EGFR-mediated crosstalk between vascular endothelial cells and hepatocytes promotes Piezo1-dependent liver regeneration

Hepatocyte proliferation is essential for recovering liver function after injury. In liver surgery, the mechanical stimulation induced by hemodynamic changes triggers vascular endothelial cells (VECs) to secrete large amounts of cytokines that enhance liver cell proliferation and play a pivotal role in liver regeneration. Piezo1, a critical mechanosensory ion channel, can detect and convert mechanical forces into chemical signals, importing external stimuli into cells and triggering downstream biological effects. However, the precise role of Piezo1 in VECs, especially in terms of mediating liver regeneration, remains unclear. Here, we report on a potential mechanism by which early changes in hepatic portal hemodynamics activate Piezo1 in VECs to promote hepatocyte proliferation during the process of liver regeneration induced by portal vein ligation (PVL) in rats. In this liver regeneration model, hepatocyte proliferation is mainly distributed in zone 1 and zone 2 of liver lobules at 24–48 h after surgery, while only a small number of Ki67-positive hepatocytes were observed in zone 3. Activation of Piezo1 promotes increased secretion of epiregulin (EREG) and amphiregulin (AREG) from VECs via the PKC/ERK1/2 axis, further activating epidermal growth factor receptors (EGFR) and ERK1/2 signals in hepatocytes and promoting proliferation. In addition, cytokines secreted by Piezo1-activated VECs can induce hepatocytes to undergo epithelial-mesenchymal transition (EMT). In the liver lobules, the expression of EGFR in hepatocytes of zone 1 and 2 is signi�cantly higher than that in zone 3. The EGFR inhibitor ge�tinib inhibits liver regeneration by suppressing the proliferation of hepatocytes in zones 1 and 2. Thus, activation of Piezo1 in VECs promotes hepatocyte proliferation, suggesting mechanical stimulation regulates hepatocyte proliferation in zones 1 and 2 during PVL-induced liver regeneration. These data provide a theoretical basis for the regulation of liver regeneration through chemical signals mediated by mechanical stimulation.


Introduction
The liver is an "injury-privileged" organ that can regenerate and recover after partial hepatectomy (PH) within a de ned period, despite the loss of up to two-thirds of liver parenchyma [1] .Nonetheless, acute liver failure is a life-threatening postoperative complication that can occur due to the substantial loss of liver parenchyma and is a signi cant limiting factor for liver surgery.Therefore, strategies aimed at promoting liver regeneration (LR) and recovery are crucial for overcoming the rate-limiting steps in the eld of liver surgery [2,3] .LR is a complex biological process involving the proliferation of parenchymal (hepatocytes) and non-parenchymal cells within the liver.The process involves a complex network of in ammatory and growth factors [1,4] .Overcoming the challenges associated with liver surgery requires an understanding of the molecular mechanisms that govern LR and identifying reliable strategies to optimize and augment LR.
The regulation of mechanical homeostasis is crucial for maintaining organ homeostasis and ensuring normal organ shape and function [5][6][7] .However, hemodynamic changes resulting from liver surgery disrupt the mechanical equilibria of cells within the liver, especially the vascular endothelial cells (VECs) [8] .Consequently, the relationship between VECs and mechanical forces has become an increasingly important area of study for LR [9] .Recent research indicates that hemodynamic changes play a critical role as "triggers and balancers" in the induction of LR following the loss of liver parenchyma [8] .Speci cally, the mechanical effects of hemodynamic changes on VECs are key to initiating and regulating LR [10] .Portal vein embolization, for example, leads to increased tensile forces on VECs, which triggers the production of high levels of interleukin-6 (IL-6), a key factor previously implicated in initiating LR [11] .Moreover, increased portal vein blood ow induces the expression of hepatocyte growth factor (HGF) by VECs, which promotes hepatocyte proliferation and hypertrophy, reduces apoptosis, and facilitates LR [12] .Blocking the increase in portal vein blood ow signi cantly delays LR [8,12] .Therefore, the stimulation of hemodynamic changes in the process of LR represents a crucial mechanism for initiating and regulating the regenerative response [4] .
Cells perceive mechanical forces predominantly through mechanosensing structures found on the cell surface [4] .Hepatocyte membranes contain Piezo1, a type of mechanosensory ion channel (MIC), that independently senses and transduces mechanochemical signals through ion ow [13][14][15] .Multiple studies have identi ed Piezo1 as a potential therapeutic target for various malignant and benign diseases [15][16][17] .It plays a critical role in maintaining homeostasis, including in epithelial cell density; transport and metabolism of calcium and iron homeostasis; regulation of bone tissue homeostasis by modulating osteoblast-osteoclast crosstalk; and innate immune homeostasis by regulating macrophage function [18- 22] .Piezo1 is also widely involved in the physiological and pathological processes of the cardiovascular and respiratory systems [20,23] , and is crucial in various malignant diseases [24][25][26] .Piezo1 activation in liver sinusoidal endothelial cells (LSEC) promotes the production of CXCL1 and the formation of neutrophil trapping nets that lead to microvascular microthrombi formation and accelerated development of liver cirrhosis in Budd-Chiari syndrome [27] .
The transient effects of hemodynamic changes induced by PH, portal vein ligation (PVL), and associated liver partition and portal vein ligation for staged hepatectomy (ALPPS) on VECs may differ from phenomena observed in Budd-Chiari syndrome.Furthermore, the role of Piezo1 in postoperative LR and its underlying mechanisms remain unclear.Therefore, in this study, we investigate the regulatory effects of Piezo1 activation in VECs on hepatocyte proliferation, and whether this activation could be a potential target for improving LR.

