Aspartate beta-hydroxylase promotes cholangiocarcinoma progression by modulating RB1 phosphorylation

Abstract

Cholangiocarcinoma (CCA) is a highly lethal and aggressive disease. Recently, IDH1/2 mutations have been identified in approximately 20% of CCAs which suggests an involvement of 2-oxoglutarate (2-OG) -dependent dioxygenases in oncogenesis. We investigated if the 2-OG dependent dioxygenase, aspartate beta-hydroxylase (ASPH) was important in tumor development and growth. Immunoassays were used to clarify how ASPH modulates CCA progression by promoting phosphorylation of the retinoblastoma protein (RB1). A xenograft model was employed to determine the role of ASPH on CCA growth. Knockdown of ASPH expression inhibited CCA development and growth by reducing RB1 phosphorylation. Expression of ASPH promoted direct protein interaction between RB1, cyclin-dependent kinases, and cyclins. Treatment with 2-OG-dependent dioxygenase and ASPH inhibitors suppressed the interaction between RB1 and CDK4 as well as RB1 phosphorylation. Knockdown of ASPH expression inhibited CCA progression and RB1 phosphorylation in vivo and they were found to be highly expressed in human CCAs. Knockdown of ASPH expression altered CCA development by modulating RB1 phosphorylation, as one of the major factors regulating the growth of these tumors.

Keywords: 2-Hydroxyglutarate; Alpha-ketoglutarate; Bile duct tumors; Cancer metabolism; IDH1 mutations.

Figure 1. The 2-OG-dependent dioxygenases are involved in cell growth of CCAs with wildtype IDH1/2

(A) Genomic sequences of HEK-293T cells and CCA cell lines, including H1, SSP25, ETK1, NEC, and TFK1. (B) Treatments with an iron chelator (DFO) and a 2-OG antagonist (DMOG) inhibited cell growth of CCA cells with wildtype IDH1/2, including H1 and SSP25. (C) Challenge with the IDH1^mut oncometabolite, (R)2-HG reduced cell growth of CCAs. *, p<0.05; **, p<0.01; ***, p<0.001.

Figure 2. Targeting the 2-OG-dependent dioxygenase, ASPH, suppressed growth, cell cycle progression and inhibited senescence of CCA cell lines

(A) Cell numbers were counted at day 4 in 3 CCA cell lines transduced with shRNA-luciferase (shLuc) or shRNA-ASPH (shASPH), including H1 and SSP25 as well as RBE with the IDH1^mut sequences. (B) Senescence associated β-gal expression and staining was used to evaluate senescence in CCA cell lines treated as indicated. (C) Cell cycle progression was analyzed in CCA cell lines 48 hours post sub-culture by using flow cytometry. *, p<0.05; **, p<0.01; ***, p<0.001.

Figure 3. Expression of ASPH is correlated with protein phosphorylation of RB1

(A) Immunoblotting blot (IB) results of ASPH, pRB1^s807, pRB1^s608, pRB1^s780, RB1, CDK2, CDK4, CDK6, cyclin D1, cyclin E, and tubulin were determined in H1, SSP25, and RBE CCA cells transduced with shLuc and shASPH. Knockdown of ASPH decreased RB1 phosphorylation. (B) Ectopic expression of ASPH for 48 hours promoted RB1 phosphorylation.

Figure 4. ASPH co-localizes and directly interacts with RB1 protein

(A) Immunofluorescence staining of RB1 and ASPH in HEK-293T cells transfected with myc-tagged ASPH and HA-tagged RB1 for 48 hours. Green fluorescence indicates the ASPH signal, red fluorescence shows RB1 localization, and DAPI is used for nuclear staining. (B) Immunohistochemical staining of ASPH and pRB^s780 was performed in human CCA tissue. The blue arrow heads indicate co-localization of ASPH and RB1. (C) IB results of myc-tagged ASPH and HA-tagged RB1 were determined in the IP products of HEK-293T cells 48 hours post transfection as indicated. HA-RB1, myc-ASPH, and tubulin were determined in the whole cell lysate (WCL) for ensuring even protein loading of inputs. (D) Knockdown of ASPH suppresses the interaction between RB1 and ASPH. ASPH, RB1, and tubulin were evaluated in the WCL of RBE cells treated with shLuc or shASPH. (E) Illustration of RB1 protein domains. (F) IB results of myc-tagged ASPH and HA-tagged RB1 in the IP products of HEK-293T cells transfected with different RB1 mutant plasmids. (G) Illustration of ASPH protein domains. (H) IB results of myc-ASPH were determined in the IP products of HEK-293T cells transfected as indicated. RB1, myc-ASPH, myc-ASPH variant 3 (ASPH^V3), myc-ASPH variant 4 (ASPH^V4), and tubulin were measured in the WCL.

Figure 5. The enzymatic activity of ASPH is involved in the protein-protein interaction between RB1 and CDK complexes

(A) IB results of CDK2, CDK4, and RB1 were determined in the IP products of HEK-293T treated as indicated. CDK2, CDK4, and ASPH were measured in the WCL. (B) IB results of cyclin D1, cyclin E, and RB1 were determined in HEK-293T cells treated as indicated. GAPDH, cyclin D1, cyclin E, and RB1 were measured in the WCL. (C) HA-RB1 and CDK4 were determined in the IP products of HEK-293T cells transfected with HA-RB1 in combination with EV, myc-ASPH, or myc-ASPH variant 3 (Myc-ASPH^v3) which does not contain the enzymatic domain of ASPH. Myc tag, CDK4, and tubulin were measured in the WCL. The WCL was collected from HEK-293T cells 48 hours post sub-culture. (D) ASPH, HA-RB1, and CDK4 were analyzed in the IP products of HEK-293T cells transfected with HA-RB1 in combination with myc-EV, myc-ASPH, or myc-ASPH^H675R which only has 20 % of enzymatic activity. The WCL was collected 48 hours post transfection. (E) The effects of DFO, DMOG, and specific ASPH inhibitor, MO-I-1151 on the protein interaction between CDK4 and RB1. HA-RB1 and CDK4 were measured in the IP products of HEK-293T transduced with myc-EV or myc-ASPH. HA-RB1 was transiently transfected in the HEK-293T-EV or HEK-293T-ASPH for 48 hours, and the treated HEK-293T cells were sub-cultured. 48 hours later, the WCL was collected from the treated HEK-293T cells which were challenged with different concentrations of DMOG, DFO, and MO-I-1151 24 hours before harvesting.

Figure 6. The ASPH specific inhibitor, MO-I-1151 suppressed RB1 phosphorylation without affecting histone methylation in contrast to other inhibitors of 2-OG-dependent dioxygenases

The pRB1^s780, RB1, and α-tubulin proteins were determined in serum starved H1 cells which were harvested 4 hours post challenges with (A) 0.5mM, 1mM DMOG, 50µM, 100µM DFO, (B) 5µM, or 10µM MO-I-1151. (C) pRB1^s780, RB1, IDH1, flag-tag, and GAPDH were measured in HEK-293T cells transfected with EV, flag-IDH1, flag-IDH1 mutated (IDH1^mut). (D) ASPH, H3K9Me2, H3K4Me3, and H3 were determined in H1 cells transduced with shLuc or shASPH. The expression levels of H3K9Me2, H3K4Me3, and H3 were evaluated in H1 cells treated with (E) 5, 10 µM MO-I-1151, (F) 0.5, 1mM DMOG, 50, or 100µM DFO for 24 hours. (G) The results of H3K9Me2, H3K4Me3, flag-IDH1, and GAPDH in HEK-293T cells transfected with EV, IDH1, or IDH1^mut for 24 hours were shown. (H) A cartoon illustrates the actions of targeting ASPH by different strategies in the protein interaction between RB1 and CDK complexes (RB1^com) as well as RB1 phosphorylation.

Figure 7. The correlations between ASPH, phosphorylated RB1, and CDKs in CCAs

(A) Relative cell growth rates were determined in H1-shLuc and H1-shASPH transduced with plenti-myc-EV, plenti-myc-CDK2, or plenti-myc-CDK4 at days 0, 1, 3, and 5. (B) ASPH, pRB1^s780, pRB1^s608, RB1, CDK2, CDK4, myc, and α-tubulin were determined in H1 cells treated as indicated. (C) Tumor growth of H1-shLuc and H1-shASPH was evaluated in vivo by calculating tumor volumes of subcutaneous tumors implanted into nude mice. (D) The representative IB results of ASPH, pRB^s780, pRB^s608, and α-tubulin were obtained in the WCL harvested from the tumors of experimental nude mice. (E) The representative images (100× and 400×) of ASPH, pRB^s608, and pRB^s780 are shown in 2 cases of human patients with CCA tumors. (F) The intensities of nuclear ASPH, cytoplasmic ASPH, nuclear pRB1^s780, cytoplasmic pRB1^s780, nuclear pRB1^s608, and cytoplasmic pRB1^s608 were determined in human CCA samples, n=163. *, p<0.05; **, p<0.01; ***, p<0.001. The student t test was used to analyze statistical difference for (A) and (B). The Kendall tau-b rank test was used to evaluate if ASPH, pRB1^s608, and pRB1^s780 have significant correlation in human CCA tumors tested for (F).