation and recommendations for use. BioData Mining. 7, 9. https://doi.org/10.1186/1756-0381-7-9 (2014).
Cunningham, F. et al. Ensembl 2022. Nucl. Acids Res. 50, D988–D995. https://doi.org/10.1093/nar/gkab1049 (2022).
Davenport, K. M. et al. An improved ovine reference genome assembly to facilitate in-depth functional annotation of the sheep genome. GigaScience. 11, giab096. https://doi.org/10.1093/gigascience/giab096 (2022).
NCBI Resource Coordinators. Database resources of the national center for biotechnology information. Nucl. Acids Res. 44, D7–D19. https://doi.org/10.1093/nar/gkv1290 (2016).
Steinhaus, R., Robinson, P. N. & Seelow, D. FABIAN-variant: Predicting the effects of DNA variants on transcription factor binding. Nucl. Acids Res. 50, W322–W329. https://doi.org/10.1093/nar/gkac393 (2022).
Khan, A. et al. Jaspar 2018: Update of the open-access database of transcription factor binding profiles and its web framework. Nucl. Acids Res. 46, D260–D266. https://doi.org/10.1093/nar/gkx1126 (2018).
Szklarczyk, D. et al. STRING v10: Protein-protein interaction networks, integrated over the tree of life. Nucl. Acids Res. 43, D447–D452. https://doi.org/10.1093/nar/gku1003 (2015).
Acknowledgements
The authors would like to thank the sheep producers who contributed to the success of this study and acknowledge AgResearch and the Animal Genomics team for access to the AgResearch Sheep Genomics 60K SNP chip. This research was funded by the Idaho Global Entrepreneurial Mission and USDA-NIFA-IDA1566 and Hatch-Multistate “Increased Efficiency of Sheep Production” project accession no. 1025808 from the USDA National Institute of Food and Agriculture.
Author information
Authors and Affiliations
Contributions
C.S., W.S. and B.M. conceived the study, W.S. and C.S. and University staff oversaw animal care, handling and data collection, G.B., K.S., J.W., S.K., C.S., B.W., W.S. and B.M. conducted research work and contributed to manuscript writing and editing. All authors read and agree to the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Becker, G.M., Shira, K.A., Woods, J.L. et al. Angular limb deformity associated with TSPAN18, NRG3 and NOVA2 in Rambouillet rams. Sci Rep 13, 16059 (2023). https://doi.org/10.1038/s41598-023-43320-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-023-43320-6
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
