Mitogen-Activated Protein Kinase BMK1 Mediates Protection in Liver Ischemia–Reperfusion Injury

Abstract

Background:

Hepatic ischemia-reperfusion injury (IRI) is a major complication occurring in hemorrhagic shock, liver resection and transplantation. BMK1, a member of MAPK family, is essential in the regulation of endothelial cell integrity and organ development.

Methods:

We explored putative role of BMK1 in human liver transplant recipients and in a murine model of partial warm liver IRI.

Results:

Liver biopsies from adult transplant patients (n=60) revealed increased intrahepatic BMK1 expression was associated with well-preserved histology and improved-graft function. In mice, the ischemia insult (90min) induced hepatocellular BMK1 expression, peaking at 6h of reperfusion. Inhibiting intrahepatic BMK1 (siRNA/chemically) made livers more vulnerable to IR- stress, as evidenced by elevated serum ALT levels, damaged hepatic architecture, increased necrosis, and inflammation. Conversely, BMK1-overexpression protected livers from IR stress in a KLF4-dependent manner. Further in vivo and in vitro experiments confirmed that BMK1-overexpression upregulated hepatic YAP mRNA/protein levels, and promoted YAP dephosphorylation, while silencing YAP abrogated the protective effects of BMK1. The hypoxia-reoxygenation in vitro assay indicated that hepatocellular BMK1 regulated YAP function via ITCH-related ubiquitination to LATS1. Hypoxia-reoxygenation stress to BMK-1-overexpressed hepatocytes resulted increased intracellular ITCH and decreased phosphorylated LATS1, while no change to the phosphorylation level of MST1. Furthermore, dual hepatocyte immunofluorescence staining revealed that BMK1 relied on the nuclear YAP-KLF4 cooperation, as hepatic deficiency of either BMK1, YAP or KLF4 enhanced the severity of hepatic IRI.

Conclusions:

This study documents the crucial role/underlying mechanism of BMK1 signaling and points to a novel target to minimize hepatocellular dysfunction in IR-stressed LT.

Introduction

Liver transplantation (LT) is the only treatment for end-stage liver disease (ESLD). However, hepatic ischemia-reperfusion injury (IRI) remains one of the major reasons for perioperative mortality and short-/long-term complications^[1]. Thus, novel therapeutic concepts to combat hepatic IRI are essential to improve LT outcomes and expand the donor organ pool available for life-saving transplantation.

The big mitogen-activated kinase-1 (BMK1), also known as extracellular signal-regulated kinase 5 (ERK5), is a member of the mitogen-activated protein kinase (MAPK) family^[2]. It contains a TEY consensus MEK (Mitogen-activated protein kinase kinase) phosphorylation motif and can be activated by the upstream kinase, MEK5^[3]. BMK1 has been implicated in cell survival, anti-apoptosis, angiogenesis, cell motility and proliferation^[3]. After activation, BMK1 is imported into the nucleus and then phosphorylates/activates transcription factors, like c-Fos, c-Myc, MEF2 (Myocyte enhancer factor 2), KLF4 (Krüppel-like factor 4) and KLF2^[4]. Sustained activation of BMK1 in human endothelial cells upregulates KLF4 activity and elicits protective gene expression pattern^[4]. In the liver, BMK1 regulates hepatocellular carcinoma cells (HCC) and modulates tumor growth/ development^[2]. However, its role in LT and hepatic IRI remains unknown.

YAP (Yes-associated protein), the downstream effector of Hippo pathway, is negatively regulated by MST1 (mammalian sterile 20-like protein kinase 1) and LATS1 (large tumor suppressor homolog 1) through phosphorylation^[5]. It regulates the transcriptional activity of TEAD (TEA domain family) in the nucleus, thereby controlling cell growth and death in many cell types^[6]. Our group has documented that YAP is an essential protein in preventing hepatocyte apoptosis and ROS-rendered tissue damage in mouse liver IRI^[7]. Although MST1 and LATS1 are the canonical upstream YAP regulators, several molecules have been identified to regulate YAP function through MST1/LATS1-dependent or independent means^[8–11]. Thus, how BMK1-YAP axis regulates liver IRI needs more exploration.

Here, we evaluated the role and underlying mechanisms of BMK1 signaling in hepatocellular protection during liver IRI. Retrospective analysis of human adult LT patients provided the evidence on the relationship between intrahepatic BMK1 levels and the graft injury after liver cold storage and warm reperfusion. Then, by constructing genetically-modified mice and using well-established hepatic warm IRI models, we explored how BMK1 protected livers from IR stress via hepatocellular YAP and KLF4 crosstalk. By providing evidence for previously unrecognized role of BMK1 in hepatic protection and illuminating the molecular mechanism, this study provides new insights into the potential therapeutic intervention against sterile inflammatory injury in organ transplantation.

Methods

Clinical liver transplantation study.

