Phb2-src interaction inhibitor and application thereof in preparation of anti-hepatoma drugs

Abstract

The application relates to the field of biological medicines, and discloses a PHB2-Src interaction inhibitor and application thereof in preparation of an anti-liver cancer drug. A new mechanism that Src-mediated PHB2 phosphorylation is taken as an oxidation-reduction sensitive molecular switch to drive liver cancer metabolic reprogramming is first discovered in the application, and based on the new discovery, the application provides a PHB2-Y34F / Y77F mutant and 10 PHB2-Src interaction inhibitors which have the effects of preventing or treating mitochondrial cristae structure damage, oxidation-reduction homeostasis disorder and metabolic reprogramming in liver cancer, and are expected to be used as anti-liver cancer drugs.

Description

Technical Field

[0001] This invention relates to the field of biomedicine, and more particularly to a PHB2-Src interaction inhibitor and its application in the preparation of anti-liver cancer drugs. Background Technology

[0002] Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths worldwide, with over 900,000 new diagnoses annually. Despite some progress in systemic therapy, such as the reported combination of atezolizumab and bevacizumab, the median overall survival for advanced HCC remains disappointingly low at less than 20 months. This limited therapeutic effect is primarily attributed to metabolic plasticity, particularly the shift from oxidative phosphorylation (OXPHOS) to glycolysis, known as the Warburg effect. This metabolic reprogramming fundamentally disrupts cellular redox homeostasis. Reduced NAD50 in HCC cells... + The NADH / S / N ratio reflects impaired electron transport and paradoxically enhances ROS production through reverse electron transport and dysfunction of complex I, leading to reducing stress. Src family kinases (SFKs), frequently activated in HCC, have become key regulators of this metabolic shift. Recent evidence suggests that Src directly phosphorylates glycolytic enzymes, enhancing their activity and driving glycolytic flux, while increased Src expression is associated with decreased expression of the oxidative phosphorylation complex and enhanced transfer potential.

[0003] In addition to metabolic alterations, HCC cells also exhibit profound mitochondrial structural abnormalities that directly affect redox homeostasis. Transmission electron microscopy revealed significantly enlarged tumor mitochondria with a sharply reduced cristae density, which is associated with impaired respiratory complex activity, increased ROS production, and poor clinical prognosis. The inhibitory protein 1 / 2 (PHB1 / 2) complex forms a ring structure on the inner mitochondrial membrane (IMM) and plays a crucial role in maintaining redox homeostasis by organizing respiratory supercomplexes and reducing electron leakage. Recent cryo-electron tomography revealed approximately 43 PHB1 / 2 complexes per cristae, functioning as a membrane scaffold. Furthermore, cardiolipin has been shown to play a key role in maintaining cristae structural stability. This mitochondrial-specific phospholipid, comprising approximately 20% of IMM lipids, preferentially accumulates at cristae junctions and tips, stabilizing respiratory supercomplexes and promoting membrane curvature. The interaction between cardiolipin and mitochondrial proteins is essential for cristae integrity, as disruption inevitably leads to cristae disorder and respiratory dysfunction. Although both the PHB1 / 2 complex and cardiolipin are essential for respiratory function, it remains unknown whether their interaction is necessary for the maintenance of the spine, and how this relationship is disrupted in cancer.

[0004] Downstream of crest structural changes, the OMA1-OPA1 proteolytic axis is a key effector system for crest remodeling. Under stress, the zinc metalloproteinase OMA1 cleaves the activator-like GTPase OPA1 from its long form (L-OPA1) to its short form (S-OPA1), with excessive cleavage leading to crest fragmentation and OXPHOS collapse. In HCC, aberrant OPA1 cleavage is associated with enhanced cell migration and metabolic reprogramming. However, the upstream signals triggering pathological OMA1 activation in hepatocellular carcinoma remain unknown. Summary of the Invention

[0005] This application provides a PHB2-Src interaction inhibitor and its application in the preparation of anti-hepatocellular carcinoma drugs. This application reveals for the first time a novel mechanism by which Src-mediated PHB2 phosphorylation acts as a redox-sensitive molecular switch driving metabolic reprogramming in hepatocellular carcinoma. Based on this, this application screened 10 compounds and the PHB2-Y34F / Y77F mutant, which have the effect of blocking Src kinase-mediated phosphorylation of PHB2 at the Y34 and / or Y77 sites, and are expected to serve as anti-hepatocellular carcinoma drugs.

[0006] The specific technical solution of this invention is as follows: First, this application discloses a PHB2-Src interaction inhibitor comprising at least one of the following compounds or pharmaceutically acceptable salts thereof; the PHB2-Src interaction inhibitor has the effect of blocking phosphorylation of PHB2 at Y34 and / or Y77 sites mediated by Src kinase.

[0007] This application reveals for the first time a novel mechanism by which Src-mediated PHB2 phosphorylation acts as a redox-sensitive molecular switch driving metabolic reprogramming in hepatocellular carcinoma. Specifically, this application is the first to discover that PHB2 (but not PHB1) specifically undergoes Src-mediated phosphorylation of Y34 and Y77 tyrosine residues in hepatocellular carcinoma cells. This phosphorylation introduces a negative charge, disrupting the PHB2-cardiolipin interaction through electrostatic repulsion, triggering the disintegration of the PHB1 / 2 complex, cytoplasmic PHB2 translocation, and OMA1 activation. The resulting crest disturbance drives a metabolic shift from oxidative phosphorylation to glycolysis, promoting liver tumor growth. Therefore, this application proposes an intervention targeting the Src-PHB2-cardiolipin axis as an effective therapeutic strategy for restoring mitochondrial function in hepatocellular carcinoma cells. Finally, this application finds that the aforementioned 10 compounds can block Src kinase-mediated phosphorylation of PHB2 at Y34 and / or Y77 sites, providing new targets and strategies for the treatment of hepatocellular carcinoma.

[0008] Preferably, the PHB2-Src interaction inhibitor is 6-acetamido-1H-benzimidazole (i.e., Compound 1) or a pharmaceutically acceptable salt thereof. 6-acetamido-1H-benzimidazole has the molecular formula C9H9N3O, a molecular weight of 175.19 Da, and is represented by SMILES as CC(=O)Nc1ccc2[nH]cnc2c1.

[0009] Second, this application discloses a PHB2 mutant, named PHB2-Y34F / Y77F, which is obtained by mutating the Y34 and Y77 sites of PHB2 to phenylalanine (inhibiting phosphorylation modification at the Y34 and Y77 sites).

[0010] Third, this application discloses the use of PHB2-Src interaction inhibitors or PHB2 mutants in the preparation of anti-hepatocellular carcinoma drugs. The drugs possess at least one of the following functions: a) restoring the mitochondrial cristae structure in hepatocellular carcinoma cells; b) maintaining redox homeostasis; c) inhibiting metabolic reprogramming.

[0011] This application discovers that PHB2-Src interaction inhibitors or PHB2-Y34F / Y77F mutants can restore mitochondrial cristae structure, maintain redox homeostasis, and inhibit metabolic reprogramming in hepatocellular carcinoma cells through the following steps: 1) blocking Src-mediated PHB2 phosphorylation and maintaining the interaction between PHB2 and cardiolipin; 2) stabilizing the PHB1 / 2 complex and PHB1 / 2 supercomplex and preventing their disintegration; 3) inhibiting the abnormal activation of OMA1 protease and reducing excessive cleavage of OPA1; 4) maintaining the integrity of mitochondrial cristae structure; 5) restoring electron transport chain function and increasing the NAD⁺ / NADH ratio; 6) reversing the metabolic shift from oxidative phosphorylation to glycolysis and inhibiting hepatocellular carcinoma tumor growth.

[0012] Fourth, this application discloses an anti-liver cancer drug, which uses a PHB2-Src interaction inhibitor and / or a PHB2 mutant as its active ingredient.

[0013] Preferably, the anti-liver cancer drug further includes a pharmaceutically acceptable carrier, excipient, or solvent.

[0014] Preferably, the drug is an oral preparation or an injectable preparation. More preferably, the dosage form of the oral preparation is a capsule, tablet, solution, or powder; and the dosage form of the injectable preparation is a vesicle, injection, or powder.

[0015] Fifth, this application discloses a method for inhibiting the proliferation of liver cancer cells in vitro, comprising: blocking phosphorylation of PHB2 at Y34 and / or Y77 sites mediated by Src kinase in cultured liver cancer cells by a PHB2-Src interaction inhibitor, thereby maintaining the interaction between PHB2 and cardiolipin in liver cancer cells; stabilizing the PHB1 / 2 complex and PHB1 / 2 supercomplex to prevent their disintegration; inhibiting the abnormal activation of OMA1 protease to reduce excessive cleavage of OPA1; maintaining the integrity of the mitochondrial cristae structure of liver cancer cells, restoring electron transport chain function, and increasing the NAD⁺ / NADH ratio; reversing the metabolic shift from oxidative phosphorylation to glycolysis; and ultimately inhibiting the proliferation of liver cancer cells in vitro.

[0016] Sixth, this application discloses the application of PHB2-Y34F / Y77F mutants or the Src-PHB2-cardiolipin regulatory axis as targets in the screening or preparation of anti-liver cancer drugs.

[0017] Based on the novel findings on the pathogenesis of hepatocellular carcinoma in this application, PHB2-Y34F / Y77F mutants or the Src-PHB2-cardiolipin regulatory axis can be used as targets during the screening or preparation of anti-hepatocellular carcinoma drugs, thereby obtaining drugs with good preventive or therapeutic effects against hepatocellular carcinoma.

