Upstream Regulator Analysis of Wooden Breast Myopathy Proteomics in Commercial Broilers and Comparison to Feed Efficiency Proteomics in Pedigree Male Broilers

Abstract

In an effort to understand the apparent trade-off between the continual push for growth performance and the recent emergence of muscle pathologies, shotgun proteomics was conducted on breast muscle obtained at ~8 weeks from commercial broilers with wooden breast (WB) myopathy and compared with that in pedigree male (PedM) broilers exhibiting high feed efficiency (FE). Comparison of the two proteomic datasets was facilitated using the overlay function of Ingenuity Pathway Analysis (IPA) (Qiagen, CA, USA). We focused on upstream regulator analysis and disease-function analysis that provides predictions of activation or inhibition of molecules based on (a) expression of downstream target molecules, (b) the IPA scientific citation database. Angiopoeitin 2 (ANGPT2) exhibited the highest predicted activation Z-score of all molecules in the WB dataset, suggesting that the proteomic landscape of WB myopathy would promote vascularization. Overlaying the FE proteomics data on the WB ANGPT2 upstream regulator network presented no commonality of protein expression and no prediction of ANGPT2 activation. Peroxisome proliferator coactivator 1 alpha (PGC1α) was predicted to be inhibited, suggesting that mitochondrial biogenesis was suppressed in WB. PGC1α was predicted to be activated in high FE pedigree male broilers. Whereas RICTOR (rapamycin independent companion of mammalian target of rapamycin) was predicted to be inhibited in both WB and FE datasets, the predictions were based on different downstream molecules. Other transcription factors predicted to be activated in WB muscle included epidermal growth factor (EGFR), X box binding protein (XBP1), transforming growth factor beta 1 (TGFB1) and nuclear factor (erythroid-derived 2)-like 2 (NFE2L2). Inhibitions of aryl hydrocarbon receptor (AHR), AHR nuclear translocator (ARNT) and estrogen related receptor gamma (ESRRG) were also predicted in the WB muscle. These findings indicate that there are considerable differences in upstream regulators based on downstream protein expression observed in WB myopathy and in high FE PedM broilers that may provide additional insight into the etiology of WB myopathy.

Keywords: feed efficiency; myopathy; proteomics; upstream regulator analysis; wooden breast.

Figure 1

A prediction legend for the interpretation of protein expression and upstream regulator and function analysis in tables and figures.

Figure 2

Network of differentially expressed proteins in the wooden breast (WB) myopathy dataset used in calculating the activation Z score for vasculogenesis (A). Overlay of the feed efficiency (FE) proteomic dataset from Kong et al. [11]; (B) indicates no commonality of differentially expressed proteins in WB vasculogenesis network Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 3

The network of differentially expressed proteins that was used to predict inhibition of necrosis in wooden breast (WB) myopathy. Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 4

The upstream regulatory network of proteins displayed in hierarchical format for Angiopoeitin 2 (ANGPT2) used in the prediction of activation of this molecule (activation Z score and p value of overlap as shown in (A)). Overlay of the feed efficiency (FE) proteomics dataset from Kong et al. [11] reveals no commonality of differentially expressed proteins between the wooden breast (WB) and FE data (B). A regulatory network for ANGPT2 indicates target proteins in the dataset that contributed to the predictions of active vasculogenesis and angiogenesis, and inhibited necrosis and apoptosis is provided in (C). Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 4

The upstream regulatory network of proteins displayed in hierarchical format for Angiopoeitin 2 (ANGPT2) used in the prediction of activation of this molecule (activation Z score and p value of overlap as shown in (A)). Overlay of the feed efficiency (FE) proteomics dataset from Kong et al. [11] reveals no commonality of differentially expressed proteins between the wooden breast (WB) and FE data (B). A regulatory network for ANGPT2 indicates target proteins in the dataset that contributed to the predictions of active vasculogenesis and angiogenesis, and inhibited necrosis and apoptosis is provided in (C). Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 5

The regulatory network of epithelial growth factor receptor (EGFR) showing downstream target proteins contributing to the prediction of activation of vasculogenesis and inhibition of necrosis and apoptosis in wooden breast (WB) myopathy. Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 6

The upstream regulatory network of transforming growth factor beta 1 (TGFB1) for wooden breast (WB) (A) and feed efficiency (FE) (B) proteomic datasets. Downstream target proteins that were differentially expressed are shown as up-regulated in pink or red or down-regulated in WB myopathy and high FE PedM broilers. Overlay of the FE proteomic dataset on the WB myopathy network (C) and of the FE data on the WB network (D) resulted in predicted activations of TGFB1 as well. Therefore, despite dissimilar proteomic landscapes of the two datasets, TGFB1 is predicted to be active in both WB myopathy and high FE.

