Ali Kishk

Drug candidates for repurposing through metabolic modelling of glioma show effectiveness in
agreement with cell lines and xenografts data


Gliomas are the most common type of malignant brain tumour, with glioblastoma having a median
survival of 15 months. Other than brain cancer, with tens of approved chemotherapies per cancer
type as first line of treatment, only five are approved for gliomas and are primarily used as a second
line after surgery and radiotherapy. In addition to histopathological methods, genetic biomarkers
enhanced the diagnosis of gliomas compared to histopathological methods alone. Among others, IDH
mutation status and 1a/19q codeletion are mainly used for subtyping glioma into glioblastoma,
astrocytoma and oligodendrioglioma. Even with the improved diagnosis of glioma subtypes, glioma
patients still suffer from chemotherapy resistance, low survival rates and increased relapse.
Metabolic reprogramming is one of the hallmarks of glioma, causing chemotherapy resistance that is
characterised by Warburg effect and, to a lesser extent, reverse Warburg effect. Drug repurposing,
identifying approved drugs for other diseases, has been established as one central element in drug
discovery to counter chemotherapy resistance. Current preclinical drug repurposing approaches have
been limited to glioblastoma, with a high failure rate in clinical trials due to non- efficacy and toxicity.
In this work, we aim to identify metabolic vulnerabilities and expand the limited anti-glioma drugs by
in silico repurposing of FDA-approved drugs utilising metabolic modelling. Metabolic modelling is a
systems biology approach that analyses cellular metabolism on a genome- scale to generate new
insights. In addition to drug repurposing in cancer, metabolic modelling has been applied in
biomarker detection, medium optimisation, strain design and others. Firstly, we build semi-curated
glioma subtype-specific models using generic metabolic reconstructions and the expression data from the cancer genome atlas of lower-grade gliomas and glioblastoma with the rFASTCORMICS tool.
Secondly, to assess the quality of the models, differences in metabolic pathways between the glioma
subtypes were identified using glioma subtype-specific models. Glutamine and fatty acid dependency
were accurately predicted as metabolic distinctions between the various glioma subtypes. Thirdly, to
repurpose FDA-approved drugs for the glioma subtypes, especially astrocytoma and
oligodendroglioma, in silico drug deletion predicted 33 single drugs and 17 drug combinations. A
literature search of the predicted drugs found that half and a quarter of the single drugs are effective
against glioma cell lines and xenografts, respectively. Moreover, two predicted drugs were found
effective as monotherapy in brain cancer clinical trials. These results show that in silico metabolic
modelling can accurately predict metabolic vulnerabilities and repurposable drugs for the glioma
subtypes that can advance drug discovery.
Keywords: Glioma, Metabolic modelling, Drug repurposing

Leave a Reply

Your email address will not be published. Required fields are marked *