Background: Central nervous system (CNS) leukemia is a negative prognostic factor in pediatric acute lymphoblastic leukemia (ALL), and up to 30-40% of relapses involve the CNS. However, little is known about the biological mechanisms that lead to dissemination and survival of leukemic blasts in the CNS. This in part reflects that previous research has employed rather crude cytospin preparation methods for classification of CNS involvement. Using a highly sensitive flow-cytometry based-approach, we previously identified leukemic blasts in the cerebrospinal fluid at diagnosis in approximately 30% of ALL patients, whereas conventional cytospin microscopy only was positive in 10% of patients ( Levinsen et al 2016 Pediatr Blood Cancer ).

Methods: To explore biological properties of leukemic cells that migrate into the CNS, we compared gene expression profiles of leukemic cells from 26 patients with CNS involvement and 138 patients without detectable blasts in their cerebrospinal fluid. CNS status was determined by flow cytometry and/or cytospin examination. Gene expression profiles were obtained by microarray analysis (Human Gene 1.0 ST GeneChip array, Affymetrix). Microarray results were validated by real-time quantitative PCR.

Results: A hierarchical cluster analysis of all 164 ALL samples in our study based on all top 10% most varying genes clearly subgrouped patients according to their B-cell precursor (BCP) versus T-cell immunophenotype. We therefore performed analyses separately for BCP and T-cell ALL patients. For patients with BCP ALL, more genes were differently expressed between patients with and without CNS involvement when patients were classified according to flow cytometry compared to conventional cytospin (68 probe sets, uncorrected P <0.05, versus 9 probe sets, uncorrected P <0.05). Noteworthy, the cysteine endopeptidase legumain ( LGMN ) demonstrated significantly higher expression in BCP ALL with CNS involvement determined by flow cytometry, but no significant difference was observed when CNS involvement was assessed by cytospin. PCR confirmed the upregulation of LGMN in BCP ALL, when comparing flow-positive samples with flow-negative samples ( P =0.006) and cytospin-negative samples ( P =0.01). In multivariate analysis, LGMN expression levels above the median were associated with an odds ratio of 9.1 for CNS involvement as determined by flow cytometry (95% CI, 1.26 to 65.50; P =0.03).

Conclusion: These data support that flow cytometry identifies CNS involvement better than conventional cytospin preparation, which is critical for a deeper understanding for the biology of CNS leukemia. Legumainplays a role in hydrolysis of proteins and small molecules and could potentially be involved in infiltration of the CNS by leukemic cells. Consequently, targeting legumain could be a novel strategy for combatting CNS leukemia.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution