Current trends and future prospects of drug repositioning in gastrointestinal oncology

Gastrointestinal (GI) cancers comprise a significant number of cancer cases worldwide and contribute to a high percentage of cancer-related deaths. To improve survival rates of GI cancer patients, it is important to find and implement more effective therapeutic strategies with better prognoses and fewer side effects. The development of new drugs can be a lengthy and expensive process, often involving clinical trials that may fail in the early stages. One strategy to address these challenges is drug repurposing (DR). Drug repurposing is a developmental strategy that involves using existing drugs approved for other diseases and leveraging their safety and pharmacological data to explore their potential use in treating different diseases. In this paper, we outline the existing therapeutic strategies and challenges associated with GI cancers and explore DR as a promising alternative approach. We have presented an extensive review of different DR methodologies, research efforts and examples of repurposed drugs within various GI cancer types, such as colorectal, pancreatic and liver cancers. Our aim is to provide a comprehensive overview of employing the DR approach in GI cancers to inform future research endeavors and clinical trials in this field.


Introduction
Gastrointestinal (GI) cancers, a group of malignancies occurring within the digestive system including colorectal cancer (CRC), esophageal cancer (EC), gastric cancer (GC), liver cancer (LC), and pancreatic cancer (PC) comprised a substantial proportion of global cancer cases.They were responsible for approximately 27.7% of cancer cases (5.5 million cases out of 18.1 million worldwide) and 35.8% of global cancer-related deaths in 2020 (Sung et al., 2021).GI cancers have different global prevalence and mortality rates.CRC is the most frequently occurring type, ranking third in prevalence and the second in cancer-related  deaths.GC, LC and EC rank as the fifth, sixth and eighth most commonly diagnosed cancers, with mortality rates ranking fourth, third and sixth, respectively.PC which has a poor prognosis with low curability rates compared to the other GI cancers, ranks 12th in frequency and the seventh in cancer death (Ferlay et al., 2021).GI cancers exhibit diverse incidence patterns across regions, with PC and CRC being more prevalent in Europe and Northern America, and GC, LC, and EC in Asia.Interestingly, the countries with high human development indicators (HDIs) report a higher incidence of PC cases (Li et al., 2021a;Huang et al., 2021).In addition, the incidence of GI cancers is influenced by several factors such as lifestyle habits and dietary choices as well as age, with elderly males having a higher risk of incidence and mortality compared to females.This emphasizes the complex interplay of environmental and biological factors in the development of GI cancers (Sung et al., 2021).
The existing standard treatment options for GI cancers comprise surgery, chemotherapy, radiation therapy, targeted therapy, and immunotherapy (Nakayama et al., 2013;Anwanwan et al., 2020;Biller and Schrag, 2021;Joshi and Badgwell, 2021).Although combination regimens offer higher response rates and improved survival rates than single-agent therapy, it is crucial to consider the toxicity profile of these regimens closely (Chakraborty and Rahman, 2012;Miller et al., 2016;Jin and Mills, 2020).Reducing GI cancer mortality involves identifying and implementing more efficient therapy strategies that lead to superior prognosis and/or fewer side effects.
Although the discovery and development of novel drugs are crucial for halting and reversing disease effects, they demand substantial funding, broad experimentation, and subsequent investigations into efficacy, pharmacokinetics, and toxicity.Moreover, only a small percentage of these drugs, approximately 5%, undergo clinical trials which may be approved for clinical use after successful results from the three phases of clinical trials (Gupta et al., 2013).Thus, developing new drugs is a costly and time-consuming procedure that requires extensive resources and expertise.Drug repurposing (DR), also known as drug repositioning, is an alternative and promising strategy for cancer treatment.It involves exploring the potential therapeutic applications of already approved drugs or withdrawn/outdated agents in the clinic (Tables 1-4; Supplementary Table S2; Figure 1).DR offers a substantial asset in drug development, uses the prior knowledge of the pharmacodynamics, pharmacokinetics and toxicity of already approved drugs which have undergone extensive animal and human studies (Gupta et al., 2013;Bertolini et al., 2015).Accordingly, repurposed compounds can be authorized for use in cancer treatment and other therapeutic applications rapidly, relying on the established safety profiles and efficacy data.Moreover, the regulatory approval process for repurposed drugs is often faster and less expensive, making DR an attractive alternative strategy for drug development (Chong and Sullivan, 2007).In this review, we aim to provide a critical overview of the current therapies for GI cancers, as well as examples of repositioned components, and the most promising candidate for drugs repurposing in GI cancers.

