EREG-secreting THBS1⁺ tissue monocytes are recruited by C5a to promote rapid liver regeneration in patients and mice during the ALPPS procedure

Introduction

Liver cancer is the second leading cause of cancer-related mortality worldwide, and hepatocellular carcinoma (HCC) accounts for 90% of cases (1,2). Surgical resection remains the mainstay of treatment for patients with HCC (3). Unfortunately, only 20% to 30% of patients with HCC can be treated by liver resection at diagnosis (4,5). One important limiting factor in the determination of tumor resectability is the volume of the future liver remnant (FLR). An inadequate FLR would result in post-hepatectomy liver failure (PHLF) and mortality for HCC patients after liver resection, especially for those with liver cirrhosis or chronic hepatitis (6). In the past, liver transplantation was the only therapy for these patients (7). However, its implementation is limited by the availability of donor liver. Associating liver partition and portal vein ligation for staged hepatectomy (ALPPS) is a recent advance in hepatobiliary surgery that gives hope of a potential cure to patients with liver tumors that have previously been considered unresectable (8). There also have been several hypertrophy-inducing interventions including portal vein embolization and liver venous deprivation (9,10). Recently, our group reported that ALPPS is a viable treatment option for selected patients with ‘unresectable’ HCC (11). The main advantage of the ALPPS strategy is that it can induce rapid liver regeneration, with the liver reaching a body-sustaining size within 7–10 days, enabling the prompt elimination of the major tumor load at the lowest risk of liver failure compared to the traditional one-stage liver resection (11,12). Although many clinical studies on ALPPS have been reported, there have not been many basic research studies published, and the mechanism(s) underlying the rapid liver hypertrophy remains unclear. Thus, it is critical to explore the relevant mechanism(s) of ALPPS-induced accelerated liver regeneration to identify a new strategy to induce the rapid growth of the remnant liver for patients with postoperative hepatic insufficiency and to further reduce the risk of PHLF after ALPPS.

The biological processes underlying liver regeneration are complex. It has been reported that the initiation of liver regeneration is driven by the innate immune system and cytokine release, and numerous studies have provided evidence that changes in the local immune microenvironment may play an important role in the process of liver regeneration (13), including neutrophils (14,15), mononuclear phagocytes (16-18), natural killer T (NKT) cells (19), natural killer (NK) cells (20,21), dendrite cells (DCs) (22), γδT cells (23), T cells (24) and B cells (25). However, limited research has focused on the mechanism underlying the ALPPS-induced rapid liver regeneration, and only a few reports using animal models have implied that the rapid liver hypertrophy in ALPPS is attributable to an increase in inflammatory cytokines—such as interleukin (IL)-6 and tumor necrosis factor (TNF)-α—resulting from liver partition (26).

Recent advances in single-cell technologies, such as single-cell RNA sequencing (scRNA-seq) and cytometry by time-of-flight (CyTOF), offer an excellent opportunity to obtain an in-depth understanding of the different cellular components and their interactions in the tissue microenvironment (27,28). Such technologies have been applied to study the biology of the healthy and diseased liver at unprecedented resolution (29). To date, there have been no reports characterizing the immune cell compositions and dynamic changes in the local microenvironment during ALPPS at the single-cell resolution level, which might help elucidate relevant mechanisms underlying the ALPPS-induced rapid liver regeneration.

Here, we used two high-dimensional single-cell technologies, CyTOF and scRNA-seq, and combined these with bulk transcriptome RNA sequencing in matched pairs of ALPPS stage I and II liver tissues from five HCC patients, to comprehensively dissect the systemic and local immune microenvironments, as well as their dynamic changes, during the ALPPS procedure. Our aim was to uncover the key inflammatory cell clusters that induce the rapid liver regeneration during the ALPPS procedure and to unearth the relevant mechanism(s). We present this article in accordance with the ARRIVE reporting checklist (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-24-391/rc).

Methods

Patients and sample collection

The study was conducted following the Declaration of Helsinki (and its subsequent amendments). The Ethics Committee of Zhongshan Hospital, Fudan University approved the protocol (No. 2024-120). Written informed consent was obtained from each participant. Human liver tissues were collected at ALPPS stage I and II, and peripheral blood samples were collected during subsequent routine clinical examinations. Five HCC patients who underwent classic ALPPS at Zhongshan Hospital of Fudan University were enrolled for CyTOF analysis, and 3 of these 5 patients with enough tissue samples were further assigned to scRNA-seq analysis (discovery cohort 1, Tables 1,2). The changes in transcription profile in liver tissues between ALPPS stage I and II (discovery cohort 2, n=12) were explored by bulk RNA-seq analysis. Tissue microarrays containing paired liver tissues from another 22 HCC patients who underwent ALPPS were constructed for further validation (validation cohort 1). Blood was also collected from these 22 patients to explore the dynamic changes in immune cells during ALPPS procedure. The dynamic changes of the key cytokines in plasma were further confirmed by enzyme-linked immunosorbent assay (ELISA) in validation cohort 2 (n=8).

Informed consent was obtained from all patients for their data to be used for research purposes. The criteria for patient selection and surgical decisions, as well as perioperative management, have been reported previously (11).

Single-cell dissociation

The liver tissues were surgically removed and washed twice before being kept in MACS Tissue Storage Solution (Miltenyi Biotec, Gladbach, Germany) until processing. For processing, samples were first washed with phosphate-buffered saline (PBS), minced into small pieces (approximately 1 mm³) on ice, and enzymatically digested with 0.4% collagenase IV (Diamond, Shanghai, China) with RPMI-1640 medium and 10% fetal bovine serum (FBS) Origin South America Gold (YEASEN, Shanghai, China) for 45 min at 37 ℃, with agitation. After digestion, samples were sieved through a 70 µm cell strainer and centrifuged at 300 g for 5 min. After the supernatant was removed, the pelleted cells were suspended in red blood cell lysis buffer and incubated for 2 min to lyse red blood cells. After they were washed with RPMI-1640 medium, the cell pellets were re-suspended in RPMI-1640 medium and re-filtered through a 35 µm cell strainer. Dissociated single cells were then stained for viability assessment using Calcein-AM (Thermo Fisher Scientific, Massachusetts, USA) and Draq7 (BD Biosciences, New Jersey, USA). The single-cell suspension was further enriched with a MACS dead cell removal kit (Miltenyi Biotec).

CyTOF analysis

Twenty samples including peripheral blood mononuclear cell (PBMC) and liver tissues from 5 patients were processed for the CyTOF analysis. A customized panel of 42 markers that encompassed a broad range of immune lineages was used (Figure S1A-S1C). The conjugated antibodies were purchased directly from the supplier. Leukocytes were washed and stained for viability with cisplatin and incubated with metal-conjugated surface-membrane antibodies for 30 min at 37 ℃. After that, the cells were fixed with fix perm buffer. Finally, cell intercalation (mixture of fixed perm buffer and iridium) was added for cellular fixation and visualization, and this procedure was allowed to proceed overnight before analysis on a Helios mass cytometer (Fludigm, San Francisco, USA). EQ Four Element Calibration Beads were used to normalize the signal. At least 1×10⁵ cell events were collected for each sample. Files (.fcs) were uploaded into Cytobank, and populations of interest were manually gated. For further analysis, a random sampling of 100,000 cells from each fcs file was performed using the CyTOFkit program on R. Visualizations based on t-distributed stochastic neighbor embedding (t-SNE) and clustering based on the FlowSOM/Renograph algorithms were then performed on these cells. Diffusion map algorithms were used to assess single-cell trajectories between monocytes and macrophages (Mo/Mφ population). Up to 2,000 cells from each subpopulation were randomly selected for analysis.

