Aspartate β-hydroxylase modulates cellular senescence through glycogen synthase kinase 3β in hepatocellular carcinoma

Abstract

Aspartate β-hydroxylase (ASPH) is an enzyme overexpressed in human hepatocellular carcinoma (HCC) tumors that participates in the malignant transformation process. We determined if ASPH was a therapeutic target by exerting effects on cellular senescence to retard HCC progression. ASPH knockdown or knockout was achieved by short hairpin RNAs or the CRISPR/Cas9 system, respectively, whereas enzymatic inhibition was rendered by a potent second-generation small molecule inhibitor of ASPH. Alterations of cell proliferation, colony formation, and cellular senescence were evaluated in human HCC cell lines. The potential mechanisms for activating cellular senescence were explored using murine subcutaneous and orthotopic xenograft models. Inhibition of ASPH expression and enzymatic activity significantly reduced cell proliferation and colony formation but induced tumor cell senescence. Following inhibition of ASPH activity, phosphorylation of glycogen synthase kinase 3β and p16 expression were increased to promote senescence, whereas cyclin D1 and proliferating cell nuclear antigen were decreased to reduce cell proliferation. The mechanisms involved demonstrate that ASPH binds to glycogen synthase kinase 3β and inhibits its subsequent interactions with protein kinase B and p38 upstream kinases as shown by coimmunoprecipitation. In vivo experiments demonstrated that small molecule inhibitor treatment of HCC bearing mice resulted in significant dose-dependent reduced tumor growth, induced phosphorylation of glycogen synthase kinase 3β, enhanced p16 expression in tumor cells, and promoted cellular senescence.

Conclusions: We have identified a new mechanism that promotes HCC growth and progression by modulating senescence of tumor cells; these findings suggest that ASPH enzymatic activity is a novel therapeutic target for HCC.

Figure 1. Effects of knockdown and knockout of ASPH on cell proliferation, anchorage-independent colony formation and cellular senescence in human HCC cell lines

The following assays were performed in human HCC cell lines expressing the control (shLuc) vs. knockdown of ASPH (shASPH) (FOCUS and Huh7), as well as control (EV) vs. knockout of ASPH (KO ASPH) (HepG2). (A) Protein expressions of ASPH in human HCC cell lines with knockdown (shASPH) or knockout of ASPH (KO ASPH) compared to the control (shLuc or EV). MTT assays showed relative absorbance at each time point over 5 days. (B) Representative photographs of anchorage-independent colony formation (left panel, 100×). The bar graphs indicated relative colony number (right panel). (C) Representative photographs of SA-β-gal staining (left panel, 200×). Senescent cells exhibited blue staining. Senescence under each condition was quantified and expressed as percent of SA-β-gal-positive cells (right panels). Mean ± S.D. of triplicate independent experiments. *P < 0.05, **P < 0.01 compared to the control cells (shLuc or EV).

Figure 2. Characterization and Evaluation of anti-tumor effects of a SMI of β-hydroxylase (MO-I-1151) on the HCC phenotype

(A) MTT assays showed relative absorbance at each concentration (ranging from 1.25 to 10 µM) of MO-I-1151, MO-I-1100 and other candidate β-hydroxylase inhibitors (MO-I-1171, 1178, 1351, 2751, 5007, 5008, 5009, 5010, 5018 and 5019) which had no biologic activity. The assay was performed after 24-hour’s exposure to DMSO or SMIs in FOCUS HCC cells. [Blue line was MO-I-1000 (the first generation SMI). red line was MO-I-1151 (the second generation SMI). Black lines represented other compounds that had no effect.] (B) MTT assays showed relative absorbance at each time point exposed to DMSO or different concentrations of MO-I-1151 (from 1.25 to 10 µM) for 5 days. (C) The bar graphs indicated relative colony numbers after 3-week exposure to DMSO or MO-I-1151 (2.5 or 5 µM). Representative photographs of anchorage-independent colony formation were shown. (D) Representative photographs of SA-β-gal staining after 24-hour’s exposure to DMSO or MO-I-1151 (5 µM) (left panel, 200×). The SA-β-gal activity was quantified and expressed as percent of SA-β-gal-positive cells (right panel). Mean ± S.D. of triplicate independent experiments. *P < 0.05, **P < 0.01 compared to DMSO treated cells.

Figure 3. The β-hydroxylase inhibitor MO-I-1151 was specific for ASPH mediated biologic activity

The following assays were performed in human HCC cell lines stably expressing the control (shLuc) vs. knockdown of ASPH (shASPH) (FOCUS and Huh7), as well as the control (EV) vs. knockout of ASPH (KO ASPH) (HepG2). (A) MTT assays showed relative absorbance after 24-hour exposure to each concentration (ranging from 1.25 to 10 µM) of MO-I-1151 or DMSO. (B) SA-β-gal staining was performed after 24-hour exposure to DMSO or MO-I-1151 (5 µM). Representative photographs of SA-β-gal staining (top, 200×). Senescence under each condition was quantified and expressed as percentage of SA-β-gal-positive cells (bottom). Mean ± S.D. of triplicate independent experiments. **P < 0.01 compared to the control cells (shLuc or EV) (A) or DMSO treated control cells (shLuc or EV) (B).