Current gastrointestinal cancers therapies, challenges, and limitations
The treatment approaches for GI cancer patients are diverse and contingent upon factors such as the patient's performance status, medical comorbidities, cancer type, stage, potential side effects and overall health (Joshi and Badgwell, 2021;Catalano et al., 2022;Ducreux et al., 2023).Treatment strategies encompass adjuvant chemotherapy, adjuvant chemoradiotherapy, preoperative chemoradiotherapy, endoscopic/colonoscopy resection, surgical resection, and perioperative chemotherapy.A comprehensive therapy regimen involves a combination of suitable treatment options to effectively address GI cancers (Sugarbaker, 2005).
Targeted therapy serves as a viable treatment option within various therapeutic regimens for GI cancers.Notably, for GC, HER2-targeted therapy (trastuzumab) and anti-angiogenesis therapy (ramucirumab) are major targeted therapies employed (Joshi and Badgwell, 2021).Ramucirumab is also considered as a targeted therapy alternative for patients with EC who have not responded well to initial treatment approaches (Fuchs et al., 2014).Erlotinib, an epidermal growth factor receptor (EGFR) blocking agent, has obtained Food and Drug Administration (FDA) approval for advanced PC patients (Moore et al., 2007).In the treatment of CRC, the commonly employed anti-angiogenesis agent, bevacizumab, in combination with chemotherapy can prolong the survival rate of advanced CRC patients (Xiong et al., 2021).Unresectable or metastatic hepatocellular carcinoma (HCC) patients commonly undergo a combination of anti-angiogenesis targeted therapy and immunotherapy as the predominant treatment strategies for HCC, while systemic chemotherapy proves ineffective in such cases (Greten et al., 2008).Immunotherapy, aimed at restoring immune system function, represents another line of therapy for patients with advanced GI cancer.An alternative treatment for metastatic CRC, HER2-positive EC, and advanced GC patients who exhibit Programmed Cell Death Ligand 1 (PD-L1) or microsatellite instability-high (MSI-H) biomarkers and are resistant to chemotherapy is pembrolizumab, an anti-PD-1 antibody, which functions by targeting the PD-1 receptor on tumor cells and preventing their evasion from the immune system.In addition, dostarlimab is another alternative option for the treatment of PCs with MSI-H or deficient mismatch repair (dMMR).Nivolumab either alone or in combination with ipilimumab is employed to treat adults with MSI-H or dMMR who have metastatic and drug-resistant CRC or EC (Zhang X. et al., 2022;Teixeira Farinha et al., 2022).It is worth noting that dostarlimab and nivolumab act upon the PD-1 receptor, whereas ipilimumab specifically targets CTLA-4.For additional information on this subject, refer to the Supplementary Material, Supplementary Table S1.
Despite the range of therapy options available for GI cancers, the high mortality rate highlights the limited efficiency of current treatments for these malignancies (Sung et al., 2021).Several challenges impede the improvement of existing therapeutic strategies, including chemotherapy resistance, tumor heterogeneity, late diagnosis, and limited efficiency of certain treatments, all contributing to treatment failure in GI cancer patients (Au et al., 2017;Parikh et al., 2019;Raziq et al., 2020).The development of drug resistance involves a complex multi-step process influenced by a variety of contributing factors.Numerous studies have shown that intensified DNA repair, apoptosis or autophagy disorders, epithelial-mesenchymal transition, inactivation of drug-metabolizing enzymes, and changes in expression or activity of membrane transporters are potential factors that promote chemotherapy resistance (Zheng, 2017).Furthermore, In the context of GI cancers, the unique characteristics of certain gastrointestinal tumors pose significant challenges to their effective treatment.For instance, the pancreas is anatomically situated in a hard-to-reach location, leading to difficulties in early diagnosis and the absence of effective screening techniques presents a significant challenge in detecting PC during its initial stages (McGuigan et al., 2018).Despite endeavors to implement personalized treatments, PC has not exhibited convincing outcomes comparable to those seen in other cancer types (Chantrill et al., 2015).The efficiency of radiotherapy in EC is limited due to the prevention of TAZ (Transcriptional Activator with PDZ-Binding Motif) ubiquitination and degradation (He et al., 2021).Furthermore, epigenetics and noncoding RNAs play a critical role in EC-related multidrug resistance and affect the effectiveness of therapies.In fact, EC's susceptibility to recurrence, metastasis, and drug resistance development after firstline treatment, highlights the urgent requirement for optimizing the medicine regimen (Liu et al., 2021;Wei et al., 2021).
3 Drug repurposing DR, is a strategic approach aimed at investigating alternative therapeutic applications of existing approved medications, beyond their originally intended uses.This method offers significant advantages in improving treatment outcomes, primarily by circumventing several essential stages of drug development.This results in reduced expenditure, shorter clinical trial durations, and mitigated risks associated with clinical trial failures due to adverse reactions (Ashburn and Thor, 2004;Chong and Sullivan, 2007;Zamami et al., 2017).
Repurposing FDA-approved medications is an efficient and inexpensive method to address oncology needs, including cancer treatment and reducing problems from existing anticancer treatments or radiation therapy (Ciociola et al., 2014).The drugrepurposing strategy has various advantages, including a shorter development period (usually 3-5 years), reduced costs (under $10 million), and higher success rates than original drug research (Morgan et al., 2011).Furthermore, DR can offer numerous benefits in overcoming therapeutic challenges by targeting various components both inside and outside of cancer cells (Würth et al., 2016).The concomitant use of multiple drugs that can target different tumor subtypes simultaneously can be a remarkable plan to overcome tumor heterogeneity (Li and Jones, 2012).Moreover, DR reveals the anticancer potential of non-oncology drugs with fewer side effects than traditional chemotherapy, making it a valuable option for cancer treatment (Crawford, 2014;Ferioli et al., 2018).Repositioned drugs can be used in combination with regular chemotherapy, and some demonstrate selectivity in causing cytotoxicity to cancer cells while sparing noncancerous cells (Foglietta et al., 2021).
Drug repositioning involves a multi-step process.Initially, drug selection can start with in silico methods like molecular docking, pathway matching, and genome-wide association studies to create a ranked list of compounds.The next step is secondary analysis, which includes experimental techniques to refine and prioritize these compounds.Tertiary analysis aims to validate these compounds using cell cultures and animal models.Finally, the chosen drugs are advanced to clinical trials, and successful ones are repurposed for new uses (Shameer et al., 2015;Kulkarni et al., 2023).The key stages of DR are elucidated in (Figure 2).