Sixty patients with end-stage liver disease who received orthotopic liver transplantation in Renji Hospital affiliated to Shanghai Jiao Tong University School of Medicine were recruited from January 2020-December 2020. Institutional research board (IRN) approved study protocol (#ChiCTR-IIR-16008040). All organ donations recorded in the electronic database contributed voluntarily, and no organ was obtained from executed prisoners or other institutionalized persons. Written informed consents were obtained from all subjects. Non-fatty donor organs were perfused with and stored in cold University of Wisconsin (UW) solution. Cold ischemia time was defined as the time from the start of perfusion to its removal from the cold storage. Hepatocellular injury was evaluated by serum alanine aminotransferase (sALT) at postoperative days (PODs) 1-7. Liver biopsies were obtained from the left lobe during back-table preparation (pre-LT: prior to implantation), and at 3h after portal reperfusion (post-LT: prior to the abdominal closure).

Mouse warm liver IRI model.

Male 6-8 weeks old wild-type (WT) mice (Jackson Laboratory, Bar Harbor, ME) on a C57BL/6 background were used in a model of partial warm hepatic IRI, as described^[7]. In the treatment groups, animals were infused at 1h prior to the onset of liver ischemia with a single dose of BMK1 inhibitor XMD8-92 (10nmol/mouse i.v., Sigma-Aldrich, St. Louis, MO) dissolved in PBS. Some animals received a single i.v. dose of shRNA or overexpression vectors (−1 or 4 weeks prior to IRI). Mice were sacrificed at various time-points after reperfusion. Liver and serum samples were collected for analysis. sALT and sAST levels were measured by IDEXX Laboratory (Westbrook, ME).

Vector production/administration.

Recombinant adeno-associated viral (rAAV) BMK1, KLF4, shBMK1, shYAP, and shKLF4 control vectors were generated according to described protocols^[12]. Mouse BMK1 and KLF4 cDNA was obtained through high-fidelity PCR amplification (PrimeSTAR, DR010A, Takara). Mouse BMK1 or KLF4 and green fluorescent protein (GFP) cDNAs were subcloned into AAV cis-plasmids containing the liver specific thyroxine-binding globulin (TGB) promoter, the woodchuck hepatitis virus post-transcriptional regulatory element (WPRE), a SV40 late polyadenylation signal and flanking inverted terminal repeats (ITR). shRNA oligos encoding target mouse sequences against BMK1 (5’-TGCTTTCCCTGCACGAGTATA-3’), YAP (5’-GAAGCGCTGAGTTCCGAAATC-3’) and KLF4 (5’-GACATCGCCGGTTTATATTGA-3’) were synthesized by Shanghai Sunbio Tech. and then cloned into AAV vector plasmids containing TGB, GFP, WPRE and a U6 promoter. Vectors carrying BMK1, KLF4, shBMK1, shYAP and shKLF4 without GFP were also constructed. All constructed vectors were then packaged into rAAV serotype 8 (rAAV8) using the triple transfection method, purified by cesium chloride centrifugation and tittered by quantitative PCR. Overexpression vectors of rAAV8-BMK1 (BMK1 OE) and rAAV8-KLF4 (KLF4 OE) were diluted in PBS and intravenously injected (5×10¹² genome copies/kg) to mice one week before IR surgery, while interference RNA vectors of rAAV8-shBMK1 (siBMK1), rAAV8-shYAP (siYAP) and rAAV8-shKLF4 (siKLF4) were intravenously injected (5×10¹² genome copies/kg) four weeks before operation. Knockdown and overexpression efficiency was validated by real-time PCR and Western blotting.

Histopathology.

Liver specimens (5μm) were stained with hematoxylin and eosin (H&E). Primary monoclonal antibodies (mAbs) against mouse neutrophils (Ly-6G) (1A8; BD Biosciences, San Jose, CA), macrophages (CD68) (FA-11; AbD Serotec, Raleigh, NC), Phospho-BMK1 (Thr218/Tyr220) and BMK1 (D23E9; Cell Signaling Technology, Danvers, MA) were used. Liver sections were evaluated blindly by counting labeled cells in 10 high-power fields (HPF).

Quantitative RT-PCR.

Quantitative PCR was performed with platinum SYBR green quantitative PCR kit (Invitrogen, Carlsbad, CA) by the QuantStudio3 (Applied Biosystems, Foster City, CA). Target genes expression was calculated by their ratios to the housekeeping gene hypoxanthine-guanine phosphoribosyl transferase (HPRT).

Western Blotting.

Western blotting was performed with liver/hepatocyte proteins (30μg/sample) and rabbit anti-mouse YAP (D8H1X), Phospho-YAP (Ser127, D9W21), Phospho-MST1 (Thr183, E7U1D), Phospho-LATS1 (Thr1079, D57D3), BMK1 (D23E9), Phospho-BMK1 (Thr218/Tyr220), ITCH (itchy E3 ubiquitin protein ligase, D8Q6D), GAPDH (D16H11), Lamin B1 (D4Q4Z) and β-actin (13E5, Cell Signaling Technology) and KLF4 mAbs (ABS1514; Millipore Sigma, St Louis, MO).

Terminal Deoxynucleotidyl Transferase-Mediated dUTP Nick End Labeling (TUNEL) Assay.

DNA fragments in liver sections, resulting from apoptosis, were detected by the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method (Klenow-FragEL DNA Fragmentation Detection Kit; Calbiochem, La Jolla, CA). TUNEL-positive cells were counted in 10 HPF/section under light microscopy.

Cell cultures.

Primary mouse hepatocytes were isolated by in situ two-stage collagenase perfusion from control or rAAV8 vectors pre-treated C57BL/6 mice and cultured with complete William’s E medium plus maintenance supplement (Life Technologies, Carlsbad, CA) for 24h. For hypoxia-reoxygenation experiment, culture medium equilibrated in 94%N₂-5%CO₂-1%O₂ or 95%O₂-5%CO₂ was used as hypoxia or reoxygenation medium, respectively. Cells were cultured with hypoxia medium at 37°C for 6h in a humidity airtight chamber equilibrated with 94%N₂-5%CO₂-1%O₂. The medium was then replaced with reoxygenation medium, and cells were further incubated at 37°C in a 95%O₂-5%CO₂ atmosphere for 8h. Cells were collected and processed for immunofluorescence staining and protein extraction, whereas supernatants were assessed for ALT. JC-1 assay (MitoProbe JC-1 assay; ThermoFisher) was applied for the mitochondria status.

ALT Release Assay.

Culture medium ALT activity was measured by ALT kit (Infinity, Thermo Fisher Scientific Inc, Canoga Park, CA) according to the manufacturer’s instructions.

Immunofluorescence.

Cultured primary hepatocytes were stained by immunofluorescence. Primary YAP (D8H1X), BMK1 (D23E9), Phospho-BMK1 (Thr218/Tyr220) or KLF4 (ABS1514; Millipore Sigma) mAbs were used and followed by donkey anti-rabbit Alexa Fluor 594 (Invitrogen) or goat anti-rabbit conjugated FITC (11-4839-81; Thermo Fisher Scientific Inc) as secondary Ab. For dual-color immunofluorescence, YAP mAb conjugating Alexa Fluor 488 (Cell Signaling Technology) was used. Cells were then treated with DAPI (Vector Labs) and evaluated blindly by counting labelled cells in 10 HPF. Superoxide levels in hepatic tissue were detected with ROS-sensing dye dihydroethidium (DHE; AnaSpec, Fremont, CA).

Statistical analysis.

Data are presented as mean±standard deviation. Normality was assessed for all variables; normally distributed data were compared using paired or unpaired Student’s t-tests, while non-normally distributed data were analyzed using the Mann-Whitney U test or Wilcoxon signed-rank test. For comparisons involving more than two groups, one-way ANOVA followed by Tukey’s post-hoc test was performed. Linear relationships were evaluated via Pearson correlation analysis, with robustness verified using Bootstrap and Benjamini–Hochberg adjustments. Statistical analyses were conducted using SPSS (version 25) and GraphPad Prism 8. Values of P<0.05 were considered to indicate statistical significance: ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05. All in vivo experiments were conducted with at least three independent individuals.

Results

Post-LT BMK1 expression negatively correlates with the hepatocellular injury in human LT

We evaluated the relationship between intra-graft BMK1 expression and hepatocellular injury in a cohort of 60 liver transplant recipients. Patients’ demographic and clinical data are shown in Table S1. Pre-transplantation (pre-LT) hepatic biopsies were collected during back-table preparation (3–10h cold storage), and post-LT biopsies were obtained approximately 3h after reperfusion (Fig. 1a). While Western blot analysis showed that BMK1 expression levels varied among recipients (Fig. 1a), the Post/Pre-LT BMK1 ratio correlated negatively with both serum ALT (sALT; Fig. 1b, R²=0.1074, p=0.0106) and serum AST (sAST; Fig. S1a) levels on POD 1. No other clinical parameters significantly correlated with the BMK1 ratio (Fig. S1d). Stratified analysis indicated that intraoperative bleeding volume and donor BMI (body mass index) influenced the slope of this relationship (Fig. S1b,1c). Meanwhile, the correlation between POD1 sALT and the BMK1 ratio lacked consistency across robustness tests, suggesting a relatively weak association (Fig. S1e,1f). Furthermore, the Post/Pre-LT BMK1 ratio was not associated with 30- or 90-day complications as graded by the Clavien-Dindo classification. To further evaluate the impact of post-LT BMK1 levels on clinical outcomes, sixty LT recipients were divided into BMK1 “decrease” (n=40) and BMK1 “increase” (n=20) group, using Post-/Pre-LT BMK1/β-actin ratio=0.99 as a threshold (Fig.1c). Interestingly, the “increased” BMK1 recipient group exhibited lower sALT levels at POD1 (Fig.1d, p<0.01) and a more rapid sALT recovery between POD1 to POD7 (Fig.1e) as compared with the “decreased” BMK1 expression group. This indicates that intra-graft BMK1 may facilitate the hepatic resistance to IR stress and mitigate the hepatocellular damage in LT recipients.

