S194: INTEGRATIVE CLINICAL PROTEOTYPING AND DRUG RESPONSE PROFILING IDENTIFIES TARGETABLE BIOLOGY UNDERLYING MYELOPROLIFERATIVE NEOPLASMS

Abstract

Background: Myeloproliferative neoplasms (MPN) are a family of hematopoietic stem cell diseases characterized by frequent driver mutations in Calreticulin (CALR) and JAK2. While patients with essential thrombocythemia (ET) have a more favorable outcome, the prognosis of patients with advanced myelofibrosis (MF) is particularly poor with limited therapeutic options. Aims: Here, we set out to elucidate biology underlying MPN and translate this into vulnerabilities that can be exploited by targeted therapies. By integration of proteomic analyses and drug response profiling we aimed to correlate MPN protein alterations directly to targeted drug responses, and subsequently validate pathway alterations using dedicated follow-up experiments. Methods: We gathered extended MPN patient and matched healthy donor (HD) cohorts (n = 119) and a follow-up MF-specific cohort (n = 44) to investigate both general and subgroup-specific alterations. We performed proteomics on isolated granulocytes, hematopoietic stem and progenitor cells (HSPCs), and T-cells, and drug screening on peripheral blood mononuclear cells (PBMCs). For follow-up experiments, we established a CALR mutant-specific antibody for use in high-content immunofluorescence imaging. Results: We identify a specific signature of disease- and mutation-specific protein alterations present in MPN patients by machine learning. Among the identified alterations, we find strong upregulation of proteins belonging to the MCM complex, a helicase involved in DNA replication, to be specifically present in a subgroup of MF patients that are clinically characterized by a worse prognosis. In these patients, a proteomic signature related to DNA replication and cell cycling is found to be elevated across blood cell types. We proof that cells from these patients are a) characterized by high proliferation rates, and b) respond well to drugs targeting this process (e.g. topoisomerase inhibitors). Furthermore, we find upregulation of proteins involved in endoplasmic reticulum (ER) stress to be significantly correlated to the CALR mutation burden across HSPCs, granulocytes, and T-cells. Using an advanced imaging approach we show that on a single cell level, the amount of ER stress corresponds to the level of CALR mutant protein expression. This results in a higher sensitivity of CALR-mutated cells to drugs targeting ER stress or the corresponding unfolded protein response (UPR). Summary/Conclusion: Using large-scale MPN patient cohorts, we are able to elucidate biology underlying MPN disease. We here show specific proteomic signatures of proliferation and ER stress to be present in subsets of MPN patients and translate this to specific biological alterations and correlated therapeutic vulnerabilities. Our findings can be instrumental to both improved stratification of response and discovery of novel targeted therapeutics for MPN patients.