Approaches used for drug repurposing in cancer therapy
To evaluate the repurposing of an existing drug as a potentially effective anti-cancer agent, a mechanistic assessment of the drug's action in preclinical models is crucial and requires systematic approaches (El-Hachem et al., 2017;Nagaraj et al., 2018).These approaches can be classified into two categories, computational and experimental, which use available data and biochemical experiments, respectively, to explore the potential of repurposing existing drugs as novel cancer treatments (Luo et al., 2017;Fernández-Torras et al., 2019).Indeed, successful DR is contingent upon the integrated and synergistic use of both approaches (Mottini et al., 2021;Palve et al., 2021).

Computational approaches
Data-driven computational approaches involve the systematic analysis of data from different sources such as gene expression, chemical structure, genotype or proteomic data, or electronic health records (EHRs), which lead to the establishment of hypotheses on DR (Hurle et al., 2013).Carla Mottini et al. (2021) thoroughly reviewed computer-aided DR strategies for cancer therapy, offering examples of this approach in cancer studies within their article (Mottini et al., 2021).Future research should improve advanced computational tools, like machine learning and artificial intelligence, to improve therapeutic efficacy and safety prediction (Prasad and Kumar, 2021).Furthermore, computational methods can be used to address current anticancer medications or radiation therapy issues in addition to cancer treatment (Dalwadi et al., 2023).The most commonly used computational approaches are discussed below.

Molecular docking
Molecular docking, a structure-based computational strategy, predicts complementarity of binding sites between the ligand (drug) and the receptor (target) (Hurle et al., 2013;Knapp, 2018;Mottini et al., 2021;Palve et al., 2021).If prior knowledge is available about a receptor target involved in cancer, multiple drugs could be investigated against that specific target, or drug libraries can be screened for a collection of target receptors to identify new interactions suitable for repurposing (Honarparvar et al., 2014;Nero et al., 2014).Dakshanamurthy and others performed molecular fit computations on a list of 3,671 FDAapproved drugs against 2,335 crystal structures of human proteins.They experimentally validated that mebendazole, an anti-parasitic agent, shows significant potential to bind to the vascular endothelial growth factor receptor 2 (VEGFR2) and effectively block angiogenesis (Dakshanamurthy et al., 2012).Furthermore, it has been predicted that levosimendan, a heart failure drug, could serve as a potential inhibitor of several kinases including RIO Kinase 1 (RIOK1), through ligand-binding site comparison and protein-ligand docking.Levosimendan shows anti-cancer activity against various cancers by directly inhibiting RIOK1 and RNA processing enzymes (Lim et al., 2019).Similarly, Arulanandam CD et al. ( 2021) suggested itraconazole as a better inhibitor of platelet-derived growth factor receptor alpha (PDGFRA) compared to other antifungal drugs for treating gastrointestinal stromal tumors (GISTs) (Arulanandam et al., 2021).

Genome-wide association studies
Genome-Wide Association Studies (GWAS) aim to identify variants associated with common genetic disorders, thereby providing new insights into the biology of the disease.The novel associations between genes and cancer through GWAS and Phenome-Wide Association Studies (PheWAS) enable the identification of new targets for existing drugs leading to the repositioning of drugs (Zhang et al., 2015;Sud et al., 2017;Kim et al., 2018).Grover et al. (2015) employed a bioinformatics strategy to match gene targets for coronary artery disease (CAD) with drug information collected from three different drug-target databases (DrugBank, Therapeutic Target Database, and PharmGKB) to identify potential therapeutics for repositioning toward treatment of CAD (Grover et al., 2015).Another valuable resource for medication indications, called Medication Indication Resource (MEDI), can validate the plausibility of inferring novel drug indications with clinical potential (Wei et al., 2013;Bejan et al., 2015).Furthermore, a genome-wide positioning systems network algorithm uncovered the antitumor activity of ouabain, an approved drug for cardiac arrhythmia and heart failure, in lung adenocarcinoma cells (Cheng et al., 2019).

FIGURE 2
The key stages of drug repurposing.The process of DR begins with the use of computational or experimental techniques to select candidate drugs.Subsequently, these selected drugs are subjected to additional validation through potential evidence and additional experimental methods.Compounds that successfully pass through this rigorous evaluation then advance to clinical trials to obtain the FDA approval.Upon approval, these drugs are introduced to the market and their labeling is updated to reflect their new applications.
GWAS results may pose challenges when applied to DR due to several reasons.Firstly, GWAS signals in gene-rich loci, where linkage disequilibrium commonly occur, can complicate the identification of the associated genes and their specific variants.Secondly, the direction of the gene variant's effect may not be immediately apparent, necessitating functional studies to determine whether they act as activators or suppressors for disease control (Sanseau et al., 2012).Moreover, GWAS results may not provide detailed pathogenic information on genetic diseases (Wang and Zhang, 2013).It is important to acknowledge that understanding of the human genome is continually evolving, and new genes may continue to be discovered (Willyard, 2018).Therefore, careful consideration and further research are necessary when leveraging GWAS data for DR endeavors.

Signature or pathway matching
The signature matching method represents an innovative and promising approach within the realm of DR, as it involves comparing the distinct characteristics of a drug with those of other drugs, disease or clinical phenotype (Hieronymus et al., 2006;Keiser et al., 2009).The drug signatures can be derived from three different sources of data including omics data, chemical structures and adverse event profiles (Pushpakom et al., 2019).As a specific example, a study employed an optimal approach including two novel benchmarking standards, namely, area under the curve (AUC)-based standard and Kolmogorov-Smirnov (KS) statistic-based standard for signature-based DR and reported homoharringtonine (HHT) as a potential agent in the treatment and prevention of LC (Yang et al., 2022).Furthermore, various omics studies (transcriptomic, proteomic, or metabolomics) in the field of cancer not only provide large-scale data for supporting DR through applying advanced bioinformatics, but also expand our knowledge of hallmarks of cancer at the molecular level (Fernandez-Banet et al., 2016;Chen B. et al., 2020).Chemical structures serve as another type of signature matching used in DR; where the comparison of the chemical features of each drug with others helps identify possible chemical similarities that could signify shared biological activity (Pushpakom et al., 2019).A recent research employed this method along with other conventional DR methods to create a multilayer network algorithm.This algorithm was used to rank drugs that possibly can be repurposed for various types of cancer, including GI cancers and to explore new therapeutic possibilities for existing drugs (Cheng et al., 2021).
The unique adverse event profile of a drug can serve as a proxy for its related phenotypic effects, and drugs with similar side effects may act on the same target protein or pathway (Dudley et al., 2011;Cheng et al., 2021).In addition, if a drug's phenotypic response resembles that of a disease, it may indicate that the drug and disease share pathways and physiological mechanisms (Pushpakom et al., 2019).Pathway and network-based approaches have been broadly used to identify drugs or drug targets with the potential for repurposing (Smith et al., 2012).Moreover, despite some of the targets identified by GWAS or other analytical networks appearing potent for drug targets, many of these genes may not be ideal in practical drug targeting applications.In such circumstances, a pathway-based approach can provide insights into genes upstream or downstream of the GWAS-associated target and could be considered as potential repurposing opportunities (Greene and Voight, 2016).In a recent study, the network-wide association study (NetWAS) technique has been employed, combining GWAS-identified genetic variant information with tissue-specific functional networks to identify disease-relevant genes more accurately than GWAS alone.Using NetWAS on the concept of hypertension and incorporating drug-target data from the DrugBank, Greene et al. (2015) observed an expansion in the number of the top target genes for the anti-hypertensive drugs compared to the GWAS (Greene et al., 2015).Additionally, pathway analysis of gene expression data obtained from a wide range of studies of human viral respiratory diseases identified 67 signaling pathways which may play important roles in respiratory viral infections (Smith et al., 2012).In summary, DR candidates can be identified through construction of drug or disease networks using gene expression patterns, protein interactions, disease pathology or GWAS data along with the signature matching studies to complement the network analysis approach (Iorio et al., 2010(Iorio et al., , 2013)).

Real World Data
Real World Data (RWD) includes information from electronic health records (EHRs) of patients, characterized by large and complex datasets (Booth et al., 2019;Eichler et al., 2019).EHRs store patient and population health information in digital format, providing a wealth of data on patient outcomes (Miriovsky et al., 2012;Luhn et al., 2019).While diagnostic and pathophysiological data including experimental and drug prescribing data are structured, a significant portion of EHR data, such as clinical descriptions of patient symptoms and signs as well as imaging data remains unstructured (Khozin et al., 2017;Booth et al., 2019).Both structured and unstructured data from patients can serve as valuable sources to identify consistent signals for drug repurposing (Hurle et al., 2013;Low et al., 2017).
The retrospective clinical trial analysis is a commonly used computational approach based on RWD, with data commonly extracted from EHRs (Gray et al., 2019;Zheng et al., 2020).Sildenafil is an interesting example of retrospective clinical analysis which led to the repurposing of a candidate molecule (Ashburn and Thor, 2004).A classical example of repurposing a noncancer drug for cancer therapy through retrospective clinical analysis is metformin, which has shown to reduce cancer mortality in a dose-dependent manner (Sadeghi et al., 2012;Chaiteerakij et al., 2016).Furthermore, Xu et al. (2015) conducted a clinical cohort study using EHRs from Vanderbilt University Medical Center and Mayo Clinic, confirming the favorable effects of metformin in cancer survival including breast, colorectal, lung and prostate cancers in independent populations (Xu et al., 2015).Other successful examples of repurposing opportunities through retrospective clinical and/or pharmacological analyses include raloxifene in breast cancer, aspirin in CRC, propranolol in osteoporosis, and valproate in acute myeloid leukemia and glioblastoma (Cavalla and Singal, 2012;Happold et al., 2016;Lübbert et al., 2020).

Phenotypic screening
Phenotypic screening, a direct method of DR, involves identifying compounds based on their effects in model systems without prior knowledge of candidate drug targets (Moffat et al., 2014(Moffat et al., , 2017;;Kim et al., 2019).This approach typically uses a wide range of cell-based in vitro assays in a 96-or 384-well format (Sala et al., 2010).Iljin and others (2009) have performed highthroughput cell-based screening of a library including 4,910 drug-like small molecules against prostate cancer and nonmalignant prostate epithelial cell lines using proliferation as the primary phenotypic criteria.They found an anticancer effect of disulfiram, a medication used to treat alcohol abuse, which was subsequently validated through genome-wide-gene expression studies (Iljin et al., 2009).Moreover, whole animal screening assays can be employed for DR which offer insights into efficient anticancer drugs, as well as pharmacokinetic and organ-toxicity potential results (Yadav et al., 2016;Baxendale et al., 2017).Cousin et al. (2014) evaluated 39 FDA-approved medications using a zebrafish model for tobacco dependence treatment and identified compounds like apomorphine and topiramate that modulate the behavioral effects of nicotine and ethanol in this model (Cousin et al., 2014).Similarly, over 26,000 small molecules were evaluated for their efficacy against leukemia using a genetically engineered T-cell reporting zebrafish model and discovered the remarkable activity of lenaldekar, against various hematologic neoplasms (Clements and Traver, 2012;Ridges et al., 2012).

Intermolecular interactions
Proteomic techniques including affinity chromatography and mass spectrometry reveal protein-protein interactions based on the intermolecular force.These techniques have been used to identify binding partners for various drugs, thus facilitating DR due to an experimentally based pharmacological analysis (Brehmer et al., 2005;Shantikumar et al., 2015).In this approach, drug treatments are administered to cells or animals, followed by a deep proteome analysis to identify protein changes (Savitski et al., 2018).An early successful example of this technique involves the validation of more than 20 cellular targets for the epidermal growth factor receptor kinase inhibitor gefitinib using mass spectrometry (Brehmer et al., 2005).
The Cellular Thermo Stability Assay (CETSA) has been introduced as a method based on altered protein thermal stabilization/destabilization in response to ligand binding which can predict drug targets when combined with thermal proteome profiling (Martinez Molina et al., 2013;Li J. et al., 2020;Mateus et al., 2020;Perrin et al., 2020).Additionally, chemical genetic approaches such as kinase drug discovery also rely on intermolecular forces and can provide insights into the relationship between binding and drug efficacy (Sloane et al., 2010;Cong et al., 2012;Wong et al., 2016).These findings can be interpreted quickly into new clinical applications or to improve drug resistance outcomes which are near-inevitable phenotypic responses to protein kinase inhibitors for the cancer treatment (Carter et al., 2005;Bago et al., 2016).Karaman et al. (2008) used a competition binding assay in vitro to evaluate 38 protein kinase inhibitors against a panel of 317 pathologically significant human protein kinases.Their analysis identified 3,175 binding interactions, revealing that some drugs such as sorafenib and dasatinib showed higher affinity to other kinase targets than their known target (Karaman et al., 2008).The development of chemical-genetic approaches, particularly nonkinase targets of small molecules planned to inhibit kinases is becoming increasingly recognized (Munoz, 2017) resulting in new repurposing opportunities for cancer treatment, such as the use of anthelmintic drug niclosamide to treat Zika virus infection (Xu et al., 2016) and the potential to treat drug-resistant pathogens (Sun et al., 2016).

Experimental approaches to validate repurposed drugs
In order to understand the mechanism of action of repurposed drugs for cancer treatment, it is often essential to evaluate and validate their effects within a comprehensive system that considers safety, dosage and toxicity before advancing to clinical trials.This evaluation can be accomplished using various models, which are divided into two main categories: in vivo and in vitro models.These models should be reproducible, cost-effective, and quickly constructed (Würth et al., 2016).

In vitro studies
In in vitro tumor models, cancer cell lines are the predominant choice, followed by primary cells, although these models may also incorporate immune cells, stem cells, and stromal cells alongside cancer cells (Martinez-Pacheco and O'Driscoll, 2021).When using these cell types for detecting repositioning activity, in vitro assays offer several advantages including the ability to examine multiple substances with distinct mechanisms of action across a wide concentration-effect range, depending on the throughput of the experiment.Furthermore, these assays provide direct knowledge about potential new disease settings and allow testing of different drugs with novel mechanisms of action (Wilkinson and Pritchard, 2015).
In vitro screening approaches have been used for medication and DR in the various GI cancer types such as CRC, PC, GC and LC.A deep understanding of cancer progression and treatment has prompted the development of accurate in vitro tumor models that better represent the physiological features of the tumor microenvironment.Consequently, recent in vitro tumor models have become increasingly complex, extending beyond monitoring primary cell behaviors such as proliferation, invasion and cytotoxicity.These advanced in vitro models recapitulate important metastasis steps including angiogenesis, extracellular matrix remodeling and tumor cell metabolism reprogramming and dormancy (Wu and Swartz, 2014).

Two-dimensional and three-dimensional cell cultures
Cell culture plates and Transwell-based models are widely employed to assess viability, apoptosis, intra/extravasation and matrix remodeling of cancer cells.Two-dimensional (2D) models have been extensively used for drug screening and drug repurposing (Hulkower and Herber, 2011;Yu et al., 2020;Tomi-Andrino et al., 2022).In recent years, three-dimensional (3D) cell culture methods have become popular as they replicate in vivo microenvironmental providing data with greater predictive value for clinical outcomes.These authentic 3D cell culture models using human cells can overcome the limitations of mice models, which in addition to their high cost and ethical implications, are not always capable of accurately mimicking human illnesses or capturing medication side effects such as liver damage (Sivaraman et al., 2005).Moreover, by simulating the cell culture environment, 3D cell cultures can encourage specific cell activity, allowing drug discovery to target cell behavior with greater precision, such as enhancing cellular motility, promoting epithelial cell proliferation and differentiation, inducing cell dormancy, supporting of stem celllike characteristics or mimicking desired microenvironment, like metastatic niches (Valastyan and Weinberg, 2011).According to their importance and applications in cancer research or DR studies, 3D cell culture models can be categorized into two major groups: Spheroids and Organoids.

Spheroids
Spheroids are cell aggressions that grow in suspension or are embedded in a three-dimensional matrix using three-dimensional culture methods.Cancer cell spheroids represent avascular tumor nodules, also known as micro-metastases and are widely used for drug screening despite being more expensive and time-consuming than 2D cell cultures.Spheroids also recapitulate interaction between cells and matrix in the tumor microenvironment and their size-dependent structure includes a necrotic central nucleus, resembling tumors with poor angiogenesis.Moreover, tumor spheroids offer valuable insights into how tumors respond to candidate drugs and combination therapies which reduces the need for animal testing and providing a more realistic representation of the tumor microenvironment (El Harane et al., 2023).

Organoids
Organoids are novel ex vivo tumor models composed of selforganized three-dimensional multicellular tissue cultures derived from stem cells, primary tissue specimens or cancer cell lines.These models are capable of mimicking the in vivo organ (Simian and Bissell, 2017).Patient-derived cancer organoids provide a promising opportunity to predict drug efficiency and treatment response in GI cancers.Furthermore, incorporating 3D cell culture technology with primary patient-derived cancer cells, molecular characterization data, or the establishment of GI organoid banks representing molecular tumor subtypes could lead to preclinical assessment of personalized drug targets to improve treatment outcomes by reducing side effects in cancer therapy (Perrone and Zilbauer, 2021).
The first CRC organoid biobank was established in 2015 for drug screening, enabling the study of gene-drug interactions to recognize potential treatment response biomarkers and understand the molecular basis of drug response (van de Wetering et al., 2015).Notably, numerous studies have reported a significant overlap between esophageal adenocarcinoma organoids and tumor response to standard chemotherapy (Donohoe and Reynolds, 2017;Li et al., 2018) and a similar approach yielded comparable outcomes using GC organoids (Verduin et al., 2021).These results highlight the benefits and advantages of 3D ex vivo models in DR and repositioned drug efficiency assessments.

