VRK2 targeting potentiates anti-PD-1 immunotherapy in hepatocellular carcinoma through MYC destabilization

Abstract

Dysregulation of MYC proto-oncogene, bHLH transcription factor (MYC) represents a common yet mechanistically unresolved driver of hepatocellular carcinoma (HCC). While MYC remains an elusive therapeutic target, developing strategies to promote its degradation emerges as a promising alternative approach. Here we show that vaccinia-related kinase 2 (VRK2) functions as a direct MYC-interacting kinase that stabilizes the oncoprotein through phosphorylation at Serine (Ser)281/293. This phosphorylation enables VRK2 to compete with the Skp1-Cullin-F-box protein complex containing FBXO24 (SCF-FBXO24) E3 ligase, thereby blocking MYC polyubiquitination and proteasomal degradation. The stabilized MYC-VRK2 complex amplifies transcriptional activation of protumorigenic programs, including the immune checkpoint programmed cell death ligand 1 (PD-L1) and VRK2 itself, establishing a self-reinforcing oncogenic circuit. Therapeutic inhibition of VRK2 in HCC models reduces MYC protein levels, suppresses tumor progression, and synergizes with anti- programmed cell death-1 (PD-1) immunotherapy. Our results reveal VRK2-mediated stabilization of MYC as a critical nexus linking hepatocarcinogenesis to immune evasion, proposing VRK2 kinase inhibition as a mechanism-based therapeutic strategy for MYC-driven HCC.

Fig. 1. Identification of VRK2 as a MYC binding partnerVRK2 and an oncogene associated with unfavorable prognosis in patients with HCC.

a Sliver-stained SDS-PAGE gel of the Co-IP products (n = 3 independent experiments). b Pearson’s correlation analysis of p-values of kinases in mass spectrometry and prognostic scores of kinases in TCGA. c The peptide fragment of VRK2 as determined by mass spectrometry. d Immunoprecipitation assay of interaction between endogenous MYC and VRK2 in MHCC97H cells were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). e The co-localization analysis for MYC and VRK2 in MHCC97H cells. Scale bar: 20 μm. The nucleus was stained with DAPI (n = 3 independent experiments). f, g The mRNA level of VRK2 in LIHC tumor stage and normal tissues in a TCGA-LIHC (RNA-seq data) (f) and LIRI-JP dataset (RNA-seq data) (g). h The mRNA level of VRK2 in LIHC was grouped according to pathological differentiation, vascular invasion in a TCGA-LIHC (RNA-seq data). i Representative IHC images of VRK2 staining of clinical HCC and paracancerous tissues (n = 123). Scale bar: 100 μm (left)/25  μm (right) (n = 3 independent experiments). j VRK2 differentially expressed in HCC and paracancerous tissues from cohort 2 (n = 120) for western blotting with the indicated antibodies, and Relative VRK2 protein level in HCC tissues (T) by comparing to their counterpart paracancerous tissues (P) after normalizing to GAPDH expression (n = 3 independent experiments). k Overall survival according to VRK2 mRNA levels expression in the TCGA-LIHC cohort, LIRI-JP cohort and Tongji cohort 1. l Progression-free survival according to VRK2 mRNA levels expression in the TCGA-LIHC cohort, LIRI-JP cohort and Tongji cohort 1. m Heatmap (left) of clinicopathologic characteristics of TCGA-LIHC HCC patients grouped by VRK2 expression (left panel). Forest plot (right panel) of the univariate and multivariate Cox proportional hazards model for overall survival. For Western blot experiments, GAPDH was used as a loading control (j). f–h P values were calculated using a student’s unpaired t-test (two-tailed). Data were presented as the mean ± SD. i, j P values were calculated using a student’s paired t-test (two-tailed). k, l P values were calculated using a Kaplan–Meier test (two-tailed). Source data are provided as a Source Data file.

Fig. 2. VRK2 promotes HCC tumorigenesis and progression in vitro and in vivo.

a Xenograft tumors from Hep3B cells (vector/VRK2-WT/VRK2-MUT) via subcutaneous injection in mice (n = 5 mice in each group). Scale bar: 1 cm. b Xenograft tumors from MHCC97H cells (sgCON/sgVRK2/sgVRK2 + VRK2 WT/sgVRK2 + VRK2 MUT) via subcutaneous injection in mice (n = 5 mice in each group). Scale bar: 1 cm. Tumor volume growth curves (middle), and tumor weight (bottom) were also quantified and compared. c Liver orthotopic models (n = 5 mice in each group) with Hep3B cells (vector/VRK2-WT/VRK2-MUT) and MHCC97H cells (sgCON/sgVRK2/sgVRK2±WT/MUT). Showing bioluminescence, gross liver/tumor images, and H&E-stained sections. Gross image Scale bar: 1 cm, H&E image Scale bar: 200 μm (left)/50 μm (right). d Bioluminescence intensity statistics of a nude mouse liver orthotopic transplantation model (n = 5 mice in each group). e Tumor volume of liver orthotopic transplantation models (n = 5 mice in each group). f Schematic diagram of different combinations (NC/VRK2 WT/VRK2 MUT) of oncogenic plasmids with luciferase injected into mice. g HTVi models (n = 7 mice in each group) showing bioluminescence imaging, gross liver/tumor images, and H&E-stained liver sections. Gross image Scale bar: 1 cm, H&E image Scale bar: 200 μm (up)/50 μm (bottom). h Liver/body weight ratio in the HTVi model (up). Bioluminescence intensity statistics of HTVi models (bottom) (n = 7 mice in each group). i Survival statistics of HTVi models. j Schematic representation of the DEN/CCL₄ HCC development model induced in VRK2 liver-specific knockout C57BL/6 J mice. k DEN/CCl₄-induced HCC models (n = 7 mice in each group) showing MRI, gross liver/tumor images, and H&E-stained liver sections. Gross image Scale bar: 1 cm, H&E image Scale bar: 200 μm (up)/50 μm (bottom). l Liver/body weight ratio in the DEN/CCL₄ model (up). Tumor volume of DEN/CCL₄ models (bottom) (n = 7 mice in each group). m Survival statistics of DEN/CCL₄ models (n = 7 mice in each group). a, b, d, e, h, l P values were calculated using a student’s unpaired t-test (two-tailed). Data were presented as the mean ± SD. i, m P values were calculated using a Kaplan–Meier test (two-tailed). Source data are provided as a Source Data file.

Fig. 3. VRK2 enhances the transcriptional activity of MYC by facilitating the phosphorylation of MYC at S281 and S293.

a sgCON or sgVRK2 MHCC97H cells were harvested for RNA-seq analysis and pathway analysis. Differentially regulated genes with more than twofold change were included in this pathway analysis (n = 3 independent experiments). b VRK2-related gene set enrichment analysis (GSEA). c, d E-box activity changes in Hep3B, HLF, MHCC97H, and HepG2 cells after VRK2 overexpression or knockout (n = 3 independent experiments). e HLF and MHCC97H cells after VRK2 overexpression or knockout were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). f Phos-tag^TM SDS-PAGE analysis of MYC extracted from MHCC97H cells treating with CIAP (n = 3 independent experiments). g LC-MS/MS spectrum showing the phosphorylation of MYC S281 and S293. h Schematic diagram showing the phosphorylation site of MYC detected by LC-MS/MS (up), and the MYC amino acidic sequence near S281 and S293 from various species (bottom). i Molecular docking model of VRK2 and MYC. j Immunoprecipitation assay of interaction between mutant HA-MYC and Flag-VRK2 in HEK293T cells with MG132 were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). k Vitro kinase assay of the protein interaction between GST-MYC WT, GST- MYC S281A or GST- MYC S293A, and Flag-VRK2 WT or Flag-VRK2 MUT. GST coding sequences were cloned onto the N-terminus of MYC. Followed by Western blot analysis with the indicated antibodies (n = 3 independent experiments). l MHCC97H cells with interference with VRK2 expression using siRNA. were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). For Western blot experiments, GAPDH was used as a loading control (e, j, i). a P values were calculated using standard accumulative hypergeometric statistical test. c, d P values were calculated using a student’s unpaired t-test (two-tailed). Data were presented as the mean ± SD. Source data are provided as a Source Data file.