[0018] Compared with existing technologies, the beneficial effects of this invention are as follows: This application reveals for the first time a novel mechanism by which Src-mediated PHB2 phosphorylation acts as a redox-sensitive molecular switch driving metabolic reprogramming in hepatocellular carcinoma. Based on this, this application intervenes in the Src-PHB2-cardiolipin axis, which can serve as an effective therapeutic strategy for restoring mitochondrial function in hepatocellular carcinoma cells. Finally, this application discovered that 10 compounds and the PHB2-Y34F / Y77F mutant can block Src kinase-mediated phosphorylation of PHB2 at the Y34 and / or Y77 sites, providing new targets and strategies for the treatment of hepatocellular carcinoma. Attached Figure Description

[0019] Figure 1 Mitochondrial structural and functional defects in hepatocellular carcinoma cells. (A) TEM analysis of mitochondrial structure in LO2 and HepG2 cell lines. The top image shows representative TEM images, with a magnified inset highlighting cristae (red outlines indicate mitochondria); scale bar: 1 μm. The bottom image shows the mitochondrial area (μm²) of n=100 mitochondria in each cell line. 2 (B) Quantitative analysis of crest density (%) and NAD in LO2 and HepG2 cell lines. + Comparison of NADH ratios. Measurement of cellular NAD. + and NADH levels to calculate NAD + / NADH ratio (n=6).

[0020] Figure 2 Mitochondrial structural and functional defects in hepatocellular carcinoma cells. (C) Mitochondrial oxygen consumption analysis. Top: Representative oxygen flux trajectory, showing respiratory complex activity measured using a substrate-inhibitor titration protocol in LO2 (black line) and HepG2 (red line) cells. G / M+ADP: Glutamate / malate + ADP of complex I; Rotenone: Inhibitor of complex I; Succinate: Substrate of complex II; Antimycin: Inhibitor of complex III; AS+TMPD: Ascorbic acid + TMPD of complex IV; AZD: Azide (inhibitor of complex IV). Bottom: Quantitative oxygen consumption rates of complexes I, II, and IV.

[0021] Figure 3 Src kinase specifically phosphorylates PHB2 at the Y34 and Y77 sites in hepatocellular carcinoma cells. (A, B) Differential phosphorylation of PHB protein in LO2 and HepG2 cell lines. (A) PHB1 phosphorylation analysis. (B) PHB2 phosphorylation analysis. Top panel: Immunoprecipitation (IP) with anti-PHB1 or anti-PHB2 antibodies, followed by detection of phosphorylated proteins (phos-PHB1 / 2) and total PHB1 / 2. IgG was used as a negative control. Bottom panel: Input lysis buffer showing total PHB1 / 2 and β-actin loading control. (C) Identification of kinases involved in PHB2 phosphorylation in HepG2 cells. PHB2-WT expression. Flag HepG2 cells were transfected with siRNAs targeting INSR, SRC, or EGFR, or as a randomized control. Top panel: Anti-PHB2 antibody IP followed by Western blot detection of phos-PHB2, total PHB2, and IgG controls. Bottom panel: Input lysis buffer showing PHB2 and β-Actin (sample loading control). (D) Effect of SRC overexpression on PHB2 phosphorylation in HepG2 cells. PHB2-WT expression... Flag HepG2 cell transduction control (SRC-NC) or SRC-overexpressing lentivirus. IP-Western blot analysis is as described in (B).

[0022] Figure 4 Src kinase specifically phosphorylates PHB2 at the Y34 and Y77 sites in hepatocellular carcinoma cells. (E) Effect of SRC knockdown on PHB2 phosphorylation in HepG2 cells. Expression of PHB2-WT Flag HepG2 cells were transfected with either a disordered control or SRC-specific shRNA. IP-Western blot analysis was performed as follows. Figure 3(B) As described in section (F). Identification of SRC phosphorylation sites on PHB2 in HepG2 cells. HepG2 cells were transfected with various Flag-tagged PHB2 mutants (Y34F, S39A, Y77F, Y34F / S39A, Y34F / Y77F, S39A / Y77F), with or without SRC overexpression. IP-Western blot analysis showed the phosphorylation levels of different PHB2 mutants. (G) KM survival curves of the Src gene in TCGA data, with log-rank test used to evaluate different groups. HR (high expression) represents the hazard ratio of the high expression group relative to the low expression group.

[0023] Figure 5 Bioinformatics prediction and analysis of PHB2 phosphorylation sites. (A) Kinase recognition motif analysis using the Scansite 4.0 algorithm. Recognition scores for Y34 and Y77 phosphorylation sites by INSR, SRC, and EGFR kinases were calculated based on the position-specific score matrix (PSSM). Scores were normalized to the 0-1 range, where >0.75 indicates high confidence and >0.85 indicates very high confidence. (B) Integrated phosphorylation site prediction analysis using multiple algorithms. Five independent algorithms were used to calculate the prediction scores for different kinases (SRC, INSR, EGFR) on PHB2 phosphorylation sites: NetPhos 3.1 (neural network-based), GPS5.0 (group-based prediction system), KinasePhos 2.0 (support vector machine-based), PhosphoSitePlus (knowledge base-based), and MustiteDeep (deep learning-based). Scores were normalized to the 0-1 range using the min-max normalization method. (C) Evolutionary conservation of PHB2 phosphorylation sites was analyzed using the ConSurf algorithm. Multiple sequence alignments were performed on 100 PHB2 homologous sequences from vertebrate species using the MUSCLE algorithm and maximum likelihood phylogenetic methods, and conservation scores (1-9, 9 = completely conserved) were calculated. (D) Conservation analysis of positively charged and polar amino acids at the N-terminus of PHB2. (E) The functional impact of PHB2 mutations was predicted using the SIFT and PolyPhen-2 algorithms. SIFT scores ranged from 0 to 1, with scores <0.05 indicating harmful mutations. PolyPhen-2 scores ranged from 0 to 1, with scores >0.85 indicating potentially harmful mutations.

[0024] Figure 6Src-mediated phosphorylation disrupts the stability of the PHB1 / 2 complex and triggers cytoplasmic mislocalization of PHB2. (A) Electrostatic surface potential analysis of PHB2 shows the effect of phosphorylation. Left panel: Electrostatic distribution of the protein surface of unphosphorylated PHB2 (Y34 and Y77 without phosphorylation modification). Right panel: Pphosphorylated PHB2 (phos-Y34 and phos-Y77) showing altered electrostatic distribution of the protein surface. Blue indicates positive charge (+5), white indicates neutral (0), and red indicates negative charge (-5). Double-framed lines represent the inner mitochondrial membrane. (B) Effect of Src-mediated phosphorylation on the formation of the PHB1 / 2 complex. PHB2-WT and phosphorylation-resistant mutants PHB2-Y34F / Y77F in HepG2 cells. Flag Blue native PAGE (BN-PAGE) analysis with or without SRC overexpression. The bottom panel shows SDS-PAGE validation of SRC, PHB1, PHB2, and β-Actin expression. (C) Analysis of PHB1 / 2 complex formation in LO2 and HepG2 cell lines. Top panel: Representative BN-PAGE showing PHB1 / 2-SC (supercomplex) and PHB1 / 2 complex formation as detected by PHB2 and AFG3L2 antibodies. Bottom panel: SDS-PAGE Western blot analysis of PHB2, AFG3L2, and β-Actin expression levels (sample control).

[0025] Figure 7 Src-mediated phosphorylation disrupts the stability of the PHB1 / 2 complex and triggers mislocalization of PHB2 cytoplasm. (D) Effect of OXPHOS inhibitors on PHB1 / 2 complex formation. Top: Representative BN-PAGE analysis of PHB1 / 2-SC and the PHB1 / 2 complex under control and treatment conditions (H2O2, rotenone, antimycin, oligomycin (OLG), and combined antimycin / oligomycin (OA)). Bottom: Representative Western blot showing protein expression levels of PHB2, AFG3L2, TIM23, and β-Actin (sample control). (E) Subcellular distribution of PHB2 under tumor microenvironment stress. Western blot analysis of PHB2 in mitochondria and cytoplasmic components of HepG2 cells treated with H2O2 (oxidative stress), CoCl2 (hypoxia mimic), rotenone (complex I inhibitor), or antimycin (complex III inhibitor).

[0026] Figure 8Src-mediated phosphorylation disrupts the stability of the PHB1 / 2 complex and triggers cytoplasmic mislocalization of PHB2. (F) Dose-dependent redistribution of PHB2 in response to increasing SRC expression levels. Western blot analysis showed subcellular localization of PHB2 in mitochondria and cytoplasmic components after SRC overexpression in HepG2 cells. HSP60 was used as a mitochondrial marker, and β-Actin as a cytoplasmic marker.

[0027] Figure 9 The PHB1 / 2 complex specifically recognizes cardiolipin via conserved positively charged residues. (A, B) Structural analysis of the PHB-cardiolipin interaction. Left panel: PHB1 structure showing key positively charged residues (K4, R41, K63) and polar residues (A) involved in cardiolipin binding. PHB2 structure highlighting key positively charged residues (K6, R17, R54, R71) and polar residues (B) involved in cardiolipin binding. Right panel: Conservation fraction analysis showing the distribution of basic, polar, acidic, and nonpolar amino acids relative to their distances from the cardiolipin head region (HR) on the inner and outer lobes of the inner mitochondrial membrane (IMM).

[0028] Figure 10 The PHB1 / 2 complex specifically recognizes cardiolipin via conserved positively charged residues. (C) Coarse-grained molecular dynamics simulations show the temporal variation of phospholipid distribution within the PHB1 / 2 ring region. Changes in the amounts of cardiolipin (CL), phosphatidylcholine (PC), and phosphatidylethanolamine (PE) on the inner and outer leaves of the IMM over a 4 μs simulation period. (D, E) Radial distribution function analysis of the interactions between PHB1 (D) and PHB2 (E) and phospholipids. The left panel shows the interactions of polar and basic amino acids with CL on the outer leaf of PHB1 / 2. The right panel shows the normalized densities of PC and PE. The interaction peaks occur at a distance of approximately 0.5 nm.