BD scRNA-seq

scRNA-seq experiment

The BD Rhapsody system (New Jersey, USA) was used to capture the transcriptomic information of single cells. Single-cell capture was achieved by a random distribution of a single-cell suspension across >200,000 microwells through a limited dilution approach. Beads with oligonucleotide barcodes were added to saturation so that a bead was paired with a cell in a microwell. Cell lysis buffer was added so that poly-adenylated RNA molecules hybridized into the beads. Beads were collected into a single tube for reverse transcription. Upon cDNA synthesis, each cDNA molecule was tagged on the 5' ends (that is, the 3' end of an mRNA transcript) with a unique molecular identifier (UMI) and cell label indicating its cell of origin. Whole transcriptome libraries were prepared using the BD Rhapsody single-cell whole-transcriptome amplification workflow. In brief, second strand cDNA was synthesized, followed by ligation of the WTA adaptor for universal amplification. Eighteen cycles of polymerase chain reaction (PCR) were used to amplify the adaptor-ligated cDNA products. Sequencing libraries were prepared using random priming PCR of the whole-transcriptome amplification products to enrich the 3' ends of the transcripts linked with the cell label and UMI. Sequencing libraries were quantified using a High Sensitivity DNA chip (Agilent, California, USA) on a Bioanalyzer 2200 and the Qubit High Sensitivity DNA assay (Thermo Fisher Scientific). The library for each sample was sequenced by HiSeq Xten (Illumina, San Diego, CA, USA) on a 150 bp paired-end run.

Single-cell RNA statistical analysis

We applied fastp with default parameter filtering the adaptor sequence and removed the low-quality reads to achieve clean data (30). Umi-tools were applied for Single Cell Transcriptome Analysis (Oxford, UK) to identify the cell barcode whitelist. The UMI-based clean data was mapped to the mouse genome (Ensemble version 92) utilizing STAR mapping with customized parameters from the UMI-tools standard pipeline to obtain the UMIs counts of each sample (31). To minimize the sample batch, we applied a downsample analysis among samples sequenced according to the mean reads per cell of each sample and finally achieved the cell expression table with a sample barcode. Cells that contained over 200 expressed genes and a mitochondria UMI rate below 20% passed the cell quality filtering. The mitochondria genes were removed from the expression table, but were used for cell expression regression to avoid the effects of the cell status for the clustering analysis and marker analysis of each cluster.

The Seurat package (version: 2.3.4, https://satijalab.org/seurat/) was used for cell normalization and regression based on the expression table according to the UMI counts of each sample and the percent of mitochondria to obtain the scaled data. A principal component analysis (PCA) was conducted based on the scaled data with all high variable genes and the top 8 principals were used for t-SNE construction. Utilizing the graph-based cluster method, we acquired the unsupervised cell cluster results based on the PCA top 8 principal, and we calculated the marker genes by the FindAllMarkers function with a Wilcox rank sum test algorithm under the following criteria: (I) log-foldchange >0.25; (II) P value <0.05; (III) min.pct >0.1. To identify the cell type detailed, cells were subjected to re-t-SNE analysis, graph-based clustering, and marker analysis.

Pseudo-time analysis

We applied the Single-Cell Trajectories analysis utilizing Monocle2 (http://cole-trapnell-lab.github.io/monocle-release) using DDR-Tree and default parameters. Before performing a Monocle analysis, we selected marker genes from the Seurat clustering results and raw expression counts of the cells that passed filtering. Based on the pseudo-time analysis, branch expression analysis modeling (BEAM Analysis) was applied for branch fate determining gene analysis.

Cell communication analysis

To enable a systematic analysis of cell-cell communication molecules, we applied a cell communication analysis based on the CellPhoneDB (32), a public repository of ligands, receptors, and their interactions. The membrane, secreted, and peripheral proteins of the cluster from different time points were annotated. Significance (P value <0.05) was calculated based on the interactions and the normalized cell matrix achieved by Seurat Normalization.

SCENIC analysis

The SCENIC analysis was run as described previously (33) on the cells that passed the filtering, using the 20-thousand motifs database for RcisTarget and GRNboost.

Animals and establishment of the mouse model

Animal experiments were performed under a project license (No. 2019-185) granted by the institutional ethics board of Zhongshan Hospital, Fudan University, in compliance with institutional guidelines for the care and use of animals. To avoid differences due to strain, age, or especially sex [estrogen has potential influence on liver regeneration (34)], only male C57/BL mice (8 weeks old; 20–25 g) were used for this study. The ALPPS surgeries were performed as reported before (35). In brief, the mice had the portal vein to the right lobe (RL), caudate lobe (CL), left lateral lobe (LLL), and left middle lobe (LML) ligated, followed by cholecystectomy and partition at the middle lobe (Figure S2A). The right median lobe (RML) was preserved as the FLR. The FLR weight to body weight ratio was about 0.489%.

Neutralizing antibodies against growth factors were purchased from R&D Systems [epiregulin (EREG): mEpiregulin Aff Pur, AF1068, R&D Systems, Minnesota, USA; amphiregulin (AREG): mAmphiregulin Aff Pu, AF989, R&D Systems]. After ALPPS surgery, the mice (n=18) were immediately injected with EREG/AREG or vehicle control intravenously and then at 6, 24, 48 and 72 h post-operation (2 µg EREG/AREG each time). At 5 days post-operation, the mice were euthanized and the livers were collected. Each group had 6 mice enrolled.

For C5a blockade, PMX-53, a specific C5aR1 antagonist, was used (HY-106178, MedChemExpress, Shanghai, China). A total of 24 mice was enrolled. After ALPPS model establishment, PMX-53 was intraperitoneally injected immediately and then at 6, 24, 48 and 72 h post-operation. The dose used was 3 mg/kg at each time point. The control group was injected with vehicle control. Sample collection was conducted on post-operation day (POD) 1 and POD 5. Each group had 6 animals sacrificed at each time point

A mouse model of PHLF was established. Considering that the development of PHLF after ALPPS is mainly caused by “small for size”, the model was established according to a previously reported protocol for an ALPPS rat model with an insufficient liver remnant (36). In our PHLF mouse model, the RL, CL, LLL, and RML were resected. The LML was the only liver remnant that remaining, and accounted for about 10% of the total liver mass, which is insufficient for mouse survival and causes severe PHLF (Figure S2B). Recombinant EREG (rmEpiregulin/CF, 1068-EP-050/CF, R&D Systems) or vehicle control was injected intravenously before the hepatectomy and then at 6, 12 and 24 h post-operatively. The time of death was recorded for a survival analysis. A total of 18 mice were enrolled (9 per group).

Multiplexed immunofluorescence (IF) staining

Multiplex staining was conducted on an ALPPS tissue microarray containing liver tissue collected at ALPPS stage I and stage II. Multiplex staining of the tissue microarray was conducted using a PANO 7-plex immunohistochemistry (IHC) kit (Panovue, Beijing, China) according to the manufacturer’s instructions. Primary antibodies against CD14 (#ab52625; Abcam, Cambridge, UK), S100A9 (MAC387; Invitrogen, California, USA), and THBS1 (#ab1823; Abcam) were sequentially used, followed by incubation with horseradish peroxidase-conjugated secondary antibody and tyramide signal amplification. The slides were then microwave heat-treated after each trichostatin A (TSA) procedure. Nuclei were stained with 4'-6'-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, Missouri, USA) after all the human antigens had been labeled.

ELISA

Plasma from the 5 patients was collected every other day after ALPPS. The concentrations of EREG and AREG in the ALPPS patient plasma was determined by ELISA. The kits, EREG (SU-B12532, KNDbio, Fujian, China) and AREG (SU-B11232, KNDbio), were used according to the manufacturer’s recommended procedure.

Cell proliferation assay

THLE-2, an immortalized normal human liver epithelial cell line, was cultivated for a cell proliferation assay. The cells were resuspended in minimum essential medium (MEM) containing 10% FBS with recombinant EREG (5 µg/mL; rmEpiregulin/CF, 1068-EP-050/CF, R&D Systems), recombinant AREG (5 µg/mL; rmAmphiregulin/CF, 989-AR-100/CF, R&D Systems), recombinant EREG plus recombinant AREG (5 µg/mL), or vehicle control and inoculated into a 96 well plate with 1,000 cells per well. After 12, 24, 36 or 48 h, the Cell Counting Kit-8 (CCK-8) reagents were added. After an additional 1-hr incubation, the optical density at 450 nm was detected.

