Abstract
Pancreatic Ductal Adenocarcinoma (PDAC) is a devastating disease with
poor clinical outcomes, which is mainly because of delayed disease
detection, resistance to chemotherapy, and lack of specific targeted
therapies. The disease’s development involves complex interactions
among immunological, genetic, and environmental factors, yet its
molecular mechanism remains elusive. A major challenge in understanding
PDAC etiology lies in unraveling the genetic profiling that governs the
PDAC network. To address this, we examined the gene expression profile
of PDAC and compared it with that of healthy controls, identifying
differentially expressed genes (DEGs). These DEGs formed the basis for
constructing the PDAC protein interaction network, and their network
topological properties were calculated. It was found that the PDAC
network self-organizes into a scale-free fractal state with weakly
hierarchical organization. Newman and Girvan’s algorithm (leading
eigenvector (LEV) method) of community detection enumerated four
communities leading to at least one motif defined by G (3,3). Our
analysis revealed 33 key regulators were predominantly enriched in
neuroactive ligand-receptor interaction, Cell adhesion molecules,
Leukocyte transendothelial migration pathways; positive regulation of
cell proliferation, positive regulation of protein kinase B signaling
biological functions; G-protein beta-subunit binding, receptor binding
molecular functions etc. Transcription Factor and mi-RNA of the key
regulators were obtained. Recognizing the therapeutic potential and
biomarker significance of PDAC Key regulators, we also identified
approved drugs for specific genes. However, it is imperative to subject
Key regulators to experimental validation to establish their efficacy
in the context of PDAC.
Introduction
Pancreatic ductal adenocarcinoma (PDAC) is a common type of cancer
originating from the pancreatic glands and is characterized by a
rapidly progressive course and a dismal prognosis [[26]1]. Combination
strategies targeting multiple signaling pathways supporting tumor
growth and propagation are active areas of contemporary research that
will likely transform the treatment paradigm of pancreatic cancer
[[27]2]. This is a highly detrimental disease with dismal clinical
outcomes, primarily attributable to delayed disease detection,
chemotherapy resistance, and absence of specific targeted therapies.
The identification of novel therapeutic targets and/or early biomarkers
for the disease has the potential to significantly improve the clinical
management of PDAC and extend the lives of patients. To develop
successful drug therapies, a deeper understanding of the molecular
mechanisms underlying drug targeting is essential. A connectivity Map
is a platform that provides information on the signaling pathways
activated by a particular drug and can serve as a valuable resource for
drug development. A promising strategy for treating PDAC involves
exploiting aberrant metabolic processes in cancer cells, particularly
PDAC cells. Cancer cells alter their metabolic pathways, a process
regulated by complex and poorly defined interplay between intrinsic and
extrinsic factors.
The incidence of PDAC is rising, and it is now one of the leading
causes of cancer-related deaths in developed countries [[28]3]. Current
treatment paradigms for PDAC, including targeted therapy, have shown
limited success in improving survival outcomes [[29]4]. The molecular
heterogeneity of PDAC and its complex tumor microenvironment contribute
to its resistance to conventional treatments and immunotherapies
[[30]5]. Retinoids, such as retinoic acid (RA), have shown potential in
maintaining normal pancreatic functions and have been explored as a
therapeutic option for PDAC [[31]6]. The use of predictive molecular
markers and cancer gene panel testing may help in selecting
personalized therapies for PDAC patients [[32]7]. Accurate diagnosis of
PDAC is crucial, as it can be easily misdiagnosed as other pancreatic
neoplasms, such as acinar cell carcinoma (ACC) or neuroendocrine tumor
(PNET). Further research is needed to overcome the challenges in PDAC
management and improve patient outcomes. Recent research is exploring a
wide range of novel therapeutic targets for PDAC, including genomic
alterations, tumor microenvironment, and tumor metabolism [[33]8]. The
rapid evolution of tumor genome sequencing technologies paves the way
for personalized, targeted therapies for PDAC [[34]9]. Immunotherapy
with immune checkpoint inhibitors has shown promise in PDAC,
particularly in tumors harboring mismatch repair deficiency (dMMR) and
high microsatellite instability (MSI-H) [[35]4]. The presence of a
dense fibrotic stroma in PDAC creates a physical barrier around the
cancer cells, hindering drug delivery and promoting tumor growth and
treatment resistance [[36]10]. Advances in surgical approaches,
immunotherapeutic approaches, and targeted therapies are being explored
to overcome the treatment refractory nature of PDAC and improve patient
outcomes [[37]11]. The tumor microenvironment (TME) plays a crucial
role in PDAC development, progression, and treatment resistance
[[38]12]. Immunotherapy, including checkpoint inhibition and
immune-based therapies, has shown promise in PDAC treatment [[39]13].
Additionally, targeted therapies that focus on specific signaling
pathways and components of the TME, such as fibro-blasts and
cancer-associated fibroblasts, are being explored.
PDAC is an aggressive and deadly cancer with limited treatment options.
Current standard of care treatments for advanced PDAC include systemic
chemotherapy regimens such as gemcitabine/nab-paclitaxel and
FOLFIRINOX, which have improved clinical outcomes [[40]14]. However,
the 5-year survival rate for PDAC remains low, highlighting the need
for new therapies [[41]15]. The tumor microenvironment (TME) of PDAC
plays a significant role in tumorigenesis and may contain promising
novel targets for therapy [[42]16]. Adjuvant chemotherapy has been
shown to offer a survival advantage over surgery alone in resected PDAC
[[43]17]. However, the addition of chemoradiation therapy (CRT) to
chemotherapy does not provide a survival advantage and is not
recommended [[44]18]. Surgical resection is the first choice for
treatment of pancreatic acinar cell carcinoma (PACC), but there is no
standard treatment option for inoperable disease. Understanding the
risk factors for PDAC can help in screening and counseling patients for
lifestyle modifications.
A disorder typically arises from disruptions within the intricate
internal web of connections [[45]19] among genes that serve related
functions, rather than from the abnormality of a single gene. This
resulted in the adoption of a systemic approach to biological issues,
founded on the principle that comprehending the involvement of
different genes/proteins is essential [[46]20] in disease initiation
and progression. It is necessary to consider the entire network of
interactions within a living system [[47]21]. In this context, it is
important to employ Network Medicine [[48]22], as this method aids in
the investigation of disease pathways and modules linked to complex
illnesses.
In this study, the protein interaction maps are analyzed through
graph/network theory to get insights about the theoretical aspect of
complex networks [[49]23]. According to the graph theory, analysis of
the topological structure of a network (PPI network in our study)
provides important information of the network [[50]24] through which
novel disease genes and pathways, biomarkers and drug targets for
complex diseases can be identified [[51]25].
Thus, the focus of our study is on the protein-protein interactions
network/graph of PDAC, constructed from differentially expressed genes
with an aim to understand the architectural principle of the
network/graph (random, small world, scale-free or hierarchical). We
further extended our study to the prediction of important key
regulators of the network which have fundamental importance due to
their activities and regulating mechanisms in the network [[52]26]. It
is expected that the findings of this study will advance our
understanding of the initiation and progression of PDAC, thereby,
strengthening different therapeutic approaches for PDAC.
This work was to establish an unbiased catalogue of change in (up and
down regulation) gene expression for the PDAC samples. we bestowed 12
samples of PDAC analysis of gene-expression from normal and cancer
patients. Finally, the comparative study of these conditions (PDAC
Cancer and Normal) has affirmed the thirty-three (33) prominent genes.
These are intricated in basic functioning of the cell like as in the
rearrangement of the cytoskeleton, tissues development and activation
of immune system. This work will be useful for the researchers to
collect evidences against PDAC working in in-vitro.
Materials and methods
Preprocessing and acquisition of dataset
The RNA-seq dataset [53]GSE171485 [[54]27] was downloaded from the NCBI
GEO database [[55]https://www.ncbi.nlm.nih.gov/].
Data analysis and visualization of differentially expressed genes
The data analyses were conducted using R Studio (4.2.2) and R. The fold
change method was employed to identify differentially expressed genes.
The fold change for a gene is calculated by subtracting the intensities
measured in two samples (control vs clinical groups). This value is
referred to as the fold change. These raw values are typically log
transformed (usually log2) [[56]28]. Another method to calculate the
fold change ratio involves dividing the two measured intensities for a
given gene in two samples. A change of at least two-fold (up or down
regulated) considered to be significant.
Screening of DEGs
We utilized the LIMMA package [[57]29] to analyze the data and identify
Differentially expressed genes (DEGs) by assessing gene expression
values. This approach employs Linear Modeling (LM) [[58]30] and
Empirical Bayes (EB) [[59]31] techniques to perform f-tests [[60]32]
and t-tests [[61]33], while simultaneously reducing standard errors.
The objective of this method is to produce results that are dependable
and reproducible, thereby enhancing the precision and reliability of
the statistical analyses. DEG analysis was conducted to compare gene
expression levels between control and clinical groups. The R package
"ggplot2" [[62]34] was employed for data visualization purposes. The
fold change criteria for each gene were established based on a
Benjamin-Hochberg adjusted p-value threshold of 0.05 and a significance
level of p ≤ 0.05.
Construction of protein-protein interaction network of PDAC
The construction of the PDAC PPI (Protein-Protein Interaction) network
was undertaken using the STRING database (The Search Tool for the
Retrieval of Interacting Genes, (<[63]http://string-db.org/>) with an
interaction score threshold of > 0.4. This approach enables the
exploration and analysis of protein-protein interactions, which can be
either physical or functional associations. These associations are
derived from text-mining of literature, co-expression analysis,
genomics context-based predictions, computational predictions, and
high-throughput experimental data, as well as the aggregation of
previous knowledge from other databases. The network was subsequently
visualized and analyzed using the Cytoscape software (version 3.6.1)
[[64]35].
Gene ontology (GO) enrichment analysis
Gene ontology terms provide a controlled vocabulary that is divided
into three categories: Molecular Function, Biological Process, and
Cellular Location. To conduct a preliminary investigation into the
functional differences of DEGs, they were submitted to DAVID (Database
for Annotation, Visualization and Integrated Discovery), an online
software (<[65]http://david.abcc.ncifcrf.gov/home.jsp>), to enrich the
set of DEGs with possible GO terms [[66]36].
Delineation of global network topological properties
The structure of the network can be examined through its topological
properties, which establish the connections between nodes and
illustrate their interactions. The topology of the network is defined
by the probability of degree distributions (P(k)), clustering
coefficient (C(k)), and neighborhood connectivity (C[N](k)), which
exhibit a power law distribution. This adherence to the power law
distribution indicates the presence or absence of scale-free properties
in the network. The network constructed for PDAC using differentially
expressed genes extracted from the RNA-seq dataset [67]GSE171485
follows the power law distribution in its probability of degree
distributions (P(k)), clustering coefficient (C(k)), and neighborhood
connectivity (C[N](k)). This is depicted in “[68]Fig 4” and confirmed
by the matrix representation of all properties [[69]37].
Fig 4. Topological properties of Network with power law fit.
[70]Fig 4
[71]Open in a new tab
(A) Probability of node degree distribution. (B) Clustering coefficient
vs degree distribution. (C) Betweenness centrality vs degree
distribution. (D) Closeness centrality vs degree distribution (E)
Neighborhood connectivity vs degree distribution. (F) Eigenvector value
vs degree distribution.
The network’s behavior is characterized by equations (see Eqs [72]1,
[73]2, [74]3) that reveal a hierarchical structure in synergy with the
scale-free nature of the network. The power law distribution was fitted
to the topological properties of the network using a standard
statistical fitting procedure.
[MATH: P(k)∼k−α<
/mi> :MATH]
(1)
[MATH: C(k)∼k−β<
/mi> :MATH]
(2)
[MATH: CN(k)∼k∅ :MATH]
(3)
The positive value in phi of connectivity parameter shows assortative
nature of the network. While, the negative value in alpha (α) of degree
distribution shows availability of each node in the network. The
negative value in beta of clustering parameter shows disassortative in
the communication between the nodes in network [[75]23].
Communities finding: LEV method
To investigate the nature and topological properties of hierarchical
networks at various levels of organization is beneficial for
understanding network behavior and uncovering the organizing principles
that govern them. There are several methods for detecting communities
within these networks, one of which is the leading eigenvector (LEV)
method. This method calculates the eigenvalue for each edge, giving
greater importance to links rather than nodes. For this study, we
applied the LEV detection method using the ‘igraph’ [[76]38] package in
R. We used this code to detect modules in the main network, sub-modules
within each module at different levels of organization, and so on,
ultimately identifying motifs (i.e. 3 nodes and 3 edges) [[77]39].
Throughout this process, we adhered to the criterion of identifying any
sub-module as a community if it contained at least one motif (defined
by G (3,3)).
TF/miRNAs screening
To identify miRNAs targeting key signature genes, we employed
MIENTURNET (MicroRNA ENrichment TURned NETwork). MIENTURNET [[78]40] is
an interactive web application designed for micro-RNA target enrichment
analysis, primarily relying on the TargetScan program for
sequence-based miRNAs target predictions [[79]41]. By utilizing the
MIENTURNET software, we successfully achieved significant functional
enrichment of the predicted miRNAs.
Drug gene interaction analysis
To identify potential drugs for treating PDAC, we utilized the DGIdb
web tool [[80]42] a database of drug-gene interactions and druggable
genes.
Results
This work provides information in RNA-seq dataset [81]GSE171485 on the
structure and function of interacting genes. The results of the
differential expression analysis, as shown in volcano plot “[82]Fig 1”,
indicate that 772 DEGs were identified. Of these, 341 genes were found
to be down-regulated and 431 were up-regulated, with the threshold
cut-offs set at log2 |FC| > = 1,P-value < 0.05, and Padj ≤ 0.05.
Fig 1.
[83]Fig 1
[84]Open in a new tab
Figure showing (A) volcano plot showing up & down regulated expressed
genes (B) Network of differentially expressed genes.
