Noonan syndrome‐associated myeloproliferative disorder with somatically acquired monosomy 7: impact on clinical decision making

Abstract

Haematological manifestations in Noonan syndrome (NS), the most common RASopathy, encompass a broad phenotypic spectrum ranging from transient monocytosis, thrombocytopenia to myeloproliferative disorder (MPD) (Roberts et al, 2013). In the majority of patients, NS-MPD is benign and shows gradual resolution of haematological abnormalities in the absence of intensive chemotherapy or haematopoietic stem cell transplantation (HSCT) (Roberts et al, 2013). The difference in clinical behaviour between NS-MPD and sporadic PTPN11-mutated juvenile myelomonocytic leukaemia (JMML) can be partly explained by differences in molecular pathogenesis. For instance, PTPN11 mutations are typically clustered in a few amino acids and show a genotype/phenotype correlation (Tartaglia et al, 2006) with some overlap between NS-MPD and JMML (Strullu et al, 2014; Tartaglia et al, 2006). Recently, genome-wide DNA methylation profiles of JMML and NS patients identified distinct methylation signatures correlating with clinical and genetic features and survival (Lipka et al, 2017; Stieglitz et al, 2017). Somatically acquired genetic lesions in NS-MPD are exceedingly rare and therefore their effect has been poorly explored. Here, we report a case of NS-MPD with loss of chromosome 7. The boy was born at 33 weeks and 1 day of gestation by means of an urgent caesarean due to hydrothorax, polyhydramnios and fetal abnormalities on ultrasound. He presented with facial dysmorphic features, pulmonary valve stenosis and massive hepatosplenomegaly. Noonan syndrome was confirmed by molecular testing of peripheral blood (PB) and cultured skin fibroblasts, identifying a germline heterozygous mutation in PTPN11 (c.182A>G). Pleural fluid contained myeloid and erythroid precursors and CD34-positive blasts, and repeated PB and bone marrow (BM) examinations showed persisting monocytosis (>1 0 9 10/l), dysplastic features in haematopoiesis and elevated numbers of myeloblasts, leading to a diagnosis of NSMPD (/Table I). Unexpectedly, array comparative genomic hybridisation, conventional karyotype and fluorescence in situ hybridisation (FISH) detected monosomy 7 in PB and BM but not in cultured skin fibroblasts, indicative for a somatic origin. While about a quarter of JMML cases show loss of chromosome 7 in addition to a RAS pathway driver mutation (Niemeyer & Kratz 2008), monosomy 7 in NS-MPD has been reported only once before in a child with a germline mutation in PTPN11 (c.1505C>T; p.Ser502Leu) (O’Halloran et al, 2017). Therefore, the prognostic value and utility for guiding therapy of somatically acquired monosomy 7 is still unknown in this context. Due to the poor general condition it was decided not to schedule HSCT. Therapy with 6-mercaptopurine (6-MP) was initiated but soon interrupted due to profound cytopenia, vomiting and respiratory deterioration. During 6-MP treatment, precursors cells and blasts briefly disappeared from the PB and reappeared as soon as white blood cells and neutrophils recovered. In addition, the proportion of cells in the PB carrying monosomy 7 decreased from 61% to 34%, as determined by FISH. Response to treatment with DNA-hypomethylating agents, such as decitabine or azacitidine, was recently demonstrated in JMML (Cseh et al, 2015). Therefore, six cycles of azacitidine 2 5 mg/kg/day for 7/28 days were given. These induced recurrent transient leucopenia and thrombocytopenia with little to no effect on the numbers of circulating precursor cells and blasts nor on the monosomy 7 burden. In two large series of genome-wide DNA methylation profiles of JMML patients, generated using Infinium HumanMethylation 450K bead chip arrays (450K; Illumina, San Diego, CA, USA), all NS-MPD cases clustered in the low methylation (LM) group, which was generally associated with low-risk disease (Lipka et al, 2017; Stieglitz et al, 2017). By contrast, JMML cases associated with monosomy 7 clustered predominantly in the intermediate methylation (IM) group (Lipka et al, 2017; Stieglitz et al, 2017). To better understand the disease biology in our patient, DNA methylation profiles of BM and PB samples at diagnosis, before, during, and after treatment with azacitidine were generated using the Infinium HumanMethylation EPIC arrays (EPIC; Illumina, San Diego, CA). A combined analysis of the EPIC profiles from our patient with the 450K profiles from 147 patients with JMML or NS-MPD registered in the European Working Group of myelodysplastic syndromes (EWOG-MDS) 1998 or EWOG-MDS 2006 trials [previously reported as the validation cohort by Lipka et al (2017)] was performed using the minfi package in R/Bioconductor (Fortin et al, 2017). This resulted in a virtual EPIC data set, containing 92% (4954/5380) of the previously identified JMMLspecific differentially methylated probes (DMP) (Lipka et al, 2017). Unsupervised hierarchical clustering of all JMML correspondence