Figure 7

The regulatory network of TGFB1 in wooden breast (WB) myopathy leading to predicted activations of angiogenesis and vasculogenesis and inhibition of apoptosis and necrosis through downstream target proteins in WB myopathy. Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 8

The upstream regulator network for NFE2L2 (nuclear factor (erythroid-derived 2)-like 2) for wooden breast (WB) myopathy (activation Z score = 2.18, 4.3 × 10⁻⁹) compared to normal muscle (A) and for the high vs. low PedM broiler feed efficiency (FE) proteomic data (activation Z score = 2.38, p = 1.54 × 10⁻⁶) (B). Overlay of the FE data on the WB myopathy data (C) and vice versa (D) both resulted in predicted activations of NFE2L2. Protein abbreviations in B and D in the FE data set are as follows: ARF1 (ADP-ribosylation factor 1), COX4I1 (cytochrome oxidase subunit IV isoform 1), EIF4G2 (eukaryotic translation initiation factor 4 gamma 2), GPX1 (glutathione peroxidase 1), PPIB (peptidylprolyl isomerase B (cyclophilin B), PSMB1 (proteosome 26S subunit beta type, 1) PSMD1 (proteosome 26S subunit non-ATPase, 1), RYR3 (ryanodine receptor 3), SOD1 (superoxide dismutase 1), UGDH (UDP-glucose 6-dehydrogenase).

Figure 9

The upstream regulatory network for RICTOR in wooden breast (WB) myopathy (A) and for feed efficiency (FE) proteomics (B). All downstream proteins were up-regulated in both datasets, but there were no common differentially expressed proteins. The regulatory network for RICTOR linking downstream molecules to the predicted inhibition of necrosis and slight activation of apoptosis in WB myopathy (C). For (A,C), protein abbreviations are defined and differential expression of proteins in WB myopathy are provided in Supplementary Table S1. Proteins in (B) are as follows: COX4I1 (cytochrome oxidase subunit IV isoform 1), NDUFB5, B8, S7 (NADH dehydrogenase (ubiquinone) Complex I assembly factor 5, factor 8, subunit 7), PPA2 (pyrophosphate (inorganic) 2, PSMB1, D1, D2 (proteosome 26S subunit, non-ATPase beta type-1) PSMD1, PSMD2 (proteosome 26S subunit, non-ATPase, 1 & non-ATPase 2), RPL12, RPL30 (ribosomal protein L12 & L30), UQCRC2 (ubiquinol-cytochrome c-reductase core protein II).

Figure 10

The upstream regulator network for PPARGC1α for WB myopathy vs. control breast muscle in commercial broilers (A) and high vs. low FE PedM (B) proteomic datasets [11]. The predicted inhibition of PGC1α in WB myopathy was associated with the predicted inhibition of Complex I of the mitochondrial electron transport chain in WB myopathy (C), whereas the predicted activation of PGC1α in high FE was associated with predicted activations of Complex I, III and IV of the mitochondrial electron transport chain (D). The regulatory network for PPARGC1A showing the downstream target molecules involved in the predictions of inhibition of vasculogenesis and oxidation of fatty acid and activation of reactive oxygen species (ROS) production and quantity of adipose tissue (E). Protein abbreviations in (A,E) are defined and differential expression of proteins are provided in Supplementary Table S1. Protein abbreviations for (D) are as follows: COX4I1 (cytochrome oxidase subunit IV isoform 1), CTSC (cathepsin C), GPX1 (glutathione peroxidase 1), NDUFB5 (NADH dehydrogenase (ubiquinone) Complex I assembly factor 5), NDUFB8 (NADH dehydrogenase (ubiquinone) beta subcomplex, 8), OMM (Outer Mitochondrial Membrane), IMM (Inner Mitochondrial Membrane).

Figure 11

The upstream regulatory networks for ARNT (A) and AHR (B). The predicted inhibition of ARNT was based on the downstream expression of proteins that are shown and was associated with predicted inhibition of apoptosis, vasculogenesis and glycolysis and prediction of activation of fibrosis. Conversely, the expression of downstream target molecules of involved in predicting inhibition of AHR were associated with predictions of enhanced angiogenesis and necrosis. The discrepancy in the prediction of active necrosis in the AHR and prediction of inhibition of necrosis in the entire dataset shown in Table 2 is based on the smaller subset of expression data for AHR compared to the overall dataset. Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 12

ESRRGamma (ESRRG or ESSRγ) was predicted to be inhibited in WB myopathy based on the expression of downstream molecules shown in this regulatory network. The regulatory network resulted in the prediction of the inhibition of vasculogenesis and activation of necroptosis. As in Figure 9, discrepancies in the predictions of these functions and those in Table 2 are due to a smaller subset of proteins in the ERRG upstream regulatory network, compared to the larger set of proteins listed for functions in Table 2. Protein abbreviations are defined and differential expression of proteins are provided in Supplementary Table S1.

Figure 13

(A) Upstream regulator interactions contributing to the overall prediction of increased vasculogenesis and decreased cell death in wooden breast muscle myopathy in commercial broilers. The combination of the predicted activities of these upstream regulators would support the prediction of increased angiogenesis and inhibition of cell death in wooden breast myopathy. (B) Arrangement of similar upstream regulators in the high vs. low feed efficiency PedM phenotype dataset revealed a scenario in which angiogenesis was not predicted to be active or inhibited while predictions were made of inhibition cell death, which is similar to that for WB myopathy.