Purpose The goal of this study was to identify second-generation mithramycin analogs that better target the EWS-FLI1 transcription factor for Ewing sarcoma. We previously established mithramycin as an EWS-FLI1 inhibitor, but the compound’s toxicity prevented its use at effective concentrations in patients. Experimental Design We screened a panel of mithralogs to establish their ability to inhibit EWS-FLI1 in Ewing sarcoma. We compared the IC50 to the maximum tolerated dose established in mice to determine the relationship between efficacy and toxicity. We confirmed the suppression of EWS-FLI1 at the promoter, mRNA, gene signature, and protein levels. We established an improved therapeutic window by using time-lapse microscopy to model the effects on cellular proliferation in Ewing sarcoma cells relative to HepG2 control cells. Finally, we established an improved therapeutic window using a xenograft model of Ewing sarcoma. Results EC-8105 was found to be the most potent analog and was able to suppress EWS-FLI1 activity at concentrations nontoxic to other cell types. EC-8042 was substantially less toxic than mithramycin in multiple species but maintained suppression of EWS-FLI1 at similar concentrations. Both compounds markedly suppressed Ewing sarcoma xenograft growth and inhibited EWS-FLI1 in vivo. Conclusions These results provide a basis for the continued development of EC-8042 and EC-8105 as EWS-FLI1 inhibitors for the clinic.

Background: Ewing sarcoma is a bone a soft tissue sarcoma with a poor overall survival. This tumor absolutely depends on the continued expression of the EWS-FLI1 transcription factor for cell survival. We are therefore focused on developing small molecules that inhibit EWS-FLI1. We have previously completed a high throughput screen that identified mithramycin as an inhibitor of EWS-FLI1 and translated this compound to the clinic in a phase I-II trial. The success of this compound in the clinic has been challenged by drug associated liver toxicity that has necessitated dose reductions. Therefore the goal of this study is to identify less toxic and-or more potent mithramycin analogs. Methods: The less toxic analog, EC8042, was identified by evaluating animal toxicity data and serum pharmacokinetics in mice and rats. In order to identify a more potent compound, a panel of more than 20 mithramycin analogs was screened using an EWS-FLI1 reporter NR0B1 luciferase construct to identify EC8105. EWS-FLI1 suppression was confirmed using quantitative PCR and western blot analysis in vitro. The ability of the drug to block EWS-FLI1 binding to chromatin was evaluated by performing chromatin immunoprecipitation in the presence and absence of drug. The relative hepatotoxicity of the analogs was modeled in vitro by comparing doses that achieve suppression of EWS-FLI1 to toxic doses of the drug in HepG2 cells and confirmed in vivo in xenograft experiments. Finally, we tested the ability of both analogs to suppress tumor growth in xenograft models of Ewing sarcoma and confirmed suppression of EWS-FLI1 using immunofluorescence of formalin fixed tissue from these experiments. Results: EC8042 shows equivalent suppression of EWS-FLI1 activity but is substantially less toxic than the parent compound, allowing higher serum levels of drug in vivo in animal models. In contrast, EC8105 is approximately 8 times more potent than the parent compound and demonstrates improved suppression of the EWS-FLI1 gene signature. Both compounds work to block EWS-FLI1 binding to chromatin. More importantly, in contrast to mithramycin, both analogs suppress EWS-FLI1 activity at concentrations that are non-toxic to HepG2 cells. These effects translate into improved suppression of Ewing sarcoma xenograft growth with a corresponding increase in mouse survival and regression of several tumors in both cohorts. Conclusions: We have identified the mithramycin analogs EC8042 and EC8105 as EWS-FLI1 inhibitors. These compounds are less toxic and more potent than the parent compound and suppress EWS-FLI1 at concentrations that do not appear to cause liver toxicity. Together these results suggest that the clinical development of these analogs is warranted. Citation Format: Christy Osgood, Nichole Maloney, Christopher G. Kidd, Meti Gebregiorgis, Luz E. Nunez, Javier Gonzalez-Sabin, Lee J. Helman, Francisco Moris, Patrick J. Grohar. Identification of mithramycin analogs with improved targeting of the EWS/FLI1 transcription factor. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1612. doi:10.1158/1538-7445.AM2015-1612

