Next Article in Journal
Therapeutic Potential of Hydrogen-Rich Water on Muscle Atrophy Caused by Immobilization in a Mouse Model
Previous Article in Journal
99mTc-Labeled FAPI SPECT Imaging in Idiopathic Pulmonary Fibrosis: Preliminary Results
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Pharmaceuticals
Volume 16
Issue 10
10.3390/ph16101435
pharmaceuticals-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Redrawing Urokinase Receptor (uPAR) Signaling with Cancer Driver Genes for Exploring Possible Anti-Cancer Targets and Drugs

by
Yu-Ching Chang
1,
Chung-Ze Wu
2,3,
Chao-Wen Cheng
4,5,
Jin-Shuen Chen
6,7,* and
Li-Chien Chang
1,8,*
1
Graduate Institute of Life Sciences, National Defense Medical Center, Taipei City 114201, Taiwan
2
Division of Endocrinology and Metabolism, Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei City 110301, Taiwan
3
Division of Endocrinology and Metabolism, Department of Internal Medicine, Shuang Ho Hospital, Taipei Medical University, New Taipei City 235041, Taiwan
4
Graduate Institute of Clinical Medicine, College of Medicine, Taipei Medical University, Taipei City 110301, Taiwan
5
Graduate Institute of Medical Sciences, National Defense Medical Center, Taipei City 114201, Taiwan
6
Department of Education and Research, Kaohsiung Veteran General Hospital, Kaohsiung City 813414, Taiwan
7
Division of Nephrology, Department of Internal Medicine, Tri-Service General Hospital, National Defense Medical Center, Taipei City 114202, Taiwan
8
School of Pharmacy, National Defense Medical Center, Taipei City 114201, Taiwan
*
Authors to whom correspondence should be addressed.
Pharmaceuticals 2023, 16(10), 1435; https://doi.org/10.3390/ph16101435
Submission received: 6 July 2023 / Revised: 1 October 2023 / Accepted: 3 October 2023 / Published: 9 October 2023
(This article belongs to the Section Pharmaceutical Technology)
Browse Figures
Versions Notes

Abstract

:
During tumorigenesis, urokinase (uPA) and uPA receptor (uPAR) play essential roles in mediating pathological progression in many cancers. To understand the crosstalk between the uPA/uPAR signaling and cancer, as well as to decipher their cellular pathways, we proposed to use cancer driver genes to map out the uPAR signaling. In the study, an integrated pharmaceutical bioinformatics approach that combined modulator identification, driver gene ontology networking, protein targets prediction and networking, pathway analysis and uPAR modulator screening platform construction was employed to uncover druggable targets in uPAR signaling for developing a novel anti-cancer modality. Through these works, we found that uPAR signaling interacted with 10 of 21 KEGG cancer pathways, indicating the important role of uPAR in mediating intracellular cancerous signaling. Furthermore, we verified that receptor tyrosine kinases (RTKs) and ribosomal S6 kinases (RSKs) could serve as signal hubs to relay uPAR-mediated cellular functions on cancer hallmarks such as angiogenesis, proliferation, migration and metastasis. Moreover, we established an in silico virtual screening platform and a uPAR–driver gene pair rule for identifying potential uPAR modulators to combat cancer. Altogether, our results not only elucidated the complex networking between uPAR modulation and cancer but also provided a paved way for developing new chemical entities and/or re-positioning clinically used drugs against cancer.

1. Introduction

The urokinase-type plasminogen activator (uPA) system embodies ligand uPA, receptor uPAR and inhibitors PAI-1 and PAI-2. It goes through its sole mediator, uPAR, a glycosyl phosphatidylinositol (GPI)-anchored membrane protein hinging three homologous cysteine-rich domains (DI, DII, DIII), to modulate many physiological/pathological processes [1]. The signaling of this system uses two different methods to transduce. In a proteolytical fashion, the untie of uPAR’s extracellular ligand-binding portion of DI and DII-DIII fragments into circulation by proteases (such as uPA) seems to serve as an autocrine signal in extracellular proteolysis and is delineated to tumor progression and inflammation [2,3]. In a non-proteolytical fashion, the uPA-bound uPAR complexing integrins to interact with receptor systems, such as receptor tyrosine kinases (RTKs) (e.g., EGFR, FGFR, PDGFR, IGFR), G protein-coupled receptors and formyl peptide receptor type 1 (FPR1) [4], is thought to behave as an intracellular signal involved in angiogenesis, cell movement and tumor progression [5,6]. In addition, uPAR is rarely expressed in normal tissues but overexpressed in some tumor tissues; these traits grant its role in diagnosis and prognosis of various malignancies [2,7]. Together, the tangling between the uPAR signaling and cancer is tight and intricate.
The foregoing relationships serve to elucidate the detail that the uPAR signaling may have a great impact on cancer biology and therapy. Some previous works targeting uPAR by monoclonal antibody, peptide-derived antagonists and small molecule inhibitors (reviewed in [8,9]), which antagonize the binding with uPA or vitronectin (Vn), showed promise at preclinical stages but failed to receive clinical approvals. On the other hand, targeting modulation of the uPAR non-proteolytic signaling to combat cancer has not been truly established due to the complexity of this intracellular signaling because the uPAR signaling remains unresolved in the post-genomic era. It is well known that several types of cancer hallmarks, such as proliferation, angiogenesis, invasion and metastasis, are affected by the uPAR signaling. Thus, how this system governs cancer cells to acquire aberrant phenotypic capabilities during malignant processes must be elucidated in order to break the deadlock in the development of uPAR-related anti-cancer agents.
Cancer driver genes are genes whose mutations can fuel cancer progression and lead to pathological changes in tumor microenvironment, angiogenesis and inflammation. Although alternative definitions and search methods are used to identify and catalogue cancer driver genes and mutations, all these genes are mutated in at least one cancer type. Therefore, cancer driver genes may represent the spot genes in any physiological signaling system that drive the oncogenicity. On the other hand, while continuous efforts based on sophisticated techniques have broadened our knowledge on this topic, many studies indicate that driver genes are often “hub” genes, playing a key role in managing complicate signal transduction. Since cancer driver genes may play the pivotal role in uPAR signaling, we propose to map out this signaling pathway with cancer driver genes for not only clarifying the crosstalk between uPAR signaling and cancer but also providing the intersection drivers in uPAR signaling as druggable hubs for developing small molecule modulators to treat cancer.
In this study, a pharmaceutical bioinformatics framework aimed to elucidate cancer driver-associated uPAR signaling via network construction, pathway analysis, machine learning and model validation was proposed. With the use of self-defined uPAR modulators to construct the uPAR signaling network via linking uPAR co-expressed cancer driver genes, the rationale for uPAR modulators as a possible strategy to reverse cancer hallmarks was verified. Understanding the cancerous uPAR signaling networks could lead to the development of a new class of anti-cancer agents.

2. Results

2.1. uPAR Modulators and Their Cancerous Signaling Networks

To identify small molecule drug candidates and uPAR-associated targets in human cancer signaling, we proposed that cancer driver genes (CDG) could be hubs to project the crosstalk between uPAR signaling and cancer. Therefore, the first step was to identify uPAR modulators from a comprehensive database, as described in Section 4, to serve as the connectors between uPAR and cancer signaling systems. Based on the data of CMap, which collects the expression profiles of 3848 genes perturbated by 2429 compounds in two ways (i.e., gene knockdown and gene overexpression), 254 uPAR modulators, including 111 stimulators and 143 suppressors, were identified by self-written data extraction scripts. Hypothetically, the interactions between these uPAR modulators and their transcriptional (mRNA) or translational (protein) targets would be comprised by the uPAR signaling networks. Because the natures of transcriptional and translational interactions are different, the construction of the uPAR networks with the protrusion of cancer driver genes were performed as follows:
In transcriptional networking, there were 3848 molecular targets identified, which were co-expressed with the 254 uPAR modulators using the same method of modulator identification (i.e., gene expression score > 0.9), and these targets and modulators were the components of the uPAR transcriptional crosstalk. Then, 299 widely accepted cancer driver genes [10] were introduced, their existence was mapped in the uPAR transcriptional signaling, and 158 of them were found in the network. By relating these associated cancer driver genes and uPAR modulators to one another, a cancerous uPAR compound–target network was built (Figure 1A). The constructed signaling system comprised 411 nodes and 5184 edges, with an average degree (the number of connections or edges the node has to other nodes) of 32.81 nodes per targets and 20.49 edges per compound, respectively. Network analysis results showed that 13 cancer driver targets (i.e., KRAS, ERBB3, EGFR, HRAS, MYC, SMAD4, AKT1, PIK3R1, JAK1, ATR, MECOM, RHOB and FGFR2) interacted with more than 90 compounds, forming the core of this cancerous uPAR signaling system, whose average node degree was 5.38. In this core signaling, oncogenes MECOM and RHOB and tumor suppressor gene ATR were linked laterally (Figure 1B). This finding suggested that the 10 cancer driver genes present in the core network might be the essential starters in the modulation of uPAR-mediated functions in cancers (Figure 1C).
On the other hand, because the translation from nucleic acid coding to protein coding is known for the existence of expression heterogeneity, the uPAR signaling based on compound–protein and protein–protein interactions could be an alternative route to modulate uPAR-mediated functions. In translational networking, the same 254 uPAR modulators were used to retrieve protein targets from PubChem and similarity ensemble approach (SEA) databases to relay 2186 experimental or putative human protein targets within the uPAR interaction network (Figure 2A). As seen in the core, 12 out of 86 cancer driver targets identified in this networking had the most abundant interactions with uPAR modulators (as ligands). Through analyzing node degree, it was shown that the top 10 cancer driver targets that interacted with uPAR modulators were EGFR, EPHA2, ERBB2, FLT3, RET, PDGFRA, KIT, ERBB4, MET and RPS6KA3 (Figure 2B). Indeed, the EGFR and ERBB families (ERBB2, ERBB3 and ERBB4) were also ranked in the top 10 cancer driver targets in the transcriptional networking, suggesting their roles in uPAR-mediated tumorigenesis processes were highly anticipated.

