Abstract

Acute promyelocytic leukemia (APL) is associated with the accumulation of promyelocytes that usually carry the t(15;17) translocation, generating the PML-RARα fusion gene. APL develops in mice only if PML-RARα expression is targeted to early myeloid cells; when PML-RARα is expressed in all hematopoietic cells, the only malignancy that develops is APL. These observations have suggested that there may be a component(s) of the early myeloid environment that is important for the actions of PML-RARα. We recently showed that murine bone marrow extracts and a human early myeloid cell line, U937, contain a serine protease activity that cleaves PML-RARα. The dominant cleaving activity was attributable to an early myeloid-associated protease, neutrophil elastase (NE). The penetrance of APL in mice expressing PML-RARα in early myeloid cells was significantly reduced in NE-deficient animals (Lane and Ley, Cell 115:305–318, 2003). To determine whether measurable functions of PML-RARα require NE, we performed a series of experiments. Using a GFP-PML-RARα fusion construct, we determined that expression of PML-RARα caused disruption of PML oncogenic domains (PODs) in several human hematopoietic cell lines, regardless of whether they expressed NE. However, when PML-RARα was expressed at high levels in cell lines that contained NE activity, substantial toxicity was observed. In contrast, PML-RARα did not cause toxicity in cell lines without NE activity. Similarly, when we expressed an NE-resistant PML-RARα cDNA (containing valine to arginine mutations at the P1 amino acid of the two preferred NE cleavage sites), toxicity was abrogated in clonogenic assays. The toxic effects of expressing full-length PML-RARα could not be recapitulated by expressing either dominant cleavage fragment alone, or in combination, which suggests that NE-induced cleavage may not directly create the “active” fragment of PML-RARα. Primary human and mouse primary APL cells contain abundant NE activity and PML-RARα-cleaving activity. However, we determined that two commonly used APL cell line models (NB4 and U937-PR9 cells) contain abundant full length PML-RARα protein, but neither contains NE activity nor PML-RARα-cleaving activity. As expected, high levels of expression of PML-RARα in PR9 cells did not cause toxicity. Using an in vitro G-CSF-dependent myeloid differentiation assay, we found that purified hematopoietic progenitors from mCGPML-RARα knock-in mice demonstrated markedly increased proliferation of early (blasts and promyelocytes) myeloid cells compared to wild type progenitors. This increase was not observed in progenitors from mCGPML-RARαNE−/− animals. The difference was not due to an intrinsic defect in myeloid development due to NE deficiency, since NE−/− progenitors developed normally in response to G-CSF in vitro. To extend these results, we performed competitive repopulation bone marrow transplants using NE+/+ and NE−/− mice as bone marrow donors; at 3 weeks after hematopoietic reconstitution, there was no difference in the contributions of either genotype to multi-lineage hematopoiesis. Together, these data strongly suggest that NE is important for several of the measurable activities of PML-RARα in early myeloid cells; we are therefore exploring pharmacologic NE inhibition as an approach to reduce the growth of APL cells. Our data also suggest that the physiologic activities of PML-RARα should be studied in early myeloid cells that contain NE activity.

Author notes

Corresponding author