<p>Table S1, PCR primers used in gene signature panel. Table S2, Evidence for selection of each gene as an EWS-FLI1 target. S3A Mean fold change in expression of NR0B1 as a function of GAPDH (2Î"Î"CT) following treatment with MMA or EC-8105. Data is the average of 3 independent experiments as measured by qPCR. S3B Mean fold change in expression of EWS-FLI1-induced targets as a function of GAPDH (2Î"Î"CT) for treatment with an siRNA targeted at the breakpoint of EWS/FLI1 for 24 h. Data is the average of 3 independent experiments as measured by qPCR S3C Mean fold change in expression of EWS-FLI1-suppressed targets as a function of GAPDH (2Î"Î"CT) for treatment with an siRNA targeted at the breakpoint of EWS-FLI1for 24 h. Data is the average of 3 independent experiments as measured by qPCR S4A: Structures and NSC numbers of mithramycin analogs S4A: Structures and NSC numbers of mithramycin analogs (cont'd). S4C. Allometric scaling estimates for dosing as a function of weight of an organism for mithramycin (MMA) and EC-8042. (see text for reference). S5A Pearson correlation analysis showing similar responses of the PPTP panel of cell lines for EC-8105 and EC-8042 relative to mithramycin. S5B IC50 values for cell lines in PPTP. Columns are all cell lines tested, exclusion of rhabdoid tumor (minus rhabdoid) and exclusion of rhabdoid tumor and acute lymphoblastic leukemia cell lines (minus rhabdoid and ALL). The statistics are shown in boxes at the bottom. S5C Graph of IC50 as a function of p53 status. Cell lines with wild-type p53 (P53 wt) were more sensitive to treatment with mithramycin (MMA) and EC-8105 than lines with mutant p53 (P53 Mut) across the entire PPTP panel of cell lines independent of histology. Note there was no difference with the non-DNA damaging analogy, EC-8042. S6 Prediction plot showing mean tumor volume as a function of days of treatment with EC-8105 (1.0 mg/kg IP, M/W/F schedule, 8 doses). Thick, dotted lines represent mean tumor growth and the individual lines in the back represent the individual tumor growth by cohort. S7 Weight of each mouse in cohort of EC-8105 treated mice (black lines) treated on a 1.5 mg/kg/dose IV Q3D X 8 doses schedule relative to weight change of control mice (grey line). S8 Survival curves for control mice (black) relative to EC-8042-treated mice (grey) treated at 24 mg/kg dose X 8 either intravenously (S8A) or intraperitoneal (S8B). Arrows indicate end of treatment. S9. Response of mice treated with 24 mg/kg/dose EC-8042 IP M, W, F X8 (S9A). Regression of tumors for every mouse mouse in the cohort. Each line represents an individual mouse. Asterisks (*) highlight treatment days. S9B Mean tumor volume (+/- SEM) of every mouse in the control cohort (3177 mm3 +/- 308.2) relative to the EC-8042-treated cohort (422.5 mm3 +/- 48) on day 11 of treatment (P < 0.0001). S10: Suppression of NR0B1 staining in cells treated with 1.5 mg/kg EC-8105 IV, 48 hours after treatment. Three representative sections are shown stained for NR0B1 (red) or DAPI (blue) since the tumors were large and showed a range of staining.</p>