Rat PVL model
Sprague-Dawley rats weighing approximately 200-240 g were procured from Charles River Laboratories (Beijing, China).All experimental protocols were approved by the Animal Experiment Center of Tsinghua University.To investigate the pattern and regularity of LR following 70% PVL, we randomly divided 28 rats into seven groups of four rats each: sham-operated, 12, 24, 48, 72, 120, and 168 h post-surgery.The surgical procedure was performed as previously described [28] .Brie y, rats were anesthetized using iso urane, and an incision was made in the anterior midline of the abdominal wall.The portal vein was carefully separated from the bile duct and hepatic artery at the location of the main trunk of the portal vein entry into the left and middle liver, and double-ligated with 4 − 0 silk suture.To explore the effect of ge tinib on PVL-induced LR, six additional rats were randomly divided into two groups of three each.Rats in the two groups received intraperitoneal administration of ge tinib (20 mg/kg) or the vehicle only, respectively, every 12 h starting 36 h before PVL and continuing until 48 h after PVL.Liver tissue samples were collected from the regenerated liver of rats.
Ultrasound imaging is a noninvasive, real-time and convenient imaging method, which is widely used in the diagnosis of liver diseases [29] .We have previously reported the detailed analysis process and results of ultrasound localization microscopy (ULM) on microvascular blood ow changes after PVL [30] .Brie y, the abdomen of the rat was shaved to provide an acoustic imaging window and sterilized using 70% alcohol.With the animal xed in the supine position on a heating pad, a bolus of 0.2 mL of 5-fold diluted Supervue MBs (NUMT, Nanjing, China) was injected through the femoral vein.ULM data acquisition was performed longitudinally at two time points for three rat: a baseline scan performed before PVL and an acute response scan performed 30 min after PVL.Please refer to our published report [30] for detailed processing.

Cell lines and cell culture
Human umbilical vein endothelial cells (HUVEC) and SK-Hep1 [31] cells were immortalized and cultured in Dulbecco's modi ed Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37°C in a 5% CO 2 incubator.Human hepatocyte cell lines LO2 and HepaRG were also cultured under the same conditions.
For cell immuno uorescence, cells were plated at a density of 4,000 cells per well in a 96-well plate and allowed to attach overnight.After experimental treatment, cells were xed with 4% paraformaldehyde for 15 min at room temperature and then permeabilized with 0.3% Triton in TBST for 10 min.Non-speci c binding was blocked with 5% donkey serum for 1 h at room temperature.Cells were then incubated with PKCα (CST: #2056, 1:50 diluted in 5% donkey serum and 0.3% Triton X-100 in TBST) overnight at 4°C.After washing with PBS for 5 min in triplicate, the cells were incubated with the secondary antibody at room temperature for 1 h.Nuclei were stained with DAPI for 15-30 min, and then cells were again washed three times with TBST for 5 min each.A high content analysis system(Germany) was used to photograph and image the cells after adding 50 µL of PBS per well.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR).
Total RNA was extracted from cell lines of different groups using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and cDNA synthesized with a reverse transcription kit (Toyobo FSQ 301, Japan) following the manufacturer's protocol.The mRNA expression levels of target genes were measured using a quantitative uorescence kit (Toyobo QPS-201, Japan) and relative RNA expression levels were normalized to b-Actin.
The 2 −ΔΔCt method was used to analyze relative gene expression levels.The primers used in this study were synthesized by Ruibiotech (Beijing, China) and their sequences are provided in Supplementary Table 1.

