Background: Despite introduction of novel agents, relapsed/refractory B-cell acute lymphoblastic leukemia (R/R B-ALL) carries dismal outcome. Inotuzumab ozogamicin (InO) is a humanized anti-CD22 monoclonal antibody conjugated to calicheamicin approved for R/R B-ALL able to obtain high CR rates but of short duration. Mechanisms underlaying InO resistance are largely unknown. Aims: To characterize the baseline differentially expressed genes in a series of R/R B-ALL patients in relation with patient response to InO in order to individuate potential pathways involved in resistance. Methods: Gene expression profile of 18 R/R B-ALL patient samples was analyzed with RNA-seq, before InO exposure. All patients received at least 1 InO course. Patient population was divided in poor/non-responders (NR) and responders (R), based on InO response. R were defined as patients with duration of response (DoR) ≥3 months after bone marrow (BM) complete remission (CR) achievement. NR were defined either refractory or with a DoR <3 months after CR achievement. The list of significant (p<0.05) differentially expressed genes (DEGs) in NR, as compared to R whose absolute fold change (FC) was ≥2, was analyzed. P values were corrected with the Benjamini-Hochberg algorithm (false discovery rate; FDR). Gene expression results were analyzed with QIAGEN Ingenuity Pathway Analysis (IPA). Results: Eighteen R/R BM samples of CD22-positive B-ALL were analyzed with RNA-seq. Eight patients were defined as NR and 10 patients as R. Patient disease characteristics in NR and R groups were homogeneous for age, sex, number of Philadelphia (Ph)-positive ALL patients, duration of 1st remission, number of previous therapy lines including hematopoietic stem cell transplantation (HSCT) and Blinatumomab. At the time of InO therapy all patients had morphological BM relapse and all were CD22-positive. Median CD22 expression percentage on leukemic blasts was 100% (range 100-100) in NR patients and 100% (range 70-100) in R patients (p=0.177). Two patients in R group had a CD22 expression blast percentage of 70% and 76%. Median CD22-fluorescent intensity (CD22-FI) in NR as compared to R was 75.38 (IQR 59.58, 89.51) and 136.51 (IQR 114.38, 151.57) respectively, with significantly lower values in NR group (p = 0.04). Overall, 370 genes were differentially expressed (p<0.05) in NR, as compared to R patients, of which 32 were significantly differentially expressed (corrected p < 0.05, FC ≥2). Thirty-one were down- and 1 was up-regulated in NR vs R. DEGs were involved in basophil differentiation, carbon dioxide transport, erythrocyte and myeloid cell development, heme metabolic and porphyrin-containing compound biosynthetic processes, cation and cellular ion homeostasis. Both IPA upstream regulator and regulator effect analysis identified the serine/threonine homeodomain-interacting protein kinase-2 (HIPK2) as predicted downregulated. The inhibition of HIPK2 was predicted to be the causal upstream condition of the under-expression of six DEGs from the set (FECH, ANK1, SCL4A1, EPOR, GATA1, HBZ) with activation Z score of -2.449 and p value of overlap = 1.02E-09. No difference in terms of HIPK2 expression was appreciated in the two groups, suggesting downregulation at post-transcriptional level. Image:Summary/Conclusion: A unique pattern of gene expression signature based on HIPK2 downregulation was identified in poor responders to InO, providing potentially important insights in mechanisms of resistance. HIPK2 downregulation needs to be further validated. CT and CP contributed equally