[0029] Figure 11 The PHB1 / 2 complex specifically recognizes cardiolipin via conserved positively charged residues. (F, G) Representative blots of cardiolipin binding assays for PHB2 (F) and PHB1 (G) mutants. Wild-type and mutant proteins (PHB2: R54T, R71T, R54T / R71T; PHB1: R41T, K63T, R41T / K63T) were incubated with cardiolipin vesicles. PHB2-ΔSPFH1 and PHB1-ΔSPFH1 served as negative controls. (H) Phospholipid vesicle co-precipitation assay. PHB1-WT and PHB2-WT bound to vesicles with different cardiolipin contents (CL:PC:PE ratios of 15:42.5:42.5, 30:35:35, 60:20:20, and 90:5:5). T: Total protein, S: Supernatant (unbound), P: Precipitate (bound).

[0030] Figure 12 The PHB1 / 2 complex drives cardiolipin aggregation and membrane invagination. (A, B) Density distribution of CL (A) and PC / PE (B) on the outer leaflets of the IMM during coarse-grained molecular dynamics simulations. The top panel shows the initial distribution from 0 to 0.1 μs, and the bottom panel shows the final distribution from 3.9 to 4.0 μs. Darker colors indicate higher densities. The PHB1 / 2 ring region is marked by a black circle.

[0031] Figure 13 The PHB1 / 2 complex drives cardiolipin aggregation and membrane invagination. (C, D) Temporal dynamics of interactions between cardiolipin and specific amino acids on the outer leaves of the IMM. (C) Barcode diagram of interactions between cardiolipin and positively charged residues (Arg41, Lys63, His55) and polar residues (Arg37, Arg54, Arg71) on PHB1. (D) Interactions between cardiolipin and positively charged residues (Asn24, Ser25, Asn29, Thr54) and polar residues (Tyr34, Gly35, Ser39, Thr42, Tyr77) on PHB2. Barcodes indicate interaction time points during a 4 μs simulation.

[0032] Figure 14 (E) Membrane curvature analysis shows PHB1 / 2-induced IMM indentation. The blue-to-white gradient represents the change in membrane height from the initial flat state (0 μs) to the final indented state (4 μs), with indentations of approximately 5.6–8.0 Å within the PHB1 / 2 ring region. (F) Cross-sectional view of membrane curvature evolution. The left panel shows the initial flat membrane state (0 μs), and the right panel shows the indented membrane (4 μs) after PHB1 / 2-induced cardiolipin aggregation. (G) Molecular model showing PHB1 / 2-mediated membrane remodeling. A 180° rotated view shows the enrichment of cardiolipin within the PHB1 / 2 ring region after simulation (colored spheres).

[0033] Figure 15 PHB2 N-terminal residues regulate cristae structure via the OMA1-OPA1 proteolytic axis. (A, B) Radial distribution function analysis of PHB1 (A) and PHB2 (B) interactions with phospholipids. The left panel shows the interactions of polar and basic amino acids on the inner leaflets of PHB1 / 2 with CL. The right panel shows the normalized densities of PC and PE. The interaction peak appears at a distance of approximately 0.5 nm.

[0034] Figure 16PHB2 N-terminal residues regulate cristae structure via the OMA1-OPA1 proteolytic axis. (C, D) Density distribution of CL (C) and PC / PE (D) on the inner leaf of the IMM during coarse-grained molecular dynamics simulations. The top figure shows the initial distribution at 0–0.1 μs, and the bottom figure shows the final distribution at 3.9–4.0 μs. Darker colors indicate higher densities. The PHB1 / 2 loop region is marked by a black circle. (E, F) Temporal dynamics of interactions between cardiolipin and specific amino acids on the inner leaf of the IMM. (E) Barcode diagram of the interaction between cardiolipin and positively charged residues (Lys4) on PHB1. (F) Interactions between cardiolipin and positively charged residues (Lys6, Arg11, Arg17) and polar residues (Gln3, Gly10, Gly15) on PHB2. Barcodes indicate interaction time points during the 4 μs simulation.

[0035] Figure 17 PHB2 N-terminal residues regulate cristae structure via the OMA1-OPA1 proteolytic axis. (G) Effect of PHB2-K6A / R17A mutation on PHB1 / 2 complex formation. HEK293T-PHB2- / - cells expressing the Flag tag PHB2-WT or PHB2-K6A / R17A underwent anti-Flag immunoprecipitation and elution, followed by blue native PAGE analysis to detect PHB1 / 2 and PHB1 / 2-SC. In: Input; E: Elution. (H, I) OMA1 and OPA1 cleavage analysis. (H) Western blot showing OMA1 isoforms (pre-pro-OMA1, L-OMA1, S-OMA1) in cells expressing PHB2-WT or PHB2-K6A / R17A. Quantitative results are shown in bar charts. (I) OPA1 cleavage analysis showing the L-OPA1 to S-OPA1 ratio and quantitative results.

[0036] Figure 18 : PHB2 N-terminal residues regulate cristae structure via the OMA1-OPA1 proteolytic axis. (J) TEM analysis of mitochondrial cristae structure. Representative images of HEK293T cells expressing PHB2-WT or PHB2-K6A / R17A, combined with controls (NC-shRNA) or OMA1 knockdown (OMA1-shRNA), and quantitative analysis of mitochondrial area and cristae density. n=100 mitochondria / group.

[0037] Figure 19OMA1-cardiolipin interaction and its functional consequences. (A) Immunoprecipitation analysis of the interaction between YME1L1 and PHB2. Immunoprecipitation was performed using anti-Flag antibody in HEK293T-PHB2- / - cells expressing PHB2-WT or PHB2-K6A / R17A, and YME1L1 and PHB2 were detected. (B) Western blot analysis of OMA1 isoforms in HEK293T-OMA1-KD cells expressing OMA1-WT or OMA1-Δ148-167 (lacking the cardiolipin-binding domain). (C) Same grouping as (B), OPA1 shearing showing changes in the L-OPA1 / S-OPA1 ratio and quantitative analysis. (D) Transmission electron microscopy analysis of crest structure in OMA1 mutants. Representative images and quantitative analysis showed that the crest structure was disordered in cells expressing OMA1-Δ148-167 compared to OMA1-WT. Each group contains n=100 mitochondria.

[0038] Figure 20 PHB2 phosphorylation accelerates metabolic reprogramming and tumor growth in hepatocellular carcinoma. (A) Effect of PHB2 phosphorylation status on glycolysis. Comparative analysis of L-lactic acid production in HepG2 cells expressing PHB2-WT, PHB2-Y34E / Y77E, or PHB2-Y34F / Y77F (n=3). (B) Analysis of mitochondrial respiratory function using the OROBOROS respiration assay. Left panel: Representative oxygen consumption trajectory. Right panel: Quantification of maximal oxygen consumption under basal and FCCP stimulation. Phosphorylation mimic mutants showed impaired respiration, while phosphorylation resistant mutants showed enhanced respiration (n=3).

[0039] Figure 21 PHB2 phosphorylation accelerates metabolic reprogramming and tumor growth in hepatocellular carcinoma. (C) Cell proliferation analysis using EdU incorporation assays. Top: Representative fluorescence microscopy images showing EdU-positive cells (red) and DAPI nuclear staining (blue) in HepG2 cells expressing PHB2-WT, PHB2-Y34E / Y77E, or PHB2-Y34F / Y77F. Scale bar: 50 μm. Bottom: Quantitative analysis showing that phosphorylation-mimicking mutations increase proliferation, while phosphorylation-resistant mutations decrease proliferation (n=3).

[0040] Figure 22PHB2 phosphorylation accelerates metabolic reprogramming and tumor growth in hepatocellular carcinoma. (D) Nude mouse xenograft tumor model. Images of mice implanted with tumors 25 days after injection of HepG2 cells stably expressing PHB2-WT, PHB2-Y34E / Y77E, or PHB2-Y34F / Y77F (n=7 mice / group). (E) Representative images of tumors harvested on day 25. (F) Tumor growth curve. Tumor volume measurements in nude mice 25 days after injection of HepG2 cells expressing PHB2-WT (black), phosphorylated mimicking PHB2-Y34E / Y77E (red), or phosphorylated resistant to PHB2-Y34F / Y77F (blue) (n=7 mice / group).

[0041] Figure 23 Small molecule design based on VD Gen and BRET high-throughput screening to identify inhibitors targeting the PHB2-Src interaction interface. (A) Molecular docking model of the PHB2-Src protein-protein interaction. The docking complex structure shows PHB2 (cyan) and Src (gray), with key interface residues highlighted. (B) Identification of druggable small molecule binding pockets adjacent to the PHB2-Src interface. Molecular target regions are marked with boxes, highlighting potential binding sites near the protein-protein interaction surface. (C) Detailed view of the small molecule binding pocket on PHB2 and surrounding amino acid residues. Key residues constituting the pocket are shown in the figure, forming a well-defined binding cavity suitable for small molecule targeting.

[0042] Figure 24 Small molecule design and BRET high-throughput screening based on VD Gen method to identify inhibitors targeting the PHB2-Src interaction interface. (D) High-throughput screening of candidate compounds was performed using BRET (bioluminescent resonance energy transfer) assays. Based on the VD Gen method, 500 candidate small molecules were designed according to the drug binding pocket. After BRET screening, dose-response curves showed the inhibitory effects of the top 10 compounds on the PHB2-Src interaction. -11 Up to 10 -2 Tests were conducted within the concentration range of M. DMSO was used as a negative control. Error bars represent the standard deviation of three repeated measurements. (E) Two-dimensional chemical structures and molecular formulas of the top 10 candidate compounds.

[0043] Figure 25NMR spectrum of 6-acetamido-1H-benzimidazole (300 MHz, DMSO-d6). Chemical shifts δ (ppm): 14.50 (br s, 1H, benzimidazole ring NH), 10.37 (s, 1H, amide NH), 9.47 (s, 1H, benzimidazole 2-H), 8.38 (d, J = 1.8 Hz, 1H, Ar-H), 7.78 (d, J = 8.9 Hz, 1H, Ar-H), 7.54 (dd, J = 8.9, 1.8 Hz, 1H, Ar-H), 2.10 (s, 3H, CH3). The integrated values ​​of each peak are consistent with the structure of the target compound, confirming the successful synthesis of the product. The solvent peak (DMSO) appears at 2.50 ppm.

[0044] Figure 26 6-Acetamino-1H-benzimidazole exhibits potent antitumor effects in a hepatocellular carcinoma xenograft model. (A) Structural information of molecule 1, 6-acetamino-1H-benzimidazole. This compound is characterized by the SMILES notation, InChI identifier, and InChIKey. The molecule contains 26 atoms (15 heavy atoms), has an atomic mass of 218.28 Da, and the molecular formula is C1. 11 H 10 N2OS, containing two rings (including two aromatic rings). (B) Physicochemical properties and ADMET characteristics of 6-acetamido-1H-benzimidazole. This compound exhibits good drug-like properties, with a QED score of 0.83, a pLDDT score of 0.9898, an SLogP of 2.59, and a TPSA of 55.12 Å. 2 One hydrogen bond acceptor, two hydrogen bond donors, three rotatable bonds, Fsp 3 =0.00, conforming to Lipinski's five rules. (C) Tumor growth curves in xenograft mouse models. Nude mice carrying HepG2-derived tumors were treated with 6-acetamido-1H-benzimidazole or solvent control (DMSO) for 25 days. Tumor volume was measured every 2 days. 6-acetamido-1H-benzimidazole treatment significantly inhibited tumor growth compared with the untreated control group and the DMSO solvent group (****P<0.0001). Data are presented as mean ± standard deviation (n=7 mice / group). (D) Representative images of tumor resection on day 25 after treatment. Tumors in the 6-acetamido-1H-benzimidazole treatment group were significantly smaller than those in the control and DMSO groups, demonstrating the potent antitumor effect of this PHB2-Src interaction inhibitor. Detailed Implementation

[0045] Example 1 I. Cells and Cell Culture Conditions: The human hepatocellular carcinoma cell line HepG2 and the normal hepatocyte cell line LO2 were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai) and identified by short tandem repeat (STR) analysis (HepG2: RRID: CVCL_0027; LO2: RRID: CVCL_6926). HEK293T cells were purchased from the American Type Culture Collection (ATCC CRL-3216; RRID: CVCL_0063). All cells were cultured in DMEM medium (Gibco) containing 10% fetal bovine serum (FBS, Gibco) and 1% penicillin-streptomycin, and kept in a humid environment at 37°C and 5% CO2. All cell lines were negative for mycoplasma using the MycoAlert assay kit (Lonza). Transient transfection was performed using Lipofectamine 3000 (Invitrogen) according to the manufacturer's instructions. For the establishment of stable cell lines, cells were transduced with lentiviral vectors and selected with 2 μg / mL puromycin for 7 days.

[0046] II. Plasmid Construction and Site-Directed Mutation: Human PHB2 cDNA was cloned into the pcDNA3.1-Flag vector using standard molecular cloning techniques. Point mutations (Y34F, Y77F, Y34E, Y77E, K6A, R17A, R54T, R71T) were introduced using the QuikChange II site-directed mutagenesis kit (Agilent Technologies) according to the manufacturer's protocol. Human SRC cDNA was cloned into the pLVX-IRES-puro lentiviral vector for overexpression studies. All constructs were validated by Sanger sequencing to confirm the absence of unexpected mutations.

[0047] III. RNA Interference: Small interfering RNAs (siRNAs) targeting human SRC (5'-CACCTTTGTGGCCCTCTATGACT-3'), EGFR (5'-GAGGAAATATGTACTACGAAAAT-3'), and INSR (5'-TCCACTATAACCCCAAACTCTGC-3'), as well as a scrambled control siRNA, were purchased from Gemma Gene (Shanghai, China). Cells were transfected with 50 nM siRNA using Lipofectamine RNAiMAX (Invitrogen), and analysis was performed 48-72 hours post-transfection. As a negative control, a pool of non-targeted scrambled siRNAs (control-siRNA) at the same concentration was used. For stable knockdown, short hairpin RNA (shRNA) sequences were cloned into the pLKO.1-puro vector, targeting human SRC (5'-CACCTTTGTGGCCCTCTATGACT-3') and OMA1 (5'-CTGTATGGAATGATGCTTTTTCA-3'). The efficiency of knockdown and overexpression was verified by detecting RNA levels using qRT-PCR or protein levels using Western blot.

[0048] IV. Western blotting and immunoprecipitation: Cells were lysed with RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) and supplemented with a mixture of protease and phosphatase inhibitors (Roche). Protein concentration was determined using the PierceBCA protein assay kit. For immunoprecipitation, 500 μg of protein lysis buffer was incubated overnight at 4°C with 2 μg of primary antibody by gentle swirl, followed by incubation with protein A / G agarose beads (SantaCruz Biotechnology) for 2 hours. The immunoprecipitate was washed three times with lysis buffer and eluted with SDS loading buffer. The following antibodies were used: anti-PHB2 (1:1000, Cell Signaling Technology #14085), anti-PHB1 (1:1000, Abcam ab75771), anti-phosphorylated tyrosine (1:1000, Cell Signaling Technology #9411), anti-SRC (1:1000, Cell Signaling Technology #2109), anti-β-actin (1:5000, Sigma-Aldrich A5441), anti-HSP60 (1:1000, Cell Signaling Technology #12165), anti-OMA1 (1:1000, Santa Cruz Biotechnology sc-515788), anti-OPA1 (1:1000, BD Biosciences 612606), and anti-AFG3L2 (1:1000, Abcam ab139503).

[0049] V. Blue Native Polyacrylamide Gel Electrophoresis (BN-PAGE): Blue native PAGE was performed as previously described: Cells were resuspended in lysis buffer (50 mM NaCl, 50 mM imidazole, 2 mM 6-aminocaproic acid, 1 mM EDTA, pH 7.0) and dissolved on ice for 15 min with 1% digitalis saponin. After centrifugation at 20,000 g for 30 min, the supernatant was replenished with Coomassie Brilliant Blue G-250 and loaded onto a 3-12% gradient native gel. Electrophoresis was performed at 4°C, initially at 100 V for 30 min, then at 200 V until completion. Proteins were transferred to a PVDF (Millipore) membrane. After blocking with 5% (w / v) skim milk powder-TBST buffer, the membrane was incubated overnight at 4°C with primary antibody, washed, and then incubated with the corresponding HRP-labeled secondary antibody for 1 h. Signals were detected by chemiluminescence using a ChemiDoc Touch imaging system.

[0050] VI. Subcellular Fractionation: Mitochondrial and cytoplasmic components were separated using a mitochondrial fractionation kit (Thermo Scientific) according to the manufacturer's instructions. Cells were homogenized in fractionation buffer using a Dounce homogenizer (20 times), followed by differential centrifugation at 750 g for 10 min and 12,000 g for 15 min. The mitochondrial pellet was washed twice with fractionation buffer. Fractionation purity was verified by Western blotting using HSP60 as a mitochondrial marker and β-actin as a cytoplasmic marker.

[0051] VII. Transmission Electron Microscopy (TEM): Cells were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) for 2 hours, followed by 1% osmium tetroxide fixation for 1 hour. They were then dehydrated using a gradient of ethanol and embedded in Spurr resin. Ultrathin sections (70 nm) were prepared using a LeicaEM UC7 microtome, stained with uranium acetate and lead citrate, and examined using a JEOL JEM-1400 transmission electron microscope at 80 kV. Images were acquired using a Gatan CCD camera. Mitochondrial area and cristae density were quantified using ImageJ software. At least 100 mitochondria were analyzed under each condition from three independent experiments.

[0052] 8. High-resolution respiration measurement: Mitochondrial oxygen consumption was measured using the Oxygraph-2k system (OROBOROS Instruments, Innsbruck, Austria) according to the substrate-inhibitor titration protocol. LO2 and HepG2 cells were harvested and resuspended in MiR05 respiration medium at 37°C. 2 mL of cell suspension (approximately 2 × 10⁻⁶ cells / mL) was added to the medium. 6 Cells were added to each O2k chamber. The following solutions were added in sequence: glutamate (10 mM) plus malate (2 mM) and ADP (2.5 mM) to measure complex I-driven respiration; rotenone (0.5 μM) to inhibit complex I; succinate (10 mM) to assess complex II activity; antimycin A (2.5 μM) to block complex III; ascorbic acid (2 mM) plus TMPD (0.5 mM) to measure complex IV respiration, followed by sodium azide (AZD, 10 mM) to inhibit complex IV. Oxygen consumption was calculated using DatLab software and normalized to cell number.

[0053] IX. L-Lactate Production Detection: Extracellular L-lactate levels were measured using an L-lactate assay kit (Beyotime Biotechnology) according to the manufacturer's protocol: 50 μL of culture supernatant from HepG2 cells expressing PHB2-WT, PHB2-Y34E / Y77E, or PHB2-Y34F / Y77F was collected and mixed with 50 μL of working solution (containing enzyme solution, substrate, and lactate detection buffer). After incubation at 37°C for 30 minutes, absorbance was measured at 450 nm. L-lactate concentration was calculated from the standard curve (0–0.5 mM) and normalized to the cell number.

