Differential expression of CD73, CD86 and CD304 in normal vs. leukemic B-cell precursors and their utility as stable minimal residual disease markers in childhood B-cell precursor acute lymphoblastic leukemia.

BACKGROUND

Optimal discrimination between leukemic blasts and normal B-cell precursors (BCP) is critical for treatment monitoring in BCP acute lymphoblastic leukemia (ALL); thus identification of markers differentially expressed on normal BCP and leukemic blasts is required.

METHODS

Multicenter analysis of CD73, CD86 and CD304 expression levels was performed in 282 pediatric BCP-ALL patients vs. normal bone marrow BCP, using normalized median fluorescence intensity (nMFI) values.

RESULTS

CD73 was expressed at abnormally higher levels (vs. pooled normal BCP) at diagnosis in 71/108 BCP-ALL patients (66%), whereas CD304 and CD86 in 119/202 (59%) and 58/100 (58%) patients, respectively. Expression of CD304 was detected at similar percentages in common-ALL and pre-B-ALL, while found at significantly lower frequencies in pro-B-ALL. A significant association (p = 0.009) was found between CD304 expression and the presence of the ETV6-RUNX1 fusion gene. In contrast, CD304 showed an inverse association with MLL gene rearrangements (p = 0.01). The expression levels of CD73, CD86 and CD304 at day 15 after starting therapy (MRD15) were stable or higher than at diagnosis in 35/37 (95%), 40/56 (71%) and 19/41 (46%) cases investigated, respectively. This was also associated with an increased mean nMFI at MRD15 vs. diagnosis of +24 and +3 nMFI units for CD73 and CD86, respectively. In addition, gain of expression of CD73 and CD86 at MRD15 for cases that were originally negative for these markers at diagnosis was observed in 16% and 18% of cases, respectively. Of note, CD304 remained aberrantly positive in 63% of patients, despite its levels of expression decreased at follow-up in 54% of cases.

CONCLUSIONS

Here we show that CD73, CD86 and CD304 are aberrantly (over)expressed in a substantial percentage of BCP-ALL patients and that their expression profile remains relatively stable early after starting therapy, supporting their potential contribution to improved MRD analysis by flow cytometry.