Chemokine receptors and CNS engraftment. (A) Flow cytometry for chemokine receptor expression in BCP-ALL. Shaded histogram represents isotype control; open histogram represents specific staining. Sample names and associated translocations are indicated. CNS+ denotes CNS engraftment, with the results of xenografting this sample into mice categorized as Y (evidence of CNS engraftment) or N (no evidence of CNS engraftment). (B) Quantitative PCR for chemokine ligand expression by cultured human primary meningeal cells (white bars) and choroid plexus (CP) epithelial cells (black bars) (both passage 3). Arbitrary expression values were derived from ΔCT. (C) Primary ALL cells from 1 CNS-3 patient (open symbols) and 4 CNS-1 matched controls (closed symbols) were retrieved from BM and meninges of xenografted mice and analyzed by flow cytometry. Contaminating murine cells were excluded by gating on human CD45. A total of 10⁵ events were analyzed when possible. Data represent adjusted mean fluorescence intensity (MFI_specific − MFI_isotype) of live leukemic cells (huCD45⁺/Draq7⁻). Bars represent means of adjusted MFIs. Differences between CNS and BM expression were analyzed using 2-tailed paired Student t tests. (D) Leukemic infiltration in NSG mice engrafted with cells expressing REH–luciferase–green fluorescent protein (GFP) and treated with the CXCR4 inhibitor AMD3100 or PBS control. Left: Leukemic engraftment in the BM as measured by numbers of GFP-positive cells (REH-luciferase-GFP) on flow cytometry. Data show mean ± standard error of the mean (SEM) in n = 7 and n = 6 mice for PBS and AMD3100 groups, respectively, analyzed by an unpaired Student t test, ***P < .001. Center: Liver infiltration quantified by counting human-CD45⁺ cells (brown diaminobenzidine-stained cells) in 8 random fields of view per section. Data show mean ± SEM for n = 4 and n = 5 mice in PBS and AMD3100 groups, respectively, analyzed by an unpaired Student t test, **P < .01. Right: Histologic analysis of murine brains from xenografts (n = 5 mice in each group). Each brain was divided into 5 segments, sections were cut from each segment, and the maximal depth of meningeal infiltrates was recorded for each section using AxioVision Rel 4.3 software (Carl Zeiss). Data show mean ± SEM. Representative histology and bioluminescent in vivo imaging from these mice, along with full experimental details are provided in supplemental Figure 3.