To get the idea of what could be the effect of DEGs can only be
possible when we have some preliminary insight in to their individual
function. As the term gene ontology (GO) enrichment gives us the
opportunity to get the basic idea about any gene. GO enrichment
analysis of significantly enriched DEGs between disease and control was
categorized into biological process, cellular components and molecular
function. Among these down-regulated DEGs, potassium ion transport,
cell-cell signaling, neuropeptide signaling pathway were most
significantly down-regulated biological process in disease “[85]Fig
2A”. While other important biological process associated with these
down-regulated DEGs were muscle contraction, cellular response to zinc
ion, response to hypoxia, negative regulation of cytosolic calcium ion
concentration and positive regulation of apoptotic process. While most
significantly down-regulated cellular components and molecular function
associated with the DEGs were extracellular region, mitochondrion,
potassium channel complex, endoplasmic reticulum, mitochondrion inner
membrane and protein binding, methyltransferase activity, structural
constituent of ribosome respectively “[86]Fig 2B and 2C”. KEGG pathways
enrichment of down-regulated genes were enriched in metabolic pathway,
peroxisome cAMP signaling pathway and chemical carcinogenesis-reactive
oxygen species “[87]Fig 2D”.
Fig 2. Gene ontology analysis for down-regulated genes of pancreatic ductal
adenocarcinoma.
[88]Fig 2
[89]Open in a new tab
Among the top 15 up-regulated DEGs, cell-cell adhesion, positive
regulation of cell migration, extracellular matrix organization, tissue
development, epithelial cell differentiation and integrin-mediated
signaling pathway were most significantly up-regulated biological
process in PDAC “[90]Fig 3A”. GO classification indicates the
up-regulation of calcium ion binding, calcium-dependent protein
binding, cadherin binding, virus receptor activity and insulin-like
growth factor I binding related molecular function in PDAC “[91]Fig
3B”.
Fig 3. Gene ontology analysis for up-regulated genes of pancreatic ductal
adenocarcinoma.
[92]Fig 3
[93]Open in a new tab
While most significantly up-regulated cellular components were plasma
membrane, extracellular exosome, and perinuclear region of cytoplasm
Figure as shown in “[94]Fig 3C”. KEGG pathways were enriched in
ECM-receptor interaction, p53 signaling pathway, pathways in cancer and
proteoglycan in cancer “[95]Fig 3D”.
PDAC network architecture reveals hierarchical scale-free features
To gain insights into the PPI network of PDAC’s structural features, we
analyzed its topology, specifically the probability of degree
distribution P(k), clustering coefficient C(k) neighborhood
connectivity C[N](k), and centrality measurements. We employed the
statistical fitting technique proposed by Clauset et al. [[96]43] to
verify that the graph’s architecture follows power-law behavior. Our
results indicate that all statistical p-values, calculated against 2500
random samplings, are greater than the critical value of 0.1, and the
goodness of fits are less than or equal to 0.35. The data points of all
the topological parameters fit power law when plotted against the
degree k of the PDAC network.
The values of the power-law exponents for each of the topological
properties of the complete network were calculated:
[MATH: PCCN∼K−αK−βK+∅;α0β0<
/mtd>∅0→0.294
0.1650.122 :MATH]
The negative values of α = 0.294; (α<2) and β = 0.165; (β<1) suggest
the hierarchical nature of the PD network, indicating the existence of
well-defined successive interconnected communities with sparsely
distributed hubs in the network. The values of the exponents of P(k),
C(k), and C[N] (k)(∅ = 0.122; (+∅ ≤ 0.5)) suggest that the network,
though not strongly hierarchical, falls into the category of a weak
hierarchical scale-free network. The negative value of β indicates that
as k increases, C(k) decreases, suggesting that nodes with a high
degree have a low tendency to cluster, further indicating a hierarchy
of hubs, in which the most densely connected hub is linked to a small
fraction of all other nodes. The power-law distribution observed in
P(k) is indicative of the scale-free nature of the small-world network,
with the negative value of α indicating that a small number of nodes
possess a high degree while the majority of nodes have a low degree,
which is consistent with the network’s scale-free behavior. The
positive value of ∅ suggests that the network exhibits assortative
mixing, where edges predominantly connect heavily connected nodes,
regulating the system’s behavior.
Further,
[MATH: CB
CC
CE
∼Kγ
Kδ
Kψ
;γ0
δ0
Ψ0
→0.3020.0951.008 :MATH]
The positive values of the exponents γ, δ, and ψ of the three
distributions C[B](k), C[c](k) and C[E](k) respectively, indicate that
the network exhibits hierarchical scale-free or fractal features. These
positive values signify that C[B](k), C[c](k) and C[E](k) increase as
the degree k increases when plotted against it “[97]Fig 4C, 4D and 4F”.
The increasing value of C[B](k) as k increases suggests that nodes with
a high degree have high C[B](k), indicating that these larger hubs have
a major influence on the information transmission in the network
compared to nodes with a low degree. Similarly, the direct
proportionality between C[c](k) and k suggests that high-degree nodes
are quick spreaders of information in the network, indicating their
high C[c](k). The positive value of ψ indicates that nodes with a high
degree have high C[E](k) as well, suggesting their influence in the
network due to their ability to spread information. The positive value
of ψ also signifies the connectedness between high-degree nodes, which
is in agreement with the assortative mixing in the network.
Through a meticulous study of these topological properties, it was
found that the PDAC network self-organizes into a scale-free fractal
state with weakly hierarchical organization.
Key regulators uncovered through clustering and tracing
Owing to the significance of regulator genes(RGs) as functional
bottlenecks in the initiation and progression of a disease by
regulating the expression of a plethora of downstream effector genes
[[98]44], we identified the most potent RGs of the PDAC network. Newman
and Girvan’s algorithm helped us untangle the PDAC network and the
network was observed to be organized in five hierarchical levels using
this algorithm “[99]Fig 5”. After tracing of the G(3,3) triangular
structure genes from top to bottom organization through these levels of
hierarchy, 33 genes were revealed to be the RGs of the PDAC network,
the criterion being their presence at every topological level “[100]Fig
5”. This agrees with the definition of RGs, according to which, RGs are
the genes/proteins which are deeply rooted from top to bottom
organization of the network. These RGs are the backbone in maintaining
a network’s stability as they capacitate the network to combat any
unacceptable alterations in it.
Fig 5. Tracking down the presence of the fundamental genes within different
Modules at different level of the network.
[101]Fig 5
[102]Open in a new tab
RNF213, EPSTI1 and XAF1 separated their way from the rest of the RGs
from the first level itself and then took the path hand in hand till
the motif level. Whereas, GAL, VIP, GNRH1,NMU, VIPR2, and GNG11 RGs all
moved into a different sub-module of the first level and took path at
the motif level. Genes MMP1, CTSG, F2RL1, F10, PLAT, F2R, PDGFA, FGF13,
PTN, APOC2, SDC1, and SDC4 moved into same module at level 2. Further,
genes MMP1, CTSG, F2RL1, F10, PLAT, and F2R makes sub-module and
belongs to same sub-module till level 4 and separated at level 5 as
triangular motif structure “[103]Table 1”.
Table 1. Network breaking mechanism to understand the gene names and their
sub-modules.
Network decomposition outline Gene names in motifs
C→C2→C2.3→C2.3.4→C2.3.4.1 F10, PLAT, F2R
C→C2→C2.3→C2.3.4→C2.3.4.2 MMP1, CTSG, F2RL1
C→C2→C2.4→C2.4.1→C2.4.1.1 PDGFA, FGF13, PTN
C→C2→C2.4→C2.4.2→C2.4.2.2 APOC2, SDC1, SDC4
C→C3→C3.3→C3.3.2→C3.3.2.3 AP2B1, AP1S3, DNM2
C→C3→C3.4→C3.4.1→C3.4.1.2 EGFL7, RASIP1, SOX18
C→C3→C3.5→C3.5.1→C3.5.1.1 CLDN2, CLDN7, CDH3
C→C3→C3.5→C3.5.2→C3.5.2.1 CLDN5, TJP1, PCDH1
C→C6→C6.2→C6.2.1→C6.2.1.1 RNF213, EPSTI1, XAF1
C→C9→C9.3→C9.3.1→C9.3.1.1 VIP, GAL, GNRH1
C→C9→C9.3→C9.3.4→C9.3.4.1 NMU, VIPR2, GNG11
[104]Open in a new tab
In the third level, PDGFA, FGF13, PTN, APOC2, SDC1, and SDC4 moved into
the same sub-module and got separated at level four into two
sub-modules. Afterward, these RGs moved separately till they reached
the motif level i.e., the 5th level. Genes AP2B1, AP1S3 DNM2, EGFL7,
RASIP1, SOX18, CLDN2, CDH3, TJP1, CLDN5 and PCDH1 clustered in same
sub-module at level 2 and got separated at levels 3, 4 and reached
motif levels “[105]Fig 6”.
Fig 6. Detecting communities within PDAC networks using leading eigenvector
(LEV) method.
[106]Fig 6
[107]Open in a new tab
Emergence of low degree node accompanied by high degree node as key
regulators
In a network, when a node’s degree is low, the node gains what strength
it has from its neighbors and thus the influence it has over the
network is a function of its neighboring degree. Whereas for high
degree nodes, the strength of the nodes comes from their large number
of connections rather than their neighboring degree [[108]45]. In
addition, a low degree bridge node, connecting two high degree nodes,
is very important in a network despite its lower degree [[109]46].
Thus, a node’s degree is not the sole determinant of its essentiality,
rather, it depends on the topological position of that node. This is
reflected in our results, where, F2RL1, F10, APOC2, CLDN2, PCDH1 which
has quite a low degree (4,4,8,11,3) respectively in the primary
network, found out to be RGs based on its ability to make it to the
last level of the organization. F2RL1 formed a motif in the last level
of the organization with another KR, MMP1, which has a fairly high
degree (36) in the primary network. Gene F10 formed a motif with PLAT
having degree 17, APOC2 formed motif with SDC1 having degree 40 and
PCDH1 formed motif with TJP1 which fairly high degree is [[110]42] in
primary network. This shows that a low degree node i.e., F2RL1, F10,
APOC2, CLDN2, and PCDH1 are also an important information propagator in
the network, contributing to serving as the backbone of the network and
functioning “[111]Table 2”.
Table 2. Topological statistical properties of low degree node accompanied by
high degree node as key regulators.
S.N. Gene name Degree Betweenness Centrality Closeness Centrality
Clustering Coefficient
1 F2RL1 4 0.0007 0.2791 0.6667
2 F10 4 0.0008 0.3041 0.1666
3 APOC2 8 0.0066 0.3030 0.2142
4 CLDN2 11 0.0009 0.3351 0.8727
5 PCDH1 3 0.0001 0.2835 0.3334
6 MMP1 36 0.0241 0.3742 0.2730
7 PLAT 17 0.0091 0.3517 0.2332
8 SDC1 40 0.0218 0.3735 0.2846
9 TJP1 44 0.0275 0.3796 0.1987
[112]Open in a new tab
Furthermore, gene-drug interaction from DGIdb were retrieved and listed
in “[113]Table 3”.
Table 3. Gene-drug interaction of the key regulators.
S.N. Gene Drug
1 F10 EDOXABAN, IDRAPARINUX SODIUM, APIXABAN, RIVAROXABAN, HEPARIN,
EDOXABAN TOSYLATE, EMICIZUMAB, BETRIXABAN, LETAXABAN, MENADIONE,
OTAMIXABAN, IDRAPARINUX, TANOGITRAN, MELAGATRAN, TINZAPARIN,
DANAPAROID, CHEMBL1271162, SEMULOPARIN, MANGIFERIN HEPTASULFATE,
THROMBIN, TRIMETHOPRIM/SULFADOXINE
2 PLAT AMINOCAPROIC ACID, ATORVASTATIN, MELPHALAN, EPOETIN BETA,
RALOXIFENE, NAPROXEN, UROKINASE, BORTEZOMIB
3 F2R RIGOSERTIB SODIUM, ATOPAXAR, VORAPAXAR SULFATE, VORAPAXAR,
ATROPINE, WORTMANNIN, ARGATROBAN, BLEOMYCIN, LEPIRUDIN, THALIDOMIDE,
MORPHINE, ALCOHOL, ASPIRIN, THROMBIN, DALTEPARIN, RUSALATIDE
4 MMP1 DOXYCYCLINE CALCIUM, APRATASTAT, DOXYCYCLINE, DOXYCYCLINE
HYCLATE, CIPEMASTAT, MARIMASTAT, PRINOMASTAT, LEUPROLIDE ACETATE,
SIROLIMUS, COLLAGENASE CLOSTRIDIUM HISTOLYTICUM, MEDROXYPROGESTERONE
ACETATE, LAMIVUDINE, LEFLUNOMIDE, HYDROCORTISONE, PENTOSAN POLYSULFATE
SODIUM, TRIAMCINOLONE, RIBAVIRIN
5 CTSG MANNITOL, CHEMBL374027
6 F2RL1 CHEMBL493076, CHEMBL494502, ROXITHROMYCIN, MINOCYCLINE,
TETRACYCLINE, ERYTHROMYCIN, CLARITHROMYCIN, DOXYCYCLINE
7 PDGFA SUNITINIB, SQUALAMINE
8 SDC1 HEPARIN, INDATUXIMAB RAVTANSINE
9 EGFL7 PARSATUZUMAB
10 CDH3 PF-06671008
11 TJP1 RISPERIDONE, GENISTEIN, ALCOHOL, DEXAMETHASONE
12 VIP DIGOXIN, LISINOPRIL, OMEPRAZOLE, RIBAVIRIN, ANDROSTANOLONE,
AZASERINE, FLUTAMIDE
13 GAL LIOTHYRONINE SODIUM, HYDROCORTISONE, PRASTERONE,
DIACETYLMORPHINE, MIRTAZAPINE
14 GNRH1 RESERPINE, RALOXIFENE, GOSERELIN, AMINOGLUTETHIMIDE,
ZALCITABINE, DAPSONE, CAPTOPRIL, CHEMBL208519, LITHIUM, LEUPROLIDE,
DITIOCARB
[114]Open in a new tab
Next, transcription factors were identified with TargetScan and listed
in “[115]Table 4”.
Table 4. Transcription factor of Key regulator genes.
S.N. Key TF Description P-value Key regulator Genes
1 NF1 neurofibromin 1 0.000266 GNRH1,PLAT
2 SP3 Sp3 transcription factor 0.001 F2R,PLAT,SOX18
3 SP1 Sp1 transcription factor 0.00127 CDH3,F10,PLAT,PTN,F2R
4 TWIST1 twist basic helix-loop-helix transcription factor 1 0.0017
F2R,MMP1
5 MYB v-myb myeloblastosis viral oncogene homolog (avian) 0.0019
NMU,CTSG
6 GATA3 GATA binding protein 3 0.002 CDH3,MMP1
7 POU2F1 POU class 2 homeobox 1 0.002 GNRH1,SDC4
8 JUN jun proto-oncogene 0.00221 PLAT,PTN,MMP1
9 PPARG peroxisome proliferator-activated receptor gamma 0.00593
MMP1,CLDN2
10 HDAC1 histone deacetylase 1 0.00684 TJP1,CLDN7
11 STAT1 signal transducer and activator of transcription 1, 91kDa
0.00945 XAF1,VIP
12 CREB1 cAMP responsive element binding protein 1 0.0108 VIP,PLAT
13 STAT3 signal transducer and activator of transcription 3
(acute-phase response factor) 0.0255 F2R,MMP1
14 RELA v-rel reticuloendotheliosis viral oncogene homolog A (avian)
0.097 SDC1,MMP1
15 NFKB1 nuclear factor of kappa light polypeptide gene enhancer in
B-cells 1 0.0981 SDC1,MMP1
[116]Open in a new tab
Using miRTarBase, microRNAs were identified as listed in “[117]Table
5”.
Table 5. miRTarBase scanned microRNAs with respect to key regulators.
S.N. microRNA Target Gene Number of interactions FDR p-value
1 hsa-miR-1-3p RNF213, F2RL1, AP1S3 4 0.619 0.0885
2 hsa-miR-335-5p F10, RNF213, VIP, CDH3, PLAT, CLDN7, XAF1, RASIP1,
SOX18 9 0.619 0.043
3 hsa-miR-526b-3p MMP1, F2RL1, F2R 3 0.619 0.114
4 hsa-miR-615-3p TJP1, DNM2, AP2B1 3 0.619 0.229
5 hsa-miR-106b-5p AP2B1, F2R, F2RL1 3 0.626 0.331
6 hsa-miR-17-5p F2RL1, F2R, TJP1 3 0.641 0.378
7 hsa-miR-124-3p DNM2, SDC4, RASIP1 3 0.666 0.513
[118]Open in a new tab
Discussion
The scarcity of knowledge regarding the genetic origins of PDAC
necessitates further investigation into its genetic aspects. Despite
the identification of several candidate genes for PDAC in recent years,
the underlying mechanism responsible for PDAC development remains
unclear. To the best of our knowledge, our in-silico study constitutes
the initial endeavor to investigate the Protein-Protein Interaction
(PPI) network of PDAC and to explore the varying contributions of the
proteins encoded by the candidate genes in regulating the entire
network through topological analysis.
Our investigation of the topological properties of the initial PDAC
network, comprising 605 nodes and 2698 edges, reveals a weak
hierarchical and scale-free fractal structure. The Girvan Newman
algorithm helps us to identify a two-tier organization of the network,
with one level representing local clustering of mostly low-degree nodes
into well-defined successive communities or modules, and the other
level representing more global connectivity in which hubs serve as
higher-order communication points between interconnected communities.
The fractal state of the network signifies self-similar organization,
while the scale-free nature contributes to network stability. These
topological properties facilitate efficient information processing
within the network.
The objective of predicting candidate genes for diseases, such as
exploring the role of gene interactions, is a fundamental goal of the
medical sciences and is crucial for effective treatment. To achieve
this, it is necessary to conduct preliminary analysis to identify
potential biomarkers. In this study, we applied differential expression
analysis on RNA-seq data of PDAC to identify transcriptomic signatures
that are characteristic of the disease, and then performed network
analysis to better understand the interactions between genes. The
results of this study can be used to identify potential biomarkers for
PDAC and contribute to the field of pharmacogenomics, which has
significant applications in drug discovery. We focused on
network-regulated genes in this study and found that the network of
classified genes from PDAC displays hierarchical characteristics,
indicating that the network is organized at the sub-module level. This
hierarchical nature of the network is important for understanding the
functional regulation of the disease. Individual gene activities are
less important in this process.
Our networks, which comprise both up- and down-regulated genes, have
led to the identification of 33 crucial regulators, including (F10^↓,
PLAT^↑,F2R^↑, MMP1^↑,CTSG^↓, F2RL1^↓,PDGFA^↓, FGF13^↓, PTN^↓,APOC2^↑,
SDC1^↑,SDC4^↑, AP2B1^↑, AP1S3^↑, DNM2^↑, EGFL7^↓, RASIP1^↓, SOX18^↓,
CLDN2^↑, CLDN7^↑, CDH3^↑, CLDN5^↓, TJP1^↑, PCDH1^↓, RNF213^↑, EPSTI1^↑,
XAF1^↑, VIP^↓, GAL^↓, GNRH1^↓, NMU^↑,VIPR2^↓, and GNG11^↓). These
regulators were identified through the analysis of motifs and module
regulation, and their biological importance, roles in network
activities and associated regulations, and potential as targets for
disease have been established. Additionally, the biological activities
and pathways in which these key regulator genes are involved have been
identified.
Studies have shown the strong potential of F10 to improve treatment
outcomes in acute myeloid leukemia, acute lymphocytic leukemia,
glioblastoma, and prostate cancer [[119]47]. The PLAT gene has been
studied in the context of cancer in several papers. PLAT, also known as
tissue-type plasminogen activator, has been found to play a role in
gefitinib resistance in non-small cell lung cancer (NSCLC) [[120]48].
Mutations and alterations in the FGFR2 gene have been found to play a
significant role in the development and progression of various types of
cancer, including endometrial cancer, breast cancer, and
gastrointestinal/genitourinary tract cancers. These alterations include
somatic hotspot mutations, structural amplifications, and fusions
[[121]49]. Pleiotrophin (PTN) is a gene that has been found to be
differentially expressed in various types of cancer, including
hepatocellular carcinoma (HCC) [[122]50], oral squamous cell carcinoma
(OSCC), ovarian cancer, and breast cancer (BrCa). Tight junction
proteins ZO-1, TJP1, TJP2, and TJP3 are scaffolding proteins that
connect trans-membrane proteins like claudins and occludin to the actin
cytoskeleton. They play a crucial role in maintaining the integrity of
tight junctions and regulating para-cellular permeability [[123]51].
F2RL1, also known as Fc receptor-like 2, has been studied in the
context of cancer. In metastatic breast cancer, decreased expression of
FCRL2 mRNA was observed in brain metastatic tissues compared to primary
breast tumors, and its expression in primary tumors was correlated with
patient survival [[124]52]. PDGFA is a protein that has been implicated
in cancer initiation and progression. It is up-regulated in several
cancers, including colorectal cancer (CRC) [[125]53]. FGF13 has been
found to play a role in cancer progression and treatment resistance. It
has been shown to be associated with tumor growth and metastasis in
pancreatic cancer [[126]54]. In human pluripotent stem cells,
disrupting TJP1 leads to the activation of bone morphogenic protein-4
(BMP4) signaling and loss of patterning phenotype [[127]55]. CTSG has
also been implicated in triple-negative breast cancer (TNBC), where it
is overexpressed and correlated with a poor prognosis. CTGF, a protein
that binds to CTSG, activates the FAK/Src/NF-κB p65 signaling axis,
resulting in the up-regulation of Glut3 and enhanced aerobic glycolysis
in TNBC cells [[128]56]. APOC2 has been identified as a potential
diagnostic biomarker for cancer detection and as an auxiliary
prognostic marker or marker for immunotherapy in certain tumor types
[[129]57]. SDC1 has been shown to play a tumor-suppressor role in CRCs
[[130]58]. In cervical cancer, SDC1 expression is associated with low
differentiation and increased lymph node metastases [[131]59]. High
SDC1 expression in cervical cancer is also correlated with a poor
prognosis [[132]60]. The AP2B1 gene has been studied in various types
of cancer. In lung cancer, the transcription factors AP2A and AP2B were
found to promote the expression of the USP22 gene, which is associated
with aggressive growth and therapy resistance [[133]61]. Patients with
higher AP1S1 expression have higher estrogen receptor gene expression,
increased risk of distant metastasis and lymph node metastasis, and
worse overall survival rates [[134]62]. DNM2 appears to be a promising
molecular target for the development of anti-invasive agents and has
shown potential in reducing cell proliferation and inducing apoptosis
in cancer cells [[135]63].
EGFL7 is a gene that has been found to play a role in cancer. It has
been identified as a driver gene for resistance to EGFR kinase
inhibition in lung cancer cells [[136]64]. In addition, RASIP1 is
negatively regulated by fork-head box O3 (FOXO3), which suppresses
DLBCL cell proliferation. FOXO3 binds to the promoter sequence of
RASIP1 and inhibits its transcription [[137]65]. In lung cancer, the
inhibition of SOX18 with a specific inhibitor called Sm4 has shown
cytotoxic effects on non-small cell lung cancer (NSCLC) cell lines,
leading to cell cycle disruption and up-regulation of p21, a key
regulator of cell cycle progression [[138]66]. In various cancers,
including ovarian, testicular, endocervical, liver, and lung
adenocarcinoma, CLDN7 is highly expressed and activates multiple
signaling pathways involved in tumor growth, migration, invasion, and
chemo-resistance [[139]67]. In OSCC, CDH3 is up-regulated and
associated with a poor prognosis, promoting migration, invasion, and
chemo-resistance in oral squamous cell carcinoma [[140]68]. CLDN5
expression levels differ significantly between cancer and normal
tissues, and it has been confirmed in multiple studies [[141]69]. CLDN5
is implicated in the oncogenesis of diverse cancer types, highlighting
its potential significance in cancer biology [[142]70]. PCDH1 enhances
p65 nuclear localization by interacting with KPNB1, activating the
NF-κB signaling pathway and promoting PDAC progression [[143]71]. PCDH1
can be used as a negative prognostic marker and a potential therapeutic
target for PDAC patients [[144]72]. In breast cancer, RNF213 is
differentially expressed in primary tumors and is correlated with
overall survival in patients with basal-like sub-type breast cancer
[[145]73]. Additionally, RNF213 knockdown disrupts angiogenesis and
sensitizes endothelial cells to inflammation, leading to altered
angiogenesis and potential links to Moyamoya disease. VIP expression
has been associated with the transcription factor ZEB1, which is known
to regulate EMT [[146]74]. In breast cancer, VIP receptor 2 (VIPR2) has
been shown to promote cell proliferation and tumor growth through the
cAMP/PKA/ERK signaling pathway [[147]75]. EPSTI1 interacts with
valosin-containing protein to activate nuclear factor
κ-light-chain-enhancer of activated B cells (NF-κB) and inhibit
apoptosis [[148]76]. EPSTI1 has also been implicated in immune
response, as it promotes the expression of viral response genes and is
associated with immune privilege and autoimmune diseases [[149]77]. NMU
is expressed at higher levels in tumor tissues compared to normal
tissues, and its expression has been associated with poor prognosis and
shorter overall survival in cancer patients [[150]78]. Studies have
shown that VIPR2 overexpression promotes cell proliferation in breast
cancer cell lines and exacerbates tumor growth in vivo [[151]75].
Furthermore, VIPR2 has been shown to form homodimers and oligomers,
which are involved in VIP-induced cancer cell migration [[152]79]. The
expression of GNG11 mRNA is down-regulated in ovarian cancer patients,
and its high expression is associated with poor prognosis [[153]80].
GNG11 may play a crucial role in the biological process of ovarian
cancer through the ECM-receptor interaction pathway [[154]81].
The study showing the role of proteinase-activated receptor 2 (PAR2) in
TGF-β1-dependent cell motility underlines the importance of PAR2 in the
tumor microenvironment. The proteinase-activated receptor 2 (PAR2) is
crucial for TGF-β1-dependent cell motility, establishing PAR2 as a key
player in PDAC invasion and metastasis [[155]82]. PAR2 is a G
protein-coupled receptor (GPCR) that senses extracellular proteases,
particularly serine proteinases, which are abundant in the PDAC
microenvironment. The interaction between PAR2 and TGF-β1 promotes
tumor cell movement, a process essential for invasion into surrounding
tissues and metastasis, which are hallmarks of PDAC. High expression of
TGF-β1 further contributes to immune evasion and the development of a
dense stromal environment, which creates barriers to effective therapy.
TGF-β1 is well-known for its dual role in cancer, functioning as a
tumor suppressor in early-stage cancers and as a promoter of invasion
and metastasis in later stages, particularly in PDAC. The cooperation
between cadherin-1 (E-cadherin) and cadherin-3 (P-cadherin), both
critical cell adhesion molecules, plays a pivotal role in determining
PDAC aggressiveness [[156]83]. While E-cadherin downregulation promotes
epithelial-mesenchymal transition (EMT), a key process in cancer
metastasis, P-cadherin overexpression in PDAC may exacerbate this
effect, leading to increased tumor invasiveness. This interplay
highlights the molecular dysregulation of adhesion mechanisms in PDAC,
facilitating the cancer’s spread to distant sites. The research showing
that 92% of PDAC cases had positive immunostaining for claudin-4 and
58% for claudin-1 [[157]84] suggests that claudins are critical markers
in PDAC pathology. Claudins are integral components of tight junctions,
which regulate paracellular permeability and maintain epithelial
barrier function. In PDAC, the overexpression of claudins, particularly
claudin-4, has been associated with tumor invasion and metastasis.
These finding positions claudin-4 as a potential biomarker for PDAC,
with therapeutic implications given its overexpression in a majority of
PDAC tumors. Claudin-4 may also influence the permeability of the tumor
microenvironment, facilitating cancer cell dissemination. The study by
T. Ito et al. [[158]85] focused on matrix metalloproteinase-1 (MMP-1),
which plays a crucial role in tissue remodeling and degradation of the
extracellular matrix (ECM). MMP-1 is frequently upregulated in cancer
and contributes to tumor invasion and metastasis by breaking down the
ECM barriers that normally confine cells. In the analysis of PDAC
tissues, MMP-1 was found to be significantly elevated, suggesting its
involvement in PDAC progression, particularly in enabling tumor cells
to invade surrounding tissues. This highlights MMP-1 as a potential
therapeutic target, where inhibiting its activity might reduce the
invasive capabilities of PDAC cells. Additionally, neuromedin U (NmU)
has been implicated in PDAC pathogenesis, providing novel insights into
its role in cancer biology [[159]86]. NmU, a neuropeptide known for its
role in immune response and regulation of smooth muscle contraction,
appears to promote tumor growth and metastasis in PDAC. The
upregulation of NmU in PDAC tissues correlates with more aggressive
tumor behavior, suggesting that NmU-targeted therapies could offer new
avenues for treating PDAC, particularly in cases where NmU expression
is high. Another significant finding in PDAC research is the role of
microRNA miR-7-5p, which acts as a tumor suppressor by targeting SOX18,
a transcription factor involved in angiogenesis and EMT [[160]87]. The
downregulation of miR-7-5p in PDAC leads to SOX18 overexpression,
promoting tumor progression through enhanced angiogenesis and increased
metastatic potential. Restoring miR-7-5p levels could inhibit tumor
growth and reduce metastasis, offering potential therapeutic strategies
focused on microRNA-based therapies in PDAC.
The neuroactive ligand-receptor interaction pathway primarily involves
neurotransmitters and their receptors, which traditionally mediate
signaling in the nervous system. However, increasing evidence indicates
that these molecules also have significant roles in tumor biology,
including PDAC. Several neurotransmitters, such as serotonin, dopamine,
and norepinephrine, are known to interact with their respective
receptors on pancreatic cancer cells, influencing signaling pathways
that promote tumor cell survival, proliferation, and metastasis. For
example: Serotonin receptors (5-HT receptors) have been found to be
overexpressed in PDAC, enhancing tumor growth. Serotonin signaling
activates downstream oncogenic pathways such as the PI3K/AKT and
RAS/ERK pathways, which promote cell proliferation and survival.
Inhibition of these receptors or signaling pathways has shown to reduce
PDAC cell proliferation, indicating a direct role of
neurotransmitter-receptor interactions in the tumor’s growth dynamics
[[161]88]. One of the hallmark features of PDAC is perineural invasion
(PNI), where cancer cells spread along nerve fibers. Neuroactive
ligand-receptor interactions are central to this process, as tumor
cells communicate with nerve cells to enhance their migration along
these neural pathways. Perineural invasion in PDAC correlates with
worse prognosis and increased metastasis, especially since neural
invasion facilitates access to distant organs [[162]89]. Cancer cells
exploit neurotransmitter signaling to promote their affinity for
nerves. This invasion is mediated by neurotrophic factors, such as
nerve growth factor (NGF) and glial cell-derived neurotrophic factor
(GDNF), which bind to their receptors (TrkA and RET, respectively) on
PDAC cells, promoting invasion toward and along nerves [[163]90].
Neuroactive ligand-receptor signaling also has implications for
cancer-related pain, a common and severe symptom in PDAC patients.
Pancreatic tumors modulate pain receptors (nociceptors) through the
release of certain neurotransmitters and inflammatory molecules,
exacerbating the experience of pain in patients. This is often due to
the activation of the transient receptor potential vanilloid 1 (TRPV1)
and other pain-related receptors expressed in pancreatic nerve fibers,
which become hyperactivated as a result of neuro-ligand interactions.
The persistent stimulation of these receptors amplifies pain signals to
the brain. Neurotransmitter signaling also contributes to immune
evasion. For instance, dopamine receptors on immune cells can suppress
the immune response by inhibiting the activity of cytotoxic T-cells and
promoting the recruitment of regulatory T-cells (Tregs), both of which
support tumor immune escape [[164]91].
The leukocyte transendothelial migration pathway refers to the movement
of leukocytes (immune cells) across the endothelial barrier from the
bloodstream into tissue, typically during immune surveillance or
inflammation. However, in PDAC, this pathway is co-opted by the tumor
to support its progression. PDAC is characterized by a highly
immunosuppressive tumor microenvironment (TME), in part due to altered
leukocyte migration. While immune cells, such as macrophages,
neutrophils, and T-cells, migrate into the tumor, they are often
reprogrammed by the tumor to support tumor growth rather than launch an
anti-tumor immune response. Tumor-associated macrophages (TAMs), which
migrate through this pathway, are frequently polarized into the M2
phenotype, which supports tumor growth, suppresses inflammation, and
promotes tissue remodeling. The M2 macrophages secrete
immunosuppressive cytokines like IL-10 and TGF-β, which inhibit
effective anti-tumor immune responses. Regulatory T-cells (Tregs),
which also migrate into the tumor via transendothelial migration,
inhibit the activity of cytotoxic T-cells and natural killer (NK)
cells, further reducing the immune system’s ability to attack the
tumor. This leads to immune evasion, a hallmark of PDAC, where the
tumor successfully suppresses immune surveillance mechanisms. The
interaction between leukocytes and endothelial cells during migration
is not only important for immune cell infiltration but also contributes
to tumor angiogenesis and metastasis. Leukocytes, particularly TAMs,
secrete factors like vascular endothelial growth factor (VEGF) and
matrix metalloproteinases (MMPs), which promote angiogenesis and
metastasis [[165]92]. Leukocytes migrating into the PDAC tumor
microenvironment contribute to a chronic inflammatory state that
paradoxically promotes tumor growth. While leukocyte infiltration
typically signifies an immune response, in PDAC, the tumor co-opts this
inflammatory response to sustain its own progression. The inflammatory
cytokines secreted by leukocytes, such as TNF-α, IL-6, and IL-1β,
promote cancer cell proliferation, survival, and resistance to
apoptosis. Chronic inflammation within the PDAC microenvironment is
associated with genetic instability, increased mutation rates, and the
activation of oncogenic pathways, such as the NF-κB and STAT3 pathways,
further accelerating tumor progression [[166]93].
Both the neuroactive ligand-receptor interaction and leukocyte
transendothelial migration pathways are intricately linked to the
aggressive phenotype of PDAC. Their roles extend beyond simple
signaling processes to become central to the immune suppression,
perineural invasion, metastasis, and tumor-promoting inflammation that
characterize PDAC(92). Neuroactive ligand-receptor interactions support
tumor growth by activating oncogenic signaling, promoting nerve
invasion, and contributing to cancer-related symptoms like pain, all of
which are directly linked to the poor prognosis seen in PDAC patients
(93). Leukocyte transendothelial migration, on the other hand, is key
in shaping the immunosuppressive microenvironment, promoting
angiogenesis, and aiding in metastasis, which are central to PDAC’s
resistance to treatment and its high metastatic potential [[167]92].
Furthermore, key regulator genes were enriched in G-protein
beta-sub-unit binding, receptor binding, serine-type endopeptidase
activity and growth factor activity as molecular function. While
important cellular process are extracellular region, plasma membrane,
and cell surface. as listed in “[168]Table 6”. Go classification of key
signature genes indicates the regulation of positive regulation of cell
proliferation, positive regulation of protein kinase B signaling,
calcium-independent cell-cell adhesion via plasma membrane
cell-adhesion molecules related biological process. Next, the pathway
crosstalk analysis explored the interactions among significantly
enriched pathways. Genes VIPR2, GAL, F2R, NMU, F2RL1, GNRH1, CTSG, VIP
were involved in neuroactive ligand-receptor interaction pathways. Key
regulator genes CLDN5, CDH3, SDC4, CLDN7, SDC1, CLDN2 were active in
Cell adhesion molecules and CLDN5, CLDN7, CLDN2 enriched in Leukocyte
trans-endothelial migration pathway.
Table 6. Enrichment analysis results for key signature genes.
Category Description Genes P-value Fold Enrichment
MF G-protein beta-subunit binding F2R, F2RL1, GNG11 0.000697 74.125
MF receptor binding EGFL7, F2R, NMU, F2RL1, PLAT 0.00433306 7.196601942
MF serine-type endopeptidase activity F10, MMP1, CTSG, PLAT 0.003455303
12.68449198
MF growth factor activity PDGFA, PTN, FGF13 0.040571326 9.076530612
CC extracellular region EGFL7, F10, MMP1, F2R, PDGFA, PLAT, PTN, GAL,
APOC2, NMU, GNRH1, CTSG, VIP, FGF13 0.00000756 4.084387709
CC plasma membrane VIPR2, F10, SDC4, F2R, AP2B1, PTN, CLDN2, GNG11,
DNM2, TJP1, CLDN5, CDH3, CLDN7, F2RL1, SDC1, CTSG, PCDH1, FGF13
0.000728 2.129213998
CC cell surface EGFL7, SDC4, F2R, PDGFA, SDC1, CTSG, PLAT 0.00038
6.868351064
BP positive regulation of cell proliferation TJP1, CLDN5, F2R, CLDN7,
PDGFA, PTN, VIP 0.00019 7.811582569
BP positive regulation of protein kinase B signaling F10, F2R, PDGFA,
F2RL1 0.003123059 13.15
BP calcium-independent cell-cell adhesion via plasma membrane
cell-adhesion molecules CLDN5, CLDN7, CLDN2 0.000506 86.88392857
KEGG Neuroactive ligand-receptor interaction VIPR2, GAL, F2R, NMU,
F2RL1, GNRH1, CTSG, VIP 0.0000453 7.486430518
KEGG CLDN5, CDH3, SDC4, CLDN7, SDC1, CLDN2 0.0000634 13.04202532
[169]Open in a new tab
Further, gene-drug interaction for regulator genes retrieved drugs
against 14 genes as listed in “[170]Table 3”. Using DGIdb, we conducted
a query to identify interactions between a predefined list of genes (33
key regulator genes) implicated in pancreatic ductal adenocarcinoma and
known drugs. The key regulator gene list was derived from this study,
and DGIdb was queried with approved drugs parameters, capturing
interaction types such as inhibition, activation, and binding. We
identified 91 drug-gene interactions involving 14 unique genes and 79
approved distinct drugs. Among these, genes MMP1, F10, F2R, GNRH1 had
the highest number of interactions, interacting with more than 20
different drugs. The interactions included various types such as
inhibition (50%), activation (30%), and binding (20%). The
identification of MMP1, F10, F2R, GNRH1 as a highly interactive gene
underscores its potential role in PDAC.
We identified 15 key transcription factors, including NF1(neurofibromin
1), NFKB1(nuclear factor of kappa light polypeptide gene enhancer in
B-cells 1), and STAT1(signal transducer and activator of transcription
3 (acute-phase response factor)), that regulate our genes of interest.
as listed in “[171]Table 4”. Some well-known transcription factors
among them are STAT1-signal transducer and activator of transcription
1, STAT3-signal transducer and activator of transcription 3
(acute-phase response factor), CREB1-cAMP responsive element binding
protein 1, RELA-v-rel reticuloendotheliosis viral oncogene homolog A
(avian), and NFKB1-nuclear factor of kappa light polypeptide gene
enhancer in B-cells 1. Notably, NFKB1 was found to regulate a network
of genes involved in positive regulation of protein kinase B signaling
process or Cell adhesion molecules pathway, highlighting its potential
role in PDAC.
MicroRNAs (miRNAs) are crucial post-transcriptional regulators that can
influence gene expression by targeting transcription factors (TFs). In
this study, we utilized the TARGET SCAN database to predict miRNAs that
regulate TFs involved in pancreatic ductal adenocarcinoma. Next, using
miRTarBase, we identified 7 key miRNAs predicted to target
transcription factors, including hsa-miR-1-3p, hsa-miR-335-5p,
hsa-miR-526b-3p, hsa-miR-106b-5p, hsa-miR-17-5p and hsa-miR-124-3p as
listed in “[172]Table 5”. Notably, hsa-miR-335-5p was predicted to
target 9 key genes with a high context score (FDR = 0.619, p-value =
0.043), suggesting a strong regulatory interaction. Additionally,
hsa-miR-1-3p was predicted to target RNF213, F2RL1, AP1S3, indicating
its potential role in regulating multiple pathways. The identification
of hsa-miR-17-5p as a regulator of F2RL1, F2R, TJP1 suggests a
significant role in positive regulation of cell proliferation. The high
context score (FDR = 0.641, p-value = 0.378) indicates a strong
interaction, which could be critical for modulating positive regulation
of cell proliferation.
In this study, we applied a systematic and extensible methodology,
which finds significant key regulators in PDAC RNA-seq data. This study
suggested that gene expression profiling can be used to differentiate
and identify patients from their healthy counterparts. However,
experimental validation of these results in large sample size would
confirm the reliableness of these key regulators and facilitate the
designing of an economical and susceptive molecular diagnostic. These
findings warrant further investigation to validate the interactions and
explore their clinical applications.
Limitation of the study
Nevertheless, further investigation is necessary to verify the
expression and function of the key regulators identified in PDAC, given
the constrained sample size.
Conclusion
We conducted a comprehensive analysis using one RNA-seq gene expression
profiles, comparing individuals with PDAC to healthy controls. Our aim
was to identify Differentially Expressed Genes (DEGs) and elucidate
their biological functions through pathway enrichment analysis.
Additionally, we explored the topological features of the gene
interaction network, uncovering significant key regulators associated
with PDAC. Moreover, we investigated approved drugs, transcription
factors and mi-RNAs targeting these key regulators. These genes are
anticipated to play a crucial role in PDAC progression. This study can
be considered as very useful in the context of personalized treatment
plan for PDAC patients. Personalized treatment in cancer represents a
paradigm shift towards individualized and precision-based approaches to
diagnosis, prognosis, and therapy selection, with the ultimate goal of
improving patient outcomes and quality of life.
Acknowledgments