Somatic gene rescue (SGR), via acquisition and selection of somatic (SM) genetic hits, neutralizes functional defects resulting from hypomorphic germline (GL) mutations. However, this process may be maladaptive and lead to the development of leukemia. Acquisition of SM gain-of-function CSF3R mutations in the context of severe congenital neutropenia (SCN) illustrates such a scenario. Here, we further explored this and similar mechanisms in a cohort of 7121 adults with myeloid neoplasia and bone marrow failure (AML=1473, MDS=2248, other=3400). CSF3R mutations were found in 97 patients. In total, there were 37 and 56 patients with SM and GL mutations, respectively; biallelic GL and SM were 4. Among SM alterations, 18 were pathogenic or likely pathogenic including 8 nonsense truncating variants in class III isoform specific region of the cytoplasmic domain and 10 classical missense mutations in the juxtamembrane region, previously defined in the context of post-SCN AML and CNL. We thus hypothesized that these CSF3R SM mutations may also represent SGR in adult patients harboring hypomorphic GL variants and investigated the presence of SCN-associated alterations ( ELANE, WAS, GFI1) and in other phosphothyrosine kinase receptors (PTKR) such as CSF1R, CSF2RA, CSF2RB, FLT3 and KIT. Indeed, we found biallelic WAS p.D264H GL and CSF3R p.A119T SM in a patient with aplastic anemia (AA). Biallelic CSF3R were present in 4 patients: a) p.Q749X GL withp.M696T SM, b) E835K GL with p.T618I SM, c) p.Y752X GL withp.Q754E SM and d) biallelic p.Y752X GL with p.Q754E GL configuration with T615A SM. Compound heterozygous configuration of CSF3R SM included CSF1R SM variants( CSF3R p.M696T SM & CSF1R p.R294Q GL in a 64 year old. patient with MPN, CSF3R p.M696T SM & CSF1R c.1626+3G>A GL in a 61 year old patient with MDS) and CSF2RB ( CSF3R p.T618I SM & CSF2RB p.V890I GL in a 69 year old female patient with AML with leukopenia). When we analyzed the remaining 66 GL CSF3R variants in 56 carriers (in addition to the aforementioned biallelic combinations), we identified 9 compound heterozygous mutants involving CSF1R SM (n=1), CSF2RA (n=1), CSF2RB (n=3), FLT3(TKD/ITD, n=4/1). Specifically, we found the following configurations and disease associations: a) CSF3R p.V372E GLCSF2RB p.P842L SM in a 49 yo. AA patient who subsequently developed AML with FLT3 TKD, b) MDS patients with CSF3R p.E149 GLCSF1R p.G413S SM, c) CSF3R p.R440Q GL& CSF2RA p.R164Q SM and d) CSF3R p.E405K GL& CSF2RB c.1152+6G>A SM and e) AML patients with CSF3R p.E405K GL& CSF2RB p.R432C SM; f) CSF3R p.L619S GLandg,h)2 AML cases with CSF3R p.E808K GL of which all 3 acquired compound heterozygosity with FLT3 TKD, i) CSF3R p.E808K GLCSF3R p.E835K GL in a 71 year old patient with MDS with FLT3ITD AML. In addition to FLT3 GOF mutations, the most common SM hits observed in CSF3RGLcarriers were ASXL1 (n=12), TET2 (n=12), and SF3B1 (n=12). Identification of somatic and GL alterations in PTKR genes other than CSF3R, led us to systematically analyze these genes as targets of GL and SM CSF3R alterations. In total, we identified 72 GL and 73 SM CSF1R variants, 7 SM and 19 GL CSF2RA and 47 GL and 69 SM CSFR2B. Among these, we found 3 patients with compound heterozygous CSF2RBSMCSF3RGL, one with CSF2RBGLCSF3RGLas described above. We also identified one patient with CSF2RASMCSF3RGL. Finally, we identified two patients with CSF1RGL and CSF3RSM, and one with CSF1RSMCSF3RGLmutations. In summary, our study provides evidence for a tremendous complexity of interaction between less penetrant GL alterations of receptor genes and somatic GOF mutations, which may in some instances, correspond to the result of SGR. It is likely that the degree of LOF corresponds to the selection pressure for somatic rescue lesions.
Read full abstract