Postnatally Acquired Mutations Underlie the Progression of Transient Leukemia to Myeloid Leukemia of Down Syndrome

Abstract INTRODUCTION . Transient Leukemia (TL; also termed Transient Myeloproliferative disorder, TMD, and Transient Abnormal Myelopoiesis, TAM) occurs in 10-30% of newborns with Down syndrome (DS). Approx. 20% of infants with TL go on to develop acute myeloid leukemia of DS (ML-DS), typically within the first four years of life. Somatic, clone-specific mutations of GATA1 are found both in the blasts of TL and ML-DS, are concordant within the same individual and thought to function as initiating event in the development of ML-DS. In contrast, additional mutations of cohesin complex and related genes (e.g. RAD21, STAG2, CTCF), epigenetic regulators (e. g. EZH2) and signal transducers (e.g. within RAS, JAK signaling pathways) have been identified only in ML-DS blasts and are thought to cooperate with mutant GATA1 in the progression from TL to ML-DS. It is not known whether these cooperating mutations already mark a minor subclone of TL blasts at birth - allowing, at least in principle, a genetic risk stratification of TL - or are acquired postnatally during the first four years of life. OBJECTIVES . We tested the functional impact of impaired function of cohesin complex genes, CTCF and EZH2 on the progression of TL to ML-DS. We asked if mutations representing putative genetic progression events were already detectable at birth in a minor clone of TL blasts or were acquired postnatally (during the first four years of life). METHODS. The spectrum of GATA1 and cooperating mutations was determined by whole exome sequencing in fractions of TL and ML-DS blasts sorted from blood and bone marrow samples of five patients who had successively developed both disorders including one with a relapse of ML-DS. Corresponding normal T lymphocyte fractions of each patient at the stage of TL and ML-DS served as controls. Numbers of blasts harboring specific mutations were quantified by digital droplet PCR (BioRad, Inc.). Primary TL cells were transduced with lentivirus encoding shRNA (pLVX-shRNA, Clontech, Inc.) to suppress expression of cohesin complex genes, CTCF and EZH2 and intrafemurally injected into 8 week old NSG recipient mice. Engraftment in the bone marrow was assessed 8 weeks later by flow cytometry and GATA1 mutational analysis and compared to TL cells transduced with control vector. RESULTS. TL blasts harbored fewer mutations than those of ML-DS. GATA1 mutations were concordant in TL and ML-DS blasts in the same patient, consistent with development of ML-DS from subclone of TL. Knockdown of RAD21 expression in primary TL blasts, mimicking loss of function mutation of a cohesin complex gene, resulted in significantly increased engraftment of transduced cells in xenograft recipients compared to controls. This finding is consistent with RAD21 loss of function mutations playing the role of a progression event. Mutations of cohesin complex genes (SMC1A, STAG2, RAD21), NRAS and other putative cooperating mutations (with mutant GATA1) were not detectable in any sample of primary TL blasts by either whole exome sequencing or digital droplet PCR. The same result was obtained with control T lymphocytes sorted from TL samples. ML-DS blasts in one case were oligo-clonal with regard to cohesin complex gene mutations. Relapse in this patient arose from a minor clone as defined by cohesin complex gene mutations; mutations of NRAS, KNASL1 and SMC1A were present in ML-DS blasts but absent at relapse. CONCLUSIONS . Increased engraftment of TL cells with suppressed RAD21 expression is consistent with a model in which RAD21 loss of function mutations function as a progression event in the development of ML-DS. Absence of detectable cohesin complex gene mutations and other putative cooperating events in TL blasts suggests these mutations are acquired during the first four years of life and do not mark a minor clone of TL blasts present at birth. Genomic screening of TL blasts at birth therefore is unlikely to predict the risk for development of ML-DS. Disclosures No relevant conflicts of interest to declare.