Animal models

All animal experiments were conformed to the relevant regulatory standards and approved by the Institutional Animal Care and Use Committee at MDACC. Sample size selection was based on literature [22]. NOD.CgPrkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice (The Jackson Laboratory) were intravenously (iv) injected with Reh-TSLPR-Luc cells or Bos-1 PDX cells. Once bioluminescence from REH reached around 10⁷ p/sec or BOS-1 number reached around 20% in blood, 10 × 10⁶ PBMC/mouse were injected I.V. for humanization, followed by stratified randomization with 5 mice in each group and treatment with 1B7/CD3 or vehicle via intraperitoneal (I.P.) injection weekly for three weeks. Leukemia burden was assessed by imaging or FACS.

Blood was collected from mice once weekly for dynamic T cell profiling. Briefly, blood was processed to single cell suspension and stained with BV510 Ghost dye (Tonbo Biosciences), anti-mouse CD45 (Invitrogen), anti-human CD45, CD3, CD19 (BD Horizon), TSLPR, CD4, CD8, CD69, CD62L CD45RO, and CD45RA (Biolegend). T cell state was determined by phenotypic markers and data were acquired on FACS and analyzed using FlowJo v10.5.

For histological analysis, formalin-fixed mouse tissues were embedded in paraffin, sectioned, stained with hematoxylin and eosin (H&E), CD3 (Serotec), or HLA-A (Abcam) for immunohistochemistry (IHC) and imaged by a Pannoramic 250 scanner.