[0054] 10. NAD + / NADH ratio measurement: using NAD + / NADH Detection Kit (Beyotime Biotechnology) Measure cellular NAD according to the manufacturer's instructions + And NADH level: approximately 1×10 6 Cells were lysed with 200 μL NAD⁺ / NADH extraction buffer, followed by centrifugation at 12,000 g for 5–10 minutes at 4°C. For total NAD... + To measure NADH, 20 μL of supernatant was mixed with 90 μL of alcohol dehydrogenase working solution and incubated at 37°C for 10 minutes to measure NAD. + The NAD⁺ concentration is converted to NADH. For NADH-only measurements, the sample is heated at 60°C for 30 minutes to decompose NAD⁺. After adding 10 μL of the colorimetric solution, the sample is incubated at 37°C for 30 minutes, and the absorbance is measured at 450 nm. The NAD⁺ concentration is calculated as [NAD⁺ concentration]. + The formula is: [Total NAD] - [NADH], and NAD is determined accordingly. + / NADH ratio.

[0055] XI. Oxygen Consumption Rate (OCR) Analysis: Mitochondrial respiration was measured using a high-resolution respiration measurement method on an Oxygraph-2k system (OROBOROS Instruments, Innsbruck, Austria). HepG2 cells stably expressing PHB2-WT, PHB2-Y34E / Y77E, or PHB2-Y34F / Y77F were harvested and analyzed at a rate of 1×10⁻⁶. 6 Cells were resuspended in 2 mL of DMEM containing 2% FBS. The cell suspension was transferred to an O2k chamber, and basal oxygen consumption was recorded under unstimulated conditions. To determine maximum respiratory capacity, the following ingredients were added in sequence: FCCP (1 μM, mitochondrial uncoupling agent), oligomycin (2.5 μM, ATP synthase inhibitor), rotenone (0.5 μM, complex I inhibitor), and antimycin A (2.5 μM, complex III inhibitor). Oxygen flux was continuously monitored in real time, and OCR values ​​were calculated and normalized to cell number.

[0056] XII. krCRISPR for Generating Conditional Gene Knockout Cells: This invention introduces krCRISPR, a strategy for efficiently generating conditional gene knockout cells using a dual-free vector. First, the epiCRISPR plasmid is modified to express SpCas9 and tTA nucleases from the EF1α promoter using a self-cleaving P2A peptide, while simultaneously expressing gRNA from the human U6 promoter. The synthetic oligonucleotide duplex encoding the gRNA can be readily cloned into the BspQI restriction site. Then, a rescue plasmid is constructed containing copGFP and a puromycin resistance gene co-expressed from the pTRE promoter, as well as a downstream exogenous PHB2 gene. Synonymous mutations are introduced into the gRNA targeting sequence to prevent Cas9 cleavage. By co-transfecting cells with the two plasmids, the endogenous PHB2 gene is effectively knocked out, while simultaneously inducing exogenous expression of both wild-type and mutant PHB2. To verify the efficiency of this method, cell proliferation, colony formation, and protein expression were measured.

[0057] XIII. Liposome Co-precipitation Assay: Liposomes were prepared by mixing cardiolipin (CL), phosphatidylcholine (PC), and phosphatidylethanolamine (PE) in chloroform at specified ratios, followed by evaporation and rehydration in binding buffer (20 mM HEPES pH 7.4, 150 mM NaCl, 1 mM DTT). Liposomes were then extruded through a 100 nm polycarbonate membrane. In addition to CL:PC:PE ratio gradients (15:42.5:42.5, 30:35:35, 60:20:20, 90:5:5), a single phospholipid group of liposomes (100% CL, 100% PE, 100% PC) was used as a control. 1 μg of purified recombinant PHB1 or PHB2 protein (wild-type or mutant, including R54T, R71T, R54T / R71T, and ΔSPFH1 of PHB2; and R41T, K63T, R41T / K63T, and ΔSPFH1 of PHB1) was incubated with liposomes (100 μg of total lipids) at room temperature for 30 min. After centrifugation at 100,000 g for 30 min, the precipitate (P) and supernatant (S) were analyzed by SDS-PAGE and Western blotting.

[0058] XIV. Coarse-grained Model of PHB1 / 2 Integration with Lipids: To investigate the dynamic changes of the protein-lipid complex formed by the PHB1 / 2 complex and the mitochondrial inner membrane bilayer, a coarse-grained system described by the Martini force field (v2.2) was generated using CHARMM-GUI. The upper lipid bilayer consisted of 900 POPCs, 700 POPEs, and 400 cardiolipin molecules, while the lower layer consisted of 918 POPCs, 720 POPEs, and 420 cardiolipin molecules. Cardiolipin molecules were modeled using a coarse-grained model with two negative charges. The simulation temperature was set to 310 K and controlled using a Nosé-Hoover thermostat. The pressure was set to 1 bar and controlled using a Parrinello-Rahman barostat. Furthermore, semi-isotropic coupling was used to control the pressure in the XY and Z directions, respectively. The Lennard-Jones (LJ) cutoff was set to 1.2 nm. Electrostatic interactions were calculated using the Particle Mesh Ewald (PME) method. The time step was set to 20 fs to constrain the protein backbone and prevent oligomer dissociation. The simulation time was 4 μs, with conformations output every 1 ns. The simulation software was GROMACS 2019.6, and the analysis software included GROMACS_tools, Mdtraj, MDAnalysis, and VMD for visualization.

[0059] 15. Analysis of Lipid Distribution and Density Around Polar and Positively Charged Residues in the IMM: Radial distribution functions (RDFs) measure the change in bead density with distance from a reference bead relative to its average density in the system. They indicate whether lipid beads are more or less likely to be found around certain protein residues. The distribution of lipids (phosphate beads) relative to their distance from protein residues (glycine main chain beads, and side chain beads of other residues) during the simulation was calculated using the mdtraj python package. The radial distribution function (RDF) or correlation function gAB(r) between type A (protein residue) and type B (phosphate bead) particles is defined as follows: where <ρB(r)> is the density of type B particles at a distance r from type A particles, and the density of type B particles is averaged over all spheres with radius rmax (1.5 nm) centered on the type A particle. The density of lipids (phosphate beads) was calculated using the MDAnalysis python package with a bin size of 5 Å. The positions of phospholipids are mapped to the xy plane, with darker colors indicating higher densities.

[0060] XVI. Cell Proliferation Assay: Cell proliferation was assessed using the Click-iT EdU Alexa Fluor 488 Imaging Kit (Invitrogen). Cells were incubated with 10 μM EdU for 2 hours, fixed with 4% paraformaldehyde, permeabilized, and subjected to click chemistry according to the manufacturer's protocol. Nuclei were counterstained with DAPI. Images were acquired using a Zeiss LSM 880 confocal microscope, and EdU-positive cells were quantified from at least 500 cells per condition.

[0061] XVII. Xenograft Tumor Model: All animal experimental procedures were approved by the Zhejiang University Animal Ethics Review Committee and conducted in accordance with the guidelines of NIH Publication No. 85-23 (1996 revised edition) (Approval No.: AIRB-2023-1448). Six-week-old male BALB / c nude mice were subcutaneously injected with 5 × 10⁻⁶ g of xenograft material. 6 HepG2 cells stably expressing PHB2-WT, PHB2-Y34E / Y77E, or PHB2-Y34F / Y77F were dissolved in 100 μl of PBS and Matrigel (1:1). Tumor volume was measured every 2 days using calipers and calculated using the formula (length × width). 2 ) / 2. On day 25, the mice were sacrificed, the tumors were removed, and they were weighed and photographed.

[0062] 18. Small Molecule Design and BRET High-Throughput Screening Based on VD Gen: Based on the structural analysis of the PHB2-Src protein-protein interaction interface, this invention employs the VD Gen (Virtual Drug Generation) method for small molecule design based on drug-binding pockets. First, using a molecular docking model of the PHB2-Src complex, druggable binding pockets at adjacent protein-protein interaction interfaces on PHB2 are identified and characterized using FPocket and SiteMap algorithms. Then, using the VDGen deep generation model, with the three-dimensional structural features of these binding pockets as input, 500 candidate small molecules with potential binding capacity are generated. During generation, the molecular weight range is set to 150-300 Da, the QED score ≥ 0.5, and compliance with Lipinski's five rules is used as constraints. The generated compounds are molecularly docked using AutoDock Vina and preliminarily ranked according to their binding free energy.

[0063] To verify the bioactivity of candidate compounds, this invention established an in vitro high-throughput screening system based on bioluminescent resonance energy transfer (BRET) technology. PHB2 was fused with NanoLuc luciferase (donor), and Src was fused with a HaloTag tag and conjugated to the HaloTag-618 fluorescent ligand (receptor). When PHB2 interacts with Src, the blue light emitted by the donor excites the neighboring receptor to produce red-shifted fluorescence. The protein-protein interaction strength was quantified by calculating the 618 nm / 460 nm emission ratio. HEK293T cells were transiently co-transfected with the PHB2-NanoLuc and Src-HaloTag expression plasmids, and seeded into 384-well white transparent plates 24 hours later. Compounds were initially screened at a starting concentration of 10 μM, and positive compounds were further screened at a concentration of 10 μM. -11 Up to 10 -2 Dose-response curves were determined for the M concentration range. Three replicate wells were set up for each concentration, with DMSO as a negative control. BRET signals were detected using a PHERAstar FSX microplate reader (BMG Labtech), and IC50 was calculated using GraphPad Prism. 50 Based on the screening results, the top 10 candidate compounds in terms of efficacy were selected for further validation, and the compound ranked first, 6-acetamido-1H-benzimidazole, was further synthesized and its in vitro and in vivo activity was validated.

[0064] 19. Statistical Analysis: Unless otherwise stated, data are expressed as mean ± standard deviation of at least three independent trials. Statistical analysis was performed using GraphPad Prism 9. Unpaired Student's t-test was used for comparisons between two groups. One-way or two-way ANOVA was used for comparisons among multiple groups, followed by Bonferroni post-hoc test. Kaplan-Meier survival analysis was performed using TCGA data, with log-rank test. P < 0.05 was considered statistically significant.

[0065] Data Analysis (1) Hepatocellular carcinoma cells exhibit severe mitochondrial cristae disorder and OXPHOS dysfunction. To investigate the structural and functional alterations of mitochondria in hepatocellular carcinoma, this invention conducted a comprehensive ultrastructural and metabolic analysis comparing the normal hepatocellular carcinoma cell line LO2 and the hepatocellular carcinoma cell line HepG2. Transmission electron microscopy revealed significant mitochondrial cristae abnormalities in HepG2 cells. Figure 1A). While LO2 cells exhibit well-organized cristae and regular lamellar structures, HepG2 cells show markedly disordered cristae with decreased density and irregular, fragmented cristae membranes. Furthermore, HepG2 mitochondria show increased cross-sectional area, consistent with mitochondrial swelling commonly observed in metabolically impaired cells.

[0066] These structural abnormalities are accompanied by significant metabolic dysfunction. The NAD⁺ / NADH ratio is a key indicator of cellular redox status and oxidative metabolism; it is sharply reduced in HepG2 cells compared to LO2 cells, suggesting impaired oxidative phosphorylation and a shift towards reductive metabolism. Figure 1 B). OROBOROS respiratory measurements confirmed severe respiratory chain defects in HepG2 cells, with reduced activity of mitochondrial respiratory chain complexes I and II (B). Figure 2 C). These findings suggest that hepatocellular carcinoma cells exhibit coordinated mitochondrial structural and functional defects, with cristae disorder closely associated with oxidative phosphorylation damage.

[0067] (2) PHB2 undergoes specific Src-mediated tyrosine phosphorylation in hepatocellular carcinoma cells. The maintenance of mitochondrial cristae structure critically depends on the forbidden protein 1 / 2 (PHB1 / 2) complex, which anchors to the inner membrane via lipid interactions to form a ring structure. Given that PHB1 / 2 dysfunction leads to cristae collapse similar to that observed in HepG2 cells in this invention, this invention investigates whether PHB1 / 2 proteins undergo aberrant post-translational modifications in hepatocellular carcinoma. Immunoprecipitation followed by phosphorylated tyrosine immunoblotting revealed that PHB2 (but not PHB1) exhibited significantly increased tyrosine phosphorylation in HepG2 cells compared to LO2 cells. Figure 3 (A and B). Importantly, the total protein levels of PHB1 and PHB2 remained unchanged between the two cell lines, indicating that the increased phosphorylation of PHB2 represents a specific regulatory event rather than a result of changes in protein abundance.

[0068] To identify the kinase responsible for PHB2 phosphorylation, this invention screened for targeted siRNAs of tyrosine kinases known to be dysregulated in hepatocellular carcinoma. Among the candidate kinases tested (insulin receptor INSR, Src kinase, and epidermal growth factor receptor EGFR), only Src knockdown significantly reduced PHB2 phosphorylation in HepG2 cells. Figure 3 C). The important role of Src in mediating PHB2 phosphorylation was further verified by gain-of-function and loss-of-function experiments. Src overexpression significantly enhanced PHB2 phosphorylation (C). Figure 3 D), while stable knockdown of Src using shRNA significantly reduced phosphorylation levels (D). Figure 4E), confirming that Src is both necessary and sufficient for PHB2 tyrosine phosphorylation in hepatocellular carcinoma cells.

[0069] To precisely locate the Src phosphorylation site on PHB2, this invention generated a set of tyrosine-phenylalanine and serine-alanine point mutants targeting the predicted phosphorylation sites. Systematic mutation analysis revealed that simultaneously mutating Y34 and Y77 to phenylalanine (Y34F / Y77F) completely eliminated Src-mediated PHB2 phosphorylation, while single mutations or combinations targeting other residues retained phosphorylation ability. Figure 4 F). These results identified Y34 and Y77 as the major Src phosphorylation sites on PHB2. The clinical relevance of this finding was highlighted by analysis of the Cancer Genome Atlas (TCGA) database, showing that high SRC expression was associated with a significant decline in overall survival in patients with hepatocellular carcinoma. Figure 4 (G), suggesting that the Src-PHB2 axis may promote disease progression.

[0070] Bioinformatics analysis further supports Y34 and Y77 as Src phosphorylation sites. Scansite 4.0 kinase recognition motif analysis showed that both Y34 and Y77 exhibited high confidence scores (>0.75) for Src kinase recognition, with Y34 showing particularly strong consensus with the Src recognition motif. Figure 5 A). Integrated predictions using five independent algorithms (NetPhos 3.1, GPS 5.0, KinasePhos2.0, PhosphoSitePlus, and MustyDeep) consistently identified Y34 and Y77 as high-probability Src phosphorylation sites, with consensus scores exceeding 0.8 for all methods. Figure 5 B). Evolutionary conservation analysis using ConSurf showed that Y34 and Y77 are highly conserved across vertebrate species, suggesting the functional importance of these residues. Figure 5 C). Furthermore, the N-terminus of PHB2 shows strong evolutionary conservation for the key positively charged and polar amino acids involved in cardiolipin binding (C). Figure 5 D), consistent with its important role in membrane anchoring. Functional impact prediction algorithms (SIFT and PolyPhen-2) show that mutations in Y34F and Y77F significantly affect the function of the PHB2 protein, supporting the functional importance of these tyrosine residues. Figure 5 E).

[0071] (3) Src-mediated phosphorylation disrupts the stability of the PHB1 / 2 complex and triggers cytoplasmic mislocalization of PHB2. After identifying the Y34 and Y77 sites of Src phosphorylation of PHB2, this invention further investigated the functional consequences of this modification on the integrity and subcellular localization of the PHB complex. Structural modeling revealed that the Y34 and Y77 residues are located near the membrane-binding interface of PHB2. Figure 6 A). Electrostatic surface potential analysis showed that the phosphorylation of these tyrosines introduced a negative charge into a region typically characterized by a positive electrostatic potential, resulting in unfavorable electrostatic repulsion against the negatively charged mitochondrial inner membrane phospholipids.

[0072] Blue natural PAGE analysis showed that Src overexpression dramatically reduced the levels of the PHB1 / 2 supercomplex and PHB1 / 2 complex in cells expressing wild-type PHB2, but had no effect on cells expressing phosphorylation-resistant PHB2-Y34F / Y77F mutants. Figure 6 B). This indicates that Y34 / Y77 phosphorylation is both necessary and sufficient for Src-induced complex disintegration. Consistent with these findings, HepG2 cells showed significantly reduced PHB1 / 2 complex levels compared to LO2 cells, and correspondingly reduced incorporation of the PHB-interacting m-AAA protease AFG3L2 into the complex. Figure 6 C).

[0073] The PHB1 / 2 complex exhibits significant sensitivity to mitochondrial stress conditions mimicking the tumor microenvironment. Oxidative stress (H2O2), inhibition of complex I (rotenone), inhibition of complex III (antimycin), or inhibition of ATP synthase (oligomycin) all lead to dissociation of the PHB1 / 2 complex, while the unrelated TIM23 translocase complex remains stable under the same conditions. Figure 7 D). This selective fragility suggests that the PHB complex acts as a sensor for mitochondrial dysfunction. Subcellular grading experiments showed that various stress conditions induced the translocation of PHB2 from mitochondria to the cytoplasm in HepG2 cells ( Figure 7 E. H2O2 (oxidative stress), CoCl2 (hypoxia mimic), rotenone, or antimycin treatment all trigger PHB2 cytoplasmic accumulation, mimicking the stressed tumor microenvironment. Dose-response analysis showed that increased Src expression levels led to a progressive redistribution of PHB2 from mitochondria to the cytoplasm, with cytoplasmic PHB2 accumulation becoming more pronounced at higher Src expression levels. Figure 8 F). The dose-dependent relationship between Src expression and PHB2 mislocalization suggests that the degree of PHB2 cytoplasmic translocation is directly related to Src kinase activity.

[0074] (4) The PHB1 / 2 complex specifically recognizes cardiolipin through conserved positively charged residues. To understand the molecular basis of PHB-membrane interactions and how phosphorylation disrupts these interactions, this invention employed structural analysis and molecular dynamics simulations. Both PHB1 and PHB2 contain conserved clusters of positively charged residues on their membrane-facing surfaces (PHB1: K4, R41, K63; PHB2: K6, R17, R54, R71), creating a favorable electrostatic environment for interactions with negatively charged phospholipids. Conservation analysis revealed an enrichment of basic and polar residues at optimal distances (0.5–1.0 nm) from the membrane interface, suggesting evolutionary conservation of lipid binding capacity. Figure 9 (A and B).

[0075] 4-microsecond coarse-grained molecular dynamics simulations showed that cardiolipin progressively and selectively accumulates in the PHB1 / 2 loop region of both lobes of the inner mitochondrial membrane, while the levels of phosphatidylcholine and phosphatidylethanolamine remain relatively constant. Figure 10 C). Radial distribution function analysis confirmed that both PHB1 and PHB2 preferentially interact with cardiolipin, showing a sharp peak at a distance of approximately 0.5 nm, indicating direct electrostatic interaction, while the interactions with PC and PE are much weaker and less structurally robust. Figure 10 (D and E).

[0076] Experiments using liposome co-precipitation assays confirmed the computational predictions. Wild-type PHB1 and PHB2 showed robust cardiolipin binding, while mutations targeting predicted cardiolipin-binding residues (PHB1: R41T, K63T; PHB2: R54T, R71T) progressively impaired binding, with double mutants showing severe defects. Figure 11 F and G). Both PHB1 and PHB2 proteins showed concentration-dependent binding, increasing with increasing phospholipid content in the liposome center, demonstrating specificity for cardiolipin-rich membrane domains. Figure 11 These results identify that the PHB1 / 2 complex recognizes cardiolipin through specific electrostatic interactions mediated by conserved basic residues.

[0077] (5) The PHB1 / 2 complex drives cardiolipin aggregation and membrane invagination. Having determined the molecular basis of PHB-cardiolipin interactions, this invention investigated the functional consequences of these interactions on membrane tissues. Molecular dynamics simulations revealed significant reorganization of cardiolipin on the outer lamina of the inner mitochondrial membrane. Although cardiolipin was initially uniformly distributed, it progressively accumulated in and around the PHB1 / 2 ring region, reaching enrichment at 4 microseconds. Figure 12 A). In contrast, PC and PE maintained a relatively uniform distribution throughout the simulation. Figure 12 B), demonstrating the specificity of PHB-mediated lipid remodeling.

[0078] Time-interaction analysis revealed persistent contact between cardiolipin molecules and specific PHB residues. Notably, PHB2 residues Y34 and Y77—Src phosphorylation sites—showed a high interaction frequency with cardiolipin, providing a structural context for why phosphorylation of these residues disrupts membrane binding. Figure 13 C and D). PHB-mediated cardiolipin aggregation has a profound impact on membrane topology. Membrane curvature analysis revealed progressive indentation within the PHB1 / 2 ring region during the simulation time. Figure 14 E and F). This localized membrane deformation specifically occurs at sites of cardiolipin accumulation, suggesting that the PHB complex shapes the membrane structure through a lipid-mediated mechanism. Three-dimensional visualization confirms the formation of cardiolipin-rich microdomains beneath the PHB ring, which may serve as the core region for cristae formation. Figure 14 G).

[0079] (6) PHB2 N-terminal residues regulate cristae structure via the OMA1-OPA1 protein hydrolysis axis While outer leaf interactions establish the localization of the PHB1 / 2 complex, this invention reveals that PHB proteins also organize cardiolipin on the inner (matrix-facing) leaves via their N-terminal residues. Radial distribution function analysis determined a strong preferential interaction between cardiolipin and the N-terminal basic residues of PHB1 (particularly K4) and PHB2 (particularly K6 and R17) on the inner leaves. Figure 15 A and B). Molecular dynamics simulations confirmed the progressive enrichment of cardiolipin around the PHB ring in the inner leaf, creating a cardiolipin enrichment domain (A and B). Figure 16 CF).

[0080] To test the functional importance of these N-terminal interactions, this invention generated a PHB2-K6A / R17A mutant targeting key cardiolipin-binding residues. This mutation severely impairs the formation of the PHB1 / 2 complex and AFG3L2 integration during reconstruction with the mutant protein in PHB2 knockout cells. Figure 17 G). The PHB2-K6A / R17A mutation triggers abnormal activation of the OMA1 protease, manifested as increased levels of S-OMA1 (active form) and excessive OPA1 cleavage, leading to a decreased L-OPA1 / S-OPA1 ratio. Figure 17 H and I). These biochemical changes translate into severe crest disorder, with reduced crest density in mutant PHB2. Significantly, OMA1 knockdown dramatically rescued the crest defects caused by the PHB2-K6A / R17A mutation, restoring crest density to near-normal levels. Figure 18 J).

[0081] The importance of PHB2 in the cardiolipin-rich domain tissue protein complex was further demonstrated by immunoprecipitation analysis, showing that the PHB2-K6A / R17A mutation does not disrupt its interaction with YME1L1 (YME1L1 is another m-AAA protease involved in mitochondrial quality control), indicating that the K6 and R17 sites in PHB2 do not affect the interaction between the PHB1 / 2 complex and YME1L1. Figure 19 A). To further test whether the cardiolipin-binding capacity of OMA1 is essential for PHB2 regulation, this invention generated an OMA1 mutant (OMA1-Δ148-167) lacking the predicted cardiolipin-binding domain. Cells expressing this deletion mutant showed constitutively elevated S-OMA1 levels compared to OMA1-WT, indicating overactivation ( Figure 19 B). Correspondingly, OMA1-Δ148-167 expression led to OPA1 over-cleavage, and the L-OPA1 / S-OPA1 ratio was significantly reduced (B). Figure 19 C). TEM analysis confirmed that the loss of the OMA1 cardiolipin-binding domain led to severe crest disorder and decreased crest density. Figure 19 D). These results indicate that disruption of the PHB2-cardiolipin interaction primarily mediates the overactivation of the OMA1 protease, leading to OPA1 cleavage and mitochondrial cristae disruption, a process unrelated to YME1L1 protease activity.

[0082] (7) PHB2 phosphorylation promotes glycolysis and accelerates tumor growth. To establish the functional significance of PHB2 phosphorylation in cancer cell metabolism and tumor progression, this invention generated HepG2 cell lines stably expressing wild-type PHB2, phosphorylation mimicking PHB2-Y34E / Y77E (simulating constitutive phosphorylation), or phosphorylation-resistant PHB2-Y34F / Y77F mutants. The phosphorylation mimic mutants exhibited significantly increased lactate production, indicating enhanced glycolytic flux, while the phosphorylation-resistant mutants showed decreased lactate production, suggesting reduced glycolysis. Figure 20 A).

[0083] OROBOROS respiratory measurements showed that phosphorylation mimicking the PHB2-Y34E / Y77E mutant severely impaired basal and maximal mitochondrial respiration, while phosphorylation resistant to the PHB2-Y34F / Y77F mutant enhanced respiratory capacity. Figure 20 B). The inverse relationship between lactate production and oxygen consumption indicates that PHB2 phosphorylation shifts cellular metabolism from oxidative phosphorylation to glycolysis. These metabolic changes are accompanied by alterations in proliferative capacity. EdU incorporation experiments showed that phosphorylation-mimicking mutations increased proliferation, while phosphorylation-resistant mutations decreased proliferation. Figure 21 C).

[0084] The effect of PHB2 phosphorylation on tumor growth was validated in vivo using a xenograft model. Nude mice subcutaneously injected with HepG2 cells expressing different PHB2 variants showed significant differences in tumor development. Phosphorylation mimicking the PHB2-Y34E / Y77E mutant significantly accelerated tumor growth. Figure 22 Conversely, phosphorylation resistance to the PHB2-Y34F / Y77F mutant significantly inhibited tumor growth. These results suggest that blocking PHB2 phosphorylation effectively suppresses hepatocellular carcinoma progression by restoring mitochondrial function and inhibiting glycolytic metabolism, identifying the Src-PHB2 axis as a potential therapeutic target.

[0085] (8) Small molecule design based on VD Gen and high-throughput screening by BRET to identify inhibitors targeting the PHB2-Src interaction interface Having established that Src-mediated PHB2 phosphorylation drives HCC metabolic reprogramming and tumor progression, this invention seeks to identify small molecule inhibitors targeting this interaction as potential therapeutic agents. Molecular docking analysis of the PHB2-Src complex revealed a well-defined protein-protein interaction interface with specific binding regions. Figure 23 A). Importantly, this invention identifies a druggable pocket near the interface that can accommodate small molecule binding without directly competing with the protein-protein interaction surface. Figure 23 B).

[0086] Detailed structural analysis of the binding pocket revealed a cavity composed of key amino acid residues, including ARG-48, GLY-46, HIS-47, ALA-65, GLY-45, LEU-64, GLU-66, ALA-49, GLU-44, GLY-67, VAL-43, HIS-69, PHE-41, LEU-68, VAL-40, and PHE-70. Figure 23 C). This pocket exhibits characteristics that favor small molecule binding, including appropriate size, depth, and a mixture of hydrophobic and polar residues that promote a variety of molecular interactions.

[0087] Using this structural information, this invention employed the Virtual Drug Generation (VD Gen) method to design 500 candidate small molecule compounds based on the three-dimensional structural features of the drug binding pocket. Furthermore, an in vitro high-throughput screening system based on BRET technology was established to verify the inhibitory activity of the compounds against the PHB2-Src interaction, identifying the top 10 compounds in terms of efficacy. Figure 24D). Dose-response analysis showed that all 10 compounds exhibited concentration-dependent inhibition of the PHB2-Src interaction, with compound 1 showing the strongest inhibitory activity, exhibiting near-complete inhibition at nanomolar concentrations. Structural analysis of these compounds revealed diverse chemical skeletons, with molecular weights ranging from 153.18 to 218.28 Da. Figure 24 E (Tables 1-4). Importantly, based on QED scores (0.38 to 0.83), high structural confidence (pLDDT score > 0.98), and compliance with Lipinski's five rules (Tables 1-4), all compounds exhibited favorable drug-like properties. ADMET property analysis revealed acceptable lipophilicity (SLogP from -0.10 to 2.59) and moderate topological polar surface area (TPSA from 41.46 to 92.52 Å). 2 The presence of these compounds and their appropriate hydrogen bonding characteristics indicate that they have good oral bioavailability potential.