Figure 4. ASPH-related cellular senescence was mediated through GSK3β in human HCC cells

(A, B) Protein expressions of ASPH, GSK3β, phosho-GSK3β, GS, phospho-GS, p16, cyclin D1 and PCNA were evaluated by Western blot analysis. (A) In the control (shLuc) or knockdown of ASPH (shASPH) human HCC cell lines (FOCUS and Huh7). (B) After 24-hour exposure to DMSO or different concentrations of MO-I-1151 (from 1.25 to 5 µM) in human HCC cell lines (FOCUS and Huh7). (C– F) The following assays were performed in the control (shLuc) or knockdown of ASPH (shASPH) in Huh7 cells with or without overexpression of CA-GSK3β. (C) MTT assay showed relative absorbance at each time point over 5 days. (D) The bar graphs represented relative colony numbers (top). Representative photographs of anchorage-independent colony formation are shown (bottom, 100×). (E) Representative photographs of SA-β-gal staining (top, 200×). SA-β-gal activity was quantified and expressed as percent of SA-β-gal-positive cells (bottom). (F) Protein expression of ASPH, GSK3β, phosho-GSK3β, GS, phospho-GS, p16, cyclin D1 and PCNA were evaluated by Western blot after 48-hour of co-transfection in Huh7 cells. Mean ± S.D. of triplicate independent experiments. **P < 0.01 compared to shLuc + EV or shASPH + EV cells.

Figure 5. ASPH inhibited phosphorylation of GSK3β from its upstream kinases by directly binding to GSK3β

The following assays were performed in HEK293 cells after 48-hour co-transfection of (A, B) HA-tagged ASPH with either WT-GSK3β or EV plasmids; as well as (C, D) WT-GSK3β with either HA-tagged ASPH or EV plasmids. (A) Protein expressions of HA, ASPH and GSK3β were confirmed by Western blot analysis, in co-transfected HEK293 cells with HA-tagged ASPH and either WT-GSK3β or EV. Antibody against α-tubulin was used as a loading control. (B) After co-immunoprecipitation using control IgG or HA-probe, immunoblotting with HA-probe and antibody against GSK3β were performed. (C) Protein expressions of HA, ASPH, GSK3β and phospho-GSK3β were confirmed by Western blot analysis, in co-transfected HEK293 cells with WT-GSK3β and either HA-tagged ASPH or EV. Antibody against α-tubulin was used as a loading control. (D) After co-immunoprecipitation using control IgG or anti-GSK3β antibody, immunoblotting with HA-probe and antibodies against AKT, p38 and GSK3β are shown.

Figure 6. Anti-tumor effects of targeting ASPH in subcutaneous and orthotopic xenograft models generated by FOCUS HCC cells

(A) Mean tumor volume was measured and plotted twice per week in each concentration of MO-I-1151 treatment (ranging from 0.1 to 10 mg/kg), compared to DMSO (n = 8). (B) Representative photographs of intrahepatic inoculated tumors 4 weeks after initiation of MO-I-1151 treatment, compared to DMSO as control (left panel). Mean tumor weight was analyzed at 4 weeks after initiation of MO-I-1151 treatment, compared to DMSO (n = 5) (right panel). (C) Protein expression of GSK3β, phospho-GSK3β and p16 in tumor lysates treated with MO-I-1151, compared to DMSO, was evaluated by Western blot (left panel). Ratio of the relative density of p-GSK3β/GSK3β or p16/Actin ratio (right panel). (D) Representative photographs of hematoxylin and eosin staining (top) and SA-β-gal staining (bottom) performed in serial sections of frozen tumor tissue samples treated with MO-I-1151, compared to DMSO (left panel, 400×). Senescence in the tumor sample was quantified and expressed as average number of SA-β-gal-positive cells (right panel). Mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.005 compared to the DMSO treated group.

Figure 7. Hypothesis of how ASPH may mediate senescence of HCC cells

(A) ASPH overexpression is followed by binding to GSK3β (purple) and inhibits the interaction and subsequent phosphorylation of GSK3β by AKT and p38 upstream kinases (blue). This mechanism keeps GSK3β constitutive active (orange) and leads HCC progression by promoting degradation of the cyclin-dependent kinase inhibitor p16 (yellow). (B) Under conditions of Inhibition such as ASPH knockdown or modulation of enzymatic activity by SMIs (MO-I-1151), AKT and p38 upstream kinases can now phosphorylate and inactivate GSK3β function (gray). The net results are stabilization of p16, and promotion of HCC cell senescence.