Response: ELA2 genotype-phenotype correlation to be expected in patients with severe congenital neutropenia

We appreciate the letter by Andersson and colleagues regarding our recent report on granulocyte colony-stimulating factor (G-CSF) response in children with severe congenital neutropenia (SCN). Although there are several causes of SCN, including ELA2, HAX1, and Gfi1 mutations, our study was designed to define the effects of ELA2 mutations on antibacterial machinery of neutrophils isolated from children under treatment with G-CSF.^1,–3  We showed abnormal elastase, myeloperoxidase, cathepsin-G, and human neutrophil peptide expression, as well as antimicrobial deficiency in neutrophils bearing 4924G>A or 4899 dele (Pro 176 Pro fs 7) mutation of ELA2.⁴ 

We agree with Andersson et al that different mutations could lead to distinctive cellular features. In our report, neutrophils from 2 SCN patients with the 4942G>A mutation display altered electrophoretic mobility of 60 kDa myeloperoxidase heavy chain and complete absence of cathepsin-G and lactoferrin, whereas cells of a patient with 4899 dele mutation show an expression pattern characterized by reduced—but still detectable—myeloperoxidase, cathepsin-G, and lactoferrin. Of note, this latter pattern resembles those observed by Andersson et al in 2 SCN patients with 92L>H and 26C>S mutations of ELA2.

Andersson et al point out that, in contrast to our report, Skokowa et al⁵  describe reduced ELA2 mRNA levels in SCN patients. This discrepancy could be due to the fact that we measured ELA2 mRNA in total hematopoietic cells, whereas the other authors analyzed the expression in isolated myeloid cells.^5,6  In any case, we believe that analysis of ELA2 mRNA may not be adequate to characterize SCN patients because it does not account for differences due to posttranslational processing of the protein. In accord with our hypothesis, Skokowa et al failed to demonstrate an association of a specific genotype with the levels of ELA2 mRNA.⁵  It is conceivable that ELA2 mutations might instead influence intracellular accumulation and trafficking of the protein.⁷ 

Patients with 4942G>A mutation display severe neutropenia, require higher G-CSF dosage, and show higher risk of developing myelodysplastic syndromes or acute myeloid leukemia⁸ ; this unfavorable outcome is observed in children presenting a poor response to the treatment and/or receiving large doses of G-CSF (higher than 8 μg/kg per day).⁹  Accordingly, in the 2 SCN patients with the ELA2 mutation 4942G>A, one presented with omphalitis and one, with severe pneumonia. These patients required prolonged intravenous antibiotic therapy and large G-CSF dosage. Because of the unsatisfactory hematologic response to this treatment, and despite the large G-CSF doses used, the patients were subjected to bone marrow transplantation. The patient bearing 4899 dele mutation, who also displayed an early onset of bacterial infections and absolute neutrophil counts at the same levels as the other 2 patients, had a more favorable outcome and is still under treatment with G-CSF. These opposite clinical outcomes may depict 2 extremes of the same disease, characterized by distinct patterns of antimicrobial peptide expression in granulocytes, and possibly, by diverse prognosis. Therefore, our study supports the hypothesis of a genotype-phenotype correlation, and we believe that when other ELA2 genotypes will be investigated at the protein level, the study of neutrophil antimicrobial peptides will probably gain clinical significance.