2.2. Essences of uPAR-Mediated Signaling (Table 1) in Cancer Pathways

Cancer is a complicated and heterogeneous diseases. The more the content of the uPAR signaling involved in tumorigeneses is clarified, the higher the chance to develop a novel and successful anti-cancer modality based on modulating this signaling. Therefore, a portion of the uPAR signal network in mediating cancer hallmarks was further extrapolated by counting the involvement of uPAR-related cancer driver genes in common cancer pathways. The network analysis was performed via inputting the cancer driver genes of each uPAR modulator, to map out their existence in the Kyoto Encyclopedia of Genes and Genomes (KEGG) [11] cancer pathways using the service provided by STRING, and the results showed that 10 out of 21 KEGG cancer pathways were comprised of uPAR-related cancer driver genes (203/254, 80%). Among common cancer pathways, these uPAR-mediated pathways, such as MAPK, JAK-STAT, PI3K-Akt, mTOR, focal adhesion, cell cycle, estrogen, VEGF, HIF-1 and apoptosis signaling pathways (Table 1), were known to govern cancer hallmarks, including angiogenesis, cell survival, cell proliferation, invasion and metastasis, suggesting that these cancer pathways not only interacted with uPAR signaling but also embedded druggable targets that could be modulated or affected by uPAR modulators. Moreover, JAK–STAT, PI3K-Akt and MAPK pathways were the main affectable pathways of uPAR modulators, while these three pathways all contained more than 6 of the top 10 uPAR modulator-associated cancer driver genes in their signal transduction routes.
Table 1. Network analysis results of the constructed translational uPAR signaling.
Table 1. Network analysis results of the constructed translational uPAR signaling.
Pathway IDKEGG Pathway DescriptionModulators
(254)		Counts of Top 10 Driver Genes
Repetition Rate
(Number/254)
hsa04010MAPK signaling pathway0.97 (246/254)7
hsa04630JAK-STAT signaling pathway0.86 (219/254)9
hsa04151PI3K-Akt signaling pathway0.96 (243/254)6
hsa04150mTOR signaling pathway0.89 (227/254)4
hsa04510Focal adhesion0.82 (208/254)4
hsa04066HIF-1 signaling pathway0.85 (216/254)3
hsa04210Apoptosis0.89 (226/254)4
hsa04370VEGF signaling pathway0.86 (219/254)4
hsa04110Cell cycle0.86 (219/254)2
hsa04915Estrogen signaling pathway0.85 (215/254)5
On the other hand, nine real uPAR modulators with proven modulating ability in cell assays [12] were employed to confirm the above analysis results. By using the data of these real modulators to reconstruct the uPAR signaling network for comparison, it was found that the reconstructed network based on nine modulators (Table 2) could interact with 8 out of 10 above-mentioned cancer pathways, including MAPK, PI3K-Akt, mTOR, focal adhesion, cell cycle, VEGF, HIF-1 and apoptosis signaling pathways (Figure 3), suggesting that our approach unbiasedly revealed the tangling essences between uPAR and cancer signaling systems. While some cancer pathways had been previously delineated to interact with the uPAR signaling system, it should be noted that this study was the first study that could address cancerous uPAR targets and pathways so comprehensively.

2.3. Virtual Screening of uPAR Modulators by Machine Learning

Based on network analysis results, it was reasoned that uPAR modulators may rep-resent a new class of anti-cancer modalities since they possess the ability to affect can-cerous signaling transduction. Therefore, a virtual screening platform aimed to identify more potential compounds with the ability to modulate the cancerous uPAR signaling for anti-cancer therapy was constructed. By using uPAR expression scores as dependent variables and cancer driver genes as features, a dataset containing 254 uPAR modulators and 185 uPAR-sham compounds with no uPAR activity was prepared from CMap and subjected to a predictive platform model. In construction, seven machine learning algorithms set at tenfold cross-validation were applied to evaluate the prediction performance of the models via indexes, including area under the ROC curve (AUC), accuracy, sensitivity and precision. The results showed that the performance indexes for each algorithm from high to low number of features—i.e., “total features” and “strictly associated features”—were all similar, suggesting the minimum features number for accurate modeling was around 30 and the prediction accuracy could achieve 0.8. (Table 3). Among these models, the neural network model possessed the highest prediction power. To further validate the feasibility of this virtual screening platform, the above-mentioned nine real uPAR modulators were again subjected to classification by these models. It was found that either the SVM model or the neural network model could accurately classify these real modulators with a right pharmacological property (Table 4). Taken together, the above results supported our hypothesis that cancer driver genes were important keystones in uPAR signaling to relay tumorigenesis processes. In summary, our study not only explained the mechanism of action of this cancerous signaling transduction but also provided druggable targets for implication in anti-cancer therapy.

2.4. Implication of uPAR Modulators in Anti-Cancer Therapy

In the study, the proposed mechanism of action of uPAR modulators was to affect uPAR-mediated cancer hallmarks, including cell proliferation, migration, angiogenesis and resisting cell death, via modulation of uPAR and its associated downstream cancer driver genes. Compared to conventional uPA/uPAR inhibitors, this type of drug action indeed favors drug discovery in the selection of appropriate drugs, since their effects can be stratified according to the upfront activities on some specific driver genes. For example, uPAR-mediated cancer hallmarks can be further divided into four signaling systems based on composed cancer driver genes, as well as preferred marker genes. (Table 5) As seen, a modulator possessing the activities on uPAR and MYC, an oncogene, could be prone to affecting cell proliferation and survival, whereas one possessing the activities on uPAR and SMAD4, a tumor suppressor gene, could affect angiogenesis and migration.
Furthermore, through surveying in-house and CMap data again, it was verified that stains, including atorvastatin, simvastatin, and lovastatin, ellipticine and pterostilbene, could be potential modulators for anti-cancer purpose, but haloperidol and phenazopyridine could be carcinogenic. (Table 6) Using the pair rule to differentiate modulators’ pharmacological activities in advance, these potential anti-cancer agents could therefore be more precisely applied to treat cancers in situ or in metastasis. This is a good fit with clinical settings where cancer patients are often diagnosed at different progression stages and need regimen options for treating various degrees of malignancy. Thus, the developing uPAR modulators are expected to be a flexible but targeting anti-cancer remedy when used with or without standard chemotherapies.

