Page 22 - CSIR-IGIB Annual Report 2020-21
P. 22
HIF signaling in cardiovascular disease markers will be studied in this model with
pathology cytokines such as TNFα and IL-1β to mimic the
inflammatory milieu.
Hypoxia inducible factor (HIF) triggers an
evolutionarily conserved response, termed the Machine learning in advancing precision
unfolded protein response (UPR). UPR aims to medicine
establish homeostasis by inducing the The potential of machine learning in advancing
expression of genes that enhance the protein precision medicine is immense. For developing
folding capacity of cells. However, in the event the methodology, we specifically targeted
of prolonged hypoxia exposure or genetic statins that are a widely used drug class in the
susceptibilities, the same cells play a key prevention of cardiovascular diseases. In
pathogenic role. Oxygen deficient environments addition, large amounts of data are available in
contribute to inflammation resulting in varied the context of the gene polymorphisms that
life-threatening vascular disease conditions affect statin metabolism. The
such as thrombosis. Hypoxia has also been pharmacogenetics related genes that have
associated with enhanced platelet reactivity in shown their effect on the statin response were
thrombosis. Thus, the relationship between HIF, retrieved from the various sources to train a
UPR and inflammatory pathways is critical in tool named Tanagra. It is adopted to make
hypoxia-induced thrombosis. predictions based on the pharmacogenetics
The activation of NLRP3 inflammasome complex analysis of the general population of patients
along with pro-inflammatory cytokine IL-1β taking statins. A dataset with a total of 249
takes place in response to hypoxia during instances and 5 attributes (Drug, Gene, SNP,
thrombosis. UPR intersects with many different Genotype, and Drug Response) were divided
inflammatory and stress signaling pathways. into two sets, training set (66%) and testing set
The preliminary investigation in our in vitro (34%). The attributes were ranked according to
thrombin-induced thrombosis model indicates a their test-value that indicated the importance of
role of Activating Transcription Factor 4 (ATF4). the difference between the groups. The
ATF4 is selectively up-regulated under hypoxic Decision tree C4.5 and Support Vector Machine
conditions in a PERK-dependent manner. The (SVM) algorithms were compared using the ROC
initial experiments demonstrate the curve that demonstrated the classifier SVM as a
upregulation of the transcription factor CHOP better model. The outcome of the tool
and the E3 ubiquitin ligase, SIAH2. CHOP is demonstrated that the variability in the
involved in ER stress-induced cytokine response to statin is majorly due to the genetic
production in macrophages while SIAH2 factors. This study could be used as a prototype
regulates Prolyl Hydroxylase 3 (PHD3) under low of developing a comprehensive tool that
oxygen concentration, thereby allowing physicians can use and take into account every
accumulation of HIF-1α. In depth investigation patient’s genetic signature associated with
will give specific ER stress markers and UPR individual pharmacokinetics of any drug and
profiles helpful in understanding thrombosis based on that an individualized treatment
disease etiology. In the future we will be regimen benefitting the patients of their time,
studying the effects of UPR markers on health and resources can be devised. We are
inflammatory responses in hypoxia-induced further trying to refine the idea with more
thrombosis using thapsigargin as an inducer of datasets and other tools such as Weka, Orange,
ER stress. Similarly, the effects of an and MILB that can be used for making
inflammatory microenvironment on UPR personalized drug response predictions based
on the genomics information of an individual.
Annual Report 2020-21 19
