In vitro evaluation of asparagine endopeptidase as a candidate biomarker of treatment failure in childhood acute lymphoblastic leukaemia

Abstract

Background Although cure rates in childhood acute lymphoblastic leukaemia (ALL) approach 90%, relapses, particularly central nervous system (CNS) relapses, pose a significant therapeutic challenge. Prevention of relapses requires better predictive biomarkers. Towards this end, global gene-expression microarray screens of primary ALL samples were performed. Overexpression of the lysosomal cysteine protease, asparagine endopeptidase (AEP) was identified in adverse-risk progenitor-B ALL genotypes. AEP overexpression in adult human epithelial cancers has been linked to metastasis and adverse prognosis. Hypothesis AEP overexpression promotes leukaemic cell infiltration of the CNS and other extramedullary sites. AEP degrades the bacterial protein, E. coli L-asparaginase (ASNase) and potentially mediates lymphoblast resistance to this key antileukaemic drug. Summary of findings Quantitative AEP transcript estimation validated microarray findings in the discovery cohort. Sample availability precluded conclusive demonstration of protein overexpression. Intracellular expression in SD1 cells, a model AEPoverepressing ALL cell line, was strikingly heterogeneous and included aberrant peripheral localisation in endolysosomal macrovesicles. Similar findings were observed anecdotally in primary lymphoblasts. In SD1 cells, precursor AEP protein was shed in microvesicles. AEP’s role in cell motility was examined in human embryonic kidney (HEK293) cells. Contrary to published reports, ectopic AEP overexpression in HEK293 cells was not associated with enhanced motility. ASNase was consistently degraded when incubated with ALL cell lysates. In overexpressing disease, ASNase degradation is potentially accelerated by AEP, resulting in inadequate drug activity during early treatment. AEP-cleaved ASNase Synopsis fragments retain known sensitising epitopes, suggesting that AEP cleavage could also potentiate formation of neutralising ASNase antibodies, compounding ASNase treatment failure. Sole substitution at an AEP cleavage site generated an AEP-resistant ASNase variant. Computational protein modelling enabled identification of an appropriate substituting amino acid that best retains drug activity. Conclusions AEP is a candidate marker of poor treatment response in childhood ALL and is presently the subject of a prospective nationwide clinical biomarker study. Its postulated role in cell motility appears to be cell-specific and dependent on the concomitant upstream expression of additional candidate pro-motility molecules. The protease also appears to be a marker of aberrant lymphoblast vesicle phenotype, an observation that is currently the focus of further studies.