In vivo studies of drug repurposing
Animal models offer a powerful tool to simulate physiological conditions and complicated interactions, as well as the responses of reactions of different cell types and tissues to chemicals (Ruggeri et al., 2014).During the drug development process, in vivo testing plays a crucial role as it allows the study of interactions between the drug and both target and non-target cells.
Based on the premise of evolutionarily conserved pathogenetic mechanisms, animal models such as zebrafish, mice, fruit flies, worms and yeast have been used to give a comprehensive understanding of the biological mechanisms underlying the effect of drug administration.Xenograft models and genetically engineered mouse cancer models were among the in vivo models used in drug repurposing (Freires et al., 2017).

Xenograft tumors
The human tumor xenograft is a widely used in vivo model where human cancer cells are transferred into immunocompromised (nude) mice through either ectopic or orthotopic implantation.Although the cell line-derived xenograft (CDX) model is considered as the goldstandard model for cancer research and investigation of anti-tumor therapies, patient-derived xenograft (PDX) tumors can also be used for this purpose (Tentler et al., 2012).
Xenograft tumors are highly used as animal models in GI cancer and DR research (Onaciu et al., 2020).In situ GI cancer can be induced by locally injecting of cell lines or implantation of tumor cells, while the metastatic models are established by injection of the tumor cells through the tail vein or into the specific organ (Morton and Houghton, 2007;Tentler et al., 2012).For decades, athymic nude mice and mouse xenograft models using human tumor cell lines have been used to study tumor progression factors.In the DR approaches, the use of xenograft tumors can be very helpful in assessing the accuracy of in vitro study results and the effectiveness of drugs on cancer cells in safe doses under in vivo situations.Doxycycline, a tetracycline-class antibiotic commonly used to treat bacterial and parasitic infections, has demonstrated to reduce tumor growth by ~80% in pancreatic tumor xenografts (Son et al., 2009;Liu H. et al., 2020).

Genetically engineered mouse models
The Genetically Engineered Mouse Model (GEMM) is another animal model for studying human cancer and conducting preclinical study of repurposed drugs to target special gene-derived GI tumors.Genetic technologies have been recently applied by an increasing number of studies to introduce oncogenes into mouse embryonic or somatic cells through tissue-specific promoters targeting the GI tract and inducing GI cancers (Kersten et al., 2017).Transgenic, gene knock-in, and gene knock-out techniques are used to modify the genetic sequence of these mice in order to transfer, mutate, delete or overexpress one or more genes associated with transformation or malignancy.For instance, transgenic mice overexpressing KRAS mutant genes can mimic pancreatic tumorigenesis.Indeed, while general single-gene modified models may not fully represent the entire process of GI tumorigenesis, it has been discovered that physiological levels of KRAS G12D induce ductal lesions that serve as putative precursors to invasive PC (Westphalen and Olive, 2012).Additional genetic modifications, such as P53 mutation, can promote tumorigenesis and metastasis (Walrath et al., 2010).Another gene that can be engineered to produce GI cancer GEMM is carbonic anhydrase, present in the basolateral membranes of gastrointestinal epithelial cells and Its overexpression has been reported in many carcinomas including GC. Mice with null mutations of the Car9 gene develop gastric hyperplasia in glandular epithelium after 1 month (Gut et al., 2002).Additionally, ApcMin/ Pten−/− mice are developed to study CRC (Stastna et al., 2019).
GEMM animals can be used to study the impact of a specific gene on tumorigenesis, or to investigate whether a de novo/ repositioned drug acts through predicted pathways or genes to control the disease (Richmond and Su, 2008).Kobayashi et al. (2017) used mogp-TAg transgenic mice for DR to target the mevalonate pathway as a key tumorigenesis pathway in ovarian cancer (Kobayashi et al., 2017).Furthermore, the effect of metformin cancer initiation and progression suppression was studied using transgenic KPC mice (Chen K. et al., 2017).This animal model was also employed to evaluate the efficiency of repurposed histone deacetylase (HDAC) and mammalian target of rapamycin (mTOR) inhibitors on PC treatment (Biermann et al., 2022).Transgenic animals can be used to evaluate safe, first-in-human (FIH) doses during preclinical studies in drug development.
For example, human-CYP3A4-expressing transgenic (Cyp3aXAV) mouse serve as practical model to evaluate the safe dosage and efficiency of CYP3A4-metabolized small-molecule drugs (Damoiseaux et al., 2022).The application of GEMM animal models can prove advantageous in specialized DR studies and allows researchers to focus on specific molecular pathways.