Fig.1.

Peri-operative hepatic BMK1 expression correlates with hepatocellular damage in human LT recipients (n=60). (a) Hepatic biopsies after 3-10h of cold storage were collected at 3h after reperfusion. Western blot-assisted pre- and post-LT BMK1 expression profiles in LT biopsies (representative 4 cases). (b) The ratio of Post/Pre-LT BMK1/β-actin correlated negatively with sALT at POD1. Pearson’s correlation coefficient (r) was used to measure the strength of the association between sALT and Post-/Pre-LT BMK1/β-actin Ratio. (c) Sixty LT recipients were divided into post/pre-LT BMK1/β-actin “decrease” (n=40) and “increase” (n=20) groups (ratio=0.99 as threshold). (d) sALT levels in LT groups at POD1. (e) sALT decreasing curve in LT groups (POD1-7). (**p<0.01, ***p<0.001).

The dynamic local BMK1 expression profile in murine liver IRI

We determined whether IR triggered the expression of endogenous BMK1 in mouse livers subjected to 90 min of warm ischemia, followed by reperfusion. After ischemia alone, BMK1 mRNA levels increased rapidly, and then declined during the early reperfusion stage, increasing later and reaching another peak at 6-12h after reperfusion, and then progressively dropping to baseline by 24h post reperfusion (Fig. 2a). Consistent with the BMK1 activation profile at the transcriptional levels, the immunofluorescence of BMK1 after 6h reperfusion indicated upregulation of intracellular BMK1 in the hepatic tissue (Fig. 2b). Meanwhile, its downstream genes, including MEF2, KLF4, KLF2 and PAK1, were all upregulated (Fig. S4a).

Fig. 2.

The BMK1 expression in murine liver IRI. (a) The hepatic BMK1 expression after sham operation or at 90min of partial warm ischemia, followed by various lengths of reperfusion. (b) Immunofluorescence staining of BMK1, f-actin and DAPI (nuclei) in the liver after sham or IRI (representative staining; magnification: x400). (c) Liver, kidney, spleen and lung tissues from mice pre-treated with rAAV8-shBMK1 vectors (−4 weeks) stained with DAPI (blue) and f-actin (red). Fluorescence screen revealed hepatocyte expression of siRNA BMK1. (d) No differences in HE staining of liver tissue from mice pre-treated with rAAV8-shBMK1 (siBMK1) and control (Control) vectors. (e-g) Mice pre-treated with rAAV8-shBMK1 (siBMK1), XMD892 (XMD) or PBS were subjected to 90min of ischemia or sham procedure. Six hours after reperfusion, liver samples were collected for analysis. (e) Western blotting detection of intrahepatic phosphorylated BMK1 (p-BMK1), KLF4 and β-actin after IRI. (f) Serum ALT levels and (g) histology (representative H&E staining; magnification: x100 and x200) after IRI. (***p<0.001; n=10-12/group).

Inhibiting intracellular BMK1 function exacerbates mouse liver IRI

To knockdown the expression of liver rBMK1, AAV serotype 8 vectors, including cloned BMK1 shRNA oligos, were generated and packaged into the virus. The rAAV8-shBMK1-GFP vectors were under the control of the liver-specific thyroxine-binding globulin gene promoter and constitutively expressed GFP. Biodistribution studies demonstrated that, after 4 weeks of treatment, rAAV8-shBMK1-GFP vectors were specifically expressed in the liver (Fig. 2c) and efficiently downregulated the hepatocellular BMK1 protein level (Fig. 2e, Fig. S2). No histological differences were observed between knockdown and control individuals (Fig. 2d). To address whether BMK1 was essential in maintaining liver homeostasis during IRI, C57BL/6 mice that received rAAV8-shBMK1-GFP (siBMK1) or BMK1-specific inhibitor XMD8-92 (XMD) treatment were put through 90min of ischemia and different periods of reperfusion. Accompanied with BMK1 inhibition, liver IRI was exacerbated, as indicated by significantly elevated serum ALT level (Fig. 2f: 8465±1228 [control] versus 20478±2565 [XMD], 18586±2969U/L [siBMK1]; p=0.0006 and 0.0005 respectively), serum AST and damaged liver structure (Fig. 2g, Fig. S3). More severe lobular edema, widespread hemorrhage and congestion/hepatocellular necrosis was observed in BMK1-inhibited livers, when compared to WT controls. This indicated that inhibiting the hepatocellular BMK1 function exacerbated the hepatic IRI.

Inhibiting BMK1 function increases cell death and inflammation response

We assessed how BMK1 inhibition affected hepatocyte apoptosis and oxidation stress after liver IRI. TUNEL test revealed higher frequency of positive cells, indicating enhanced hepatocyte apoptosis in siBMK1 and XMD treated groups (Fig. 3a and Fig. S4b). Meanwhile, stronger hepatocellular oxidative stress after liver IRI was detected in BMK1 inhibited groups, as evidenced by more intense DHE (Dihydroethidium) hepatocyte staining (Fig. 3b). Associated with increased cell death, enhanced macrophage (CD68⁺) and neutrophil (Ly6G⁺) infiltration was also observed in IR stressed livers (Fig. 3c and Fig. S4b). Meanwhile, the monocyte-derived proinflammatory cytokines (TNF-α, IL-1β and IL-6) were markedly elevated after inhibiting intracellular BMK1 function, while IL-10 level was suppressed (Fig. 3d).