Preparation of conditioned medium
To prepare conditioned medium (CM), VECs (HUVEC and SK-Hep1) were seeded into 10 cm dishes at approximately 30-40% density.After overnight incubation, FBS-free DMEM containing Yoda1 (10 µM) or vehicle was added according to experimental requirements to activate Piezo1 of the VECs for approximately 5-6 h.To exclude a potential non-speci c effect of Yoda1, cells were then washed three to ve times with PBS and the medium changed to FBS-free DMEM.Cells were then incubated for 20-24 h then collected and centrifuged at 2000 rpm for 10 min at 4°C.The supernatant was retained as CM and used to culture hepatocytes, with the addition of drugs according to experimental requirements.Six CM were produced, HCM: conditioned media from HUVEC with DMEM, HCM DMSO : conditioned media from HUVEC with treatment by DMSO, HCM Yoda1 : conditioned media from HUVEC with Piezo1 activated by Yoda1, SCM: conditioned media from SK-Hep1 with DMEM, SCM DMSO : conditioned media from SK-Hep1 with treatment by DMSO, SCM Yoda1 : conditioned media from SK-Hep1 with Piezo1 activated by Yoda1.

Cell proliferation assay
Hepatocytes were subjected to different experimental conditions, and then 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) solution was added to the culture medium with a working concentration of 5 mg/mL.After 6 h of incubation, the upper medium was removed, and crystals formed in the lower medium were dissolved using 150 µL DMSO.After 10 min of shaking, absorbance was measured at a wavelength of 490 nm.

Cell morphology observation
Hepatocytes subjected to different experimental conditions were xed with absolute methanol for 15 min, washed with PBS three times for 5 min each, and stained with crystal violet for 10-15 min.The cells were again triple-washed with PBS for 5 min, and their morphological changes were recorded by microscopy after adding PBS.

RNA sequencing assay
When HUVEC and SK-Hep1 cells plated in 10 cm cell culture dishes reached 30-40% con uence, culture dishes were treated with 10 µM Yoda1 in FBS-free DMEM for 6 h.Three independent samples of two cell lines in each group (DMSO and Yoda1) were used for RNA sequencing by Annoroda Gene Technology (Beijing, China).To understand the potential signal pathway of Yoda1-induced gene expression in HUVEC and SK-Hep1 cell lines, we performed gene set enrichment analysis (GSEA) using software obtained online (http://www.gsea-msigdb.org/gsea/index.jsp).To gain further insight into Yoda1-induced differences in gene expression changes, the online website bioladder (https://www.bioladder.cn/web/#/pro/cloud) was used to analyze gene expressions in the two cell lines of different groups.Log 2 [mRNA fold change] was used to identify differentially expressed mRNAs, with the calculated value of <-1.5 or > 1.5 deemed statistically signi cant (p ≤ 0.05).The online bioinformatics database (DAVID Bioinformatics Resources 6.8, NIAID/NIH, https://david.ncifcrf.gov/tools.jsp) was used to analyze the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways based on the upregulated gene list.The human ligand-receptor interaction pairs list was obtained from the website CellTalkDB (http://tcm.zju.edu.cn/celltalkdb/download.php).The overlapping gene lists of upregulated genes in the Yoda1-induced group of HUVEC and SK-Hep1 cell lines with ligand genes were again subjected to KEGG analysis using the DAVID website.

Enzyme-linked immunosorbent assay (ELISA)
Approximately 40 x 10 4 cells were seeded into a six-well plate overnight and incubated with Yoda1 to induce Piezo1 activation.The supernatant from VECs was collected and centrifuged to remove particles for further analysis.Levels of AREG and EREG were measured using AREG (EK0304) and EREG (EK1394) ELISA kits from Boster Biological Technology (Wuhan, China) following the manufacturer's protocol.Samples were run in duplicate and three independent experiments were conducted.

Statistical analysis
All experimental data are presented as means ± standard error from at least three independent experiments.Statistical analyses were performed using GraphPad Prism 8 software.To compare data between two different treatment groups, a two-tailed, unpaired Student's t-test was used.Differences were considered to be statistically signi cant when P ≤ 0.05.

Results
3.1 Portal vein hemodynamic alterations accompany PVL-induced regional LR in rats and Piezo1 is highly expressed in VECs.
To investigate the impact of changes in portal vein blood ow on LR, we used a rat PVL model.Figure 1A illustrates the experimental setup and surgical procedure.Initially, PVL leads to liver tissue atrophy on the ligated side and regeneration on the non-ligated side (Supplementary Fig. 1A).The regenerated liver tissue regained approximately 90% of the whole liver weight in 120-168 h (Fig. 1B).During early regeneration stages, the distribution of Ki67-positive hepatocytes was zone-dependent.Speci cally, Ki67positive hepatocytes were concentrated primarily in zones 1 and 2, with a signi cantly lower proportion observed around central veins (Fig. 1C and Supplementary Fig. 1B).Notably, in zone 1, Ki67-positive hepatocytes increased rapidly at 24-72 h, followed by a gradual decrease (Fig. 1D and Supplementary Fig. 1C).
We explored the hemodynamic changes of microvascular blood ow after PVL using ULM.Postoperative microvascular blood ow velocity signi cantly increased at 30 min after PVL (Fig. 1E, F).Piezo1, an important MIC responsive to various mechanical stimuli such as shear stress and tension, is expressed in many tissues, including liver tissue (Supplementary Fig. 2A).To clarify the expression level of Piezo1 in different liver cell types, we queried its expression in different liver cells through multiple searchable single-cell sequencing databases.Expression levels of Piezo1 are high in VECs (Fig. 1G, Supplementary Fig. 2B, C).Additionally, high expression of Piezo1 in liver VECs of mice, pigs and monkeys was found (Supplementary Fig. 2D).Cell sequencing of LSEC in the GEO database suggests that Piezo1 expression is higher than other MIC (Fig. 1H).Therefore, we further explored the effect of Piezo1 activation in VECs on hepatocyte proliferation and the corresponding mechanisms.

Conditioned media (CM) from VECs with active Piezo1 can promote proliferation and epithelialmesenchymal transition (EMT) of hepatocytes in vitro
To investigate the potential of Piezo1 activation in VECs to induce hepatocyte proliferation, we collected CM from VECs where Piezo1 had been activated by Yoda1 (a speci c agonist of Piezo1); HCM Yoda1 , CM from HUVEC cells; SCM Yoda1 , and CM from SK-Hep1 cells all promoted the proliferation of hepatocytes (Fig. 2A, B).Interestingly, we observed a concomitant weakening of intercellular connections and the disappearance of colony growth, with a gradual scattering of cells (Fig. 2C and Supplementary Fig. 3A).We next evaluated the expression of E-cadherin (CDH1, an epithelial marker) and Vimentin (VIM, a mesenchymal marker) in hepatocytes using qRT-PCR; HCM Yoda1 and SCM Yoda1 signi cantly downregulate CDH1 and upregulate VIM expression (Fig. 2D and Supplementary Fig. 3B).Taken together, our results indicate that the activation of Piezo1 in VECs contributes to the proliferation and EMT of hepatocytes.
Next, we employed shRNA to knock down the expression of Piezo1 in VECs (Fig. 2E, Supplementary 3C, D).The knockdown of Piezo1 resulted in a weakened ability of the CM to promote hepatocyte proliferation (Fig. 2F and Supplementary 3E).Additionally, we observed that hepatocyte distribution transformed from scattered to colony growth (Fig. 2G and Supplementary 3F).The qRT-PCR analysis showed an increase in CDH1 expression and a decrease in VIM expression levels in hepatocytes (Fig. 2H and Supplementary 3G), indicative of EMT.These results collectively suggest that Piezo1 activation in VECs promotes hepatocyte proliferation and EMT in vitro.

