Introduction
AML presenting with diabetes insipidus (DI) is an infrequent but well-documented phenomenon, with cytogenetic aberrations involving chromosome (chr) 3 and/or chr7 (de la Chapelle and Lahtinen, Europ. J. Haematol. 1987). The EVI1/PRDM3 gene (chr3) is often affected, resulting in its overexpression by enhancer hijacking. EVI1 repositioning in AML is associated with dysmegakaryopoesis, thrombocytosis and a dismal prognosis. We present the first case of AML with t(1;2)(p36;p21) associated with DI. The t(1;2) is a rare genetic alteration, described in 7 hematological neoplasms, including 3 MDS/CMML and 1 AML.
Patient and Methods
A44-year-old female was diagnosed with AML in March 2023, presenting with leukocytosis, thrombocytosis and dysmegakaryopoesis strikingly reminiscent of EVI1 AML. Cytogenetics disclosed a t(1;2) and monosomy 7 as sole aberrations. Remarkably, the patient (pt) reported a newly diagnosed DI 6 months before the AML diagnosis. Induction therapy with CPX-351 failed to achieve disease control, 10-day decitabine/venetoclax resulted in bone marrow blast reduction to 20%. Eight weeks after AML diagnosis an allogeneic stem cell transplantation from a sibling donor was performed, resulting in a hematologic and molecular remission of 14+ months duration, with control of the DI.
Enhancer hijacking was detected using the pyjacker tool (Sollier et al., in prep.), processing WGS and RNA-seq data to identify monoallelic and highly overexpressed genes around the breakpoint.
For DNA methylation analyses, HumanMethylation450 array data from 194 AML samples were downloaded from TCGA. Differential methylation between EVI1-overexpressing and other TCGA AML sample groups was computed at the level of CpG sites using RnBeads. CpG sites with absolute mean beta-value difference > 0.2 and FDR-adjusted p-value < 0.05 were considered differentially methylated. All differentially methylated CpG sites with overlapping probes on the 450K and EPIC array were used for clustering of TCGA samples together with the PRDM16-overexpressing sample using Ward's hierarchical clustering method.
Results
We compared PRDM16 mRNA expression levels of our pt to a cohort of 25 AML samples (UFMC), noting that only one other case showed higher expression (harboring a NUP98-NSD1 fusion, known to cause high PRDM16 levels), and confirmed the massive PRDM16 overexpression by her AML blasts in a second AML cohort (DKFZ). PRDM16 (on chr1), located near the t(1;2) breakpoint, a zinc finger transcription factor with histone methyltransferase (HMT) activity, is involved in AML pathogenesis. PRDM16 expression was monoallelic, indicating that its overexpression was driven by the translocation. At the breakpoint region on chr2, we identified two putative enhancers in myeloid cells. Ranking hematopoietic enhancers by their super-enhancer (SE) potential, we determined that these are among the strongest hematopoietic SEs.
Previously, the t(1;2) had been hypothesized to cause PRDM16 overexpression driven by juxtaposition to the THADA promoter which, however, in our pt is not part of the region translocated to PRDM16. The role of the SEs on chr2 in the wild-type state is unclear, but they are located in the same topologically associated domains as ZFP36L2, a gene with high expression and an important role in hematopoietic cells; hence, they may be responsible for the activation of this gene. Next, to determine possible epigenetic similarities between PRDM16 and EVI1-overexpressing AML, we performed DNA methylome analyses on the PRDM16 overexpressing pt sample, integrated with publicly available methylation data (Lugthart et al., Blood 2011). Indeed, CpG sites differentially methylated in EVI1 AML carried a similar methylation pattern also in the PRDM16-overexpressing index pt.
Conclusions
The clinical presentation of our pt strikingly phenocopies the syndrome of AML with EVI1 overexpression. Since the HMTs PRDM16 and EVI1/PRDM3 are structurally and functionally very closely related we speculate that PRDM16 is shaping a specific EVI1/PRDM3-like DNA methylation pattern in this AML. The t(1;2) causes massive PRDM16 overexpression, most likely due to ZFP36L2 enhancer hijacking. The clinical and cytogenetic features, PRDM16 overexpression, and lack of EVI1 overexpression point towards a previously unrecognized functional association of PRDM16 activation with paraneoplastic diabetes insipidus.
Lübbert:Cheplapharm: Research Funding; Syros Pharmaceuticals: Honoraria; Otsuka: Honoraria; Janssen: Research Funding; AbbVie: Honoraria.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal