ALPPS surgery induces macroscopic division of labor with metabolic lobe support to accelerate liver regeneration - experimental studies

Abstract

Background: Hepatic surgery rests on the unique, however limited regenerative capacity of the liver. Metabolic duties may be a key factor regulating proliferative abilities of hepatocytes. During regeneration, division of labor leads to spatial separation of proliferation from metabolic tasks, suggesting these opposing activities compete for space. Here, we exploited two-stage-hepatectomy mouse models to explore whether metabolic needs constrain the regenerative capacity of the liver.

Materials and methods: Mice were subjected to sham or associating liver partition and portal vein ligation for staged hepatectomy (ALPPS) surgery (a two-stage-hepatectomy renowned for accelerated regeneration), which leaves one fast-regenerating lobe (FLR) plus portally ligated lobes (LLs) that do not grow but have an intact arteriocentral flow. FLR and LLs were analyzed by omics approaches. Functional surgery was applied to create ALPPS variants with differing metabolic capacity.

Results: The FLR and the adjacent LL displayed a similar metabolite profile which however completely diverged during the major FLR growth phase. Combined transcriptomics-metabolomics and histology assigned proliferative activities explicitly to the FLR, while LLs were enriched with metabolic tasks, establishing macroscopic division of labor. In ALPPS variants differing in ligated volume, FLR growth was increased or reduced with gain or loss of metabolic capacity, respectively, revealing control of regeneration through metabolic duties. Notably, FLRs of slow-growing variants had upregulated metabolic activities, reflecting plastic adaptation to the increased metabolic pressure coming with little ligated volume. Transcriptomics disclosed macroscopic division of labor also in human ALPPS regeneration.

Conclusion: LLs act as auxiliary livers after ALPPS, enabling the FLR to focus on growth. Our findings demonstrate the functional requirement for division of labor during regeneration. This transient division roots on plastic behavior of the different lobes during ALPPS regeneration and reveals how metabolic needs define the liver's regenerative capacity.

Keywords: cellular plasticity; functional surgery; liver regeneration; metabolic capacity; posthepatectomy liver failure; two-stage-hepatectomy.

Figure 1.

Metabolite changes and energy metabolism in the growing and the adjacent ligated lobe after ALPPS. (A) Similarity assessment of sham, RML, and FLR metabolite profiles postsurgery. (B) Tissue metabolite counts in sham liver, the RML, and the FLR at various times postsurgery. (C) Plasma membrane fluidity changes over time. Fluidity [F] was calculated from levels of phosphatidylcholine [PC], phosphatidylethanolamine [PE], and sphingomyelin [SM] as derived from lipidomics using the following formula: F = PC/(PE + SM). (D) Hepatic acylcarnitine, triglyceride, and (E) ketone levels over time. Ketones may replace glucose as a peripheral energy provider during periods of postsurgical hypoglycemia^[2]. (F) Hepatic cardiolipin (mitochondrial marker) content. (G) ATP5A1 (mitochondrial marker) immunohistochemistry. n = 5–6/group except for RML/FLR metabolites at 8 hours (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. In multiple comparisons, * relates to sham vsRML, # to sham vsFLR, and $ to RML vsFLR. FLR, future liver remnant; RML, right median lobe.

Figure 2.

Relative enrichment of metabolic activities within the RML. (A) Relative upregulation of pathways in the RML and FLR as deduced from combined metabolomics-transcriptomics data. The pathway impact score combines statistical significance with topological weight to provide an objective estimate of pathway importance. (B) Glycogen staining and albumin/HNF1A immunohistochemistry in the RML and FLR. (C) ABCB11/CYP7A1 immunohistochemistry and taurine counts (metabolomics peak area) in the RML and FLR. (D) UGT1A1 immunochemistry and Cyp3a11/13 expression after ALPPS. (E) CPS1 and OTC immunochemistry. (F) HNF4A immunochemistry. The right squares show magnification (2×) of the left squares. RML–FLR comparisons show lobe pairs from the same liver. Pairs from two mice are shown for (E) and (F). n = 5–6/group for transcriptomics, n = 5–6/group for metabolomics (ex. RML/FLR metabolites at 8 hours with n = 3), n = 4–6/group for histology. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. In multiple comparisons, * relates to sham vsRML, # to sham vs FLR, and $ to RML vs FLR. FLR, future liver remnant; RML, right median lobe.

Figure 3.

Functional gain/loss ALPPS variants and their regenerative capacity. (A) Schematic presentation of ALPPS variants with a varying proportion of ligated volume. (B) FLR weight gain in the four ALPPS variants at 2, 4, 8, and 24 hours postsurgery. Associated proliferative counts (pHH3) are shown for 24 hours. (C) Lipid content in ALPPS variants 24 hours postsurgery as assessed by PLIN2 (lipid droplet marker) intensity. All ALPPS variants had PLIN2 intensities significantly elevated over sham. n = 6/group. *P < 0.05, **P < 0.01, and ***P < 0.001. In multiple comparisons, * relates to sham vs RML, # to sham vs FLR, and $ to RML vs FLR. FLR, future liver remnant; RML, right median lobe; RLL right lateral lobe; CL, caudate lobe; LLL, left lateral lobe.

Figure 4.

Comparison of metabolic activities at 8 hours postsurgery in FLRs from ALPPS variants that differ in ligated volume. (A) ATP5A immunochemistry. (B) Glycogen staining and albumin immunochemistry. (C) ABCB11/CYPA1 immunochemistry. (D) UGT1A1 immunochemistry. (E) CPS1 and OTC immunochemistry. (F) HNF4A immunochemistry. Right squares show magnification (2×) of left squares. For all, fast-growing (87L) and slow-growing (26L) FLRs from two different animals are shown. n = 4–6/group for histological assessment (G) Serum metabolite counts. Total metabolite counts in sera from mice subjected to three different ALPPS variant surgeries (87L, standard ALPPS 55L, 26L) are shown. Note the enrichment of metabolites in the slowly regenerating 26L variant. n = 4–6/group. *P <0.05, and **P <0.01. FLR, future liver remnant.

Figure 5.

Immunofluorescence for the metabolic (HNF4A) and proliferative (CCND1) state of the FLRs from ALPPS variants with different ligated volume. HNF4A (blue) and CCND1 (yellow) immunofluorescence is shown for fast-growing (87L) and slow-growing (26L) FLRs (each from two different animals). Nuclear counterstain (DAPI) was not included into merged images for better visibility. Note the inverse expression of HNF4A and CCND1 expression in the 87L and 26L FLRs, respectively. n = 4–5/group. FLR, future liver remnant.

Figure 6.

Transcriptomics analysis of ligated lobes and future liver remnants from patients undergoing ALPPS surgery. (A) Similarity (PCA clustering) analysis of differential genes for prestage 1, poststage 1, and prestage 2. Note the dissociation of the LLs from the FLRs after stage 1 transection (n = 6/group). (B) Pathway enrichment analysis of differentially expressed genes at poststage 1 and prestage 2 (n = 6/group). Hardly any differences were observed for prestage 1 (not shown). FLR, future liver remnant; LL, ligated lobe; PCA, principal component analysis.