Integrated transcriptomic and proteomic analysis identifies protein kinase CK2 as a key signaling node in an inflammatory cytokine network in ovarian cancer cells

Abstract

We previously showed how key pathways in cancer-related inflammation and Notch signaling are part of an autocrine malignant cell network in ovarian cancer. This network, which we named the "TNF network", has paracrine actions within the tumor microenvironment, influencing angiogenesis and the immune cell infiltrate.The aim of this study was to identify critical regulators in the signaling pathways of the TNF network in ovarian cancer cells that might be therapeutic targets. To achieve our aim, we used a systems biology approach, combining data from phospho-proteomic mass spectrometry and gene expression array analysis. Among the potential therapeutic kinase targets identified was the protein kinase Casein kinase II (CK2).Knockdown of CK2 expression in malignant cells by siRNA or treatment with the specific CK2 inhibitor CX-4945 significantly decreased Notch signaling and reduced constitutive cytokine release in ovarian cancer cell lines that expressed the TNF network as well as malignant cells isolated from high grade serous ovarian cancer ascites. The expression of the same cytokines was also inhibited after treatment with CX-4945 in a 3D organotypic model. CK2 inhibition was associated with concomitant inhibition of proliferative activity, reduced angiogenesis and experimental peritoneal ovarian tumor growth.In conclusion, we have identified kinases, particularly CK2, associated with the TNF network that may play a central role in sustaining the cytokine network and/or mediating its effects in ovarian cancer.

Keywords: inflammatory cytokine networks; microenvironment; ovarian cancer; systems biology; therapeutics.

Figure 1. Phosphoprotein profiling of human ovarian cancer cell lines

A. A network of putative constitutively active kinases in IGROV-1 cells was created using Genego's Metacore pathway analysis tool. Interactions between any two objects are depicted by an arrow indicating the direction of interaction. B. Heatmap containing three samples in each group of statistically significantly differentially phosphorylated proteins were generated using IGROV-1 and IGROV-Mock transfected versus two independently isolated clones of IGROV-1 cells stably transfected with shRNA to CXCR4 (I and II). Increases in phosphorylation are shown as pseudocolour red and decreases in yellow. C. Western blot analysis of selected kinases associated with the phenotype of high cytokine expression and downstream effect on phospho-STAT3 and NOTCH signaling in malignant cells.

Figure 2. Summary of compounds whose transcriptional response is similar to that induced by the knock-down of the TNF Network

A. Drug-network based classification using the gene expression signature in shCXCR4 cells. The expression signature of cells in which the TNF network has been inhibited is connected to a given drug if the profile of differential expression following the knock-down is significantly similar to the transcriptional response elicited by that drug in human cancer cell lines. Two drugs are connected to each other if their transcriptional response is similar according to the same criteria. Colors indicate groups of densely interconnected nodes enriched for a given mode of action, common drug targets or therapeutic application. B. Included are documented effects for each drug in light blue and shown in red are the drugs known to target protein kinase CK2. The dark blue bars indicate the extend of similarity between the transcriptional responses elicited by the listed compounds and that of the knock-down of CXCR4 (shCXCR4). This is defined as the inverse of the distance reported in Supplementary Table 1, and normalised in [0,1]. According to this metric the compound eliciting the most similar response to shCXCR4 is 8-azaguanine and CK2 is a recurrent target among those of the top-similar to shCXCR4.

Figure 3. Effects of specific inhibition of CK2 on the cytokine network in EOC

A. mRNA gene expression of CK2 by real time RT-PCR, B. Western blot analysis C. proliferation and D. cytokine production after 48 hours of transient transfection with a pool of 3 oligos targeting CK2 in IGROV-1 and SKOV3ip1 cells (mean ± SEM, **, P<0.01, *, P<0.05). E. Cytokine and growth factor expression levels were measured in cell culture supernatants after 48 hours inhibition of CK2 with 5 μM CX-4945. F. Western blot analysis of constitutive CK2 activity by its substrate phospho-Akt (Ser129) and downstream effects after treatment with CX-4945 in a dose dependent manner after 24hour treatment with and without CX-4945 (mean ± SEM, ***, P<0.001, **, P<0.01, *, P<0.05). G. Cytokine and growth factor expression levels in cell culture supernatants of primary ovarian cancer cells from ascites after 48 hours inhibition of CK2 with 5 μM CX-4945.

Figure 4. Effects on apoptosis and proliferation of HGSC cell lines after specific inhibition of CK2 in vitro

Apoptotic and proliferative effect of CK2 inhibitor CX-4945 on A. IGROV-1 and SKOV3ip1 cells, in a dose dependent manner by Western blot analysis of apoptotic marker cleaved PARP and by cell numbers on plastic over a period of 6 days, and B. representative pictures of H&E stained sections in a 3D model after 6 days treatment, respectively. C. Quantification of cell growth inhibition and D. mRNA expression of cytokines and growth factors after treatment of SKOV3ip1 cells with CX-4945 in the 3D model (mean ± SEM, **, P<0.01, *, P<0.05).

Figure 5. Effects of specific inhibition of CK2 on xenograft growth in vivo

A. Quantification of bioluminescence from tumors (n=6 mice per group) at different time points (*, P <0.05) and B. mRNA expression of cytokines in primary tumors (n=3) by real time RT-PCR after treatment with CK2 inhibitor CX-4945. C. Representative pictures of confocal images (magnification x20) from tumors following injection of TRITC-lectin and Ki67 stained paraffin sections by immunohistochemistry after 42 days treatment with 75mg/kg CX4945. D. Quantification vascular area and proliferative index. Columns, mean in each group in 10 randomly selected areas of tumor sections (mean ± SEM, ***, P<0.001).