Repositioning of FDA-approved anti-cancer drugs for fatal blast crisis CML by integrated high-throughput genomic and artificial intelligence-based drug discovery analyses of pan-leukemic genetic abnormalities: Implications in post-COVID-19 era

Background: Chronic Myeloid Leukemia (CML) is a type of cancer that affects the blood and bone marrow. Although treatable in initial chronic phase (CP-CML) and accelerated phase (AP-CML), in its advanced stage, known as the blast crisis phase (BC-CML), it becomes very aggressive and challenging to treat, with overall survival (OS) ranging from few months to less than 2 years, making it almost a fatal manifestation. Therefore, the treatment of BC-CML is one of the biggest challenges in modern cancer medicine. Repositioning existing FDA-approved drugs for other hematological malignancies offers a promising approach to address the aggressive blast crisis phase of CML. Therefore, objective of this study was to carry out analyses of druggable pan-leukemic genetic abnormalities in BC-CML by employing highthroughput genomic and artificial intelligence (AI)-based drug discovery tools and find out FDA-approved drugs targeting these pan-leukemic genetic abnormalities for drug repositioning in BC-CML. Patients & Methods: The study included 141 CML patients (123 CP-CML as control groups; 6 AP-CML and 12 BC-CML as experimental groups). Most of the patients received imatinib mesylate (IM) as first-line treatment. All response criteria were per European LeukemiaNet (ELN) guidelines 2020. Whole exome sequencing (WES) was carried out to find out druggable gene mutations and the druggability of the mutated genes was determined using the online Artificial intelligence (AI) tool www.pandrugs.com. SAS/STAT software version 9.4 was used for data analysis (SAS Institute Inc., Cary, NC, USA). For statistical computing, the R package was employed (Vienna, Austria). The study was approved by the ethical committee of KAIMRC and carried out per the guidelines of the Helsinki Declaration Results: During course of the study, 18 (12.8%) patients progressed to AP-CML while 12 (8.5%) to BC-CML finally. Due to overall poorer response to TKIs and higher mortality rate (75%, vs 8.1% in CP-CML) of BC-ML patients, investigations were carried out to find out additional druggable pan-leukemic genetic abnormalities in study subjects. Overall WES coverage was about 110X. WES showed overall 67 pan-leukemic genes mutated in advanced phase CML patients. AP-CML had on average 1644 variants, whereas BC-CML had 2531 variants, with a 54% gain in mutations from AP-CML to BC-CML (P< 0.000001). Among AML-/ALL- related mutated genes, in addition to ABL gene, were NPM1 (%1.98), DNMT3A (%1.86), PML (%1.82), AKT1 (%1.62), CBL (%1.30), JAK2 (%0.71), TET2 (%0.59), IDH1 (%0.32), BCL2, FGFR3 and GATA2. The FDA approved drugs targeting these mutations are Venetoclax, Bortezomib, Doxorubicin, Mitoxantrone, Tretinoin, Quizartinib, Decitabine, Azacitidine, Arsenic Trioxide, and Ivosidenib. Conclusions: NGS analysis of AP- & BC-CML found mutations in many pan-leukemic genes, particularly in genes associated with AML and ALL. The frequencies of these mutations is much higher than previously reported. This shows a huge genetic similarity between BC-CML and other myeloid/lymphoid leukemias. FDA-approved anti-leukemic and various novel experimental drugs under active clinical trials are available for many of the gene mutations reported in this study. We conclude that our approach of highthroughput sequencing can help in finding pan-leukemic druggable gene mutations in almost every BC-CML patient and provide a practical guidance for drug repurposing as well as gateway to under-clinical-trial novel experimental drugs to individualize fatal blast crisis CML treatment.