Fig. 3.

Inhibiting BMK1 promoted cell death, ROS accumulation and inflammatory response after liver IRI (90min ischemia and 6h reperfusion). (a) TUNEL-assisted detection of hepatic necrosis/apoptosis (dark stained nuclei) in ischemic liver lobes (magnification: x200). (b) ROS-sensing dye dihydroethidium (DHE) staining (strong red fluorescence indicating massive ROS production in BMK1-inhibited groups). (c) CD68⁺ macrophages and Ly6G⁺ neutrophils in IR-stressed liver lobes (white arrow indicate positive staining, magnification: x200). (d) qRT-PCR assisted detection of proinflammatory cytokines TNF-α, IL-6, IL-1β and IL-10 (data normalized to HPRT gene expression). (***p<0.001; n=4-6/group).

KLF4 dependence of BMK1 protection against liver IRI

Since previous studies revealed that activation of BMK1 could promote the function of KLF4, we would like to figure out whether the hepatic protection of BMK1 during IRI was KLF4 dependent. We constructed the rAAV8-KLF4-GFP vector to achieve constitutive high-level liver KLF4 expression. Fluorescence screen indicated overexpression of KLF4 in livers as early as 1 week after rAAV8-KLF4-GFP vector injection (Fig. 4a). Mice that were pre-treated with rAAV8-shBMK1, rAAV8-KLF4-GFP and empty vector were then subjected to liver IR surgery (90min ischemia and 6h reperfusion). Overexpression of KLF4 significantly ameliorated IR injury, as evidenced by decreased sALT level (Fig. 2b, 7922±903 [control], 19455±2899 [BMK1 KD] vs 2134±452 [KLF4 OE], p=0.0006 and 0.0008, respectively) and preservation of liver histology (Fig. 2d). Furthermore, overexpressing BMK1 could achieve comparable liver protection against IR injury (Fig. 2b,2d). Then, we explored whether the protective effect of BMK1 was KLF4-dependent. Mice were pre-treated with interference RNA vectors of rAAV8-shBMK1 or rAAV8-shKLF4 to knockdown BMK1 or KLF4 expression at 4 weeks before IR surgery. Then, mice were administrated with rAAV8-KLF4 or rAAV8-BMK1 vectors to overexpress KLF4 or BMK1, respectively at 1 week before surgery. Western blotting had proven the efficiency of knockdown vs overexpression, and also indicated that disruption of BMK1 signaling could silence KLF4 (Fig. 4c). Further, overexpression of KLF4 ameliorated liver IRI in BMK1 KD recipients, while KLF4 knockdown restored IRI in BMK1 OE mouse recipients (Fig. 4b,4d). Besides, inhibiting KLF4 also exacerbated intrahepatic expression of inflammatory genes, while overactivation of KLF4 depressed IR-triggered inflammatory response, even in BMK1 KD mice. This indicated that the BMK1-mediated hepatocellular protection against IR injury was KLF4-dependent.

Fig. 4.

BMK1 cytoprotective effect in liver IRI is KLF4-dependent. Mice were pre-treated with rAAV8-shBMK1 or rAAV8-shKLF4 vectors (−4 weeks) to establish BMK1 KD or KLF4 KD phenotype groups. To overexpress BMK1 or KLF4, groups of mice were pretreated (−1 week) with rAAV8-BMK1-GFP or rAAV8-KLF4-GFP vectors. Both pre-treated or control mice were subjected to liver IRI procedure (90min ischemia and 6h reperfusion) followed by liver tissue analyses (+6h). (a) Liver KLF4 overexpression by fluorescent screening. (b) Serum ALT levels. (c) Western blot-assisted detection of intrahepatic phosphorylated BMK1 (p-BMK1), KLF4 and β-actin. (d) Liver tissue histology (representative H&E staining; magnification: x100 and x200) after IRI. (e) qRT-PCR-assisted detection of TNF-α, IL-6 and IL-1β (Data normalized to HPRT gene expression). (***p<0.001; n=4-6/group).

BMK1 inhibition suppresses YAP function

We have previously shown that YAP prevented liver IRI by diminishing cell death and oxidative stress^[7]. Here, we assessed whether YAP function was regulated by BMK1. Inhibiting BMK1 by siRNA or chemical inhibitor dramatically decreased intrahepatic YAP mRNA and protein expression (Fig. 5a, c). Interestingly, blocking BMK1 could also increase the phosphorylation level of YAP, which indicated that BMK1-mediated regulation on YAP might happen at both transcriptional and post-translational levels. Meanwhile, the downstream YAP genes, like CYR61, CTGF and MnSOD, were also downregulated after BMK1 blockage, while the upstream YAP genes, like NF2, MST1 and LATS1 were not impaired (Fig. 5d, Fig. S5). As mentioned above, inhibiting BMK1 function resulted in a more severe hepatic IRI, while knockdown of hepatic YAP, using rAAV8-shYAP, caused similar hepatocellular injury (Fig. 5b). Thus, BMK1 might regulate the YAP function in the mechanism of hepatic IRI.

Fig. 5.

Inhibiting BMK1 suppressed YAP function after liver IRI. (a) Decreased hepatic YAP expression after inhibiting BMK1 (siRNA or chemical inhibitor). (b) Serum ALT levels. (c) Western blot-assisted detection of YAP, phosphorylated YAP (p-YAP) and β-actin. (d) Expression of YAP downstream genes (CYR61, CTGF and MnSOD). (N.S. represented no significance; ***p<0.001; n=4-6/group)

BMK1 inhibition facilitates YAP exclusion from hepatocyte nuclei

As BMK1-mediated regulation of YAP might happen at both transcriptional and post-translational levels, we then applied in vitro hepatocyte hypoxia-reoxygenation model to explore the underlying regulatory mechanisms. Primary hepatocytes and endothelial cells were extracted from WT (Ctrl) and rAAV8-shBMK1 (siBMK1) pre-treated mice and stressed with hypoxia-reoxygenation insult in vitro. As positive control, one group of hepatocytes extracted from WT mice were also pre-treated with XMD at 1h before hypoxia-reoxygenation insult. BMK1 and phosphorylated BMK1 in hepatocytes was detectable and readily provoked after insult (Fig. 6a). Inhibiting BMK1 activity made hepatocytes more susceptibility to hypoxia-reoxygenation stress, as indicated by elevated supernatant ALT levels (Fig. 6b: 6.2±1.1 versus 13.3±2, 12.4±1.9U/L; p=0.0004). Although hypoxia-reoxygenation insult induced YAP expression and promoted nuclear translocation, BMK1 inhibition significantly diminished the YAP abundance in stressed hepatocytes (Fig. 6c). This confirmed that the hepatic BMK1-mediated regulation of YAP occurred at the post-translational level.

Fig. 6.

BMK1 inhibition abolished YAP function in hepatocytes. Primary hepatocytes were isolated from mice pre-treated with rAAV8-shBMK1 (siRNA) or empty vectors (Ctrl) and cultured in vitro for 24h. The cells of XMD group were pre-treated with XMD8-92 at 1h prior to treatment. After 6h of hypoxia and 8h of reoxygenation, the cell cultures were analyzed. (a) Immunofluorescence staining of BMK1, phosphorylated BMK1 (p-BMK1), f-actin and DAPI (nuclei) in hepatocytes. (b) Supernatant ALT. (c) Immunofluorescence staining of YAP, f-actin and DAPI (nuclei) (magnification: x200). (***p<0.001; n=4-6/group).

BMK1 regulates YAP function by mediating LATS1 ubiquitination

We further explored the underlying mechanism of how BMK1 regulated YAP function during hepatic IRI injury by using in vitro experiments. Mice were pre-treated with rAAV8-BMK1 or rAAV8-shYAP to overexpress BMK1 (BMK1 OE) or knockdown YAP (YAP KD). Primary hepatocytes from these mice were then extracted and subjected to the hypoxia-reoxygenation insult. Consistent with in vivo data, BMK1 OE hepatocytes were more resistant to hypoxia-reoxygenation stress with lower supernatant ALT levels and well preserved mitochondrial function (Fig. 7a,7c). Besides, BMK1 OE also promoted the expression of YAP and its downstream genes (Fig. 7d). However, BMK1-mediated cytoprotection could be diminished by YAP KD, as exhibited by increased supernatant ALT levels and mitochondrial dysfunction (Fig. 7a,7c). Even though BMK1 cascade had no effect on the mRNA MST1 and LATS1 levels (Fig. S5), BMK1 overactivation in hepatocytes could significantly downregulate the phosphorylated LATS1 protein level, but not phosphorylated MST1 (Fig. 7b), which indicated a post-transcriptional regulation of BMK1 on LATS1. ITCH is a HECT class E3 ubiquitin ligase and a negative regulator of LATS1 through ubiquitination^[13]. MAPK pathway had been identified as an important regulator of ITCH^[14,15]. Western blotting indicated that cytoplasmic ITCH level was significantly upregulated by BMK1 overactivation, as compared with controls. Meanwhile, inhibition of YAP had no effect on ITCH function (Fig. 7b). Therefore, BMK1 may regulate hepatocyte YAP function using a mechanism by mediating LATS1’s ubiquitination through ITCH.

Fig. 7.

BMK1 regulated YAP function by ubiquitinating LATS1. Primary hepatocytes isolated from mice pre-treated with rAAV8-BMK1 (BMK1 OE), rAAV8-shYAP (YAP KD) or empty vectors (Ctrl) were subjected to hypoxia-reoxygenation insult. Supernatant and cells were collected for analysis. (a) Supernatant ALT levels. (b) Western blot-assisted detection of BMK1, KLF4, ITCH, phosphorylated MST1 (p-MST1), phosphorylated LATS1 (p-LATS1), YAP, phosphorylated YAP (p-YAP) and β-actin. (c) JC-1 staining (green fluorescence indicating mitochondria depolarization, red fluorescence indicating intact mitochondria membrane) (magnification x200). (d) Expression of YAP, CYR61, CTGF and MnSOD in hepatocytes after hypoxia-reoxygenation insult. (*p<0.001; n=4-6/group).

Inhibiting BMK1 function impairs the transcriptional cooperation between YAP and KLF4

KLF4 is a transcriptional factor that is directly regulated by BMK1 ^[4]. Hepatocytes from rAAV8-shKLF4 (siKLF4) pre-treated mice were also vulnerable to hypoxia-reoxygenation insult (Fig. 8a). Since both YAP and KLF4 were regulated by BMK1, we then studied their cooperation. Dual immunofluorescence staining of YAP and KLF4 revealed the overlapping expression pattern in hypoxia-reoxygenation stressed hepatocytes (Fig. 8c). Notably, silencing either YAP or KFL4 excluded the other factor from the hepatocyte nuclei and impaired the downstream gene expression of YAP and KLF4 (Fig. 8b, c and Fig. S6), which eventually increased hypoxia/reoxygenation-mediated hepatocytes necrosis (Fig. 8a). Similarly, blocking hepatocellular BMK1 function led to exclusion of YAP and KLF4 from the nuclei. These data indicate that YAP and KLF4 cooperated under BMK1 regulation to protect liver against IRI.

Fig. 8.

Inhibiting BMK1 cascade diminished the cooperation between YAP and KLF4 in hepatocyte nuclei. Primary hepatocytes from mice pre-treated with rAAV8-shYAP (siYAP), rAAV8-shBMK1 (siBMK1), rAAV8-shKLF4 (siKLF4) or empty vectors (Ctrl) were subjected to the hypoxia-reoxygenation insult. (a) Supernatant ALT levels. (b) Western blot-assisted detection of nuclear KLF4, YAP and Lamin B1, and cytoplasmic KLF4, YAP and GAPDH. (c) Immunofluorescence staining of KLF4 (red), YAP (green) and DAPI (blue, nuclei) in hepatocytes after hypoxia-reoxygenation insult. (N.S. represented no significance; ***p<0.001; n=4-6/group)

Discussion

The present study unraveled the previously unrecognized modulatory role and mechanism of BMK1 in combating IR-hepatocellular damage and maintaining homeostasis in IR-stressed human and murine livers.

One of the major consequences of IRI is the delayed recovery of allograft function, which increases the overall mortality after transplantation^[16]. Moreover, the severe shortage of donor organs promotes the usage of marginal liver, which is even more susceptible to IRI^[17]. One approach to reduce the incidence of allograft dysfunction related mortality is to identify livers tending to suffer more severe damage after transplantation, which could provide more individualized management strategy. In our study, we enrolled sixty adult OLT recipients and analyzed the relationship between intra-graft BMK1 expression and the recovery trajectory of liver function after transplantation. Our result revealed that the upregulation of hepatic BMK1 may ameliorate graft IRI, which facilitated a more rapid function recovery after transplantation. What’s more, increase of hepatic BMK1 level before IR treatment could also protect hepatocytes from IRI in murine model. This inspires us that hepatic BMK1 could not only serve as a predictive marker for liver recovery, but also could be a potential therapeutic target to reduce IRI related dysfunction.

High BMK1 expression levels were found in cirrhotic liver, hepatocellular carcinoma cells, and other digestive duct cancer tissues^[2,18,19]. Different from the wide expression in cardiovascular endothelial cells, our study revealed that liver BMK1 was mainly expressed in hepatocytes. After IR insult, BMK1 and its downstream genes were immediately triggered in stressed hepatocytes. Of note, silencing BMK1 function exacerbated, while overexpression of BMK1 significantly attenuated IR injury. In this study, the BMK1 chemical inhibitor XMD8-92, which blocks hepatic BMK1 phosphorylation, exacerbated hepatic IRI in a mouse model. Considering the potential off-target effects of XMD8-92, we additionally employed a more specific genetic knockdown model to further validate the protective role of BMK1 in hepatocellular IRI. Others also found that BMK1 maintained endothelial cell homeostasis through anti-oxidant and anti-apoptotic responses^[4,20]. In our study, inhibiting BMK1 exacerbated IR-mediated liver damage, increased hepatocytes death, promoted intracellular oxidation stress and aggravated the inflammatory response, confirming the essential role of BMK1 in maintaining hepatocyte homeostasis following IR insult^[21,22].

