Abstract
Abstract 2493
Introduction.
Glucocorticoids (GC) such as prednisolone (Pred) and dexamethasone (Dex) are critical drugs in multi-agent chemotherapy protocols used to treat acute lymphoblastic leukemia (ALL). The NOD/SCID ALL xenograft mouse model is a clinically relevant model in which the mice develop a systemic leukemia which retains the fundamental biological characteristics of the original disease. Here we report the results of a study evaluating the NOD/SCID xenograft model to investigate GC-induced gene expression.
Methods.
Cells from a GC-sensitive xenograft derived from a child with B-cell precursor ALL were inoculated into NOD/SCID mice. Engraftment, defined as the proportion of human vs mouse CD45+ cells in the peripheral blood, was monitored by serial weekly tail-vein sampling. When engraftment levels reached >50%, the mice were randomised and treated with either dexamethasone 15 mg/kg or vehicle control by intraperitoneal injection, and harvested at 0, 8, 24 or 48 h thereafter. The 48 hour groups received a second dose of vehicle or Dex at 24 hours. At the defined timepoints, the mice were euthanized and lymphoblasts harvested from the spleen. RNA was extracted, amplified and hybridised onto Illumina WG-6 V3 chips. The data was pre-processed using variance-stabilisation transformation, and quantile normalisation. Differential expression was determined using limma by comparing all treated groups to time 0, with the positive False Discovery Rate correction for multiple testing. Hierarchical clustering was used to compare groups to each other. The stability of results when reducing the number of replicates was assessed using the Recovery Score method. Functional analysis was performed using gene set enrichment analysis (GSEA) and comparison to publicly available microarray data using parametric GSEA.
Results.
The 8 hour Dex-treated timepoint had the most number of significantly differentially expressed genes (see Table), with fewer observed at the 24 and 48 hour Dex-treated timepoints. There was minimal significant differential gene expression across the time-matched controls. At the 8 hour timepoint, ZBTB16, a known GC-induced gene, was the most significantly upregulated gene. Other significantly differentially expressed genes included TSC22D3 and SOCS1, both downstream targets of the glucocorticoid receptor (upregulated), and BCL-2 and C-MYC (downregulated). GSEA at 8 hours revealed a significant upregulation of catabolic pathways and downregulation of pathways associated with cell proliferation, particularly C-MYC. GSEA at 24 hours revealed enrichment of pathways associated with NF-kB. Replicate analysis revealed that at the 8 hour Dex treated timepoint, a dataset with high signal and differential expression, using data from 3 replicates instead of 4 resulted in excellent recovery scores of >0.9. However at other timepoints with less signal very poor recovery scores were obtained using 3 replicates. We compared our data to publicly available datasets of GC-induced genes in ALL (Schmidt et al, Blood 2006; Rainer et al, Leukemia 2009) using parametric GSEA, which revealed that the 8 hour gene expression data obtained from the NOD/SCID xenograft model clustered with data from primary patient samples (Schmidt) rather than the cell line data (Rainer). The 24 and 48 hour datasets clustered separately from all other datasets by this method, reflecting fewer and predominantly downregulated gene expression at these timepoints.
Conclusions:
The NOD/SCID xenograft mouse model provides a reproducible experimental model system in which to investigate clinically-relevant mechanisms of GC-induced gene regulation in ALL; the 8 hour timepoint provides the highest number of significantly differentially expressed genes; time-matched controls are redundant and excellent recovery scores can be obtained with 3 replicates.
Disclosures:
No relevant conflicts of interest to declare.
Global Brain Transcriptome Analysis of a Tpp1 Neuronal Ceroid Lipofuscinoses Mouse Model
