RNA-Seq transcriptome profiling in three liver regeneration models in rats: comparative analysis of partial hepatectomy, ALLPS, and PVL

Abstract

The liver is a unique organ that has a phenomenal capacity to regenerate after injury. Different surgical procedures, including partial hepatectomy (PH), intraoperative portal vein ligation (PVL), and associated liver partition and portal vein ligation for staged hepatectomy (ALPPS) show clinically distinct recovery patterns and regeneration. The observable clinical differences likely mirror some underlying variations in the patterns of gene activation and regeneration pathways. In this study, we provided a comprehensive comparative transcriptomic analysis of gene regulation in regenerating rat livers temporally spaced at 24 h and 96 h after PH, PVL, and ALPPS. The time-dependent factors appear to be the most important determinant of post-injury alterations of gene expression in liver regeneration. Gene expression profile after ALPPS showed more similar expression pattern to the PH than the PVL at the early phase of the regeneration. Early transcriptomic changes and predicted upstream regulators that were found in all three procedures included cell cycle associated genes (E2F1, CCND1, FOXM1, TP53, and RB1), transcription factors (Myc, E2F1, TBX2, FOXM1), DNA replication regulators (CDKN1A, EZH2, RRM2), G1/S-transition regulators (CCNB1, CCND1, RABL6), cytokines and growth factors (CSF2, IL-6, TNF, HGF, VEGF, and EGF), ATM and p53 signaling pathways. The functional pathway, upstream, and network analyses revealed both unique and overlapping molecular mechanisms and pathways for each surgical procedure. Identification of molecular signatures and regenerative signaling pathways for each surgical procedure further our understanding of key regulators of liver regeneration as well as patient populations that are likely to benefit from each procedure.

Figure 1

(a) Venn diagrams representing the differentially expressed genes specific or common among ALPPS, PH and PVL at 24 h and 96 h, respectively. (b) Sources of variation in the data matrix. The x-axis shows the components of the 2-way ANOVA model and the y-axis shows the mean signal to noise ratio. (c) Unsupervised PCA analysis. Different colors indicate different surgical types and shapes indicate different time points. (d) Two-dimensional hierarchical clustering of genes, that are significant in at least one surgical procedure (with respect to sham group) at any of the two time points, and samples. The figure shows the most associated GO biological processes for each cluster of genes. Red and green denote highly and weakly expressed genes, respectively.

Figure 2

Bar chart of the most significant GO biological processes that are associated with up- and down-regulated genes for each surgical procedure at 24h (a) and 96h (b). X-axis represents the statistical significance of the enrichment (–log10(p-value)). Color-coding represents different surgical types.

Figure 3

Gene interaction network, functional, and canonical pathway analyses (a–c, respectively) at 24h and (d–f, respectively) at 96h of common DEGs among ALPPS, PH and PVL. Red and green denote up- and down-regulated genes, respectively, and grey indicates direct interactors of the DEGs. The sizes of nodes (in a) are proportional to their betweenness centrality values. Straight and dashed lines (in d) represent direct or indirect gene to gene interactions, respectively. The color intensity is correlated with fold change. X-axis (in bar graphs) indicates the significance (–log10(p-value)) of the functional/pathway association. The threshold line represents a P value of 0.05.The network (in d) and functional/pathway analyses were generated using IPA (QIAGEN Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis).

Figure 4

The upstream regulator analyses. (a) Venn diagrams representing the overlap of predicted upstream regulators that are unique or common among ALPPS, PH and PVL at 24 h (top) and 96 h (bottom). (b,c) Inter-procedural variations at 24 h and 96 h, respectively. (d–f) Temporal variations after ALPPS, PH and PVL, respectively. Heatmaps for union of top 15 activated upstream regulators in ALPPS, PVL and PH at 24 h and 96 h are used. Color scale shows significance by –log10(p-value). (g) Heatmap that displays surgical procedure as well as temporal changes for the top 50 activated/inhibited upstream regulators. The upstream regulators are ordered according to significance in ALPPS procedure. The color-key is for the z-score. Data were analyzed through the use of IPA (QIAGEN Inc., https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis).

Figure 5

Mechanistic networks of upstream regulators and their predicted relationship by IPA®. The networks for top upstream regulators: E2F1, HGF and TP53 at 24 h and TGFB1, Vegf, and TP53 at 96 h are shown.