Background: De novo or acquired resistance to glucocorticoids (GCs), the common therapeutic regimens for remission-induction in pediatric B-precursor acute lymphoblastic leukemia, confronts treatment success and is a strong predictor for disease relapse. Discovery of predictive markers and underlying mechanisms of GC resistance could lead to new treatment and prognostic modalities. Aims: This study aims: (1) to demonstrate the association between CD9 expression and GC sensitivity in childhood BCP-ALL; (2) to dissect the underlying mechanisms driving GC resistance; and (3) to discover the possible therapeutic modalities for restoring GC susceptibility. Methods: Ex vivo drug sensitivity profiling to prednisone (Pred) and dexamethasone (Dex) was performed on 6 BCP-ALL cell lines and 18 patient samples with differential CD9 expression. Association of CD9 expression with patient responses to Pred prophase was analyzed in a pediatric BCP-ALL cohort (n=182). Functional impact of CD9 on GC susceptibility was dissected by gain- and loss-of-function experiments in BCP-ALL cells. Mechanistic basis of CD9-driven GC response was investigated through genomics, transcriptomics, and proteomics. Reversal of GC resistance by pathway-targeted agents were explored on BCP-ALL cells and xenograft models. Results: Distinctive differences in the drug sensitivity pattern were found between CD9- and CD9+ BCP-ALL cell lines, with the former showing resistance to GCs but not other chemotherapeutic agents. Concurrently, CD9- primary lymphoblasts exhibited resistance to Dex (IC50s: 2,761 vs. 31 nM) and Pred (IC50s: 8,185 vs. 268 nM) compared with CD9+ samples (positivity defined by the presence of ≥ 20% CD9+ blasts). More Pred poor responders (defined by the presence of ≥ 1×10 9/L circulating lymphoblasts on day 8) were found in CD9- than CD9+ patients (19.4% vs. 6.2%, P=0.018). Subgroup analyses revealed more poor responders with CD9- phenotype were enriched in patients with older age (60% vs. 3.7%, P=0.008), male gender (21.7% vs. 6.7%, P=0.045), higher white blood cell count (66.7% vs. 15.4%, P=0.004), and in those with BCR-ABL1 translocation (100% vs. 12.5%, P=0.024), or not otherwise cytogenetically specified BCP-ALL (30.8% vs. 5.8%, P=0.019). Univariate and multivariate analyses further confirmed CD9 negativity as an independent predictor for GC resistance (OR=5.1, P=0.009). Overexpression of CD9 in CD9- cells (SEM, KOPN-8) by lentiviral transduction significantly enhanced GC sensitivity by 2.8-62.3-fold ( P<0.024), whereas CD9 knockout in CD9+ cells (697) using CRISPR-Cas9 decreased GC susceptibility ( P<0.003). No significant differences in mRNA or protein expression of the GC receptor NR3C1 were identified between CD9+ and CD9- cells, as well as its isoforms or mutants. The cytoplasm-to-nucleus translocation of GC receptor was consistent among cell lines regardless of CD9 expression status, but CD9 was physically interacted with NR3C1. Upon GC stimulation, the presence of CD9 increased the number of DEGs (110 vs. 82 DEGs) and magnitude of GC-responsive genes exclusively detected in the background of CD9 (1.3-18.6-fold; e.g. AKAP13, STAT5A, EZR). CD9- cells showed constitutive activation of the MAPK pathway and could be preferentially restored for GC susceptibility by the MEK1/2 inhibitor trametinib in vitro. BCP-ALL xenograft modeling revealed that, compared to single agent Dex treatment, combination with trametinib significantly reduced the leukemic load in animal transplanted with CD9- SEM but not CD9+ BV-173 cell lines. Ex vivo drug testing further showed that the resistance to GCs in some CD9+ cases could be partially reversed by combination of trametinib and the JAK inhibitor ruxolitinib due to the activation of STAT5. Conclusions: This study provides a definitive linkage of CD9 with GC sensitivity in childhood BCP-ALL mediated through a NR3C1-independent mechanism. CD9- cases could be prone to trametinib combination therapy, whereas CD9+ cases might need a third drug such as ruxolitinib to counteract GC resistance.