UC Davis

ABSTRACT OF DISSERTATIONInvestigating the Effects of a Small-Molecule Allosteric Inhibitor of Core-Binding Factor Subunit Beta in Human Osteosarcoma Osteosarcoma (OS) is the most common primary malignant bone tumor in humans, however, there have been no new successful therapies that improve long-term survival in several decades. This indicates the need for novel therapeutics and targets for treating OS, particularly in reducing its metastatic potential which is the leading cause of osteosarcoma-related deaths. Core Binding Factor is a heterodimeric transcriptional complex, comprised of proteins runt-related transcription factor 2 (RUNX2) and core-binding factor subunit beta (CBFb), that promotes transcription of genes related to bone formation, osteoblast differentiation, and bone mineralization. RUNX2 protein expression is deregulated in OS tumors which contributes to chemoresistance, increased proliferation, and defective differentiation, suggesting it could be a therapeutic target in OS. This study utilizes small molecule allosteric inhibitor of CBFb, AI-14- 91, to disrupt the interaction of RUNX2 and CBFb, thereby reducing RUNX2 binding to target genes, to evaluate the CBFb and RUNX2 interaction as a pharmacological target in OS. We hypothesized that AI-14-91 will reduce the malignant phenotype of human OS cell lines, show favorable pharmacokinetic and pharmacodynamic qualities in vivo, and effectively reduce the development of cell line-derived xenografts (CDXs) in vivo. In vitro cell-based assays were performed with a panel of human OS cell lines treated with AI-14-91. RUNX2 knockout (KO) and CBFb KO cell lines were also generated from the parental U2OS wild-type (WT) by CRISPR/Cas9 editing, to compare the effects of AI-14-91 treatment versus the loss of each component of the Core Binding Factor complex. Concentration- and time- ii dependent proliferation assays were performed with all cell lines, demonstrating AI-14-91 IC50 values ranging from approximately 13-25 μM across the panel. Clonogenic assays showed AI-14- 91 treatment significantly reduced colony formation and colony size in most cell lines. AI-14-91 disrupted the cell cycle progression in some of the cell lines as well with a significant decrease of cell number in G0/G1 phase and an increase in S phase upon 48-hour exposure to 20 μM AI-14- 91. Induction of apoptosis by AI-14-91 was quantified with caspase-3/-7 activity fluorometric assay and western blotting with cleaved caspase-3 and cleaved PARP-1. Apoptotic and necrotic cell populations were evaluated in U2OS WT cells by annexin V/PI staining analyzed by flow cytometry. All of these showed no significant induction of apoptosis. Lastly, a transwell invasion and migration assay demonstrated that AI-14-91-treated LM7 and U2OS cells had reduced invasive ability. These anti-proliferative and -clonogenic effects as well as disruptions in cell cycle were seen in the U2OS WT and both KO cell lines, suggesting off-target effects. Batch 3’ Tag-Seq RNA-seq was performed on parental U2OS WT cells treated with 5 and 20 μM AI-14-91 for 6 and 48 hours, U2OS WT cells treated with DMSO vehicle control, and on untreated RUNX2 KO and CBFb KO cell lines. Differential gene expression (DGE), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, Gene Ontology (GO), and Gene Set Enrichment Analysis (GSEA) were the analyses performed to evaluate pertinent transcriptional pathways altered by inhibitor treatment, and to compare the effects of inhibitor treatment with knockout of putative inhibitor targets. RNA-seq data plotted on multidimensional scaling plots showed that inhibitor treatment produced a unique transcriptional profile when compared to the CBFb KO cell lines or DMSO treated cells. The DGE analysis revealed that loss of CBFb and 6- hour treatment with AI-14-91 produced unique expression changes in U2OS WT cells and that loss of RUNX2 had a smaller effect on gene expression than loss of CBFb. In the AI-14-91 versus iii DMSO comparison, KEGG pathway analysis showed enrichment in cancer pathways where RUNX2 is associated with metastasis to the bone. GO analysis highlighted terms and genes that would support the anti-tumor effects seen in the in vitro experiments such as pathways in regulation of cell migration and mitotic cell cycle. These are further supported by specific changes in expression of genes such as CDKN1C, BRINP1 and BMP4. These pathways that were altered support the anti-tumor properties of AI-14-91 and reduced metastatic potential seen in U2OS WT cells. Lastly, we determined concentrations and pharmacokinetic (PK) parameter values for both serum and lung tissue samples after a single intraperitoneal injection of 100 mg/kg AI-14-91 over a time course of 0-24 hours in nude mice. There were cell line growth-inhibiting concentrations of AI-14-91 found in the blood, exceeding the determined in vitro IC50 values (the highest being 23 μM). This effective concentration was seen in the blood for 1 hour post-injection. The measured concentration in the lung was not as high but reached 10 μM at its peak. Both peaks in concentration were seen with an early time of peak concentration (Tmax) of 15 and 30 min, respectively. The half-life of AI-14-91 in the serum was comparable to that of published literature and exceeded 10 hours in the lung. With a pulmonary metastatic CDX nude mouse model utilizing the LM7 cell line and live bioluminescent imaging, we were unable to demonstrate that a 6-day treatment period with AI-14-91 can reduce the rate of CDX growth. This study is still ongoing at this time and will continue to be monitored for CDX development. If successful future CDX development occurs, the efficacy of AI-14-91 would be evaluated by bioluminescent signal and metastatic burden in the lung along with survival curve analysis. Through these experiments, we have demonstrated that AI-14-91 has desirable anti-tumor properties in multiple OS cell lines, alters gene transcription that would support these properties, iv and has favorable pharmacokinetic parameters in vivo. The effects that AI-14-91 exerted in multiple human OS cell lines were also seen in U2OS cells without CBFb protein, the target of the small-molecule inhibitor. This suggested there are alternate mechanisms of action of this compound in OS, which began to be investigated here, but should be continued to be expanded upon. Additionally, this PK study could inform future studies that further characterize the ADMET properties of AI-14-91 and utilize this compound in pulmonary CDX models in mice. Overall, the small-molecule inhibitor of CBFb, AI-14-91, deserves further investigations that characterize its mechanism of action, anti-tumor properties, and efficacy in models of OS. v