Piezo1 activation promotes the upregulation of the ERBB and MAPK signaling pathways in VECs.
Differential mRNA expression between vehicle-and Yoda1-treated HUVEC and SK-Hep1 cells at 6 h (Supplementary Fig. 4) revealed the presence of 790 upregulated and 570 downregulated genes in HUVEC and 1472 upregulated and 1022 downregulated genes in SK-Hep1 cells after Yoda1 treatment for 6 h (Fig. 3A).KEGG pathway enrichment analysis revealed that pathways related to MAPK and ERBB were enriched in both HUVEC and SK-Hep1 cell lines after Yoda1 treatment (Fig. 3B).The overlapping gene list of upregulated genes in Yoda1-induced HUVEC and SK-Hep1 cell lines with ligand genes from the online website CellTalkDB were further analyzed using KEGG analysis on the DAVID website.
3.4 PKC and ERK1/2 contribute to regulation of the expression of HBEGF, EREG and AREG in VECs under Piezo1 activation qRT-PCR results indicate that Yoda1 induces an increase in the mRNA expression of HBEGF, EREG, and AREG in VECs (Fig. 4A and Supplementary Fig. 5A).Furthermore, Piezo1 knockdown led to an obvious inhibition of the HBEGF, EREG, and AREG expression in VECs treated with Yoda1 (Fig. 4B and Supplementary Fig. 5B).In HCM Yoda1 and SCM Yoda1 , we found that the levels of AREG and EREG proteins were signi cantly elevated (Fig. 4C and Supplementary Fig. 5C), and Piezo1 knockdown led to a noticeable inhibition of EREG, AREG secretion in HUVEC (Fig. 4D).Mechanical forces have been shown to increase HBEGF secretion in various cell types, and its role in promoting hepatocyte proliferation and LR is well established.We investigated whether the MAPK family played a role in Yoda1-induced HBEGF, EREG and AREG expression in VECs.First, four speci c inhibitors with different targets were utilized to examine the regulation of HBEGF, AREG, and EREG by MAPK and downstream signaling pathways, namely Rav (ravoxertinib, a selective inhibitor of ERK1/2), SB (adezmapimod SB 203580, a selective inhibitor of p38 MAPK), SP (SP600125, a selective inhibitor of JNK), and SR11302 (a selective inhibitor of AP-1, a downstream target of JNK).The q-RT-PCR results showed that Rav could signi cantly inhibit the expression of HBEGF, EREG, and AREG, while SB could only inhibit the expression of AREG and EREG.However, SP and SR11302 had little effect on the expression of the three genes (Fig. 4E).WB results suggested that Yoda1 induces an increase in ERK1/2 phosphorylation (Supplementary Fig. 5D), and that knockdown of Piezo1 reduces this increase (Fig. 4F and Supplementary Fig. 5E).
Piezo1 activation can activate multiple protein kinases such as PKA, PKB and PKC.To determine which protein kinase mediates the enhanced phosphorylation of ERK1/2, we conducted a series of experiments.
The nonselective protein kinase inhibitor AM 2282 was used to block the activation of PKA and PKC, which signi cantly inhibited ERK1/2 activation and the expression of HBEGF, AREG, and EREG induced by Yoda1 in VECs (Fig. 4G, H and Supplementary Fig. 5F, G).Moreover, the PKC-speci c inhibitor GO 6983 also blocked Yoda1-induced ERK1/2 activation and the expression of HBEGF, AREG, and EREG (Fig. 4I, J ).To further con rm the role of PKC in the biological effects induced by Yoda1, we used Phorbol 12myristate 13-acetate (PMA), an established PKC activator, to promote ERK1/2 phosphorylation and elevated expression of HBEGF, EREG, and AREG in VECs (Supplementary Fig. 6A-D).These results suggest that the activation of PKC and ERK1/2 play a crucial role in the enhanced expression levels of HBEGF, AREG and EREG resulting from Piezo1 activation induced by Yoda1.
To further elucidate the relationship between Yoda1-induced Piezo1 activation and PKC activation, we performed immuno uorescence staining to examine the changes in PKCa, a critical member of the classical PKC family.First, we con rmed Piezo1 is involved in regulating the expression of PKCa.qRT-PCR and WB results demonstrated that Piezo1 knockdown did not affect the mRNA and protein expression of PKCa in HUEVC and SK-Hep1 cells (Supplementary Fig. 6E).Membrane translocation is a vital feature of PKCa activation.Yoda1 rapidly induces the membrane translocation of PKCa in both HUEVC and SK-Hep1 cells (Fig. 4K and Supplementary Fig. 6F).In summary, our results indicate that PKCα and ERK1/2 participate in regulating the expression of HBEGF, EREG and AREG in VECs where Piezo1 is activated by Yoda1 (Fig. 4L).

Hepatocyte proliferation and partial EMT induced by AREG and EREG depend on EGFR and ERK1/2
Based on RNA sequencing data from mice liver tissue in the GEO database (GSE169242), the expression level of EGFR is highest in the ERBB family (Supplementary Fig. 5A).We thus hypothesized that CM from VECs with Yoda1-activated Piezo1 (CM Yoda1 ) could activate EGFR and promote the proliferation and EMT of hepatocytes.WB results revealed that the CM Yoda1 signi cantly increases the phosphorylation levels of EGFR in LO2 and HepaRG cells (Fig. 5A).The promoting effect of CM Yoda1 on the proliferation of LO2 and HepaRG cells was blocked by ge tinib, an EGFR-speci c inhibitor (Fig. 5B).Morphological observations indicate that the morphology of hepatocytes was restored by ge tinib (Fig. 5C and Supplementary Fig. 7B).The qRT-PCR results showed that ge tinib partially reverses the mRNA expression level of CDH1 but has no signi cant effect on the expression of VIM (Fig. 5D and Supplementary Fig. 7C).These results suggest that CM Yoda1 can promote the proliferation and partial EMT of hepatocytes through the EGFR signaling pathway, which is characterized by the loss of epithelial characteristics with little effect on the expression of mesenchymal components.Additionally, the activation of Piezo1 in VECs may lead to the production of other cytokines, such as the IL6 family, TGF-β, and other cytokines, which could contribute to the promotion of hepatocyte proliferation and the occurrence of EMT.
Both AREG and EREG showed similar results to HCM Yoda1 and SCM Yoda1 in terms of hepatocyte morphological changes and proliferation (Fig. 5E and Supplementary Fig. 7D).qRT-PCR results indicate that AREG and EREG signi cantly decrease CDH1 expression in LO2 cells and have little effect on VIM; both have relatively minor effects on CDH1 and VIM expression in HepaRG cells (Fig. 5F).The observed differences in the response of LO2 and HepaRG cells to AREG and EREG likely re ect differences in the a nity of AREG and EREG for EGFR and in the basal state of the two cell lines.The qRT-PCR results indicate mRNA expression levels of CDH1 in HepaRG cells are signi cantly lower than that in LO2 cells.In addition, the levels of AREG and EREG were approximately 160 and 18,000 times higher, respectively, in HepaRG cells than in LO2 cells (Fig. 5G).

