Does SPRED1 contribute to leukemogenesis in juvenile myelomonocytic leukemia (JMML)?

To the editor:

Dr Pasmant and colleagues recently reported a child with a neuro-cardio-facial-cutaneous (NCFC) syndrome caused by a germline SPRED1 mutation.¹  The child developed acute myeloid leukemia.¹  The relation between the molecular pathology of NCFC syndromes and that of myeloid malignancies in children is well-documented. NCFC syndromes include Costello, Noonan, LEOPARD, and cardiofaciocutaneous syndromes, neurofibromatosis type 1 (NF-1),²  and the SPRED1-related syndrome whose phenotype closely resembles that of NF-1.³  All of these conditions are secondary to germline mutations in genes encoding components of the Ras mitogen–activated protein kinase (MAPK) pathway.²  Somatic mutations in some of these genes play a fundamental role in the pathogenesis of juvenile myelomonocytic leukemia (JMML),⁴  as we and others have shown for PTPN11,⁵ KRAS,⁶  and NF1.⁷ 

SPRED1 encodes a protein that is highly expressed in hematopoietic cells and negatively regulates Ras-MAPK signaling by suppressing Raf activation.⁸  The phenotypic similarity of the SPRED1 syndrome to NF-1, the fact that the NF1 gene product is a negative regulator of Ras-MAPK, the observation that children with NF-1 are at highly increased risk of JMML, and the recent letter by Pasmant et al¹  (exemplifying a link between SPRED1 disease and myeloid leukemia in children) led us to hypothesize that somatic mutations in SPRED1 might occur in cases of JMML that lack mutations in other JMML-associated Ras-related genes.

To test this possibility, we sequenced the SPRED1 gene in granulocyte DNA from 23 JMML patients without mutations in PTPN11, KRAS/NRAS, or CBL and without an NF-1 phenotype. All children were enrolled in the European Working Group of MDS in Childhood (EWOG-MDS) studies 98 or 2006. Approval was obtained from the Freiburg University institutional review board for these studies. Informed consent was provided according to the Declaration of Helsinki. The analyses comprised the entire coding sequence of SPRED1. No mutations were discovered in the 23 JMML samples, whereas 2 known exonic synonymous single nucleotide polymorphisms, c.291 G>A and c.1044 T>C, were identified in all 23 cases. At position c.291 the genotype was G/A in 4 cases and A/A in 19 cases, whereas the variation c.1044 T>C showed the genotype C/T in 5 cases and C/C in 18 cases. These findings are in accordance with documented allele frequencies in whites. We also reproduced the known intronic polymorphisms c.424-18 G>A (G/G, 12 cases; A/A, 2 cases; A/G, 9 cases) and c.424-8 C>A (C/A, 6 cases; A/A, 17 cases). In addition, 3 known intronic polymorphisms, c.684 + 49_684 + 50insTTAA/−, c.684 +50_684 + 51insT/− and c.684 + 53_684 + 54insTA/, were found in 5 JMML specimens. In all 5 instances the variations were linked on one allele and the other allele retained wild-type sequence.

In summary, we found no evidence of leukemogenic SPRED1 involvement in JMML cases negative for mutations in PTPN11, KRAS/NRAS, or CBL and without NF-1 features. The assumption put forward by Dr Pasmant and colleagues, that germline SPRED1 mutations predispose children to leukemia, is certainly plausible. However, the absence of SPRED1 mutations in an early childhood leukemia such as JMML indicates that the putative link between SPRED1 lesions and childhood myeloid malignancies requires further clarification.