Neutropenia-Associated Mutations of PFAAP5, a Novel Protein Mediating Transcriptional Repressor Interaction between Gfi1 and Neutrophil Elastase.

Mutations in ELA2, encoding the protease neutrophil elastase, cause cyclic neutropenia and are also the most common cause of the Kostmann syndrome of congenital neutropenia. Mutations in the transcriptional repressor and intrinsic regulator of hematopoiesis, Gfi1, are another rare cause of congenital neutropenia. ELA2 is a target of Gfi1; consequently, mutation of Gfi1 leads to over-expression of neutrophil elastase. Defects in a predicted feedback circuit offer a possible explanation for the periodic oscillations of cyclic neutropenia, but how neutrophil elastase and Gfi1 regulate hematopoiesis with respect to each other, and what additional factors may be involved, has remained undefined.

To identify other components of a potential hematopoietic feedback circuit, we performed yeast two-hybrid screens with Gfi1 and neutrophil elastase using bone marrow cDNA libraries. We found 25 proteins that interacted with Gfi1. Only two proteins, however, interacted with the isoform of neutrophil elastase containing an intact carboxyl terminal extension (deletion of which is the most frequently occurring type of mutation in congenital neutropenia). Remarkably, one of the two, the previously uncharacterized protein PFAAP5, was among the 25 found to also interact with Gfi1. PFAAP5 may therefore play a unique role in mediating interactions between neutrophil elastase and Gfi1.

We further studied PFAAP5. Immunofluorescent localization with an epitope-tagged construct transfected into cultured cells demonstrates its localization in the nucleus. To determine if PFAAP5 influences Gfi1’s activities as a transcription factor in the presence or absence of neutrophil elastase, we performed transient transfection assays with a Gfi1 responsive reporter and PFAAP5 siRNA knockdown in NIH3T3 cells, which express PFAAP5, but not Gfi1 or neutrophil elastase, endogenously. Not surprisingly, transfected neutrophil elastase demonstrated no effect as a transcriptional repressor by itself. Unexpectedly, however, transcriptional repression by transfected Gfi1 increased with co-expression of neutrophil elastase. Knockdown of PFAAP5 had no effect on transfected Gfi1 repressor activity alone, but knockdown of PFAAP5 blocked the ability of co-expressed neutrophil elastase to increase Gfi1 repressor activity. Therefore, neutrophil elastase can increase Gfi1 repressor function, but only in the presence of PFAAP5.

Finally, we evaluated PFAAP5 as a candidate gene in cases of congenital neutropenia in which ELA2 or Gfi1 mutations are absent. One homozygous and two different heterozygous coding sequence substitutions were detected in four of 230 unrelated patients, yet absent among double the number of controls. The mutations prevent PFAAP5 from interacting with either neutrophil elastase or Gfi1 in the yeast two-hybrid assay. Thus, mutations of neutrophil elastase, Gfi1, or PFAAP5 may each similarly disrupt a common circuit in which Gfi1 represses ELA2 and neutrophil elastase feeds back to cooperate with Gfi1 in regulating its own transcription.