Treatment with 8F4, An Anti-PR1/HLA-A2 T Cell Receptor-Like Antibody, Reduces Established Acute Myeloid Leukemia and Eliminates Leukemia Stem Cells in Mice

Abstract 766

PR1 (VLQELNVTV) is an HLA-A2-restricted leukemia-associated peptide from proteinase 3 and neutrophil elastase that is recognized by PR1-specific cytotoxic T lymphocytes that contribute to cytogenetic remission of myeloid leukemia. We developed a high affinity T cell receptor (TCR)-like mouse monoclonal antibody (8F4) that binds to a conformational epitope of the PR1/HLA-A2 complex. Flow cytometry and confocal microscopy of 8F4-labeled cells showed significantly higher PR1/HLA-A2 expression on AML blasts compared with normal leukocytes. Moreover, 8F4 mediated complement dependent cytolysis of AML blasts and Lin⁻CD34⁺CD38⁻ leukemia stem cells (LSC), but not normal leukocytes. To investigate in vivo biological effects 8F4 on established leukemia, we established xenografts of primary human HLA-A2-positive AML in sublethally irradiated NOD.Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NSG) mice. Leukemia engraftment was monitored in peripheral blood by flow cytometry. Mice with established PR1/HLA-A2-expressing leukemia were treated with twice-weekly intravenous injections of 200 μg 8F4 or isotype control antibody. Flow cytometry and histology analysis of tissues was used to assess leukemia burden and level of engraftment. After 5 weeks of treatment AML was reduced 300-fold in bone marrow of 8F4-treated mice compared to isotype-treated control animals (0.07 ± 0.06% hCD45+cells versus 20.4 ± 4.1%, n=5 mice per group). Moreover, leukemia stem cells (LSC, CD34+CD38−Lin-) were no longer detected in bone marrow of 8F4-treated mice, compared to 0.88 ± 0.24% in isotype-treated mice. Equally, AML was evident in the liver and spleen of isotype-treated mice (1.1 ± 0.16% and 0.32 ± 0.17%, respectively), but was undetectable in 8F4-treated mice (p<0.001). Similar results were obtained with AML from two additional patients, one with secondary AML (CMML) and one with AML-M7. Bone marrow contained 6.2 ± 3.0% (n=3) AML versus 41 ± 15% (n=2 mice; p=0.06) in the first case and 0.16 (n=1) versus 7.0 ± 4.1 (n=2) in the second case after 2–3 weeks of twice-weekly injection. To confirm 8F4-mediated elimination of LSC, we performed secondary transfer experiment with 1×10⁶ bone marrow cells from 8F4- and isotype-treated mice, transplanted into recipient NSG mice, irradiated with 250 cGy. AML was undetectable in mice that received bone marrow from 8F4-treated animals versus 4.1 ± 2.4% (n=4) in bone marrow of mice that received cells from isotype- treated mice, determined at 16 weeks after secondary transfer.

Because PR1/HLA-A2 expression on normal hematopoietic cells (HSC) is similar to LSC in AML patients, we sought to determine whether 8F4 treatment of NSG mice xenografted with CD34-selected umbilical cord blood resulted in elimination of xenograft. Fourteen weeks after transplant stable chimerism (4.1 - 7.7% hCD45+ cells) was established, mice were treated with 50 μg 8F4 intravenously and peripheral blood was monitored weekly for chimerism. Human CD45+ cells decreased to 0.35 – 0.95% by week 1, but increased to 1.9 – 2.1 % hCD45+ cells at week 3. Bone marrow at week three contained myeloid (CD13+CD33+) and lymphoid (CD19+) cells showing that while 8F4 has off- target effects against normal hematopoietic cells, HSC are preserved. This is consistent with our previous studies that showed no 8F4-mediated effect on colony formation of normal bone marrow cells. In conclusion, these results show that anti-PR1/HLA-A2 monoclonal antibody 8F4 is biologically active in vivo and selectively eliminates LSC, but not normal HSC. This justifies continued study of 8F4 as a novel therapy for AML.