Statistical analysis

Statistical analysis was performed using the Student’s t-test. When comparing ALPPS stage I vs. stage II, a paired t-test was applied. For data evaluated over multiple timepoints, each individual test was compared with day 0 using paired t-test. When comparing two independent groups, a two-sample t-test was used. Data were shown as mean ± standard deviation. The survival analysis of PHLF mice was compared by log-rank test. All statistical analyses were performed using SPSS version 24.0 for Mac (IBM Corp., Armonk, USA). A P value of <0.05 was considered statistically significant.

Results

CyTOF revealed a distinct immune cell atlas during ALPPS-induced rapid liver regeneration

To depict the dynamic changes in the peripheral and local liver immune microenvironments during liver regeneration induced by ALPPS, CyTOF was performed using a customized panel of antibodies targeting 42 surface markers. Tissue samples of paired liver tissues (N=10, paired, P1–P5) and PBMC samples (N=10, paired, P1–P5) from five HCC patients in stage I and II of ALPPS were evaluated (Figure 1A and Table 1). Through a bioinformatic analysis of the high-dimensional single-cell proteomics profiles examined by CyTOF, a total of 2,257,053 CD45⁺ cells were retained and subsequently clustered into 45 discrete cell populations with different immune marker expression (see the heatmap in Figure 1B and Figure S1A-S1C). These 45 clusters were classified into B cells, CD4⁺ T cells, CD8⁺ T cells, γδT cells, NKT cells, NK cells, DC, monocytes/macrophages (Mo/Mφ population) and granulocytes (GRAN) (Figure 1C and Figure S1A,S1B).

Figure 1 Single-cell atlas of human ALPPS. (A) The flowchart of the overall study design. scRNA-seq and CyTOF were performed on liver and blood samples. The results of the bioinformatic analysis were validated by evaluating a separate clinical cohort and an animal model. (B) A heatmap showing the mean expression levels of the selected marker genes for all 45 clusters. Clusters were from PBMC and liver and were collected during two ALPPS stages. (C) A t-SNE plot representing all clusters annotated as certain cell types. (D,E) t-SNE plots of the CyTOF data from PBMC obtained from patients during stage I (D), PBMC during stage II (D), FLR during stage I (E), and FLR during stage II (E). (F) Boxplot showing the proportions of immune cell lineages. (G) The column chart shows the proportions of major cell types of PBMC. (H) The column chart shows the proportions of major cell types of FLR. Comparisons were made for each cell type at ALPPS stages I and stage II. *, P<0.05; **, P<0.01; ***, P<0.001 (actual P values were reported in results). ALPPS, associating liver partition and portal vein ligation for staged hepatectomy; CyTOF, cytometry by time-of-flight; DC, dendrite cells; DL, deportalized lobe; ELISA, enzyme-linked immunosorbent assay; FLR, future liver remnant; Gran, granulocytes; Mo, monocytes; Mφ, macrophages; NK, natural killer; NKT, natural killer T; OD, optical density; PBMC, peripheral blood mononuclear cell; POD, post-operation day; scRNA-seq, single-cell RNA sequencing; t-SNE, t-distributed stochastic neighbor embedding.

To explore the role of circulating factors on ALPPS-induced liver regeneration, we first viewed the changes of the immune cells present in the PBMC during ALPPS. At ALPPS stage I, lymphoid cells comprised the majority of the PBMC lineages, as revealed by CyTOF (61.33%, Figure 1D-1G and Figure S3A-S3C). Among all immune cell types, the most dominant cells were T cells, which accounted for 36.38% of the total PBMC. The second most dominant cells were monocytes (30.80%), followed by NK cells (18.69%), B cells (6.26%), γδT cells (2.45%), and DCs (1.62%). Compared with ALPPS stage I, there was a trend toward an increase in monocytes (from 30.80% to 40.12%) and a decrease in NK cells (from 18.69% to 13.51%) and DCs (from 1.62% to 0.94%) in stage II, although these differences failed to reach statistical significance (Figure 1G and Figure S3D). We further evaluated the dynamic changes in circulating monocytes, as well as other cell lineages, by obtaining routine blood data from a larger patient cohort (validation cohort 1, Figure S3E). In agreement with the data from the original cohort, circulating monocytes increased during the regeneration period (Figure S3E) as did neutrophils, implying that the myeloid cells might take part in the process of rapid liver regeneration. Notably, the number of lymphocytes was decreased according to the blood routine examination (Figure S3E).

We further analyzed the local immune cell atlas, as well as the dynamic changes in residual liver tissues, during ALPPS-induced liver regeneration. The Mo/Mφ population in the residual liver can be clearly distinguished as tissue monocytes and macrophages by CyTOF as the different distribution in t-SNE, although it is generally considered to be difficult to classify by other means (Figure 1C,1E). A significant increase in tissue monocytes (P=0.03) was observed in the local immune environment during the ALPPS-induced liver regeneration, and the proportions of both macrophages (P=0.04) and DC (P=0.03) were decreased. In contrast, there were no significant differences in the distribution of CD4⁺ T cells, CD8⁺ T cells, γδT cells, B cells, NK cells, NKT cells, or GRAN (Figure 1E,1H).

Since myeloid and lymphoid cells have distinct origins and functions, we further analyzed these two cell lineages separately. We found that the proportion of tissue monocytes in the total myeloid cells was increased significantly, while that of macrophages was decreased (Figure 2A). The mean proportion of tissue monocytes in the total myeloid cells was 13.36% at ALPPS stage I and significantly increased to 29.78% at ALPPS stage II (P=0.04, Figure 2A). Meanwhile, the proportion of macrophages in the total myeloid cells was decreased in all five patients during ALPPS-induced liver regeneration (mean, 27.94% for ALPPS stage I vs. 8.60% for ALPPS stage II, P=0.03; Figure 2A). We also analyzed the changes in the proportion of lymphoid cells by CyTOF analysis, but no significant changes were observed in the residual liver tissues at ALPPS stage II (Figure S4A-S4C). We also confirmed these finding by IHC using tissue microarrays in validation cohort 1 (Figure S4A).

Figure 2 Monocytes are enriched by the ALPPS-induced liver regeneration. (A) t-SNE plots of clusters annotated as myeloid cells according to CyTOF data. The color represents the expression level of CD14. (B-D) Diffusion map of the Mo/Mφ population. (B) Shows the three-dimensional diffusion map. (C) Shows a two-dimensional diffusion map, color-coded by clusters. (D) Shows a two-dimensional diffusion map of PBMC and FLR at both stage I and stage II. (E) The expression levels of selected marker genes used for clustering and mononuclear-phagocyte-specific markers. (F) Bar chart representing the ratios of monocytes and macrophages at stage I and stage II. (G) Multiplexed immunofluorescence staining showing the distribution of monocytes (CD14⁺S100A9⁺) and macrophages (CD14⁺S100A9⁻). The monocyte-specific plot illustrates the increase in monocytes from ALPPS stage I to stage II. *, P<0.05 (actual P values were reported in results). ALPPS, associating liver partition and portal vein ligation for staged hepatectomy; CyTOF, cytometry by time-of-flight; DC, dendrite cells; FLR, future liver remnant; Mo, monocytes; Mφ, macrophages; PBMC, peripheral blood mononuclear cell; t-SNE, t-distributed stochastic neighbor embedding.

The ALPPS procedure leads to a shift from a macrophage-enriched to monocyte-enriched local liver immune microenvironment

A diffusion map was used to clarify the potential relationship and mutual transformation between monocytes and macrophages during rapid liver regeneration induced by ALPPS (Figure 2B-2D) (37). Three main branches of the Mo/Mφ population were identified by the three-dimensional diffusion scatters plot (Figure 2B). Circulating monocytes were included as a reference. Among them, the first branch expands along diffusion component one to mostly contain circulating monocytes and clusters 38, 42, 43, 44, and 45 in the FLR. These clusters have a similar distribution as the circulating monocytes in diffusion scatter plots (Figure 2C,2D). However, they were characterized by high levels of CD11b and intermediate levels of HLA-DR and CD68 (Figure 2E), and this phenotype further proved that these clusters were derived from circulating monocytes. The second branch along diffusion component two mainly consists of clusters 32, 33, 34, 35 and 37, and the third branch along diffusion component three was almost exclusively formed by cluster 36 (Figure 2B,2C). Both of these were branches found to have high levels of CD163 or CD206 (Figure 2E) and therefore were identified as macrophages, implying that there may be a distinctive polarization route of tissue monocytes to macrophages.