[0088] (9) 6-Acetamido-1H-benzimidazole inhibits the growth of hepatocellular carcinoma in vivo. Based on its superior efficacy and favorable drug-like properties in the BRET screening experiment, compound 1 (named molecule 1, i.e., 6-acetamido-1H-benzimidazole) was selected for further study. Its synthetic route is as follows: Starting from 5-nitrobenzimidazole, it was synthesized via a four-step reaction involving Boc protection, nitro reduction, acetylation, and deprotection. The synthetic route is as follows: The specific preparation steps include: Boc protection of nitrogen atom: 5-Nitrobenzimidazole (200 mg, 1.22 mmol) was dissolved in tetrahydrofuran (10 mL), and di-tert-butyl dicarbonate (321 mg, 1.47 mmol, 1.2 equivalents) and triethylamine (255 µL, 1.5 equivalents) were added sequentially at room temperature, and the mixture was stirred for 1 h. The reaction solution was concentrated under reduced pressure to give a mixture 1 of tert-butyl 5(6)-nitro-1H-benzimidazole-1-carboxylic acid ester (243 mg, yield 75.6%), a white solid, which was used directly in the next step without purification. LC-MS (ESI): m / z = 264.1 [M+H] + .

[0089] Catalytic hydrogenation reduction of nitro groups: Mixture 1 (82 mg, 0.31 mmol) was dissolved in methanol (10 mL), and 10% Pd / C (15 mg) was added as a catalyst. The reaction was carried out at 50°C for 1.5 h under a hydrogen atmosphere to reduce the nitro group to an amino group. The reaction solution was filtered through diatomaceous earth to remove the catalyst, and the filtrate was concentrated under reduced pressure to obtain mixture 2 (67 mg, yield 92.2%) of tert-butyl 5(6)-amino-1H-benzimidazole-1-carboxylic acid ester, a white solid that could be used directly in the next step without purification. LC-MS (ESI): m / z = 234.2 [M+H] + .

[0090] Acetylation of amino groups: Mixture 2 (67 mg, 0.24 mmol) was dissolved in tetrahydrofuran (5 mL), and acetic anhydride (37 mg, 0.36 mmol, 1.5 equivalents) was added to carry out the acetylation reaction at 50°C for 10 min. The reaction solution was concentrated under reduced pressure and purified by silica gel column chromatography (dichloromethane / methanol = 50:1 → 30:1) to give tert-butyl 5(6)-acetamido-1H-benzimidazole-1-carboxylic acid ester mixture 3 (62 mg, yield 78.4%) as a white solid. LC-MS (ESI): m / z = 276.1 [M+H] + .

[0091] Deprotection of the Boc group: Mixture 3 (61 mg, 0.225 mmol) was dissolved in dichloromethane (5 mL), and trifluoroacetic acid (0.1 mL) was added for acidic deprotection. The reaction was carried out at room temperature for 5 min. The reaction solution was concentrated under reduced pressure to obtain the target product 6-acetamido-1H-benzimidazole (37 mg, yield 95.1%), which was a pale yellow solid. LC-MS (ESI): m / z = 176.1 [M+H]⁺;¹H NMR (300 MHz, DMSO-d6) δ 14.50 (br s, 1H), 10.37 (s, 1H), 9.47 (s, 1H),8.38 (d, J = 1.8 Hz, 1H), 7.78 (d, J = 8.9 Hz, 1H), 7.54 (dd, J = 8.9, 1.8Hz, 1H), 2.10 (s, 3H).

[0092] 6-Acetamido-1H-benzimidazole was selected for further characterization and in vivo validation. NMR results fully demonstrated the accuracy and efficiency of the 6-acetamido-1H-benzimidazole synthesis. Figure 25 6-Acetamido-1H-benzimidazole is a thiophene-containing amide derivative with the molecular formula C1. 11 H 10N2OS, with a molecular weight of 218.28 Da ( Figure 26 A). This compound exhibits excellent physicochemical properties, including a QED score of 0.83 indicating high drug-likeness, good lipophilicity (SLogP = 2.59), and an appropriate molecular size and polarity suitable for cell penetration (TPSA = 55.12 Å). 2 () Figure 26 B). Importantly, the compound fully complies with Lipinski's five rules, indicating good potential for oral bioavailability.

[0093] The meanings of the above-mentioned physicochemical property indicators are as follows: QED (Quantitative Estimate of Drug-likeness) is an indicator for comprehensively evaluating the drug-likeness of a compound. Its value ranges from 0 to 1; a higher value indicates that the compound is closer to the physicochemical properties of marketed drugs. A QED score ≥ 0.5 is generally considered to indicate good drug-likeness. SLogP (lipophilicity parameter) is the logarithm of the octanol-water partition coefficient calculated using the atomic contribution method, reflecting the lipophilicity of the compound. An appropriate SLogP value (usually 1-5) is beneficial for the compound to penetrate cell membranes. TPSA (Topological Polar Surface Area) is the sum of the surface areas of all polar atoms (mainly oxygen and nitrogen atoms) in a molecule, expressed in Å. 2 TPSA is measured in Å units. The lower the TPSA value, the stronger the compound's ability to penetrate cell membranes and the blood-brain barrier. The TPSA of orally administered drugs is typically less than 140 Å. 2 The Lipinski five rules are classic rules for predicting the oral bioavailability of compounds. These rules include: molecular weight ≤ 500 Da, SLogP ≤ 5, number of hydrogen bond donors ≤ 5, and number of hydrogen bond acceptors ≤ 10. Compounds meeting these rules generally have good oral absorption properties. All parameters of molecule 1 are within the ideal range, indicating its potential as a candidate molecule for oral drug delivery.

[0094] To evaluate the therapeutic potential of 6-acetamino-1H-benzimidazole in vivo, this invention established a xenograft model in nude mice using HepG2 hepatocellular carcinoma cells. Animals were randomly assigned to three groups: an untreated control group, a DMSO solvent control group, and a 6-acetamino-1H-benzimidazole treatment group. Tumor volume was monitored during the 25-day treatment period. Figure 26 C). Notably, treatment with 6-acetamido-1H-benzimidazole resulted in significant tumor growth inhibition compared to the two control groups. Although tumors in the control and DMSO groups progressively grew to approximately 400–450 mm by day 25. 3However, the tumors in the 6-acetamido-1H-benzimidazole treatment group remained significantly smaller, reaching only about 100 mm. 3 (P<0.0001).

[0095] Visual examination of the resected tumor on day 25 confirmed the significant antitumor effect of 6-acetamido-1H-benzimidazole. Figure 26 (D) Tumors in the 6-acetamino-1H-benzimidazole treatment group were significantly smaller than those in the control and DMSO groups, with a consistent reduction in tumor volume observed in all animals in the 6-acetamino-1H-benzimidazole treatment group. The comparable tumor size between the untreated control and the DMSO solvent control group suggests that the observed antitumor effect is specifically attributable to 6-acetamino-1H-benzimidazole rather than a non-specific solvent effect. These results demonstrate that pharmacological disruption of the PHB2-Src interaction using 6-acetamino-1H-benzimidazole effectively inhibits hepatocellular carcinoma growth in vivo, validating this interaction as a druggable therapeutic target. Combined with the findings of this invention establishing the mechanism by which Src-mediated PHB2 phosphorylation plays a role in driving metabolic reprogramming and tumor progression, these data support the development of PHB2-Src interaction inhibitors as a novel therapeutic strategy for hepatocellular carcinoma.

Claims

1. A PHB2-Src interaction inhibitor, characterized in that: Includes at least one of the following compounds or their pharmaceutically acceptable salts; The PHB2-Src interaction inhibitor blocks phosphorylation of PHB2 at Y34 and / or Y77 sites mediated by Src kinase.

2. The PHB2-Src interaction inhibitor as described in claim 1, characterized in that: It is 6-acetamido-1H-benzimidazole or a pharmaceutically acceptable salt thereof.

3. A PHB2 mutant, characterized in that: Named PHB2-Y34F / Y77F, it is derived from the mutation of phenylalanine at the Y34 and Y77 sites of PHB2.

4. The use of the PHB2-Src interaction inhibitor as described in claim 1 or 2, or the PHB2 mutant as described in claim 3, in the preparation of anti-hepatocellular carcinoma drugs.

5. The application as described in claim 4, characterized in that: The drug has at least one of the following functions: a) Restore the mitochondrial cristae structure in liver cancer cells; b) Maintain redox homeostasis; c) Inhibit metabolic reprogramming.

6. An anti-liver cancer drug, characterized in that: The active ingredient is the PHB2-Src interaction inhibitor as described in claim 1 or 2 and / or the PHB2 mutant as described in claim 3.

7. The anti-liver cancer drug as described in claim 6, characterized in that: The anti-liver cancer drugs also include pharmaceutically acceptable carriers, excipients, or solvents.

8. The anti-liver cancer drug as described in claim 6 or 7, characterized in that: The drug is an oral or injectable formulation.

9. A method for inhibiting the proliferation of liver cancer cells in vitro, characterized in that... include: By using PHB2-Src interaction inhibitors to block Src kinase-mediated phosphorylation of PHB2 at Y34 and / or Y77 sites in cultured hepatocellular carcinoma cells, the interaction between PHB2 and cardiolipin in hepatocellular carcinoma cells is maintained; the PHB1 / 2 complex and PHB1 / 2 supercomplex are stabilized to prevent their disintegration. Inhibiting abnormal activation of OMA1 protease to reduce excessive cleavage of OPA1; maintaining the structural integrity of mitochondrial cristae in hepatocellular carcinoma cells, restoring electron transport chain function, and increasing the NAD⁺ / NADH ratio; reversing the metabolic shift from oxidative phosphorylation to glycolysis; and ultimately inhibiting the proliferation of hepatocellular carcinoma cells in vitro.

10. Application of PHB2-Y34F / Y77F mutants or Src-PHB2-cardiolipin regulatory axis as targets in the screening or preparation of anti-hepatocellular carcinoma drugs.

Images