Effective Targeting of Leukemic Cells in Children with B-Precursor Acute Lymphoblastic Leukemia Treated with Anti-CD22 (Epratuzumab). A Children’s Oncology Group (COG) Study.

Identifying new approaches to treating relapsed ALL is a top priority because these patients fare poorly with current retrieval strategies. COG protocol ADVL04P2 is a phase I/II feasibility pilot study for children with relapsed B-precursor ALL that uses epratuzumab, a humanized anti-CD22 monoclonal antibody, in combination with conventional re-induction chemotherapy. Children whose blasts are CD22 positive first receive 4 doses of 360 mg/m2 epratuzumab as a single agent on days minus(−)14, −10, −6 and −2. Subsequently, they receive epratuzumab weekly in combination with standard re-induction chemotherapy starting on day 0. We assessed the effectiveness of epratuzumab targeting by by assaying CD22 expression on residual leukemic blasts. Peripheral blood was obtained from 15 patients 24 hours after administration of the first dose (day −13), and again at day −6 and day 0 and stained with the combination of CD10/CD22/CD45/CD19 in four color flow cytometry to permit assessment of CD22 on analytically isolated leukemic blasts. We determined quantitative expression of CD22 with a calibration kit (Quantibrite, BDBiosciences, San Jose, CA) using two PE-conjugated anti CD22 antibodies directed against different epitopes of the CD22 molecule: clone RFB4 (Caltag, Burlingame CA), directed against the same epitope as epratuzumab, and clone SHCL-1 (BDBiosciences), directed against a non-cross-reacting epitope of CD22. RFB4 binding was decreased by more than 99% within 24 hours after administration of the first dose in all patients, indicating rapid targeting of epratuzumab to leukemic cells. In all but 4 patients, levels remained low at all subsequent time points; in 2 patients expression was restored to about 15% of baseline levels by day 0, and 2 additional patients had very small (1–2%) subsets of blasts with significant expression, one at day −6 and one at day 0. In contrast to RFB4, SHCL-1 identified residual CD22 antigen expression, but SHCL-1 binding was decreased by an average of 70–75% in 14 of 15 patients at all three post-epratuzumab timepoints. The one patient with no change in SHCL-1 binding was a patient with MLL-rearranged ALL with very low levels of CD22 expression at diagnosis, and poor response to therapy. Because binding of epratuzumab to CD22 has been shown to result in antigen internalization in vitro, we interpret our findings as demonstrating in vivo internalization of CD22 with epratuzumab binding, though we cannot exclude either shedding of antigen or conformational change to block SHCL-1 sites. We conclude that epratuzumab rapidly targets CD22 on ALL blasts in vivo, though some blasts in some patients may escape over time. Targeting is associated with a physical change in CD22 antigen expression in most patients, most likely through internalization.