Modelling and analysis of glucocorticoid action in childhood acute lymphoblastic leukaemia

Othman, Aisha

Abstract

Acute lymphoblastic leukaemia (ALL) is characterised by the abnormal proliferation of white
blood cells. Glucocorticoids (GCs) induce apoptosis in leukaemia through the glucocorticoid
receptor (GR) and are the main treatment for ALL. The obstacle to achieving ALL remission
is resistance to GCs, highlighting the need to uncover the causes of the therapy resistance.
Previous use of systems biology led to the construction of a logical model, GR interactome
named GEB052 which consisted of 52 nodes and 241 interactions, where nodes represented
genes or proteins that interact with GR. The model was linked to GC input and cell
death/inflammation outputs and was used to predict how components can react with each other
to induce physiological functions, disease mechanisms, and drug responses.
Towards improved models’ predictive capacity, here microRNAs (miRs), which are key factors
of gene regulation, were added to the model. Target scan and DIANA tools were used to
identify miRs that target proteins in the GEB052 model, after literature-based validation. 26
miRs were added to the GEB052 model, Cytoscape was used to visualise and CellNetAnalyzer
to analyse the model. A new model named GAO078 which has 78 nodes with 274 interactions
was built. All miRs added had an inhibitory effect on the target nodes. In silico knockdown of
GR that mimics mutation, and logical steady state analysis (LSSA) was conducted in both wild-
type GR and mutant (miRs ON and OFF) scenarios. This demonstrated that the correct
prediction rate obtained through model stimulation was 65.7% in GAO078 (when the miRs are
OFF) that is higher than in GEB052 model (56.6%). Additional analysis (semi-quantitative
Signal Transduction Score Flow Algorithm -STSFA) resulted in a high correct prediction
(80%) compared to LSSA. Model also predicted that ARHGAP35, NR2F2 and STAT5B genes
have an inhibitory effect on GR, and through DRUGSURV database tyrosine kinase inhibitor
(TKI) dasatinib (DAS), that is an approved drug targeting STAT5B, was identified and used to
treat leukaemia cells through the lab-based experiments to validate GAO078 model
predictions.
Cell viability assays were used to test synthetic glucocorticoid Dexamethasone (DEX) and
DAS effects on CEM-C7-14, CEM-C1-15, MOLT-4 and SUP-B15 ALL cell lines. The data
obtained using MTS assay indicated that DEX was effective in reducing the viability of CEM-
C7-14 but not CEM-C1-15 and to a lesser extent MOLT-4 and SUP-B15. DAS inhibited
proliferation of ALL cells to varying extent, and the two medications used together had a larger
inhibitory effect on cell death in some cell lines studied. Measurement of Sub-G1 and apoptosis
rates suggested that in CEM-C7-14 cells most cell death is due to DEX-induced apoptosis,
14
whereas in other cell lines alternative cell death pathways may be employed. Therefore, model
predictions that inhibiting STAT5B using DAS should increase GR activity have been partially
validated and uncovered cell-specific effects of such treatment. Furthermore, this investigation
indicated the potential benefits of treating GC-resistant leukaemia with DAS.
To investigate microenvironment effects and cell death pathways, conditioned media (CM)
from bone marrow cell line was used. CM treatment showed an inhibitory trend on Sub-G1
and apoptosis in most cell lines. Studies of cell death markers revealed a complex pattern when
BIM, RIPK1 and LC3 protein levels were analysed in ALL cell lines treated with DEX, DAS,
and their combination, with CM showed an inhibitory trend in these proteins’ expression.
Quantitative Real-time-PCR (qRT-PCR) assay was then used to study the GR and BIM gene
expression. DEX treatment resulted in increased GR expression in CEM-C7-14 and SUP-B15
whereas a decrease was observed in CEM-C1-15 and MOLT-4. The BIM gene was expressed
more in all DEX-treated cell lines except in MOLT-4 cells which exhibited a low expression
pattern.
Another prediction from the model identified several miRs that feature frequently in the model
and have clinical significance. Preliminary experiments using qRT-PCR assay demonstrated
that after DEX treatment, the expression of miR-27a and miR-9 was highly upregulated in GC-
resistant CEM-C1-15, and the expression of miR-20a was also high in MOLT-4. CM and SEVs
caused the substantial increase in miR-27 and miR-9 expression in CEM-C1-15 cells, whereas
the high expression in miR-20a was observed in SUP-B15 cells, suggesting a potential
mechanism for increased resistance. The findings of this study highlight the significance of
comprehending the complex relationships among GR signalling, the tumour microenvironment
(TME), and miRs in order to design efficacious ALL therapies.

• Thesis Type: Thesis
• Online Publication Date: May 29, 2025
• Deposit Date: May 20, 2025
• Publicly Available Date: May 30, 2027
• Award Date: May 29, 2025