Ewing Sarcoma is caused by a pathognomonic genomic translocation that places an N-terminal EWSR1 gene in approximation with one of several ETS genes (typically FLI1). This aberration, in turn, alters the transcriptional regulation of more than five hundred genes and perturbs a number of critical pathways that promote oncogenesis, cell growth, invasion, and metastasis. Among them, translocation-mediated up-regulation of the insulin-like growth factor receptor 1 (IGF-1R) and mammalian target of rapamycin (mTOR) are of particular importance since they work in concert to facilitate IGF-1R expression and ligand-induced activation, respectively, of proven importance in ES transformation. When used as a single agent in Ewing sarcoma therapy, IGF-1R or mTOR inhibition leads to rapid counter-regulatory effects that blunt the intended therapeutic purpose. Therefore, identify new mechanisms of resistance that are used by Ewing sarcoma to evade cell death to single-agent IGF-1R or mTOR inhibition might suggest a number of therapeutic combinations that could improve their clinical activity. Male non-obese diabetic (NOD)-SCID-IL-2Rgnull mice were used to generate EW5 explants (2 mm). Mice bearing subcutaneous tumors were randomized into treatment and control groups when their tumors reached a diameter of 6 mm and received ridaforolimus (MK-8669, mTOR inhibitor, 5mg/kg per dose, once weekly), dalotuzumab (MK-0646, IGF-1R inhibitor monoclonal antibody, 0.5mg IP twice weekly), a placebo control (sterile buffer) or a combination of both treatments. Animals were treated either until their tumors reached 1500 mm3 in volume. RPPA profiling of EW5 xenografts treated in vivo either with MK-0646, MK-8669, control and combination and compared each other were performed simultaneously using the same array. Lysates were processed, spotted onto nitrocellulose-coated FAST slides, probed with 179 validated primary antibodies, and detected using a DakoCytomation-catalyzed system with secondary antibodies. MicroVigene software program (VigeneTech) was used for automated spot identification, background correction, and individual spot-intensity determination. Expression data was normalized for possible unequal protein loading, taking into account the signal intensity for each sample for all antibodies tested. Log2 values were media-centered by protein to account for variability in signal intensity by time and were calculated using the formula log2 signal – log2 median. Principal component analysis was used to check for a batch effect and feature-by-feature two-sample t-tests were used to assess differences between treatment and control groups. We also used feature-by-feature one-way analysis of variance (ANOVA) followed by the Tukey test to perform pair comparisons for all groups. Beta-uniform mixture models were used to fit the resulting p value distributions to adjust for multiple comparisons. The cutoff p values and number of significant proteins were computed for several different false discovery rates (FDRs). Biostatistical analyses comparing two groups were performed using an unpaired t-test with Gaussian distribution followed by the Welch correction. To distinguish between treatment groups, we used one-way ANOVA with the Geisser-Greenhouse correction. Differences with p values <0.05 were considered significant. Within clustered image maps (CIM), unsupervised double hierarchical clustering used the Pearson correlation distance and Ward’s linkage method as the clustering algorithm to link entities (proteins or genes) and samples.