The dynamic changes of monocytes and macrophages in the FLR were further explored. At ALPPS stage I, macrophages constituted the main Mo/Mφ population (accounting for 65.19%), while the main population changed to tissue monocytes (accounting for 77.99%) at ALPPS stage II (Figure 2F). The enrichment of tissue monocytes at ALPPS stage II was confirmed in validation cohort 1 by IF staining (CD14⁺ S100A9⁺ cells, 1.84% in ALPPS stage I vs. 5.34% in ALPPS stage II; P=0.03, Figure 2G).

scRNA-seq identified an increase in THBS1+ tissue monocytes during liver regeneration

To further investigate the role of tissue monocytes in ALPPS-induced liver regeneration, six paired liver tissues from three patients in ALPPS stage I and II (P2, P4 and P5) were collected for scRNA-seq analysis, yielding data from 22,989 cells (stage I, 12,642 cells; stage II, 10,347 cells; Figure S5A-S5E and Figure S6A-S6D). All of the cells were classified into 18 clusters and then assigned to 11 cell types, including T cells (CD3D), B cells (IGHG1), liver sinusoidal endothelial cells (LSECs; VWF), GRAN (S100A8), cells of the Mo/Mφ population (HBB), hepatocytes (ALB), fibroblasts (APOC3), DC (MYL9), cholangiocytes (KRT19), and erythroid cells (FCGR3B).

The Mo/Mφ population in the scRNA-seq data were extracted for further analysis. Referring to their transcriptional profile, these cells were classified into four populations: tissue monocytes, tissue resident macrophages, DC, and proliferating cells (Figure 3A-3C and Figure S7A-S7C). Further, the 11 subtypes were displayed in the t-SNE map, including Kupffer cells (KC, MP0), CD1C⁺ DC (MP1), MYH9⁺ tissue monocytes (MYH9⁺ tissue monocytes, MP2), monocyte-derived macrophages (MoMFs, MP3), THBS1⁺ tissue monocytes (THBS1⁺ tissue monocytes, MP4), ITGAL⁺ tissue monocytes (ITGAL⁺ tissue monocytes, MP5), CLEC9A⁺ DC (MP6), cycling MP (MP7), SAM1 (MP8), CTSK⁺ macrophages (CTSK⁺ MAC, MP9), and SAM2 (MP10) (Figure 3A,3B and Figure S7A-S7C).

Figure 3 Annotation of the Mo/Mφ population from human livers of ALPPS stage I and stage II. (A) t-SNE plot showing the Mo/Mφ population clustered into 11 phenotypes, indicated by the color-coded legend. (B) The expression levels of selected marker genes in each cluster. (C) Pseudotime-ordered analysis of the Mo/Mφ population. The trajectories from tissue monocytes to Kupffer cells or monocyte-derived macrophages are separate. (D) Cluster-specific plots illustrating heterogeneity within different patients, except MP 1, 4, 6. (E) Bar charts showing activated pathways according to the GO and KEGG pathway analyses for THBS1⁺ tissue monocytes. (F) The results of a QuSAGE gene enrichment analysis for THBS1⁺ tissue monocytes. (G) Multiplexed immunofluorescence staining showing the distribution of THBS1⁺ tissue monocytes (CD14⁺ S100A9⁺THBS1⁺), other monocytes (CD14⁺S100A9⁺THBS1⁻) and macrophages (CD14⁺S100A9⁻THBS1⁻). Comparisons were made for THBS1⁺ tissue monocytes at ALPPS stage I and stage II. *, P<0.05 (actual P values were reported in results). ALPPS, associating liver partition and portal vein ligation for staged hepatectomy; DC, dendritic cell; GO, Gene Ontology; IF, immunofluorescence; KC, Kupffer cell; KEGG, Kyoto Encyclopedia of Genes and Genomes; MAC, macrophage; Mo, monocytes; MoMF, monocyte-derived macrophage; MP, Mo/Mφ population; Mφ, macrophages; SAM, scar-associated macrophage; TM, tissue monocyte; t-SNE, t-distributed stochastic neighbor embedding.

Clusters MP2, MP4, and MP5 were classified as tissue monocytes based on their high expression levels of markers including CD14 and S100A9, and clusters MP0, MP3, MP8, and MP10 were presumed to be macrophages due to their high expression of C1QA. Specifically, MP0 was speculated to be Kupffer cells because these cells had high expression of MARCO, CD163, TLR4, TIMD4, etc., while MP3 were classified as MoMFs due to their high expression of S100A9 (38,39). MP8 and MP10 were defined as scar-associated macrophages (SAM) due to their expression of specific markers CD9 and TREM2 (40). MP4 and MP5 were categorized as THBS1⁺ tissue monocytes and ITGAL⁺ tissue monocytes, respectively.

To longitudinally dissect the phenotypic dynamics and cell transitions of Mo/Mφ during rapid liver regeneration, we constructed the pseudotime differentiation trajectory of the Mo/Mφ subsets by using Monocle (Figure 3C). The resulting transcriptional programs, along with transitional states, could be categorized into three modules. The pseudotime analysis showed that MP4 and MP5 were similar tissue monocytes that were located at the starting point of the trajectory axis, whereas KCs and MoMFs were mainly enriched at the middle and differentiated ends of the axis (Figure 3C and Figure S7B).

Among the various clusters, only THBS1⁺ tissue monocytes were increased in all three patients from stage I to stage II, and the average of THBS1⁺ tissue monocyte ratio in the Mo/Mφ population exhibited about a three-fold change in stage II compared with stage I (Figure 3D). Consistent with the CyTOF findings (Figure 2F), the scRNA-seq data also showed an enrichment of tissue monocytes in the residual livers at ALPPS stage II, which had high expression of SIRPA (CD172a) (Table S1).

Using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses, it was found that the genes upregulated in THBS1⁺ tissue monocytes were enriched in pathways related to high inflammatory activity, including the inflammatory response, toll-like receptor signaling, regulation of cell proliferation, NF-kappa B signaling pathway, TNF signaling pathway, MAPK signaling pathway, as well as pro-regeneration pathways related to liver regeneration including growth factors, positive regulation of angiogenesis, regulation of cell proliferation, VEGF signaling, etc. (Figure 3E and Figure S8A). A quantitative analysis of the gene expression also revealed relatively high gene set activity for pathways including pro-inflammatory, hypoxia inducible factor (HIF)-related genes, cytokines, and growth factors in the THBS1⁺ tissue monocytes (Figure 3F).

The results of multiplexed IF staining (THBS1, S100A9, CD14, DAPI) for liver samples from 22 ALPPS patients also showed that the ratio of THBS1⁺ tissue monocytes to the overall monocytes increased from 4.87% to 9.39% between stage I and stage II during liver regeneration (P=0.01, Figure 3G).

As revealed by QuSAGE, the gene set activity of growth factors, including EREG and AREG, was highly enriched in THBS1⁺ tissue monocytes (Figure 3F, Figure 4A and Figure S8A). The EGFR is conventionally known to play a crucial role in hepatocyte proliferation and liver regeneration (41). EREG and AREG are both EGRF ligands, and were highly expressed in THBS1⁺ tissue monocytes, while in particular, EREG was almost exclusively expressed in the Mo/Mφ population, especially in THBS1⁺ tissue monocytes (Figure 4B and Figure S8B,S8C).

Figure 4 Proliferation phenotype of THBS1⁺ tissue monocytes. (A) t-SNE plot of the distribution of the Mo/Mφ population at ALPPS stage II. Results are color-coded according to the growth factor gene expression level. MP, Mo/Mφ population. (B) Violin and t-SNE plots of the EREG expression in THBS1⁺ tissue monocytes, i.e., THBS1⁺ tissue monocytes. (C) Network of hepatocyte receptor interactions (left panel) and bubble plot of monocyte cluster ligand and hepatocyte receptor interactions (right panel). (D) Bubble plot showing the expression of receptors in monocyte clusters. (E) Boxplot representing a comparison of the peak concentrations of EREG in the plasma of patients at ALPPS stages I and II. *, P<0.05. (F) Line chart showing the OD450 values of hepatocytes in the EREG group and control group following CCK-8 incubation. (G) In the liver tissue from the murine model of ALPPS, regeneration was impaired after systemic neutralization of EREG. (H) Survival analysis of EREG rescue versus vehicle treatment for mice with PHLF. *, P<0.05; **, P<0.01; ***, P<0.001 (actual P values were reported in results). ALPPS, associating liver partition and portal vein ligation for staged hepatectomy; CCK-8, Cell Counting Kit-8; EREG, epiregulin; FLR, future liver remnant; Mo, monocytes; Mφ, macrophages; MP, Mo/Mφ population; PHLF, post-hepatectomy liver failure; TM, tissue monocyte; t-SNE, t-distributed stochastic neighbor embedding.