Fig. 4. VRK2 reduces FBXO24-mediated polyubiquitination degradation of the MYC K323 by activating phosphorylation of MYC at 281 and 293.

a, b CHX chase assays in Hep3B (VRK2 overexpression) and MHCC97H (VRK2 knockout) cells treated with CHX (20 μM). Quantified MYC levels normalized to GAPDH (bottom). c Co-IP of VRK2, MYC, and Ub in HLF cells overexpressing VRK2 and transfected with MYC-tagged Ub plasmids, with MG132 (10 μM, 4 h). d Co-IP of Flag-VRK2, HA-MYC, and MYC-tagged Ub K48 in HEK293T cells (MG132-treated). e Western blot analysis of HA-MYC in HEK293T were transfected with Flag-FBXO24 WT, Flag-FBXO24^ΔF-box and HA-MYC plasmids. f, g Immunoprecipitation assay of interaction among MYC, VRK2, or FBXO24 in HLF (up) and MHCC97H cells (bottom). h Co-localization analysis of MYC and FBXO24 in MHCC97H cells. Scale bar: 20 μm. The nucleus was stained with DAPI (n = 3 independent experiments). i Western blot analysis of VRK2, MYC, and FBXO24 in knockout VRK2 MHCC97H cells with interference with FBXO24 expression using siRNA. j Co-IP of MYC, VRK2, or FBXO24 in HEK293T with CIAP treatment. k Co-IP of Flag-VRK2, HA-FBXO24, and MYC-tagged MYC in HEK293T (MG132-treated). l Co-IP of VRK2, FBXO24, and MYC in HLF cells overexpressing oe-VRK2 WT/MUT (MG132-treated). m Co-IP of VRK2, FBXO24, and MYC in MHCC97H cells with sgCON/sgVRK2 (MG132-treated). n–p Co-IP of Flag-FBXO24, MYC-tag Ub, HA-MYC WT or HA-MYC S281A/S281D (n) or HA-MYC S293A/S293D (o) or HA-MYC S281A + S293A(2SA)/S281D + S293D(2 SD) (p) in HEK293T cells and incubated with MG132 (10 μM) for 4 h. For Western blot experiments, GAPDH was used as a loading control (a–e, i–p). The cells involved in the above experiments were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). Source data are provided as a Source Data file.

Fig. 5. VRK2 regulates MYC target genes and promotes HCC progression by phosphorylating MYC at S281 and S293.

a Venn plots of RNA-seq and CUT-Tag shows the number of differential gene results for si-VRK2 and si-MYC in MHCC97H cells. b qRT-PCR analysis of MYC targets genes in MHCC97H cells after gradient transfection of the si-NC, si-VRK2, si-MYC, si-MYC + MYC WT, si-MYC + MYC S281A, si-MYC + MYC S281D, si-MYC + MYC S293A or si-MYC + MYC S293D siRNA or plasmids (n = 3 independent experiments). c, d Heatmaps (c) and averaged plots (d) of MYC CUT-tag signal intensities (normalized against spike-in control and sequencing depth) around peak centers in MHCC97H cells. e IGV views of MYC at PD-L1 (left) and CCND1 (right) in MHCC97H cells. f Luciferase activities of VRK2 promoter reporter vectors in MHCC97H cells after MYC overexpression or silence (n = 3 independent experiments). g CUT-RUN assays followed by qRT-PCR and analysis was conducted in MHCC97H cells (n = 3 independent experiments). h Schematic diagram of different combinations (MYC WT/ MYC S281A/ MYC S281D/ MYC S293A/ MYC S293D) of oncogenic plasmids (Myc/sgTp53/SB100) with luciferase injected into C57BL/6 J mice (n = 7 mice in each group). i Representative images are presented that show the liver with tumor images. Gross image Scale bar: 1 cm, H&E image Scale bar: 200 μm. j Tumor numbers of HTVi models (n = 7 mice in each group). b, f, j P values were calculated using a student’s unpaired t-test (two-tailed). Data were presented as the mean ± SD. Source data are provided as a Source Data file.

Fig. 6. AZD7762, a small molecular inhibitor of VRK2, can reduces HCC growth.

a Molecular docking of AZD7762 with VRK2 (PDB:2V62). b IC50 assay demonstrated high inhibitory potency and selectivity of AZD7762 for VRK2 (n = 6 independent experiments). c Enzyme activity assay confirmed VRK2 selectivity (n = 6 independent experiments). d In vitro kinase assay showed AZD7762 inhibited MYC phosphorylation, proteins were harvested for western blotting with the indicated antibodies (n = 6 independent experiments). e CCK-8 assays in VRK2-knockout MHCC97H cells treated with DMSO/AZD7762 for 12 h (n = 6 independent experiments). f Representative images of colony formation and quantification (left/right panel) post-treatment (n = 6 independent experiments). g Representative EdU assay image of MHCC97H cells after VRK2 knockout, which were treated with DMSO or AZD7762 for 12 h and quantification (left/right panel) post-treatment (n = 6 independent experiments). h qRT-PCR analysis of PD-L1 expression, which was treated with DMSO or AZD7762 for 12 h (n = 3 independent experiments). i Western blot analysis of PD-L1 expression, which were treated with DMSO or AZD7762 for 12 h, cells were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). j Flow cytometry analysis of PD-L1 expression, which was treated with DMSO or AZD7762 for 12 h (n = 3 independent experiments). k Schematic of DEN/CCL₄-induced HCC model in WT C57BL/6J mice (n = 7 mice in each group) and dosing strategy. l Liver tumor images: H&E, Oil Red, Ki67, PD-L1 and CD8 staining of isolated liver tissues. Gross image Scale bar: 1 cm, IHC image Scale bar: 200 μm. m Western blot analysis in mice HCC tissues. After prophylactic and therapeutic of AZD7762, cells were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). n Tumor volume of the indicated models (n = 7 mice in each group). o Tumor numbers of indicated models (n = 7 mice in each group). p, q Serum AST/ALT levels in models (n = 7 mice in each group). r Body weight measured. After prophylactic and therapeutic of AZD7762 indicated models (n = 7 mice in each group). For Western blot experiments, GAPDH was used as a loading control (i, m). c, e, g, h, n–r P values were calculated using a student’s unpaired t-test (two-tailed). Data were presented as the mean ± SD. Source data are provided as a Source Data file.

Fig. 7. Blockade of VRK2 enhances the therapeutic effect of Anti-PD-1mAb in a mouse model of HCC.

a Schematic representation of the DEN/CCL₄ HCC development model induced in WT C57BL/6J mice (n = 5 mice in each group) and drug delivery strategy. b DEN/CCL₄ model in C57BL/6J mice with representative images are presented that show the MRI images, liver with tumor images and H&E, Oil Red, Ki67, PD-L1, and CD8 staining of isolated liver tissues. Gross image Scale bar: 1 cm, IHC image Scale bar: 200 μm. c Schematic representation of the HTVi HCC development model (Myc/sgTp53) induced in WT C57BL/6J mice (n = 5 mice in each group) and drug delivery strategy. Gross image Scale bar: 1 cm, IHC image Scale bar: 200 μm. d HTVi model in C57BL/6J mice with representative images are presented that show the bioluminescence, liver with tumor images and H&E, Oil Red, Ki67, PD-L1, and CD8 staining of isolated liver tissues. e, f Flow cytometry analysis of CD3⁺, CD8⁺, CD69⁺CD8⁺, IFN-γ⁺CD8⁺, and PD-1⁺CD8⁺ in TILs from DEN/CCL₄ model or HTVi model tumors tissues in C57BL/6J mice (n = 5 mice in each group). e, f P values were calculated using a student’s unpaired t-test (two-tailed). Data were presented as the mean ± SD. Source data are provided as a Source Data file.

Fig. 8. VRK2-MYC-PD-L1 axis in patients with HCC.

a Western blot analysis o in 50 HCC and paracancerous tissues from cohort 3 patients, cells were harvested for western blotting with the indicated antibodies (n = 3 independent experiments). b Correlation analysis of the MYC, p-MYC S281, p-MYC S293, and VRK2 protein levels in HCC cohort 3. c Representative IHC images of VRK2, MYC and PD-L1 staining of clinical HCC cohort 1. Scale bar: 100 μm (left)/25 μm (right). d, e IHC staining score of PD-L1, MYC, and VRK2 has a high correlation in HCC cohort 1. f mIHC staining analysis of PD-L1, MYC, VRK2, CD8, and GZMB has a high correlation in HCC cohort 1. Scale bar: 100 μm (left)/25 μm (right). g–j Kaplan–Meier analysis of OS (g, h) and PFS (i, j) for different VRK2/MYC expressions in TCGA-LIHC datasets and Cohort 1. For Western blot experiments, GAPDH was used as a loading control (a). b, e P values were calculated using a Spearman correlation analysis (two-tailed). g–j P values were calculated using a Kaplan–Meier test (two-tailed). Source data are provided as a Source Data file.

Fig. 9. The schematic diagram.

Proposed model underlying the roles of VRK2-MYC-PD-L1 axis in promoting tumor progression and immune evasion in HCC. A schematic diagram was designed using Adobe Illustrator.