<div>Abstract<p><b>Purpose:</b> The goal of this study was to identify second-generation mithramycin analogues that better target the EWS-FLI1 transcription factor for Ewing sarcoma. We previously established mithramycin as an EWS-FLI1 inhibitor, but the compound's toxicity prevented its use at effective concentrations in patients.</p><p><b>Experimental Design:</b> We screened a panel of mithralogs to establish their ability to inhibit EWS-FLI1 in Ewing sarcoma. We compared the IC₅₀ with the MTD established in mice to determine the relationship between efficacy and toxicity. We confirmed the suppression of EWS-FLI1 at the promoter, mRNA, gene signature, and protein levels. We established an improved therapeutic window by using time-lapse microscopy to model the effects on cellular proliferation in Ewing sarcoma cells relative to HepG2 control cells. Finally, we established an improved therapeutic window using a xenograft model of Ewing sarcoma.</p><p><b>Results:</b> EC-8105 was found to be the most potent analogue and was able to suppress EWS-FLI1 activity at concentrations nontoxic to other cell types. EC-8042 was substantially less toxic than mithramycin in multiple species but maintained suppression of EWS-FLI1 at similar concentrations. Both compounds markedly suppressed Ewing sarcoma xenograft growth and inhibited EWS-FLI1 <i>in vivo</i>.</p><p><b>Conclusions:</b> These results provide a basis for the continued development of EC-8042 and EC-8105 as EWS-FLI1 inhibitors for the clinic. <i>Clin Cancer Res; 22(16); 4105–18. ©2016 AACR</i>.</p></div>

<p>Table S1, PCR primers used in gene signature panel. Table S2, Evidence for selection of each gene as an EWS-FLI1 target. S3A Mean fold change in expression of NR0B1 as a function of GAPDH (2Î"Î"CT) following treatment with MMA or EC-8105. Data is the average of 3 independent experiments as measured by qPCR. S3B Mean fold change in expression of EWS-FLI1-induced targets as a function of GAPDH (2Î"Î"CT) for treatment with an siRNA targeted at the breakpoint of EWS/FLI1 for 24 h. Data is the average of 3 independent experiments as measured by qPCR S3C Mean fold change in expression of EWS-FLI1-suppressed targets as a function of GAPDH (2Î"Î"CT) for treatment with an siRNA targeted at the breakpoint of EWS-FLI1for 24 h. Data is the average of 3 independent experiments as measured by qPCR S4A: Structures and NSC numbers of mithramycin analogs S4A: Structures and NSC numbers of mithramycin analogs (cont'd). S4C. Allometric scaling estimates for dosing as a function of weight of an organism for mithramycin (MMA) and EC-8042. (see text for reference). S5A Pearson correlation analysis showing similar responses of the PPTP panel of cell lines for EC-8105 and EC-8042 relative to mithramycin. S5B IC50 values for cell lines in PPTP. Columns are all cell lines tested, exclusion of rhabdoid tumor (minus rhabdoid) and exclusion of rhabdoid tumor and acute lymphoblastic leukemia cell lines (minus rhabdoid and ALL). The statistics are shown in boxes at the bottom. S5C Graph of IC50 as a function of p53 status. Cell lines with wild-type p53 (P53 wt) were more sensitive to treatment with mithramycin (MMA) and EC-8105 than lines with mutant p53 (P53 Mut) across the entire PPTP panel of cell lines independent of histology. Note there was no difference with the non-DNA damaging analogy, EC-8042. S6 Prediction plot showing mean tumor volume as a function of days of treatment with EC-8105 (1.0 mg/kg IP, M/W/F schedule, 8 doses). Thick, dotted lines represent mean tumor growth and the individual lines in the back represent the individual tumor growth by cohort. S7 Weight of each mouse in cohort of EC-8105 treated mice (black lines) treated on a 1.5 mg/kg/dose IV Q3D X 8 doses schedule relative to weight change of control mice (grey line). S8 Survival curves for control mice (black) relative to EC-8042-treated mice (grey) treated at 24 mg/kg dose X 8 either intravenously (S8A) or intraperitoneal (S8B). Arrows indicate end of treatment. S9. Response of mice treated with 24 mg/kg/dose EC-8042 IP M, W, F X8 (S9A). Regression of tumors for every mouse mouse in the cohort. Each line represents an individual mouse. Asterisks (*) highlight treatment days. S9B Mean tumor volume (+/- SEM) of every mouse in the control cohort (3177 mm3 +/- 308.2) relative to the EC-8042-treated cohort (422.5 mm3 +/- 48) on day 11 of treatment (P < 0.0001). S10: Suppression of NR0B1 staining in cells treated with 1.5 mg/kg EC-8105 IV, 48 hours after treatment. Three representative sections are shown stained for NR0B1 (red) or DAPI (blue) since the tumors were large and showed a range of staining.</p>