THBS1+ tissue monocytes release EREG to stimulate the liver regeneration induced by ALPPS

To further explore the role of THBS1⁺ tissue monocytes in promoting liver regeneration, we analyzed cell-cell interactions by observing ligand-receptor pairs using the CellphoneDB on the scRNA-seq data. Strong bidirectional communication was observed between THBS1⁺ tissue monocytes and hepatocytes (Figure 4C,4D). In particular, strong EREG-EGFR interactions between THBS1⁺ tissue monocytes and hepatocytes were observed. These interactions were previously reported to play key roles in promoting hepatocyte proliferation (41). This finding was also supported by the observation that the positive regulation of MAPK cascade, a proliferative pathway that can be stimulated by growth factors, was significantly activated in hepatocytes at ALPPS stage II (Figure 3E).

Plasma from ALPPS patients were also collected, and an ELISA for EREG in plasma was performed (validation cohort 2). We found that the circulating EREG concentration was gradually increased from ALPPS stage I to stage II (P=0.045, Figure 4E and Figure S9A, the concentration of ARAG was also examined seeing in Figure S9B,S9C). Furthermore, in vitro CCK-8 experiments also showed that the administration of EREG could significantly promote the proliferation of a human hepatocyte cell line, THLE-2 (P<0.05, Figure 4F and Figure S10). A mouse model of ALPPS (using the RML as the FLR, 30% of the total liver volume) was then constructed. The FLR weight/body weight ratio at stage II was 2.1% in the control group, while it was only 1.5% after the administration of an EREG-neutralizing antibody (P=0.02, Figure 4G). We also constructed a PHLF mouse model (the LML was used as an insufficient liver remnant, 10% of the total liver volume) and the role of EREG in treating PHLF was explored using this model (Figure 4H). Our in vivo experiments showed that the administration of EREG to PHLF mice could improve their survival compared with the control group. The median survival time after PHLF was prolonged from 22 to 36 hours by the administration of EREG (P=0.048, Figure 4H). In summary, the above results revealed that THBS1⁺ tissue monocytes could induce the proliferation of hepatocytes and liver regeneration via EREG secretion.

THBS1+ tissue monocytes were recruited by hepatocytes via the C5a/C5aR1 interaction

The results of bulk RNA-seq showed the pathway of monocyte chemotaxis based on a GO analysis, which correlated with our observation above (Figure 5A). To further explore the mechanism underlying the enrichment of THBS1⁺ tissue monocytes in the FLR after ALPPS stage I, CellphoneDB was used to analyze the cell-cell interactions between THBS1⁺ tissue monocytes and other cell types by observing ligand-receptor pairs (Figure 5B). Strong cell-cell communications between hepatocytes and THBS1⁺ tissue monocytes were observed through the C5-C5AR1 axis. After examining the receptors expressed by THBS1⁺ tissue monocytes and looking for the corresponding ligands, we found that hepatocytes can recruit circulating monocytes through the C5a-C5aR1 interaction (Figure 5B). Furthermore, a scRNA-seq analysis revealed that the complement component, C5a, was specifically expressed in hepatocytes, and C5aR1 was widely expressed among the Mo/Mφ population, with relatively high expression in THBS1⁺ tissue monocytes (Figure 5C,5D).

Figure 5 The potential role of C5a in the recruitment of THBS1⁺ tissue monocytes. (A) The results of a GO analysis of bulk FLR tissue. (B) Bubble plots of THBS1⁺ tissue monocyte receptor interactions (left) and hepatocyte ligand interactions (right) generated according to the CellphoneDB. The color is coded by the mean intensity of the interaction, and the size of the bubble is determined by the −log₁₀P. (C) t-SNE plot showing clusters generated by scRNA-seq data. The color is coded by the expression level of C5a. (D) t-SNE plot showing MP clusters generated by scRNA-seq data, and a violin plot showing the C5aR1 expression in all Mo/Mφ populations. (E) GO and KEGG analyses of genes up-regulated in stage II hepatocytes as revealed by scRNA-seq. FLR, future liver remnant; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; Mo, monocytes; MP, Mo/Mφ population; Mφ, macrophages; scRNA-seq, single-cell RNA sequencing; t-SNE, t-distributed stochastic neighbor embedding.

A GO analysis of scRNA-seq data showed that the genes upregulated in hepatocytes of the FLR in stage II were enriched in signaling pathways that involved complement activation, and complement activation via the classical pathway, compared with hepatocytes from the FLR in stage I (Figure 5E). Further, IF staining of tissue from 22 patients who underwent ALPPS revealed that C5aR1 was expressed in tissue monocytes (C5aR1⁺CD14⁺ cells) (Figure 6A), while hepatocytes in the FLR at stage II showed significantly higher expression of C5a compared with those at stage I (validation cohort 1, P=0.01; Figure 6B).

Figure 6 THBS1⁺ tissue monocytes were recruited through C5a. (A) Multiplexed immunofluorescence staining, showing the distribution of C5aR1⁺ monocytes (CD14⁺C5aR1⁺). (B) Immunohistochemical staining of C5a at ALPPS stage I (upper left) and stage II (lower left). The H-score shows the up-regulation of C5a expression in hepatocytes from ALPPS stage I to stage II. *, P<0.05. (C) Liver samples from the murine ALPSS model on POD1 and POD5 (left). The boxplot represents comparisons of the FLR to body weight ratio between the control group and PMX-53 group on POD1 and POD5. (D) Immunohistochemical staining of Ki-67 for the control group and PMX-5 group on POD1 and POD5 (left). The boxplot shows comparisons of the ratios of Ki-67⁺ cells between the control and PMX-53 group on POD1 and POD5 (right). (E) Boxplot showing the EREG expression in the control and PMX-53 group on POD1 and POD5. (F) Schematic diagram of the ALPPS-induced liver regeneration. *, P<0.05 (actual P values were reported in results). ALPPS, associating liver partition and portal vein ligation for staged hepatectomy; EREG, epiregulin; IF, immunofluorescence; IHC, immunohistochemistry; FLR, future liver remnant; POD, post-operation day.

To further explore the role of C5a during liver regeneration in vivo, treated the mouse ALPPS model (RML as FLR, 30% of total liver volume) with the antagonist of C5aR1 (PMX53) and evaluated the change in the FLR mass based on the FLR weight/body weight ratio (Figure 6C). There were no significant differences between the PMX53 group and control group on POD1 (1.19% vs. 1.24%, Figure 6C). However, the FLR weight/body weight ratio at stage II (POD 5) was 2.03% in the control group, while it was only 1.55% after the application of PMX53 (P=0.03, Figure 6C). The proliferation index of hepatocytes was also evaluated by Ki67 staining, which revealed that the average ratio of Ki67 hepatocytes/all hepatocytes in the PMX53 group on POD5 was significantly reduced to 5.96% compared with 13.97% in the control group (P=0.04, Figure 6D). The transcriptional level of EREG in the FLR tissues were also significantly reduced in the PMX group compared with that in control group on POD5 (P=0.03, Figure 6E).

Discussion

The emerging use of ALPPS in clinical application could induce rapid liver regeneration in HCC patients, allowing them the chance to undergo radical tumor resection (8). As a treatment option for unresectable HCC, ALPPS has its advantages compared to liver transplantation. ALPPS is not limited by liver donor availability. There’s also no concern about disease progression during the waiting period for a liver donor. Additionally, patients who undergo ALPPS do not need to receive immunosuppressive therapy in the later stages. However, despite the promising results of this procedure, few researchers have examined the potential mechanism(s) underlying the rapid liver regeneration induced by ALPPS, despite the need to identify better strategies to deal with decompensation of liver function (42,43). Our present study comprehensively dissected the circulating and local immune cell composition and the dynamic changes during ALPPS-induced rapid regeneration by using CyTOF and scRNA-seq analysis, providing a glimpse into the architecture of the liver microenvironment at single-cell resolution. In the present study, we focused on immune cells to explore the mechanism of ALPPS-induced liver regeneration.

Several studies have shown that the immune system can regulate the repair and regeneration of the liver parenchyma after injury (44). In different liver regeneration models (after hepatectomy or drug-induced liver injury), various immune cells were found to play different roles. In patients with ALPPS, the liver tissue of the FLR is a T-cell enriched microenvironment. Monocytes were the second most enriched immune cell component in the liver microenvironment, and a shift from a macrophage-dominant to a monocyte-dominant liver microenvironment was observed, with a significant enrichment of monocytes at stage II.

Normally, tissue repair is accompanied by a large accumulation of mononuclear phagocytes at the site of injury (45). However, a large number of macrophages reside in the liver, and surface markers are often difficult to distinguish among different phenotypes of mononuclear phagocytes. However, through mass spectrometry flow clustering and diffusion map analysis, tissue monocytes and macrophages can be classified and clearly defined, providing more information about their functions. Although the mononuclear phagocyte ratio in immune cells remained unchanged during ALPPS-induced liver regeneration, we found that after the cells were reclustered, the proportion of tissue monocytes significantly increased, even surpassing the tissue-resident macrophages in stage II. The change in proportion could be due to the simultaneous occurrence of an increase in monocytes and a decrease in macrophages. The reduction in macrophages was also reported in other models of liver regeneration (46). The diffusion map suggests that the phenotype of the tissue monocytes in our study was closer to that of peripheral blood monocytes. It is speculated that this group of tissue monocytes may be recruited from the peripheral blood. In fact, in peripheral blood, monocytes exhibited an increase in absolute numbers during the ALPPS induced liver regeneration (Figure S3E). Other studies have shown that monocyte recruitment occurs in the early stage of liver aseptic injury and plays a role in repairing liver damage (47). This recruitment and aggregation phenomenon was maintained for at least 2–3 weeks after surgery in our study. These observations show that these tissue monocytes may play an important role in the whole process of liver regeneration induced by ALPPS.

It is currently believed that there are two different developmental branches of macrophages in the liver. One of the tissue-resident macrophage branches (mainly KC) is derived from the differentiation of progenitor cells during the development of the yolk sac and/or fetal liver (48). The maintenance of these cells depends on their self-renewal (49). The other branch is differentiated from monocytes recruited from the peripheral blood, which are called MoMF. In this study, both the scRNA-seq results and CyTOF results showed two subpopulations of macrophages with different phenotypes, which are speculated to represent KC and MoMF, respectively. Based on their phenotypes, we distinguished these two lineages by S100A9, which is marker commonly used to distinguish these two subpopulations (50). Since there was a significant enrichment of tissue monocytes after ALPPS stage I, we further dissected the exact phenotypes of the tissue monocytes in the liver regeneration induced by ALPPS. The scRNA-seq results indicated that THBS1⁺ tissue monocytes accounted for the highest proportion of cells, and the population exhibited an upward trend in all patients. Some studies have demonstrated that THBS1 is necessary for tissue repair (51). A recent study showed that THBS1⁺ Mreg cells expand in steatotic liver disease (SLD)-associated HCC (52). This would refer to the macrophage-monocyte shift to be linked with the HCC as a primary disease. The mentioned THBS1⁺ Mreg cells have a certain connection to our target cells. However, we believe that the influence of tumors on these cells is limited. Firstly, ALPPS-induced liver regeneration is a very rapid process (7–14 days), whereas the impact of tumors on immune cells requires a certain amount of time. Secondly, the ALPPS procedure isolates the FLR from the tumor, making it difficult for the tumor to directly induce significant changes in monocytes within the FLR. Additionally, other researchers have also proposed that the relevant monocytes contribute positively to liver regeneration (53,54). In the present study, it was found that THBS1⁺ tissue monocytes highly expressed epidermal growth factors, especially EREG, which was almost exclusively expressed in this cluster. Past studies have found that EREG, the ligand of EGFR, is highly expressed in the serum of patients with liver injury (55). High expression of EREG has also been observed in liver regeneration induced by liver resection or liver injury (55,56). EGFR activation can lead to the activation of its downstream MAPK signaling pathway and promote hepatocyte proliferation (57). The increase in the concentration of epidermal growth factors, including EREG, in the serum of ALPPS patients was observed after stage I in the present study. Supporting these findings, we observed that blocking EREG inhibited the FLR regeneration after ALPPS, suggesting that the application of EREG could also potentially serve as a rescue method for PHLF.

The complement system is involved in the process of tissue repair and regeneration (58). In this study, the scRNA-seq analysis indicated that the complement pathway was activated in hepatocytes after stage I of ALPPS. It was further found that C5 is highly expressed in hepatocytes. Studies have shown that C5a is also able to recruit monocytes. For example, after C5a was neutralized by antibodies, the recruitment of CD11b⁺CD45^hi monocytes was greatly reduced (59). The present study confirmed that C5a is highly expressed by hepatocytes stimulated by ALPPS, and is almost exclusively secreted by hepatocytes. However, the initiator of the complement system during ALPPS is still unknown. It has been reported that the activation of the complement system is a reaction to liver local injury, systemic inflammation, and regeneration. During ALPPS, all three of these processes would serve as potential activators. The most essential regulators of the complement response should be identified in future studies. The present receptor-ligand interaction analysis also suggested that C5a acts on THBS1⁺ monocytes through C5aR1 and plays a role in their recruitment and activation. At the same time, it was observed that the volume of the FLR was significantly lower after PMX-53 treatment compared to the control group, indicating that blocking C5a reduces ALPPS-induced liver regeneration. Immunohistochemical staining for Ki67 also indicated a reduction in the proliferation of hepatocytes after C5aR1 was blocked. In addition, the down-regulation of EREG and AREG was also observed after the inhibition of C5a, suggesting that C5a plays an important role in recruiting and activating THBS1⁺ monocytes to secrete EREG. Whether the secretion of EREG is directly downstream of C5aR1 activation needs further validation.

It has been reported that the liver volume increase after ALPPS overestimates the true functional increase in the colorectal dominant profile (60). All patients in this study have HCC. Regrettably, the number of active hepatocytes obtained from single-cell sequencing in this study was limited due to the limited sample volume, and the resulting data were difficult to analyze in detail regarding the hepatocyte status. Previous reports from our center (11) showed that hypertrophy in the normal liver and the advanced fibrosis liver was different based on the light microscopy of FLR samples. The hypertrophy of FLR in the normal liver is attributed to both regeneration and increased size of hepatocytes, whereas hypertrophy of FLR in the advanced fibrosis liver mainly relies on an increase in the size of hepatocytes. Further mechanistic studies are needed to clarify.

The hypothesis of the present study is summarized in Figure 6F. In brief, we found that the complement system is activated after the ALPPS procedure. C5a, one of the most pivotal components, is synthesized in hepatocytes and secreted, where it acts as a chemotactic factor for monocytes. Large amounts of monocytes would infiltrate into the FLR and reorganize the local mononuclear phagocytes, shifting the balance from a macrophage-enriched into monocytes-enriched liver microenvironment. The increased tissue monocytes (called THBS1⁺ tissue monocytes) are the exclusive source of EREG, which plays a vital role in the ALPPS-induced hepatocyte proliferation.

In this study, a comprehensive transcriptional analysis of the liver immune microenvironment of ALPPS patients was performed. Functional experiments validating the mechanism of THBS1⁺ tissue monocytes promoting ALPPS-induced liver regeneration were explored, and several noteworthy findings were uncovered. Nevertheless, there are shortcomings of our study that should be kept in mind. First of all, due to limited ALPPS surgery conduction, the sample size of the cohorts in our study is not very large. Consequently, the study of molecular mechanisms is not deep enough. Similarly, due to the limited sample size, factors affecting liver regeneration such as previous illnesses, previous therapy, or liver damage cannot be analyzed. In addition, the role of EREG in ALPPS-induced liver regeneration needs to be further verified using additional techniques in a larger number of patients. Incorporating multiple time points and pathological conditions is also needed. A comprehensive comparison between ALPPS and port vein embolization can help further examine the pertinent mechanisms.

Conclusions

Collectively, our study found that the complement pathway was activated in hepatocytes after the ALPPS procedure. The hepatocytes subsequently secreted C5a, which then recruited THBS1⁺ tissue monocytes. These monocytes could further release growth factors such as EREG to promote hepatocyte proliferation in the residual liver. These findings may be useful not only to design a strategy to boost liver regeneration but also to identify promising therapeutic targets that might be used in the treatment of acute liver failure in the future.

Acknowledgments

We thank all the patients who participated to this study. We also thank NovelBio Ltd. and PLT Tech Ltd. for providing technical support.

Footnote

Reporting Checklist: The authors have completed the ARRIVE reporting checklist. Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-24-391/rc

Data Sharing Statement: Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-24-391/dss

Peer Review File: Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-24-391/prf

Funding: This work was jointly supported by the Original Discovery Program of National Natural Science of China (No. 82150004), the National Natural Science Foundation of China grants (Nos. 82072715 and 82341027), the Eastern Talent Program (Leading Project), the Shanghai Municipal Health Commission (Nos. 201940075 and 2022LJ005), the Shanghai Science and Technology Commission (Nos. 21140900300 and 22S31901800), the Project from Shanghai Hospital Development Center (No. SHDC2023CRD025), the Projects from Science Foundation of Zhongshan Hospital, Fudan University (Nos. 2021ZSCX28, 2020ZSLC31), Shanghai Municipal Science and Technology Major Project, and the Projects from Science Foundation of Shanghai Geriatrics Center, Zhongshan Hospital, Fudan University Minhang Meilong (No. YQ2004-003).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-24-391/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted following the Declaration of Helsinki (and its subsequent amendments). The Ethics Committee of Zhongshan Hospital, Fudan University approved the protocol (No. 2024-120). Written informed consent was obtained from each participant. Animal experiments were performed under a project license (No. 2019-185) granted by the institutional ethics board of Zhongshan Hospital, Fudan University, in compliance with institutional guidelines for the care and use of animals.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Zhou M, Wang H, Zeng X, et al. Mortality, morbidity, and risk factors in China and its provinces, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2019;394:1145-58. [Crossref] [PubMed]
2. Singal AG, Kanwal F, Llovet JM. Global trends in hepatocellular carcinoma epidemiology: implications for screening, prevention and therapy. Nat Rev Clin Oncol 2023;20:864-84. [Crossref] [PubMed]
3. Syaiful RA, Mazni Y, Siagian NKP, et al. Surgical resection for hepatocellular carcinoma: a single-centre's one decade of experience. Ann Med Surg (Lond) 2024;86:1289-96. [Crossref] [PubMed]
4. Llovet JM, Pinyol R, Yarchoan M, et al. Adjuvant and neoadjuvant immunotherapies in hepatocellular carcinoma. Nat Rev Clin Oncol 2024;21:294-311. [Crossref] [PubMed]
5. Gordan JD, Kennedy EB, Abou-Alfa GK, et al. Systemic Therapy for Advanced Hepatocellular Carcinoma: ASCO Guideline Update. J Clin Oncol 2024; [Crossref] [PubMed]
6. Clavien PA, Petrowsky H, DeOliveira ML, et al. Strategies for safer liver surgery and partial liver transplantation. N Engl J Med 2007;356:1545-59. [Crossref] [PubMed]
7. European Association for the Study of the Liver. EASL Clinical Practice Guidelines on liver transplantation. J Hepatol 2024;81:1040-86. [Crossref] [PubMed]
8. de Santibañes E, Clavien PA. Playing Play-Doh to prevent postoperative liver failure: the "ALPPS" approach. Ann Surg 2012;255:415-7. [Crossref] [PubMed]
9. Chebaro A, Buc E, Durin T, et al. Liver Venous Deprivation or Associating Liver Partition and Portal Vein Ligation for Staged Hepatectomy?: A Retrospective Multicentric Study. Ann Surg 2021;274:874-80. [Crossref] [PubMed]
10. Korenblik R, Heil J, Smits J, et al. Liver regeneration after portal and hepatic vein embolization improves overall survival compared with portal vein embolization alone: mid-term survival analysis of the multicentre DRAGON 0 cohort. Br J Surg 2024;111:znae087. [Crossref] [PubMed]
11. Wang Z, Peng Y, Hu J, et al. Associating Liver Partition and Portal Vein Ligation for Staged Hepatectomy for Unresectable Hepatitis B Virus-related Hepatocellular Carcinoma: A Single Center Study of 45 Patients. Ann Surg 2020;271:534-41. [Crossref] [PubMed]
12. Chan A, Zhang WY, Chok K, et al. ALPPS Versus Portal Vein Embolization for Hepatitis-related Hepatocellular Carcinoma: A Changing Paradigm in Modulation of Future Liver Remnant Before Major Hepatectomy. Ann Surg 2021;273:957-65. [Crossref] [PubMed]
13. Pepe-Mooney BJ, Dill MT, Alemany A, et al. Single-Cell Analysis of the Liver Epithelium Reveals Dynamic Heterogeneity and an Essential Role for YAP in Homeostasis and Regeneration. Cell Stem Cell 2019;25:23-38.e8. [Crossref] [PubMed]
14. Yang W, Tao Y, Wu Y, et al. Neutrophils promote the development of reparative macrophages mediated by ROS to orchestrate liver repair. Nat Commun 2019;10:1076. [Crossref] [PubMed]
15. Brandel V, Schimek V, Göber S, et al. Hepatectomy-induced apoptotic extracellular vesicles stimulate neutrophils to secrete regenerative growth factors. J Hepatol 2022;77:1619-30. [Crossref] [PubMed]
16. Selzner N, Selzner M, Odermatt B, et al. ICAM-1 triggers liver regeneration through leukocyte recruitment and Kupffer cell-dependent release of TNF-alpha/IL-6 in mice. Gastroenterology 2003;124:692-700. [Crossref] [PubMed]
17. Amemiya H, Kono H, Fujii H. Liver regeneration is impaired in macrophage colony stimulating factor deficient mice after partial hepatectomy: the role of M-CSF-induced macrophages. J Surg Res 2011;165:59-67. [Crossref] [PubMed]
18. Melgar-Lesmes P, Edelman ER. Monocyte-endothelial cell interactions in the regulation of vascular sprouting and liver regeneration in mouse. J Hepatol 2015;63:917-25. [Crossref] [PubMed]
19. Martin-Murphy BV, Kominsky DJ, Orlicky DJ, et al. Increased susceptibility of natural killer T-cell-deficient mice to acetaminophen-induced liver injury. Hepatology 2013;57:1575-84. [Crossref] [PubMed]
20. Graubardt N, Fahrner R, Trochsler M, et al. Promotion of liver regeneration by natural killer cells in a murine model is dependent on extracellular adenosine triphosphate phosphohydrolysis. Hepatology 2013;57:1969-79. [Crossref] [PubMed]
21. Bi J, Zheng X, Chen Y, et al. TIGIT safeguards liver regeneration through regulating natural killer cell-hepatocyte crosstalk. Hepatology 2014;60:1389-98. [Crossref] [PubMed]
22. Connolly MK, Ayo D, Malhotra A, et al. Dendritic cell depletion exacerbates acetaminophen hepatotoxicity. Hepatology 2011;54:959-68. [Crossref] [PubMed]
23. Liang Q, Liu Z, Zhu C, et al. Intrahepatic T-Cell Receptor β Immune Repertoire Is Essential for Liver Regeneration. Hepatology 2018;68:1977-90. [Crossref] [PubMed]
24. Tumanov AV, Koroleva EP, Christiansen PA, et al. T cell-derived lymphotoxin regulates liver regeneration. Gastroenterology 2009;136:694-704.e4. [Crossref] [PubMed]
25. Behnke K, Zhuang Y, Xu HC, et al. B Cell-Mediated Maintenance of Cluster of Differentiation 169-Positive Cells Is Critical for Liver Regeneration. Hepatology 2018;68:2348-61. [Crossref] [PubMed]
26. Dhar DK, Mohammad GH, Vyas S, et al. A novel rat model of liver regeneration: possible role of cytokine induced neutrophil chemoattractant-1 in augmented liver regeneration. Ann Surg Innov Res 2015;9:11. [Crossref] [PubMed]
27. Ramachandran P, Matchett KP, Dobie R, et al. Single-cell technologies in hepatology: new insights into liver biology and disease pathogenesis. Nat Rev Gastroenterol Hepatol 2020;17:457-72. [Crossref] [PubMed]
28. Sahaf B, Rahman A, Maecker HT, et al. High-Parameter Immune Profiling with CyTOF. Methods Mol Biol 2020;2055:351-68. [Crossref] [PubMed]
29. Saviano A, Henderson NC, Baumert TF. Single-cell genomics and spatial transcriptomics: Discovery of novel cell states and cellular interactions in liver physiology and disease biology. J Hepatol 2020;73:1219-30. [Crossref] [PubMed]
30. Chen S, Zhou Y, Chen Y, et al. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 2018;34:i884-90. [Crossref] [PubMed]
31. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 2013;29:15-21. [Crossref] [PubMed]
32. Vento-Tormo R, Efremova M, Botting RA, et al. Single-cell reconstruction of the early maternal-fetal interface in humans. Nature 2018;563:347-53. [Crossref] [PubMed]
33. Aibar S, González-Blas CB, Moerman T, et al. SCENIC: single-cell regulatory network inference and clustering. Nat Methods 2017;14:1083-6. [Crossref] [PubMed]
34. Chaturantabut S, Shwartz A, Evason KJ, et al. Estrogen Activation of G-Protein-Coupled Estrogen Receptor 1 Regulates Phosphoinositide 3-Kinase and mTOR Signaling to Promote Liver Growth in Zebrafish and Proliferation of Human Hepatocytes. Gastroenterology 2019;156:1788-1804.e13. [Crossref] [PubMed]
35. Almau Trenard HM, Moulin LE, Padín JM, et al. Development of an experimental model of portal vein ligation associated with parenchymal transection (ALPPS) in rats. Cir Esp 2014;92:676-81. [Crossref] [PubMed]
36. Dili A, Lebrun V, Bertrand C, et al. Associating liver partition and portal vein ligation for staged hepatectomy: establishment of an animal model with insufficient liver remnant. Lab Invest 2019;99:698-707. [Crossref] [PubMed]
37. Chevrier S, Levine JH, Zanotelli VRT, et al. An Immune Atlas of Clear Cell Renal Cell Carcinoma. Cell 2017;169:736-749.e18. [Crossref] [PubMed]
38. Antoniades CG, Khamri W, Abeles RD, et al. Secretory leukocyte protease inhibitor: a pivotal mediator of anti-inflammatory responses in acetaminophen-induced acute liver failure. Hepatology 2014;59:1564-76. [Crossref] [PubMed]
39. Triantafyllou E, Pop OT, Possamai LA, et al. MerTK expressing hepatic macrophages promote the resolution of inflammation in acute liver failure. Gut 2018;67:333-47. [Crossref] [PubMed]
40. Ramachandran P, Dobie R, Wilson-Kanamori JR, et al. Resolving the fibrotic niche of human liver cirrhosis at single-cell level. Nature 2019;575:512-8. [Crossref] [PubMed]
41. Komposch K, Sibilia M. EGFR Signaling in Liver Diseases. Int J Mol Sci 2015;17:30. [Crossref] [PubMed]
42. Langiewicz M, Graf R, Humar B, et al. JNK1 induces hedgehog signaling from stellate cells to accelerate liver regeneration in mice. J Hepatol 2018;69:666-75. [Crossref] [PubMed]
43. Borger P, Schneider M, Frick L, et al. Exploration of the Transcriptional Landscape of ALPPS Reveals the Pathways of Accelerated Liver Regeneration. Front Oncol 2019;9:1206. [Crossref] [PubMed]
44. Caldwell CC, Okaya T, Martignoni A, et al. Divergent functions of CD4+ T lymphocytes in acute liver inflammation and injury after ischemia-reperfusion. Am J Physiol Gastrointest Liver Physiol 2005;289:G969-76. [Crossref] [PubMed]
45. Mescher AL, Neff AW, King MW. Inflammation and immunity in organ regeneration. Dev Comp Immunol 2017;66:98-110. [Crossref] [PubMed]
46. Zigmond E, Samia-Grinberg S, Pasmanik-Chor M, et al. Infiltrating monocyte-derived macrophages and resident kupffer cells display different ontogeny and functions in acute liver injury. J Immunol 2014;193:344-53. [Crossref] [PubMed]
47. Dal-Secco D, Wang J, Zeng Z, et al. A dynamic spectrum of monocytes arising from the in situ reprogramming of CCR2+ monocytes at a site of sterile injury. J Exp Med 2015;212:447-56. [Crossref] [PubMed]
48. Fan X, Lu P, Cui XH, et al. Repopulating Kupffer cells originate directly from hematopoietic stem cells. Stem Cell Res Ther 2023;14:351. [Crossref] [PubMed]
49. Hashimoto D, Chow A, Noizat C, et al. Tissue-resident macrophages self-maintain locally throughout adult life with minimal contribution from circulating monocytes. Immunity 2013;38:792-804. [Crossref] [PubMed]
50. Antoniades CG, Quaglia A, Taams LS, et al. Source and characterization of hepatic macrophages in acetaminophen-induced acute liver failure in humans. Hepatology 2012;56:735-46. [Crossref] [PubMed]
51. Tatematsu Y, Khan Q, Blanco T, et al. Thrombospondin-1 Is Necessary for the Development and Repair of Corneal Nerves. Int J Mol Sci 2018;19:3191. [Crossref] [PubMed]
52. Giraud J, Chalopin D, Ramel E, et al. THBS1(+) myeloid cells expand in SLD hepatocellular carcinoma and contribute to immunosuppression and unfavorable prognosis through TREM1. Cell Rep 2024;43:113773. [Crossref] [PubMed]
53. Nachmany I, Nevo S, Edelheit S, et al. Myeloid derived suppressor cells mediate hepatocyte proliferation and immune suppression during liver regeneration following resection. Genes Immun 2024;25:483-91. [Crossref] [PubMed]
54. Nachmany I, Bogoch Y, Sivan A, et al. CD11b(+)Ly6G(+) myeloid-derived suppressor cells promote liver regeneration in a murine model of major hepatectomy. FASEB J 2019;33:5967-78. [Crossref] [PubMed]
55. Tomita K, Haga H, Mizuno K, et al. Epiregulin promotes the emergence and proliferation of adult liver progenitor cells. Am J Physiol Gastrointest Liver Physiol 2014;307:G50-7. [Crossref] [PubMed]
56. Komurasaki T, Toyoda H, Uchida D, et al. Mechanism of growth promoting activity of epiregulin in primary cultures of rat hepatocytes. Growth Factors 2002;20:61-9. [Crossref] [PubMed]
57. Berasain C, Avila MA. The EGFR signalling system in the liver: from hepatoprotection to hepatocarcinogenesis. J Gastroenterol 2014;49:9-23. [Crossref] [PubMed]
58. Huang YF, Zhu ZJ. The role of innate immune cells in post-hepatectomy liver regeneration. Hepatobiliary Surg Nutr 2023;12:267-8. [Crossref] [PubMed]
59. Katschke KJ Jr, Xi H, Cox C, et al. Classical and alternative complement activation on photoreceptor outer segments drives monocyte-dependent retinal atrophy. Sci Rep 2018;8:7348. [Crossref] [PubMed]
60. Chan KS, Shelat VG, Low HM, et al. Is the extent of functional liver remnant increase truly "functional"? A single-institution case series of patients with Associating Liver Partition and Portal vein ligation for Staged hepatectomy (ALPPS). Clin Exp Hepatol 2023;9:28-36. [Crossref] [PubMed]