Page 69 - TD-4-1

Tumor Discovery Drug repurposing for pancreatic cancer via AI

proliferation and survival. The IL-17 signaling pathway resistance to chemotherapy. Research has demonstrated
involves proteins such as activator protein 1 (AP-1) that overexpression of MYC enhances the resistance of
and tumor necrosis factor receptor-associated factor 6 PDAC cells to chemotherapy, thereby reducing treatment
51
(TRAF6), which work together to regulate the expression efficacy. By regulating c-MYC to restore cellular function,
of downstream genes. Additionally, tumor necrosis factor the sensitivity of PDAC cells to existing treatments can be
α (TNF-α) is a key cytokine involved in immune responses enhanced, thereby improving therapeutic outcomes.
and inflammation regulation. Signaling through the IL-17 Forkhead box O3 is a TF that plays a crucial role in
pathway often triggers the production of TNFα, thereby regulating cell proliferation, apoptosis, and metabolism. It
52
impacting the inflammatory responses. The activation of has been shown to significantly involve in the development
the IL-17 signaling pathway is regulated by genes including and progression of PDAC. Dysregulation of FOXO3 is
53
IL17B, IL17RB, TRAF6, AP-1, and TNFα. closely linked to the onset and progression of PDAC, with
Through KEGG pathway enrichment analysis and its expression in PDAC being much higher than in normal
54
annotation of the core GWGENs for both PDAC and tissues, although the precise mechanism remains unclear.
healthy controls, as shown in Figure 2, we obtained the Scholars speculate that FOXO3 may be upregulated or
core signaling pathways and their downstream target degraded in cancer tissues, leading to its accumulation in the
genes. This analysis provides a deeper understanding of nucleus and increased expression levels. In normal tissues,
55
the carcinogenic mechanisms of PDAC. By comparing FOXO3 is transcribed at low levels and rapidly degraded.
the core signaling pathways and their downstream target In our study, FOXO3 regulates several target genes, and
genes between PDAC and healthy controls, we identified our goal is to restore FOXO3 to normal expression levels
significant biomarkers as drug targets, thus facilitating to mitigate downstream cellular dysfunction.
the discovery of more effective multi-molecular drugs for The TP53 gene encodes the tumor suppressor protein
treating PDAC. Furthermore, recent studies have suggested p53, one of the most commonly mutated genes in PDAC.
56
that miRNAs may be the answer to cancer treatment. Based TP53 regulates the expression of many genes involved in
on the core GWGENs of healthy controls, we identified cell cycle control, apoptosis, and senescence, as well as
miRNA LET-7C. According to recent research, miRNA influencing the immune microenvironment of PDAC.
LET-7C regulates key genes such as Ras, STAT1/3, and Inactivation of the TP53 gene disrupts these processes,
c-MYC in the core signaling pathways of PDAC. 45 leading to altered immune responses. Research indicates
56
3.4. Selection of drug targets by investigating that the loss of TP53 expression allows cells harboring
the pathogenesis of PDAC and systematic drug the KRASG12D mutation to survive, promoting tumor
57
repurposing via deep neural network-based drug- formation and metastasis. Our goal is to restore cellular
target interactions model function to normal by targeting mutated TP53, thereby
achieving therapeutic effects.
Based on the core signaling pathways of PDAC patients and After selecting these important biomarkers of PDAC
the abnormal downstream cellular dysfunctions compared carcinogenesis as potential drug targets, we obtained DTI
with the healthy controls in Figure 2, we selected key 31 32 33
biomarkers, such as c-MYC, FOXO3, and TP53, as drug data from databases such as KEGG, BIDD, UniProt,
34
37
36
35
targets. Our aim is to identify potential multi-molecular DrugBank, PubChem, ChEMBL, and STITCH. Using
these data, we trained a DNN-based DTI model to predict
drugs that could restore the abnormal expression of these molecular drugs for these targets. The training dataset
key biomarkers to normal levels for the treatment of PDAC.
consists of 180,315 drug-target interactions, including
c-MYC, a TF located downstream of the MAPK 80,291 experimentally verified drug-target interactions
signaling pathway, is important in regulating cell growth, and 100,024 unverified interactions. To address the
proliferation, and apoptosis. Research indicates that potential imbalance between the two categories in the
46
overexpression of c-MYC is strongly associated with the training dataset, we randomly selected an equal number
development of PDAC and plays a significant role in its early of drug-target interactions from each database. Before
transformation. Excessive activation of c-MYC promotes training the DNN-DTI model, the interaction data were
47
the proliferation of PDAC cells, thereby accelerating transformed and standardized, followed by dimensionality
tumor growth. Moreover, c-MYC inhibits apoptosis reduction using PCA, which reduced the transformed
48
in PDAC cells, further enhancing tumor progression. features from 1,359 to 996. This step is necessary to
49
Thus, downregulating c-MYC can effectively inhibit match the input layer dimensions of the model. As shown
PDAC cell proliferation and survival, offering therapeutic in Figure 3, the input layer of the DNN-DTI model
potential. Additionally, c-MYC contributes to PDAC consists of 996 nodes, followed by four hidden layers of
50
Volume 4 Issue 1 (2025) 61 doi: 10.36922/td.4709