3. Discussion

In this study, the whole uPAR signaling pathway, via linking associated target genes/proteins retrieved from several large public databases (i.e., CMap, PubChem Bio-Assay, SEA, etc.), was successfully constructed. The transformation of this cancerous signaling to a uPAR-centered drug-target network focused on therapeutic application was demonstrated. Through assessing the constructed transcriptional and translational pathways, a predictive machine learning model, using cancer driver genes as its features for identifying so-called uPAR modulators from market drugs, was established. In addition, the mechanism by which the uPAR signaling system changes cancer hallmarks was elucidated. Nevertheless, it is the first study to show there are about 10 cancer pathways associated with uPAR signaling.
The primary function of uPAR is the localization of the proteolytic activity of uPA on the extracellular surface to degrade the extracellular matrix and promote cell migration [19]. Lately, a large body of evidence has implicated that uPAR has an intracellular role in the regulation of various biological processes that contribute to cancer hallmarks. According to comprehensive reviews [4,6], uPAR interacts with more than 42 proteins and behaves as a central mediator in the epithelial–mesenchymal transition (EMT) process relating cell proliferation, differentiation, migration and survival. Crosstalk between uPAR and fMLP receptors (FPRs) regulates cell migration in vitro and in vivo [3,20]. The interaction of CXCR4, which is strongly upregulated in various malignancies, with uPAR-associated proteins drives disseminating cells toward metastatic sites [21], and the complexing of uPA with uPAR regulates EMT to promote pancreatic cancer progression [22]. In addition, the silencing of uPAR depressed hypoxia-induced EMT in multiple cancer cells, but the overexpression of uPAR mimicked EMT under normoxia [23]. Moreover, a positive loop connecting the uPA/uPAR system with TGF-β, where TGF-β could increase the expression of uPA and uPAR, activates plasminogen to plasmin, then triggers latent TGF-β [24]. However, it should be noted that all these studies and reviews could not reveal the whole landscape of uPAR signaling in mediating many physiological and pathological processes. Therefore, in the study, we proposed this pharmaceutical bioinformatics approach to elucidate the detailed crosstalk between uPAR and cancer, to highlight novel anti-cancer targets and/or drugs, as this signaling system is essential in cancer biology [25,26].
In the transcriptional networking, the top 10 cancer driver genes identified to be kin to the uPAR signaling are AKT1, KRAS, HRAS, PIK3R1, JAK1, EGFR, MYC, ERBB3, FGFR2 and SMAD4. Three of them, EGFR, ERBB3 and FGFR2, belong to the class of receptor tyrosine kinases (RTKs), and RTKs are transmembrane proteins that control important cellular processes such as cell growth, survival, and differentiation. It is well documented that dysregulation of RTKs is linked to the pathogenesis of many diseases, notably cancer, making RTKs a therapeutic target for treating tumors [27]. EGFR and ERBB3 both belong to the ERBB family, which were reported by activating the PI3K pathway for transmission signal [28]. The use of monoclonal antibodies (mAbs) or small molecule tyrosine kinase inhibitors (TKIs) to inhibit ERBBs is the most common approach to providing anti-cancer effects. Cetuximab and Panitumumab are FDA-approved mAbs against EGFR, and Erlotinib and Gefitinib are EGFR-TKIs first-line therapies for patients with non-small cell lung cancer (NSCLC). In contrast to their therapeutic efficacy, however, these medicines suffer emerged resistance and rapid tumors regrowth [29]. Previous data indicated that uPAR induced the resistance to anti-cancer agents via the EGFR/p-AKT signaling pathway, as observed in gefitinib-resistant human lung adenocarcinoma cells [30] and vemurafenib-resistant melanoma cells [31].
PIK3R1 and AKT1 both belong to the PI3K pathway, which is the most frequently altered pathway in cancer. PIK3R1 (the code for p85α) was the most abundant isoform to be expressed broadly in normal tissues [32], functioning as a tumor suppressor. The deficiency of PIK3R1-induced hepatocellular carcinogenesis [33] but complete loss of PIK3R1 impaired mouse survival [34]. On the other hand, AKT1, also named threonine kinase 1, was phosphorylated by PI3K as a critical node in the PI3K pathway. AKT1 dysregulation leads to tumorigenesis processes, including cancer cell proliferation, growth and survival [35], tumor angiogenesis [36], recruitment of inflammatory cells required for the tumor microenvironment [37] and resistance to apoptosis [38]. It was previously demonstrated that silencing of uPAR reduced PI3K/AKT downstream signaling activation in rheumatoid arthritis [39] and astrocytoma [40].
KRAS and HRAS are in the RAS/MAPK pathway to relay signals from outside the cell to the nucleus to instruct cell growth, proliferation and differentiation [41]. RAS proteins are believed to be molecular switches and cycle between “on” and “off” conformations by binding to GTP and GDP, respectively [42]. Among the RAS protein family, KRAS is the most frequently mutated protein in human cancer, followed by NRAS and HRAS [42]. It is evident that overexpressed uPAR could be related to oncogenic features such as adhesion in RAS-mutated NSCLC and CRC cells [43].
Except for the above-mentioned seven cancer driver genes, MYC played a role in regulation of protein-coding and non-coding genes to affect cellular functions, which include cell proliferation, differentiation, survival and immune surveillance. In relation to uPAR signaling, MYC downregulated the uPA system to reduce cell motility and invasiveness in gastric cancer [44], while silencing uPAR was associated with decreased c-MYC expression in melanoma and colon cancer [45,46]. On the other hand, JAK1 could behave as an oncogene or tumor suppressor under certain conditions. It played a critical role in different stages of cancer progression [47] and was essential for transduction of IL6-class inflammatory cytokine signaling; it was also involved in mediation of the oncogenic activation of STAT3 in mammary cancer cells [48]. It is documented that the uPAR signaling complex recruited JAK/STAT signaling components to transmit activation signals via STAT1 phosphorylation [49]. Therefore, the potential utility of uPAR signaling in anti-cancer therapy simply cannot exclude the involvement of JAK/STAT signaling. There are some JAK1 inhibitors used in clinics, such as Ruxolitinib, a selective inhibitor of JAK1 and JAK2, for the treatment of myelofibrosis and polycythemia [50]. However, there are limited clinical trials for examining the efficacy of JAK inhibitors on solid tumors so far. Based on the tight tangling between uPAR signaling and cancer, JAK inhibitors might be potential anti-cancer repurposing drugs when used in combination with chemotherapy, radiotherapy, immunotherapy or other targeted agents (e.g., the EGFR inhibitor Erlotinib) [51].
Lastly, SMAD4 as a transcription factor was the central mediator of TGF-β signaling, and it broadly interacted with many classical pathways, such as MAPK, PI3K/AKT and WNT/β-catenin, forming a complex network responsible for a wide range of cellular processes, such as proliferation, differentiation, apoptosis and migration, as well as cancer initiation and development [52]. In cancer signaling, SMAD4 was a tumor suppressor whose deficiency commonly occurred in pancreatic, colorectal, cholangiocarcinoma and many other less common cancers and was associated with pathological stages [53]. Whether SMAD4 plays an essential role in transduction of uPAR signaling remained unsolved, but it was reported that SMAD4 could regulate the expression of PAI-1 [54].
Indeed, the top ten cancer drug genes could be utilized to derive a uPAR-centered cancer hallmarks map that clearly illustrated the main cellular functions mediated by uPAR signaling (Table 7 and Figure 4). Based on the numbers of connections, it could be identified that angiogenesis, resisting cell death, proliferative signaling, invasion and metastasis would be the target indications when developing uPAR modulators for anti-cancer therapies. This is also in good agreement with empirical experience that indicates an inhibitor blocking the uPA/uPAR signaling pathways may exhibit its function in regulating the EMT process. On the other hand, as indicated in Section 2, the paired expressions of uPAR and some specific cancer driver genes (i.e., MYC and SMAD4) can be utilized to further differentiate drug effects for precise medication to revert cancerous phenotypic change. This use facilitates the drug development process from screening of candidate drugs toward appropriate indications. For example, statins, experimentally proven to be uPAR modulators by us [12], were demonstrated to possess anti-proliferation and anti-metastasis activities at concentrations higher than 10 μM in several studies [12,13,55]. Also, statins were shown to exhibit biphasic angiogenic effects at lower concentrations of 0.1–10 μM [56]. Given MYC and SMAD4 are known to be modulated by statins [57,58], the phenotypic switch caused by high and low doses of the same uPAR modulator may become an interesting research topic to elucidate whether these cancer driver genes could also serve as surrogate markers in anti-cancer therapies.
In the analysis of translational networking, RTKs were again found to play important roles in the transduction of uPAR signaling to mediate cellular functions, suggesting our approach was useful in the elucidation of complicate cancer biology whether it is genomic data or not. On the other hand, RPS6KA3, identified as one of the top 10 cancer driver targets in the uPAR translational networking, belongs to a family called ribosomal S6 kinases (RSKs). It worked as a downstream substrate of extracellular signal-regulated kinases (ERK) to function in the RAS/MAPK pathway, regulating many physical and pathological processes [59,60]. In the presence of RPS6KA3 boosters [61], cell migration was significantly decreased while silencing uPAR. Nevertheless, this is the first time that the uPAR signaling system, via bridging two intracellular kinase systems, i.e., RTKs and RSKs, was shown to execute and exhibit its functions.
Moreover, the pathway enrichment analysis results suggested that approximately 80% of uPAR-mediated signaling crosstalked to the identified 10 cancer pathways, where the MAPK and PI3K-AKT pathways were evident to be the most augmented. The involvement of these two pathways in cancer was likely through the complexing of uPAR with integrins/RTKs, as discussed above. The MAPK pathway was tightly regulated by phosphatases to communicate bidirectionally with other pathways, including the PI3K/AKT/m-TOR pathway. Dysregulation of the MAPK and PI3K-AKT pathways could bring uncontrolled cell proliferation, and such a machinery was already labeled as a part of the cancer hallmarks [62]. Given the close relationships of uPAR with these identified cancer pathways, our study should help illustrate which druggable targets in uPAR signaling are useful for reversing cancer hallmarks. Consequently, the potential uPAR modulator proposed by this study could be a new class of anti-cancer agents for effectively treating cancers, since how these compounds change cancer hallmarks, such as via sustaining proliferative signaling, resisting cell death, inducing/accessing vasculature and activating invasion and metastasis, etc., is much clearer now.
Finally, a machine learning screening platform based on cancer driver genes for the identification of potential uPAR modulators was established. This model could be used for in silico drug screening, but also for mechanism of action elucidation based on the intracellular uPAR signaling system, which was composed of about 30 cancer driver genes that impelled its versatile functionalities. While market drugs could have more certain safety profiles than investigating compounds, our approach could offer the profound benefits as finding a candidate and deciphering its mechanism of action at the same time. Compared to conventional uPA/uPAR inhibitor development, this indeed provides a paved way to efficiently identify small molecule anti-cancer agents that are either new synthetic compounds or old market drugs.

4. Materials and Methods

In the study, an integrated pharmaceutical bioinformatics approach (Figure 5) that combined modulator identification, driver gene ontology networking, protein targets prediction and networking, pathway analysis and uPAR modulator screening platform construction was employed to address the crosstalk between uPAR signaling and cancer, as well as to explore the druggable targets in uPAR signaling for potential anti-cancer applications. Interactome data, including transcriptional (mRNA), translational (protein) and perturbation (modulator) data retrieved from multiple public databases (Table 8), were used to mine the uPAR networking and to decipher its essences in modulating cancer. Cancer driver genes used for this framework study were adopted mainly from authoritative literature sources [10]. Since the first layer data (uPAR modulator) were generated from a database covering merely 3848 targets, the mapping ratio of driver to non-driver genes was around 1:12.

4.1. Modulator Identification

To address the crosstalk between uPAR signaling and cancer, we first used the gene perturbation data from the Connectivity Map (CMap) [65] database to identify compounds, namely the uPAR modulator, as its perturbation score of uPAR gene exceeding 0.9. In total, 2429 compounds collected in CMap were subjected to modulator extraction using self-written computer scripts.

4.2. Target/Cancer Driver Gene Ontology Networking

Two procedures were used to map out the uPAR signaling with cancer driver genes, the identification of uPAR co-activated cancer driver genes from the CMap perturbation data and connection of all these identified targets, to assemble the uPAR–driver genes network. The identification of active (co-expressed) cancer driver genes from the perturbation data of uPAR modulators was the same as that used above, in the identification of uPAR modulators (i.e., perturbation score exceeding 0.9), whereas the network construction and analysis of uPAR signaling was performed by means of Cytoscape 3.8.2, which is a popular visualization interface for probing complicated biological interactions [66]. The type of networking is called the uPAR transcriptional network.

4.3. Therapeutic Target Prediction, Networking and Pathway Analysis

To interpret the possible functional interaction between uPAR signaling and cancer, an alternative uPAR-mediated target–pathway (T-P) network was also constructed in this study. This involved retrieving the protein targets that experimentally or putatively interacted with the uPAR modulators, from the PubChem BioAssay [63] and similarity ensemble approach (SEA) [64] databases, and filtering out non-cancer driver protein targets; the enriched T–P network of the translational uPAR-signaling was constructed by means of STRING-DB (11.0), with targets tagged Homo sapiens taxonomy and confidence scores greater than 0.7. For pathway analysis, the extraction of the representative annotations and cancer pathways of the top 10 clusters was performed by using CytoHubba [67]. This is called the uPAR translational network.

4.4. Data Mining and Predictive Model Construction

To further explore the usefulness of uPAR modulator in cancer therapy, an in silico screening platform based on the expressions of cancer driver genes as features and modulator activity as dependent variables was constructed. Except for the 254 uPAR modulators identified in the above process, the negative dataset, including 185 uPAR-sham modulators, was prepared from the CMap compounds pool with uPAR perturbation scores between −0.2 to 0.2. The model construction was performed by machine learning using seven algorithms implemented in Orange [68], which included logistic regression, support vector machine (SVM), random forest, naïve Bayes, neural network, gradient boosting and k-nearest neighbors (kNN). In this procedure, 70% of the compounds were set for training and 30% for testing using a relevance-based feature selection (CFS) algorithm and tenfold cross-validation for model construction. A confusion matrix was applied to calculate the sensitivity, precision and overall accuracy of each algorithm.

5. Conclusions

In the study, the cancer driver genes and/or targets that compose the uPAR signaling system were successfully identified. The crosstalk between uPAR signaling and cancer pathways could be illustrated through kinase signaling (i.e., RTKs and RSKs) and intracellular cancer signal pathways, such as MAPK and PI3K-AKT pathways, to communicate and to affect cellular phenotypes (i.e., cancer hallmarks) during tumorigenesis and progression. With elucidation of the complicate networking between uPAR modulation and cancer biology, our results nevertheless provide the strongest context for further developing uPAR modulators as a new class of anti-cancer medicines.

Author Contributions

All authors contributed significantly. Y.-C.C. gathered all results and wrote the manuscript. Y.-C.C. and C.-Z.W. conducted and discussed the analysis of the experiments. C.-W.C., J.-S.C. and L.-C.C. supervised the whole process of study and edited the manuscript. L.-C.C. and J.-S.C. developed the experimental concept and designed the process of the whole study. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Defense Medical Center (NDMC), Taiwan, under grant numbers MAB-110005, MAB-111037, and MAB-112027.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

All data were retrieved from publicly accessible databases, as indicated in Table 8.

Acknowledgments

All authors acknowledge the help of An-Rong Lee, School of Pharmacy, NDMC, in manuscript preparation.

Conflicts of Interest

All authors declare no conflict of interest.

References

  1. Blasi, F. Proteolysis, Cell Adhesion, Chemotaxis, and Invasiveness Are Regulated by the u-PA-u-PAR-PAI-1 System. Thromb. Haemost. 1999, 82, 298–304. [Google Scholar] [CrossRef] [PubMed]
  2. Rasch, M.G.; Lund, I.K.; Almasi, C.E.; Hoyer-Hansen, G. Intact and cleaved uPAR forms: Diagnostic and prognostic value in cancer. Front. Biosci. Landmark 2008, 13, 6752–6762. [Google Scholar] [CrossRef] [PubMed]
  3. Bifulco, K.; Votta, G.; Ingangi, V.; Di Carluccio, G.; Rea, D.; Losito, S.; Montuori, N.; Ragno, P.; Stoppelli, M.P.; Arra, C.; et al. Urokinase receptor promotes ovarian cancer cell dissemination through its 84–95 sequence. Oncotarget 2014, 5, 4154–4169. [Google Scholar] [CrossRef]
  4. Eden, G.; Archinti, M.; Furlan, F.; Murphy, R.; Degryse, B. The Urokinase Receptor Interactome. Curr. Pharm. Des. 2011, 17, 1874–1889. [Google Scholar] [CrossRef] [PubMed]
  5. Blasi, F.; Carmeliet, P. uPAR: A versatile signalling orchestrator. Nat. Rev. Mol. Cell Biol. 2002, 3, 932–943. [Google Scholar] [CrossRef] [PubMed]
  6. Smith, H.W.; Marshall, C.J. Regulation of cell signalling by uPAR. Nat. Rev. Mol. Cell Biol. 2010, 11, 23–36. [Google Scholar] [CrossRef] [PubMed]
  7. Leurer, C.; Rabbani, S.A. Plasminogen Activator System—Diagnostic, Prognostic and Therapeutic Implications in Breast Cancer; IntechOpen: London, UK, 2015. [Google Scholar] [CrossRef]
  8. Montuori, N.; Pesapane, A.; Rossi, F.W.; Giudice, V.; De Paulis, A.; Selleri, C.; Ragno, P. Urokinase type plasminogen activator receptor (uPAR) as a new therapeutic target in cancer. Transl. Med. UniSa 2016, 15, 15–21. [Google Scholar] [PubMed]
  9. Yuan, C.; Guo, Z.; Yu, S.; Jiang, L.; Huang, M. Development of inhibitors for uPAR: Blocking the interaction of uPAR with its partners. Drug Discov. Today 2021, 26, 1076–1085. [Google Scholar] [CrossRef]
  10. Bailey, M.H.; Tokheim, C.; Porta-Pardo, E.; Sengupta, S.; Bertrand, D.; Weerasinghe, A.; Colaprico, A.; Wendl, M.C.; Kim, J.; Reardon, B.; et al. Comprehensive Characterization of Cancer Driver Genes and Mutations. Cell 2018, 173, 371–385.e18. [Google Scholar] [CrossRef]
  11. Kanehisa, M.; Furumichi, M.; Tanabe, M.; Sato, Y.; Morishima, K. KEGG: New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. 2017, 45, D353–D361. [Google Scholar] [CrossRef]
  12. Wei, L.; Chen, J.-S.; Lin, H.-T.; Wu, C.-Z.; Zheng, C.-M.; Chang, Y.-C.; Chang, L.-C.; Lin, Y.-F. Atorvastatin from target screening attenuates endothelial cell tube formation and migration by regulating urokinase receptor-related signaling pathway and F/G actin. J. Chin. Med. Assoc. 2017, 80, 86–95. [Google Scholar] [CrossRef]
  13. Crescencio, M.E.; Rodriguez, E.; Paez, A.; Masso, F.A.; Montano, L.F.; Lopez-Marure, R. Statins inhibit the proliferation and induce cell death of human papilloma virus positive and negative cervical cancer cells. Int. J. Biomed. Sci. 2009, 5, 411–420. [Google Scholar] [PubMed]
  14. Miller, C.M.; O’Sullivan, E.C.; McCarthy, F.O. Novel 11-Substituted Ellipticines as Potent Anticancer Agents with Divergent Activity against Cancer Cells. Pharmaceuticals 2019, 12, 90. [Google Scholar] [CrossRef] [PubMed]
  15. Shin, H.J.; Han, J.M.; Choi, Y.S.; Jung, H.J. Pterostilbene Suppresses both Cancer Cells and Cancer Stem-Like Cells in Cervical Cancer with Superior Bioavailability to Resveratrol. Molecules 2020, 25, 228. [Google Scholar] [CrossRef] [PubMed]
  16. Obrador, E.; Salvador-Palmer, R.; Jihad-Jebbar, A.; Lopez-Blanch, R.; Dellinger, T.H.; Dellinger, R.W.; Estrela, J.M. Pterostilbene in Cancer Therapy. Antioxidants 2021, 10, 492. [Google Scholar] [CrossRef] [PubMed]
  17. Chiang, J.Y.; Lei, F.J.; Chang, H.J.; Wei, S.T.; Wang, C.C.; Huang, Y.C.; Wang, H.L.; Chuang, C.F.; Hu, S.Y.; Hsieh, C.H. Haloperidol Instigates Endometrial Carcinogenesis and Cancer Progression by the NF-kappaB/CSF-1 Signaling Cascade. Cancers 2022, 14, 3089. [Google Scholar] [CrossRef]
  18. National Toxicology Program. Phenazopyridine Hydrochloride: CAS No. 136-40-3. In 15th Report on Carcinogens; National Toxicology Program: Research Triangle Park, NC, USA, 2021. [Google Scholar]
  19. Binder, B.R.; Mihaly, J.; Prager, G.W. uPAR–uPA–PAI-1 interactions and signaling: A vascular biologist’s view. Thromb. Haemost. 2007, 97, 336–342. [Google Scholar]
  20. Resnati, M.; Pallavicini, I.; Wang, J.M.; Oppenheim, J.; Serhan, C.N.; Romano, M.; Blasi, F. The fibrinolytic receptor for urokinase activates the G protein-coupled chemotactic receptor FPRL1/LXA4R. Proc. Natl. Acad. Sci. USA 2002, 99, 1359–1364. [Google Scholar] [CrossRef]
  21. Montuori, N.; Bifulco, K.; Carriero, M.V.; La Penna, C.; Visconte, V.; Alfano, D.; Pesapane, A.; Rossi, F.W.; Salzano, S.; Rossi, G.; et al. The cross-talk between the urokinase receptor and fMLP receptors regulates the activity of the CXCR4 chemokine receptor. Cell. Mol. Life Sci. 2011, 68, 2453–2467. [Google Scholar] [CrossRef]
  22. Huang, C.; Xie, D.; Cui, J.; Li, Q.; Gao, Y.; Xie, K. FOXM1c promotes pancreatic cancer epithelial-to-mesenchymal transition and metastasis via upregulation of expression of the urokinase plasminogen activator system. Clin. Cancer Res. 2014, 20, 1477–1488. [Google Scholar] [CrossRef]
  23. Gupta, R.; Chetty, C.; Bhoopathi, P.; Lakka, S.; Mohanam, S.; Rao, J.S.; Dinh, D.H. Downregulation of uPA/uPAR inhibits intermittent hypoxia-induced epithelial-mesenchymal transition (EMT) in DAOY and D283 medulloblastoma cells. Int. J. Oncol. 2011, 38, 733–744. [Google Scholar] [CrossRef] [PubMed]
  24. Laurenzana, A.; Biagioni, A.; Bianchini, F.; Peppicelli, S.; Chillà, A.; Margheri, F.; Luciani, C.; Pimpinelli, N.; Del Rosso, M.; Calorini, L.; et al. Inhibition of uPAR-TGFβ crosstalk blocks MSC-dependent EMT in melanoma cells. J. Mol. Med. 2015, 93, 783–794. [Google Scholar] [CrossRef]
  25. Mahmood, N.; Mihalcioiu, C.; Rabbani, S.A. Multifaceted Role of the Urokinase-Type Plasminogen Activator (uPA) and Its Receptor (uPAR): Diagnostic, Prognostic, and Therapeutic Applications. Front. Oncol. 2018, 8, 24. [Google Scholar] [CrossRef] [PubMed]
  26. Chounta, A.; Ellinas, C.; Tzanetakou, V.; Pliarhopoulou, F.; Mplani, V.; Oikonomou, A.; Leventogiannis, K.; Giamarellos-Bourboulis, E.J. Serum soluble urokinase plasminogen activator receptor as a screening test for the early diagnosis of hepatocellular carcinoma. Liver Int. 2015, 35, 601–607. [Google Scholar] [CrossRef] [PubMed]
  27. Paul, M.K.; Mukhopadhyay, A.K. Tyrosine kinase—Role and significance in Cancer. Int. J. Med. Sci. 2004, 1, 101–115. [Google Scholar] [CrossRef] [PubMed]
  28. Arteaga, C.L.; Engelman, J.A. ERBB Receptors: From Oncogene Discovery to Basic Science to Mechanism-Based Cancer Therapeutics. Cancer Cell 2014, 25, 282–303. [Google Scholar] [CrossRef] [PubMed]
  29. Hirsh, V. Next-Generation Covalent Irreversible Kinase Inhibitors in NSCLC: Focus on Afatinib. BioDrugs 2015, 29, 167–183. [Google Scholar] [CrossRef]
  30. Zhou, J.; Kwak, K.J.; Wu, Z.; Yang, D.; Li, J.; Chang, M.; Song, Y.; Zeng, H.; Lee, L.J.; Hu, J.; et al. PLAUR Confers Resistance to Gefitinib Through EGFR/P-AKT/Survivin Signaling Pathway. Cell. Physiol. Biochem. 2018, 47, 1909–1924. [Google Scholar] [CrossRef]
  31. Laurenzana, A.; Margheri, F.; Biagioni, A.; Chillà, A.; Pimpinelli, N.; Ruzzolini, J.; Peppicelli, S.; Andreucci, E.; Calorini, L.; Serratì, S.; et al. EGFR/uPAR interaction as druggable target to overcome vemurafenib acquired resistance in melanoma cells. EBioMedicine 2019, 39, 194–206. [Google Scholar] [CrossRef]
  32. Cortés, I.; Sánchez-Ruíz, J.; Zuluaga, S.; Calvanese, V.; Marqués, M.; Hernández, C.; Rivera, T.; Kremer, L.; González-García, A.; Carrera Ana, C. p85β phosphoinositide 3-kinase subunit regulates tumor progression. Proc. Natl. Acad. Sci. USA 2012, 109, 11318–11323. [Google Scholar] [CrossRef]
  33. Taniguchi, C.M.; Winnay, J.; Kondo, T.; Bronson, R.T.; Guimaraes, A.R.; Alemán, J.O.; Luo, J.; Stephanopoulos, G.; Weissleder, R.; Cantley, L.C.; et al. The Phosphoinositide 3-Kinase Regulatory Subunit p85α Can Exert Tumor Suppressor Properties through Negative Regulation of Growth Factor Signaling. Cancer Res. 2010, 70, 5305–5315. [Google Scholar] [CrossRef] [PubMed]
  34. Fruman, D.A.; Snapper, S.B.; Yballe, C.M.; Davidson, L.; Yu, J.Y.; Alt, F.W.; Cantley, L.C. Impaired B Cell Development and Proliferation in Absence of Phosphoinositide 3-Kinase p85α. Science 1999, 283, 393–397. [Google Scholar] [CrossRef] [PubMed]
  35. Sabbineni, H.; Alwhaibi, A.; Goc, A.; Gao, F.; Pruitt, A.; Somanath, P.R. Genetic deletion and pharmacological inhibition of Akt1 isoform attenuates bladder cancer cell proliferation, motility and invasion. Eur. J. Pharmacol. 2015, 764, 208–214. [Google Scholar] [CrossRef] [PubMed]
  36. Somanath, P.R.; Razorenova, O.V.; Chen, J.; Byzova, T.V. Akt1 in endothelial cell and angiogenesis. Cell Cycle 2006, 5, 512–518. [Google Scholar] [CrossRef] [PubMed]
  37. Fruman, D.A.; Rommel, C. PI3K and cancer: Lessons, challenges and opportunities. Nat. Rev. Drug Discov. 2014, 13, 140–156. [Google Scholar] [CrossRef] [PubMed]
  38. Testa, J.R.; Tsichlis, P.N. AKT signaling in normal and malignant cells. Oncogene 2005, 24, 7391–7393. [Google Scholar] [CrossRef] [PubMed]
  39. Liu, Y.; Pan, Y.F.; Xue, Y.-Q.; Fang, L.-K.; Guo, X.-H.; Guo, X.; Liu, M.; Mo, B.-Y.; Yang, M.-R.; Liu, F.; et al. uPAR promotes tumor-like biologic behaviors of fibroblast-like synoviocytes through PI3K/Akt signaling pathway in patients with rheumatoid arthritis. Cell. Mol. Immunol. 2018, 15, 171–181. [Google Scholar] [CrossRef] [PubMed]
  40. Gondi, C.S.; Kandhukuri, N.; Dinh, D.H.; Gujrati, M.; Rao, J.S. Down-regulation of uPAR and uPA activates caspase-mediated apoptosis and inhibits the PI3K/AKT pathway. Int. J. Oncol. 2007, 31, 19–27. [Google Scholar] [CrossRef]
  41. Prior, I.A.; Lewis, P.D.; Mattos, C. A Comprehensive Survey of Ras Mutations in Cancer. Cancer Res. 2012, 72, 2457–2467. [Google Scholar] [CrossRef]
  42. Pylayeva-Gupta, Y.; Grabocka, E.; Bar-Sagi, D. RAS oncogenes: Weaving a tumorigenic web. Nat. Rev. Cancer 2011, 11, 761–774. [Google Scholar] [CrossRef]
  43. Mauro, C.D.; Pesapane, A.; Formisano, L.; Rosa, R.; D’Amato, V.; Ciciola, P.; Servetto, A.; Marciano, R.; Orsini, R.C.; Monteleone, F.; et al. Urokinase-type plasminogen activator receptor (uPAR) expression enhances invasion and metastasis in RAS mutated tumors. Sci. Rep. 2017, 7, 9388. [Google Scholar] [CrossRef] [PubMed]
  44. Alfano, D.; Votta, G.; Schulze, A.; Downward, J.; Caputi, M.; Stoppelli, M.P.; Iaccarino, I. Modulation of cellular migration and survival by c-Myc through the downregulation of urokinase (uPA) and uPA receptor. Mol. Cell. Biol. 2010, 30, 1838–1851. [Google Scholar] [CrossRef] [PubMed]
  45. Laurenzana, A.; Chillà, A.; Luciani, C.; Peppicelli, S.; Biagioni, A.; Bianchini, F.; Tenedini, E.; Torre, E.; Mocali, A.; Calorini, L.; et al. uPA/uPAR system activation drives a glycolytic phenotype in melanoma cells. Int. J. Cancer 2017, 141, 1190–1200. [Google Scholar] [CrossRef] [PubMed]
  46. Biagioni, A.; Chillà, A.; Del Rosso, M.; Fibbi, G.; Scavone, F.; Andreucci, E.; Peppicelli, S.; Bianchini, F.; Calorini, L.; Li Santi, A.; et al. CRISPR/Cas9 uPAR Gene Knockout Results in Tumor Growth Inhibition, EGFR Downregulation and Induction of Stemness Markers in Melanoma and Colon Carcinoma Cell Lines. Front. Oncol. 2021, 11, 663225. [Google Scholar] [CrossRef] [PubMed]
  47. Qureshy, Z.; Johnson, D.E.; Grandis, J.R. Targeting the JAK/STAT pathway in solid tumors. J. Cancer Metastasis Treat. 2020, 6, 27. [Google Scholar] [CrossRef] [PubMed]
  48. Wehde, B.L.; Rädler, P.D.; Shrestha, H.; Johnson, S.J.; Triplett, A.A.; Wagner, K.-U. Janus Kinase 1 Plays a Critical Role in Mammary Cancer Progression. Cell Rep. 2018, 25, 2192–2207.e5. [Google Scholar] [CrossRef] [PubMed]
  49. Koshelnick, Y.; Ehart, M.; Hufnagl, P.; Heinrich, P.C.; Binder, B.R. Urokinase Receptor Is Associated with the Components of the JAK1/STAT1 Signaling Pathway and Leads to Activation of This Pathway upon Receptor Clustering in the Human Kidney Epithelial Tumor Cell Line TCL-598. J. Biol. Chem. 1997, 272, 28563–28567. [Google Scholar] [CrossRef] [PubMed]
  50. Bose, P.; Verstovsek, S. JAK2 inhibitors for myeloproliferative neoplasms: What is next? Blood 2017, 130, 115–125. [Google Scholar] [CrossRef]
  51. Huynh, J.; Etemadi, N.; Hollande, F.; Ernst, M.; Buchert, M. The JAK/STAT3 axis: A comprehensive drug target for solid malignancies. Semin. Cancer Biol. 2017, 45, 13–22. [Google Scholar] [CrossRef]
  52. Connolly, E.C.; Freimuth, J.; Akhurst, R.J. Complexities of TGF-β targeted cancer therapy. Int. J. Biol. Sci. 2012, 8, 964–978. [Google Scholar] [CrossRef]
  53. Zhao, M.; Mishra, L.; Deng, C.-X. The role of TGF-β/SMAD4 signaling in cancer. Int. J. Biol. Sci. 2018, 14, 111–123. [Google Scholar] [CrossRef] [PubMed]
  54. Dennler, S.; Itoh, S.; Vivien, D.; ten Dijke, P.; Huet, S.; Gauthier, J.M. Direct binding of Smad3 and Smad4 to critical TGF beta-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. EMBO J. 1998, 17, 3091–3100. [Google Scholar] [CrossRef] [PubMed]
  55. Sanfelice, R.A.; da Silva, S.S.; Bosqui, L.R.; Miranda-Sapla, M.M.; Barbosa, B.F.; Silva, R.J.; Ferro, E.A.V.; Panagio, L.A.; Navarro, I.T.; Bordignon, J.; et al. Pravastatin and simvastatin inhibit the adhesion, replication and proliferation of Toxoplasma gondii (RH strain) in HeLa cells. Acta Trop. 2017, 167, 208–215. [Google Scholar] [CrossRef] [PubMed]
  56. Weis, M.; Heeschen, C.; Glassford, A.J.; Cooke, J.P. Statins have biphasic effects on angiogenesis. Circulation 2002, 105, 739–745. [Google Scholar] [CrossRef]
  57. Rao, P.S.; Rao, U.S. Statins decrease the expression of c-Myc protein in cancer cell lines. Mol. Cell Biochem. 2021, 476, 743–755. [Google Scholar] [CrossRef] [PubMed]
  58. Kodach, L.L.; Bleuming, S.A.; Peppelenbosch, M.P.; Hommes, D.W.; van den Brink, G.R.; Hardwick, J.C. The effect of statins in colorectal cancer is mediated through the bone morphogenetic protein pathway. Gastroenterology 2007, 133, 1272–1281. [Google Scholar] [CrossRef] [PubMed]
  59. Gawecka, J.E.; Young-Robbins, S.S.; Sulzmaier, F.J.; Caliva, M.J.; Heikkilä, M.M.; Matter, M.L.; Ramos, J.W. RSK2 protein suppresses integrin activation and fibronectin matrix assembly and promotes cell migration. J. Biol. Chem. 2012, 287, 43424–43437. [Google Scholar] [CrossRef]
  60. Houles, T.; Roux, P.P. Defining the role of the RSK isoforms in cancer. Semin. Cancer Biol. 2018, 48, 53–61. [Google Scholar] [CrossRef]
  61. Doehn, U.; Hauge, C.; Frank, S.R.; Jensen, C.J.; Duda, K.; Nielsen, J.V.; Cohen, M.S.; Johansen, J.V.; Winther, B.R.; Lund, L.R.; et al. RSK is a principal effector of the RAS-ERK pathway for eliciting a coordinate promotile/invasive gene program and phenotype in epithelial cells. Mol. Cell 2009, 35, 511–522. [Google Scholar] [CrossRef]
  62. Hanahan, D.; Weinberg, R.A. Hallmarks of Cancer: The Next Generation. Cell 2011, 144, 646–674. [Google Scholar] [CrossRef]
  63. Kim, S.; Chen, J.; Cheng, T.; Gindulyte, A.; He, J.; He, S.; Li, Q.; Shoemaker, B.A.; Thiessen, P.A.; Yu, B.; et al. PubChem 2019 update: Improved access to chemical data. Nucleic Acids Res. 2019, 47, D1102–D1109. [Google Scholar] [CrossRef] [PubMed]
  64. Keiser, M.J.; Roth, B.L.; Armbruster, B.N.; Ernsberger, P.; Irwin, J.J.; Shoichet, B.K. Relating protein pharmacology by ligand chemistry. Nat. Biotechnol. 2007, 25, 197–206. [Google Scholar] [CrossRef] [PubMed]
  65. Subramanian, A.; Narayan, R.; Corsello, S.M.; Peck, D.D.; Natoli, T.E.; Lu, X.; Gould, J.; Davis, J.F.; Tubelli, A.A.; Asiedu, J.K.; et al. A Next Generation Connectivity Map: L1000 Platform and the First 1,000,000 Profiles. Cell 2017, 171, 1437–1452.e17. [Google Scholar] [CrossRef] [PubMed]
  66. Shannon, P.; Markiel, A.; Ozier, O.; Baliga, N.S.; Wang, J.T.; Ramage, D.; Amin, N.; Schwikowski, B.; Ideker, T. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. 2003, 13, 2498–2504. [Google Scholar] [CrossRef] [PubMed]
  67. Chin, C.-H.; Chen, S.-H.; Wu, H.-H.; Ho, C.-W.; Ko, M.-T.; Lin, C.-Y. cytoHubba: Identifying hub objects and sub-networks from complex interactome. BMC Syst. Biol. 2014, 8, S11. [Google Scholar] [CrossRef] [PubMed]
  68. Demšar, J.; Curk, T.; Erjavec, A.; Gorup, C.; Hocevar, T.; Milutinovic, M.; Možina, M.; Polajnar, M.; Toplak, M.; Staric, A.; et al. Orange: Data Mining Toolbox in Python. J. Mach. Learn. Res. 2013, 14, 2349–2353. [Google Scholar]
Pharmaceuticals 16 01435 g001
Figure 1. Constructed uPAR modulators–targets networks and module analysis of their interactions. (Transcriptional signaling) (a) Compounds–targets network and its amplified view. Dark green nodes represent compounds; other color nodes represent interacted targets being cancer driver genes. Each edge represents the interaction between them. Node size was proportional to its interaction degree. (b) The core interactions of this network, which comprised 13 cancer driver genes, was constructed using STRING with PPI score > 0.7. (c) The interaction network of top ten uPAR-related cancer driver genes was prepared by Cytohubba (ver. 0.1) in Cytoscape. Node color denotes interaction degree (red for high degree, orange for intermediate degree, and yellow for low degree).
Figure 1. Constructed uPAR modulators–targets networks and module analysis of their interactions. (Transcriptional signaling) (a) Compounds–targets network and its amplified view. Dark green nodes represent compounds; other color nodes represent interacted targets being cancer driver genes. Each edge represents the interaction between them. Node size was proportional to its interaction degree. (b) The core interactions of this network, which comprised 13 cancer driver genes, was constructed using STRING with PPI score > 0.7. (c) The interaction network of top ten uPAR-related cancer driver genes was prepared by Cytohubba (ver. 0.1) in Cytoscape. Node color denotes interaction degree (red for high degree, orange for intermediate degree, and yellow for low degree).
Pharmaceuticals 16 01435 g001
Pharmaceuticals 16 01435 g002
Figure 2. Constructed uPAR modulators–targets network and module analysis of their interactions. (Translational signaling) Node color denotes gene type (pink for oncogene, light blue for tumor suppressor gene, and purple for bifunctional gene) (a) Compounds–targets network. Dark green nodes represent compounds; other color nodes represent interacted cancer driver genes. Each edge represents the interaction between them. Node size was proportional to its interaction degree. (b) The interaction network of top 10 uPAR-related cancer driver genes were prepared by Cytohubba (ver. 0.1) in Cytoscape.
Figure 2. Constructed uPAR modulators–targets network and module analysis of their interactions. (Translational signaling) Node color denotes gene type (pink for oncogene, light blue for tumor suppressor gene, and purple for bifunctional gene) (a) Compounds–targets network. Dark green nodes represent compounds; other color nodes represent interacted cancer driver genes. Each edge represents the interaction between them. Node size was proportional to its interaction degree. (b) The interaction network of top 10 uPAR-related cancer driver genes were prepared by Cytohubba (ver. 0.1) in Cytoscape.
Pharmaceuticals 16 01435 g002
Pharmaceuticals 16 01435 g003
Figure 3. Interactome between real uPAR modulators and their prospective targets. (a) Compounds–targets–pathways network. Dark green nodes represent compounds, red nodes mean pathways, and others represent interacted targets being cancer driver genes. (b) Main KEGG cancer pathways were affected by real uPAR modulators (from lab testing).
Figure 3. Interactome between real uPAR modulators and their prospective targets. (a) Compounds–targets–pathways network. Dark green nodes represent compounds, red nodes mean pathways, and others represent interacted targets being cancer driver genes. (b) Main KEGG cancer pathways were affected by real uPAR modulators (from lab testing).
Pharmaceuticals 16 01435 g003
Pharmaceuticals 16 01435 g004
Figure 4. Main cellular functions (cancer hallmarks) and cancer driver genes mediated by the uPAR signaling system. Blue rectangles and colored ovals denote mediated cancer hallmarks and driver genes, respectively.
Figure 4. Main cellular functions (cancer hallmarks) and cancer driver genes mediated by the uPAR signaling system. Blue rectangles and colored ovals denote mediated cancer hallmarks and driver genes, respectively.
Pharmaceuticals 16 01435 g004
Pharmaceuticals 16 01435 g005
Figure 5. Workflow of this integrated pharmaceutical bioinformatics study.
Figure 5. Workflow of this integrated pharmaceutical bioinformatics study.
Pharmaceuticals 16 01435 g005
Table 2. Real uPAR modulators and their affectable cancer driver genes.
Table 2. Real uPAR modulators and their affectable cancer driver genes.
Active ModulatorsCancer Driver Genes
Suppression
HomatropineAKT1, ARHGAP35, ATR, AXIN2, BTG2, CCND1, CDKN1B, CHEK2, DICER1, DNMT3A, EGFR, EPAS1, ERBB2, ERBB3, FAT1, FBXW7, FGFR2, HRAS, JAK1, KRAS, MECOM, MSH6, MYD88, PDS5B, PMS1, RAC1, RASA1, RET, RHOA, RHOB, RXRA, SF3B1, SMAD4, TCF7L2, ZBTB20, ZFP36L1
HydralazineARHGAP35, B2M, BTG2, CCND1, CDK4, CDKN2C, CHEK2, EGFR, ERBB3, FLT3, KRAS, MET, NIPBL, PIK3CA, PIK3R1, PLCB4, PLCG1, RAC1, RASA1, RET, RPS6KA3, SMAD4, SMARCB1, WT1, ZBTB20
SalbutamolARHGAP35, ATR, BRAF, CCND1, CTNNB1, EGFR, ERBB3, ERBB4, FGFR2, HRAS, IDH2, KRAS, MAP2K4, MECOM, MEN1, MTOR, MYC, PRKAR1A, RAC1, RHOA, SF3B1, SMAD4, SMARCB1, TRAF3
SpironolactoneARHGAP35, AXIN2, CDKN2C, EGFR, ERBB2, ERBB3, FOXQ1, GNA13, HRAS, NRAS, RHOB, SMAD4, TSC2, ZFP36L2
Stimulation
AspirinAPOB, ARHGAP35, ATXN3, AXIN2, B2M, CHEK2, EGFR, ERBB4, FLT3, FOXQ1, GNA13, GNAS, HRAS, IDH1, JAK1, MAP2K4, MECOM, MET, MYC, PIK3CA, PIK3R1, POLRMT, PPP2R1A, PSIP1, RASA1, RB1, RET, RHOB, RPL22, SMARCB1, TCF7L2, TSC1, TSC2, WT1
AtorvastatinARID5B, ATM, ATR, BRCA1, CDKN1A, CDKN2C, CTNND1, ERBB3, FLT3, IL6ST, IRF2, KRAS, MECOM, MEN1, MTOR, MYC, NRAS, PIK3R1, PIK3R1, PIM1, RHOB, RPS6KA3, RXRA, SF1, SMAD2, TSC2, WT1, ZBTB20, ZFP36L1, ZMYM2, ZNF133
MethotrexateAKT1, ARHGAP35, ATR, AXIN2, BRCA1, CHEK2, EGFR, ERBB3, KRAS, MAX, MECOM, MEN1, MTOR, MYC, PIK3CA, PIK3CG, RHOA, RXRA, TCF12, TCF7L2
QuinineATM, BRCA1, CDKN2C, CHEK2, ERBB3, ERBB4, FGFR2, GNAS, HRAS, IRF6, MAP2K4, MAX, MECOM, MET, MTOR, NRAS, PIK3CG, PIK3R1, PRKAR1A, RAC1, RB1, SMARCB1, TNFAIP3, TSC2, TXNIP, USP9X, VHL, ZFP36L2
SimvastatinCDK4, CDKN1A, CDKN2C, ERBB3, IL6ST, KRAS, MAP2K4, MYC, MYD88, RAC1, RHOB, RXRA, TXNIP, ZFP36L1, ZMYM2
Top 10 cancer driver genes identified in this study are highlighted in bold.
Table 3. Performance of various machine learning algorithms on model construction.
Table 3. Performance of various machine learning algorithms on model construction.
ClassifierPerformance
AUCAccuracySensitivityPrecision
Total Features (218)
kNN0.8500.7270.7270.717
SVM0.9210.7750.7750.770
Random Forest0.8530.7020.7020.692
Neural Network0.9200.7880.7880.784
Naive Bayes0.8680.6980.6980.708
Logistic Regression0.9120.7780.7780.776
Gradient Boosting0.8910.7300.7300.722
Strictly Associated Features (31)
kNN0.8720.7380.7380.731
SVM0.9170.7740.7740.770
Random Forest0.8770.7280.7280.720
Neural Network0.9130.7780.7780.779
Naive Bayes0.9050.7470.7470.748
Logistic Regression0.8960.7490.7490.749
Gradient Boosting0.9050.7550.7550.750
The number of features for prediction models. At each feature selected set, the highest prediction performances are highlighted in bold.
Table 4. Classification results of real modulators using virtual screening platform.
Table 4. Classification results of real modulators using virtual screening platform.
Total Features (218)Strictly Associated Features (31)
SVMNeural NetworkSVMNeural Network
uPAR activity9999
No activity0000
Table 5. Stratification of uPAR-mediated cancer hallmarks, driver genes and preferred markers.
Table 5. Stratification of uPAR-mediated cancer hallmarks, driver genes and preferred markers.
HallmarksDriver GenePreferred Marker
proliferative signalingAKT1, EGFR, ERBB3, FGFR2, HRAS, JAK1, KRAS, MYCMYC
resisting cell deathAKT1, EGFR, ERBB3, FGFR2, HRAS, JAK1, KRAS, MYCMYC
angiogenesisAKT1, EGFR, HRAS, KRAS, SMAD4SMAD4
invasion and metastasisAKT1, EGFR, ERBB3, HRAS, KRAS, MYC, PIK3R1, SMAD4PIK3R1, SMAD4
Table 6. Identified uPAR modulators and their possible implications.
Table 6. Identified uPAR modulators and their possible implications.
Predicted
Drug		CMap Expression Score *Drug EffectsReferred
Validation
uPARMYCSMAD4
Statins (atorvastatin)88.31(KD)99.15(KD) Anti-cancerIn-house data, [12,13]
Ellipticine91.97(KD)90.64(KD) Anti-cancer[14]
Pterostilbene95.19(KD)94.29(KD) Anti-cancer[15,16]
haloperidol94.07(OE) 98.59(KD)Carcinogenic[17]
phenazopyridine92.97(OE) 94.41(KD)Carcinogenic[18]
* Data retrieved from CMap; KD denotes knock down and OE denotes over express.
Table 7. Associations of top 10 uPAR-related cancer driver genes to cancer hallmarks.
Table 7. Associations of top 10 uPAR-related cancer driver genes to cancer hallmarks.
Driver GeneGene DescriptionHallmarks
AKT1AKT serine/threonine kinase 1proliferative signaling, evading growth suppressors, invasion and metastasis, angiogenesis, resisting cell death, deregulating cellular metabolism, genome instability and mutations
EGFRErbB (epidermal growth factor) receptor family,
epidermal growth factor receptor		proliferative signaling, avoiding immune destruction, invasion and metastasis, angiogenesis, resisting cell death, deregulating cellular metabolism
ERBB3ErbB (epidermal growth factor) receptor family,
v-erb-b2 erythroblastic leukemia viral oncogene homolog 3		proliferative signaling, invasion and metastasis, resisting cell death
FGFR2Type V RTKs: FGF (fibroblast growth factor) receptor family, fibroblast growth factor receptor 2proliferative signaling, resisting cell death
HRASRAS subfamily,
v-Ha-ras Harvey rat sarcoma viral oncogene homolog		proliferative signaling, tumor-promoting inflammation, invasion and metastasis, angiogenesis, genome instability and mutations, resisting cell death, avoiding immune destruction
JAK1Janus kinase (JakA) family,
Janus kinase 1		proliferative signaling, avoiding immune destruction, resisting cell death
KRASRAS subfamily,
v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog		proliferative signaling, enabling replicative immortality, tumor-promoting inflammation, invasion and metastasis, angiogenesis, resisting cell death, deregulating cellular metabolism
MYCBasic helix-loop-helix proteins,
v-myc myelocytomatosis viral oncogene homolog		proliferative signaling, angiogenesis, avoiding immune destruction, genome instability and mutations, deregulating cellular metabolism, resisting cell death, invasion and metastasis, enabling replicative immortality
PIK3R1Phosphatidylinositol kinases, phosphoinositide-3-kinase, regulatory subunit 1 (alpha)evading growth suppressors, enabling replicative immortality, invasion and metastasis
SMAD4SMADs, SMAD family member 4evading growth suppressors, tumor-promoting inflammation, invasion and metastasis, angiogenesis, genome instability and mutations
Table 8. Publicly accessible databases used in this pharmaceutical bioinformatics study.
Table 8. Publicly accessible databases used in this pharmaceutical bioinformatics study.
ResourceDescriptionWebsiteRef.
Translational (protein)
PubChemA web-based informatics environment for data from small molecules and their biological activities.https://pubchem.ncbi.nlm.nih.gov/ (accessed on 1 July 2023)[63]
Similarity ensemble approach (SEA)An open resource related to proteins based on the set-wise chemical similarity among their ligands.http://sea.bkslab.org/ (accessed on 1 July 2023)[64]
Transcriptional (gene)
Connectivity Map (CMap)A public catalog of gene expression data collected from human cells treated with chemical compounds and genetic reagentshttps://clue.io/cmap (accessed on 1 July 2023)[65]
KEGGA curated database collecting comprehensive data including genes, reactions, pathways, drugs and diseases, for studying functions and utilities of the biological systemshttp://www.kegg.jp/kegg/ (accessed on 1 July 2023)[11]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Chang, Y.-C.; Wu, C.-Z.; Cheng, C.-W.; Chen, J.-S.; Chang, L.-C. Redrawing Urokinase Receptor (uPAR) Signaling with Cancer Driver Genes for Exploring Possible Anti-Cancer Targets and Drugs. Pharmaceuticals 2023, 16, 1435. https://doi.org/10.3390/ph16101435

AMA Style

Chang Y-C, Wu C-Z, Cheng C-W, Chen J-S, Chang L-C. Redrawing Urokinase Receptor (uPAR) Signaling with Cancer Driver Genes for Exploring Possible Anti-Cancer Targets and Drugs. Pharmaceuticals. 2023; 16(10):1435. https://doi.org/10.3390/ph16101435

Chicago/Turabian Style

Chang, Yu-Ching, Chung-Ze Wu, Chao-Wen Cheng, Jin-Shuen Chen, and Li-Chien Chang. 2023. "Redrawing Urokinase Receptor (uPAR) Signaling with Cancer Driver Genes for Exploring Possible Anti-Cancer Targets and Drugs" Pharmaceuticals 16, no. 10: 1435. https://doi.org/10.3390/ph16101435

APA Style

Chang, Y. -C., Wu, C. -Z., Cheng, C. -W., Chen, J. -S., & Chang, L. -C. (2023). Redrawing Urokinase Receptor (uPAR) Signaling with Cancer Driver Genes for Exploring Possible Anti-Cancer Targets and Drugs. Pharmaceuticals, 16(10), 1435. https://doi.org/10.3390/ph16101435

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop