Quantitative Proteomic Analysis Implicates An Altered B-Cell Differentiation State as a Mechanism Underlying GC-Resistance in An Acute Lymphoblastic Leukaemia (ALL) Cell Line Model.

Glucocorticoids (GC) are pivotal agents used in the treatment of childhood acute lymphoblastic leukaemia (ALL). GC-resistance is a significant prognostic indicator of a poor treatment outcome in childhood ALL, but the underlying molecular basis remains unclear. Previous studies using cell lines have identified mutation/deletion of the glucocorticoid receptor (GR) as a mechanism of GC-resistance. However, genetic aberration of the GR is rare in clinical samples¹. This disparity may be due to the mismatch repair deficient status of many ALL cell lines which consequently have a greater likelihood of acquiring mutations under GC-selection. We have used a discovery proteomics approach for hypothesis generation on potential mechanisms for resistance. To achieve this, we compared a well-characterized mismatch repair proficient GC-sensitive cell line, PreB 697, and a GC-resistant sub-clone (R3F9) both bearing wildtype GR, in a comparative proteomics experiment using 4-channel isobaric tagging for relative and absolute quantitation (the iTRAQ approach).

Cells were treated with either vehicle control or 0.1μM dexamethasone for 24 hours and subjected to subcellular fractionation to prepare a nuclear fraction. Each sample was labelled with a distinct isobaric tag for relative quantification and analysed by 2-dimensional liquid chromatography/ tandem mass spectrometry. The proteins were identified and relatively quantified using Protein Pilot software (Applied Biosystems). Ratios were calculated for dexamethasone-treated ‘versus’ control vehicle for each cell line and an ITRAQ ratio of greater than or equal to ± 1.2 or less than 0.8 fold change were considered to be differentially expressed.

The comparative dataset highlighted two transcription factors which are involved in B-cell differentiation, PAX5 and IRF4, to be differentially expressed in the PreB 697 compared to the R3F9 cell line. The GC-resistant R3F9 cell line had reduced PAX5 and IRF4 protein expression compared to the parental cell line and this was further validated in other GC-resistant sub-clones derived from the PreB 697 cell line by western blot analysis. The reduced PAX5 level in the GC-resistant cell lines was not due to monoallelic loss, as measured by a QRT-PCR method or mutation as determined by DHPLC analysis of ‘hot-spot’ exons. In addition, PAX5 mRNA levels were not significantly altered, thus suggestive of a post-transcriptional mechanism for PAX5 protein reduction. To test the direct role of PAX5 in GC-resistance, we reduced PAX5 mRNA and protein levels using RNA interference in the parental GC-sensitive, PreB 697 cell line. PAX5 protein levels were reduced by at least 80% and were maintained for 48 hours post-transfection. The PreB 697 cell line was transfected with siRNA directed to PAX5 using electroporation, the cells were allowed to recover for 24 hours and the levels of cell kill were assessed in response to a 48 hour incubation with 1 μM dexamethasone by Annexin V staining and the MTS assay. Paradoxically, PAX5 knockdown increased GC-sensitivity (mean 60.4% apoptosis, S.D. 16.8, N=3) in comparison to a non-specific siRNA (mean 31.0% apoptosis, S.D. 5.2, N=3) but did not influence sensitivity to either vincristine or daunorubicin. Thus, this response was specific to glucocorticoids.

Using a proteomic approach we have shown alterations in PAX5 protein levels are associated with a GC-resistant phenotype which an mRNA-based technology would fail to detect. Modulation of PAX5 in ALL cells may influence the response to GC-therapy. It is known that GC-sensitivity alters during B-cell development, with early lymphoid precursors being highly sensitive and more mature B cells being highly resistant to GC-induced apoptosis. We propose that reduced PAX5 protein levels may reflect an altered differentiation state of the sub-clones of PreB 697 which are associated with a GC resistant phenotype. ¹Irving et al, Cancer Res, 2005

²Schmidt et al, FASEB, 2006

No relevant conflicts of interest to declare.