Functional and gene network analyses of transcriptional signatures characterizing pre-weaned bovine mammary parenchyma or fat pad uncovered novel inter-tissue signaling networks during development

Abstract

Abstract Background The neonatal bovine mammary fat pad (MFP) surrounding the mammary parenchyma (PAR) is thought to exert proliferative effects on the PAR through secretion of local modulators of growth induced by systemic hormones. We used bioinformatics to characterize transcriptomics differences between PAR and MFP from ~65 d old Holstein heifers. Data were mined to uncover potential crosstalk through the analyses of signaling molecules preferentially expressed in one tissue relative to the other. Results Over 9,000 differentially expressed genes (DEG; False discovery rate ≤ 0.05) were found of which 1,478 had a ≥1.5-fold difference between PAR and MFP. Within the DEG highly-expressed in PAR vs. MFP (n = 736) we noted significant enrichment of functions related to cell cycle, structural organization, signaling, and DNA/RNA metabolism. Only actin cytoskeletal signaling was significant among canonical pathways. DEG more highly-expressed in MFP vs. PAR (n = 742) belong to lipid metabolism, signaling, cell movement, and immune-related functions. Canonical pathways associated with metabolism and signaling, particularly immune- and metabolism-related were significantly-enriched. Network analysis uncovered a central role of MYC, TP53, and CTNNB1 in controlling expression of DEG highly-expressed in PAR vs. MFP. Similar analysis suggested a central role for PPARG, KLF2, EGR2, and EPAS1 in regulating expression of more highly-expressed DEG in MFP vs. PAR. Gene network analyses revealed putative inter-tissue crosstalk between cytokines and growth factors preferentially expressed in one tissue (e.g., ANGPTL1, SPP1, IL1B in PAR vs. MFP; ADIPOQ, IL13, FGF2, LEP in MFP vs. PAR) with DEG preferentially expressed in the other tissue, particularly transcription factors or pathways (e.g., MYC, TP53, and actin cytoskeletal signaling in PAR vs. MFP; PPARG and LXR/RXR Signaling in MFP vs. PAR). Conclusions Functional analyses underscored a reciprocal influence in determining the biological features of MFP and PAR during neonatal development. This was exemplified by the potential effect that the signaling molecules (cytokines, growth factors) released preferentially (i.e., more highly-expressed) by PAR or MFP could have on molecular functions or signaling pathways enriched in the MFP or PAR. These bidirectional interactions might be required to coordinate mammary tissue development under normal circumstances or in response to nutrition.