EGFR mediates hepatocyte proliferation in vitro and LR in vivo.
In light of the distinct baseline states of the two hepatocyte cell lines under investigation, we selected to focus our inquiry on LO2 cells to explore the mechanisms of AREG and EREG on the promotion of hepatocyte proliferation and partial EMT.Both AREG and EREG enhance the phosphorylation of EGFR in LO2 cells (Fig. 6A), and proliferative and partial EMT effects are abrogated by ge tinib treatment, which also restores morphological changes (Fig. 6B and Supplementary Fig. 8A).Our results suggest that AREG and EREG's roles in promoting LO2 cell proliferation and partial EMT are, to some extent, dependent on EGFR.With ERK1/2 serving as an important mediator of the ERBB signaling pathway (Fig. 6C), Rav can partially block the promoting effects of AREG and EREG on LO2 cell proliferation (Fig. 6D), while qRT-PCR assays reveal that Rav has a signi cant effect on the recovery of CDH1 (Supplementary Fig. 8B).The effects of AREG and EREG on p-ERK1/2 are not blocked by ge tinib (Fig. 6E), indicating the effect of AREG and EREG on LO2 cells is mediated through the EGFR and ERK1/2 signaling pathways, respectively (Fig. 6F).Notably, EGFR activation primarily mediates LO2 cell proliferation while ERK1/2 activation mediates partial EMT.
To investigate the potential role of EGFR in the distribution of proliferating hepatocytes during the early stages of PVL-induced LR, we rst examined the expression of EGFR in human liver tissue using an online database (https://www.proteinatlas.org) and found that its expression is relatively higher in hepatocytes in zones 1 and 2, and relatively lower in zone 3 (Supplementary Fig. 9).To further explore the distribution of EGFR expression in hepatic lobules, we performed immuno uorescence staining on human and rat liver tissue sections.EGFR expression was higher in zone 1 and zone 2 hepatocytes and relatively lower in zone 3 (Fig. 6G and 6H).This "non-central venous periphery" distribution of EGFR is similar to the distribution of Ki67-positive hepatocytes in the regenerating liver induced by PVL at 24-48 h.To investigate whether EGFR mediates the regional distribution of proliferating cells in the liver during the early stage of regeneration induced by PVL, we used ge tinib to inhibit EGFR activity (Fig. 6I).The drug and vehicle were administered intraperitoneally every 12 h starting 36 h before PVL and continuing until 48 h after PVL in the ge tinib and control groups.Ge tinib inhibited PVL-induced hepatocyte proliferation, as evidenced by the reduced positive rates of Ki67 and PCNA in one regenerating unit (within the range of PV-CV) in the ge tinib group (Fig. 6J-L).Taken together, these ndings suggest that activation of EGFR may mediate PVL-induced proliferation of zone 1 and zone 2 hepatocytes during LR.

Discussion
Hemodynamic changes are a common feature of liver surgery, and VECs are most closely connected to blood ow in the liver [32] .Hemodynamic changes induced by surgery inevitably cause signi cant functional change in VECs, including LSECs.Previous research has demonstrated the involvement of LSECs in modulating the initiation and cessation of LR and in regulating the proliferation of hepatocytes during the regeneration process [10,33,34] .Extreme PH-induced liver cell damage can induce postoperative LR disorders and liver failure [2,35] , which are important factors limiting liver surgery.However, controlled portal vein blood ow at 3.2 times higher than the baseline can promote hepatocyte proliferation and reduce hepatocyte necrosis [36] .Therefore, moderate hemodynamic changes can activate mechanosensing structures in VECs, leading to the production of cytokines, which play a crucial role in promoting LR [8,10,11] .
The present study demonstrates that activation of Piezo1 in VECs promotes hepatocyte proliferation through the release of HBEGF, EREG, and AREG.The regulatory mechanisms of these growth factors vary across different cell types and stimuli.Previous studies suggest that AP1 is involved in the increased expression of HBEGF induced by tension [37] .However, our ndings indicate that the expression of HBEGF in VECs induced by Yoda1 is not blocked by speci c inhibitors of JNK and AP1, suggesting that AP1 does not mediate the increase of HBEGF expression in this context.Similarly, several studies have shown that YAP1 regulates the expression of AREG and EREG [38][39][40][41] , while we found that the YAP1 inhibitor vertepor n did not inhibit Yoda1-induced increases in AREG and EREG expression (data not shown).
Instead, we observed that Yoda1-induced EREG and AREG expression was blocked by inhibitors of the ERK1/2 and P38 MAPK pathways, while Rav and SB had different effects on HBEGF expression.
Literature suggests that Piezo1 activation induces calcium in ux [18,42] , leading to the activation of PKC and ERK1/2.PKC inhibitors were able to block Yoda1-induced increases in ERK1/2 phosphorylation and expression levels of HBEGF, EREG, and AREG, indicating that the expression of these growth factors in VECs was regulated by the Piezo1/PKC/ERK1/2 axis.
Blood ow changes can stimulate the expression of various growth and in ammatory factors from VECs, among which HGF and IL6 are the more common cytokines regulating LR [11,43] .HBEGF plays a crucial role in LR and hepatocyte proliferation [44] .The transcriptional expression of HBEGF rapidly rises at 1.5 h post PH, peaks at 6 h, and continues to increase for up to 72 h.Notably, the expression of HBEGF is primarily observed in non-parenchymal cells, namely Kupffer cells and LSECs, rather than hepatocytes [44] .
Although overexpression of HBEGF does not affect baseline liver size, it accelerates the restoration of liver weight during the process of LR following PH [45] .AREG and EREG are also important ligands of EGFR and play a pivotal role in hepatocyte proliferation [46][47][48] .The expression level of AREG rapidly increases after PH in rats, promoting hepatocyte proliferation and DNA synthesis through the ERK1/2 and AKT signaling pathways [49] .In liver transplantation models, the induction of AREG expression signi cantly increases with 50% liver volume transplantation but is not signi cantly increased with 30% liver volume transplantation, resulting in disrupted LR [48] .Despite with greater mechanical stimulation in the 30% liver transplantation model, AREG expression is signi cantly inhibited compared to 50% liver transplantation model.The con icting literature results indicate that the regulation of AREG expression by mechanical stimulation is complex.In addition to being affected by blood ow, cholestatic liver injury, such as primary biliary cirrhosis and primary sclerosing cholangitis, increase expression levels of AREG to protect liver cells from damage [50] .The expression level of EREG increases rapidly after PH, but the abnormal expression of EREG does not affect the process of LR [51] .In vitro studies have shown that EREG can promote the proliferation of primary hepatocytes and hepatic progenitor cells [46,47] .Mechanical stimulation can induce increased expression of HBEGF in various cell types, including VECs [52][53][54] .LSECs are one of the main sources of HBEGF during regeneration [44] .Mechanical stretch can stimulate the transcriptional level of AREG in epidermal stem cells [55] , but whether the activation of Piezo1 in VECs can promote the expression of AREG is still unclear.In this study, we reveal for the rst time that activation of Piezo1 in VECs regulates the expression of HBEGF, EREG, and AREG, and that this is the mechanism by which EREG and AREG promote hepatocyte proliferation and partial EMT.These results provide a theoretical basis for further understanding and exploration of the role of VECs in promoting hepatocyte proliferation and LR under mechanical stimulation.
The activation of the EGFR is crucial for LR [56,57] .In a mouse model of PH, zone 2 hepatocytes displayed initial proliferation, which gradually spread toward zones 1 and 3. Notably, activation of the EGFR signaling pathway was critical for hepatocyte proliferation [58] .Despite its high expression level in the ERBB family, the distribution of EGFR remains poorly understood.Our in vitro study demonstrated that Piezo1 activation promotes hepatocyte proliferation by secreting EGFR ligands.Could changes in mechanical stimulation of VECs after PVL regulate the distribution of proliferating hepatocytes during the regeneration process?The distribution of EGFR was similar to that of Ki67-positive cells in the early stages of LR.In vivo experiments revealed that EGFR inhibition signi cantly reduces Ki67-and PCNApositive staining in hepatocytes in zones 1 and 2, impairing LR.Our in vitro experiments suggest that mechanical stimulation may regulate the regional distribution of hepatocyte proliferation during regeneration through Piezo1-mediated secretion of EGFR ligands from VECs.
Several models exist on the impact of changes in blood ow on LR, including PVL, PH, ALPPS, smallvolume liver transplantation, and partial liver transplantation.However, few studies have investigated the distribution of hepatocyte proliferation during regeneration.Our study found that activating Piezo1 in VECs upregulated the expression of EREG and AREG, which mediated hepatocyte proliferation through EGFR (Fig. 7).In vivo ndings suggest that the regional distribution of EGFR may mediate the spatial distribution of hepatocyte proliferation during the early stages of regeneration.This study is the rst to link spatial changes in proliferating hepatocytes during liver regeneration with the mechanical stimulation-endothelial cell system.
In conclusion, we investigated the effects and mechanisms of Piezo1 activation in VECs on hepatocyte proliferation for the rst time.Our results indicate that Piezo1 activation in VECs leads to the upregulation of HBEGF, EREG, and AREG gene expression in the ERBB signaling pathway through the PKCα-ERK1/2 axis.This, in turn, activates EGFR in hepatocytes, promoting cell proliferation.Our in vivo and in vitro ndings suggest that cytokines secreted by mechanically stimulated VECs partially in uence the distribution of hepatocyte proliferation during LR.These results provide partial molecular evidence on the hemodynamically-induced regulation of LR by VECs and offer a theoretical basis for developing more effective methods to promote LR by reducing mechanical damage, such as reducing portal vein blood ow and pressure.