We have proposed that BMK1 mediates anti-oxidant response through KLF2, MEF2, and Nrf2 ^[23,24]. BMK1-activated Nrf2 upregulates the expression of anti-oxidant genes, defends overproduced ROS and rescues mitochondrial function^[25]. Here, we found a novel anti-oxidation regulation mechanism of BMK1 through YAP. Our previous studies had revealed that YAP-mediated anti-oxidative genes, such as MnSOD, catalase and GSHPxs, were essential in maintaining hepatocyte homeostasis after IR stress. Herein, we found that inhibiting hepatocyte BMK1 suppressed the expression and transcriptional activity of YAP. Both in vivo and in vitro experiments suggested that inhibiting BMK1 facilitated the phosphorylation and cytoplasmic translocation of YAP. Thus, the decreased expression of anti-oxidative genes exacerbating hepatocellular oxidative stress/cell death after IR, has revealed a novel regulation of hepatocyte homeostasis through BMK1-YAP signaling pathway.

Our present study has also revealed that the BMK1-mediated regulation of YAP occurred at the post-translational level. Ubiquitin-dependent proteolysis controls turnover/abundance of transcriptional factors and other regulatory proteins, including MAPK pathway^[14]. Protein ubiquitylation depends on a three-component enzyme system, namely, ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzymes (E2), and ubiquitin ligases (E3). ITCH is a HECT (Homologous to E6-AP C Terminus) domain E3 ligase that serves as a catalytic intermediate in the transfer of ubiquitin chains from the E2 to the substrate^[26,27]. Since the catalytic activity of ITCH is directly stimulated by phosphorylation, previous studies had shown that JNK and ERK1/2 could activate ITCH after extracellular stimulation and facilitate protein turnover^[14,15]. Here, we found that overactivation of hepatocellular BMK1 was associated with increased intracellular ITCH levels following hypoxia–reoxygenation, accompanied by a reduction in phosphorylated LATS1, an upstream regulator of YAP. Given that ITCH has been reported to function as an E3 ubiquitin ligase for LATS1^[13], its upregulation may facilitate the ubiquitination and degradation of phosphorylated LATS1, thereby relieving its inhibitory effect on YAP. Consequently, increased levels of dephosphorylated YAP translocate to the nucleus, conferring enhanced resistance to hypoxia–reoxygenation injury. Although further studies are required to elucidate the precise mechanisms by which BMK1 regulates ITCH, our findings suggest that BMK1 may modulate YAP activity through a post-transcriptional regulatory pathway.

KLF4 has been recently confirmed as a BMK1 downstream mediator, which elicits vasoprotective responses in endothelial cells^[4]. A genome-wide scale screen revealed co-localization of YAP and KLF4 and indicated that YAP and KLF4 worked as partner transcription factors to promote goblet cell differentiation in the intestine^[28]. Our results also confirmed YAP and KLF4 overlapped in hepatocyte nuclei. The hypoxia-reoxygenation insult induced YAP and KLF4 accumulation in the nuclei and upregulated their downstream genes. Interestingly, either silencing YAP or KLF4 eliminated the other from the hepatocyte nuclei and suppressed the expression of both factors’ downstream genes. This indicates a cooperative transcriptional regulation of YAP and KLF4. Apart from TEAD, a number of transcriptional factors have been shown to collaborate with YAP, such as FoxO1, AP-1, SMAD and p73^[6,10,28]. Here, we revealed a novel cooperation pathway between YAP and KLF4 during liver IRI in maintaining hepatocyte integrity/homeostasis and document that both were regulated by BMK1 in the stressed liver.

In summary, this is the first study to document the cytoprotective function and underlying mechanism of BMK1 signaling in liver IRI. BMK1 promoted cooperation between YAP and KLF4 in the hepatocyte nuclei and upregulate YAP/KLF4-mediated anti-oxidative, anti-apoptotic and anti-inflammatory genes. Our findings provide new insights into the pathogenesis in and identify potential molecular targets for therapeutic intervention against sterile inflammatory hepatocellular damage in liver transplantation.

List of Abbreviations:

ALT

alanine aminotransferase

BMK1

big mitogen-activated kinase-1

DHE

dihydroethidium

DMSO

Dimethyl sulfoxide

ERK5

extracellular signal-regulated kinase 5

HCC

hepatocellular carcinoma cells

HECT

Homologous to E6-AP C Terminus

H&E

Hematoxylin and eosin

HP

hepatocyte

HPRT

hypoxanthine-guanine phosphoribosyl transferase

IRI

ischemia-reperfusion injury

ITCH

itchy E3 ubiquitin protein ligase

Ken

kenpaullone

KLF4

Krüppel-like factor 4

LATS1

large tumor suppressor homolog 1

LPA

1-oleoyl lysophosphatidic acid

LSEC

liver sinusoidal endothelial cells

MAPK

mitogen-activated protein kinases

mRNA

messenger RNA

MEK5

mitogen-activated protein kinase kinase 5

MST1

mammalian sterile 20-like protein kinase 1

ROS

reactive oxygen species

TEAD

TEA domain family

TNF

tumor necrosis factor

TUNEL

terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling

VP

verteporfin

WT

wild type

YAP

yes-associated protein