Chemically induced gastrointestinal tumors
Chemical agents can be used to induce GI cancers in mice, potentially leading to mutations in relevant human cancer genes.The most well-established chemically induced GC model is produced by the administration of N-Methyl-N-nitrosourea (MNU) (Tomita et al., 2011;Ji et al., 2020;Rabben et al., 2021b).MNU is an N-nitroso compound mostly generated by anaerobic gut bacteria in the presence of nitrates and nitrites (Sobko et al., 2005;Zhuang et al., 2017).Furthermore, 1,2-dimethylhydrazine (DMH) and its metabolite, azoxymethane (AOM), are the most commonly used chemical compounds and carcinogens to induce CRC in mice.It has also shown that intraperitoneal injection of azaserine in rats induces metastatic pancreatic acinar cell carcinoma, although 10% of animals develop tumors in other organs (Rosenberg et al., 2009).Another method to produce a chemically induced PC model is through topical application of benzopyrene which induces adenocarcinoma (Kong et al., 2020).This group of animal models is widely employed in DR studies for GI cancers.Kannen et al. (2012) used C57BL/6 mice exposed to the carcinogen N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) as a colonic carcinogen mouse model to evaluate the antiproliferative effect of fluoxetine, an antidepressant medicine, on colon cancer (Kannen et al., 2012).In a similar study, the inhibitory effects of Liuwei Dihuang Pill (LDP) were investigated on MNU-induced gastric tumorigenesis in diabetic mice (Zhuang et al., 2017).
5 Repurposed drugs for gastrointestinal cancers treatment

Colorectal cancer
Researchers in the field of Gl cancer treatment are exploring innovative therapeutic strategies through drug repurposing, with an extensive focus on leveraging existing medications for the management of CRC.Among the studies of significance, Antona et al. (2022) have recently showed that spiperone, an approved drug for schizophrenia treatment, triggers apoptosis in CRC cells by activating phospholipase C, disrupting intracellular calcium balance, inducing irreversible endoplasmic reticulum stress, causing lipid metabolism changes, and damaging the Golgi apparatus (Antona et al., 2022).Furthermore, ten small molecules/drugs were identified via bioinformatic techniques for treating CRC, specifically targeting the upregulated Tissue Inhibitor of Matrix Metalloproteinases-1 (TIMP1) gene; these included established agents like formaldehyde and paclitaxel, as well as promising new drug candidates (Leng et al., 2022).Another study discovered that ebselen effectively inhibits Autophagy related protease 4B (ATG4B) through oxidative modification.This was based on the FDA-approved drug library, using Fluorescence Resonance Energy Transfer (FRET)-based highthroughput screening and gel-based analysis.The study showcased the potential of ebselen as an anti-CRC agent by influencing autophagy and tumor suppression (Xie et al., 2022).Mao et al. (2023) developed an efficient method for identifying repurposed drugs for CRC using organoid-based screening and computational drug prediction.Out of 335 tested drugs, 34 showed anti-CRC effects, with distinct transcriptome patterns including differentiation induction, growth inhibition, metabolism inhibition, immune response promotion, and cell cycle inhibition.Validation in patient-derived organoid-based xenograft (PDOX) systems demonstrated the anticancer effectiveness of drugs like fedratinib, trametinib, and bortezomib.(Mao et al., 2023).It has been revealed that the antifungal agent oxiconazole induces antitumor effects in CRC cells by inhibiting autophagy through downregulation of peroxiredoxin-2 (PRDX2), leading to growth suppression, and suggests its potential therapeutic use in combination with oxaliplatin for CRC treatment (Shi J. et al., 2022).Moreover, Dhakal et al. (2022) investigated the cytotoxic effects of the anti-anginal drug perhexiline and its enantiomers on CRC cells and demonstrated their ability to induce apoptosis and reduce cell viability in both monolayers and spheroids, as well as patient-derived organoids (Dhakal et al., 2022).Liñares-Blanco et al. (2020) proposed abemaciclib, an inhibitor of the CDK4/6 protein, as a promising option for the treatment of colon cancer (Liñares-Blanco et al., 2020).Furthermore, mebendazole, an antihelminthic medication used to treat gut worm infections, exhibited a cytotoxic effect on the RKO and HCT-116 colon cancer cell lines.Mebendazole was evaluated against a panel of kinases to determine the mechanism of its cytotoxic effect, which indicated significant inhibitory action against Abl and BRAF proteins.Additionally, in a case study of a patient with resistant metastatic colon cancer, twice-daily therapy with the normal antihelminthic dose of mebendazole led to a substantial reduction in metastasis (Nygren et al., 2013).A comprehensive investigation introduced several promising repurposing drugs (crizotinib, arsenic trioxide, vorinostat, dasatinib, estramustine, and tamibarotene) for CRC by prioritizing candidate genes obtained from the GWAS data (Zhang et al., 2015).Zhao P et al. (2022) highlighted the integration of metabolomics and transcriptomics as a powerful approach to gain insights into the antitumor mechanism of tadalafil, a phosphodiesterase type 5 (PDE5) inhibitor, in patients with CRC (Zhao et al., 2022).The anti-cancer effect of niclosamide was confirmed in nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice implanted with human CRC xenografts from patients with metastatic CRC and remarkable results were obtained with a non-lethal dose of 200 mg/kg (Osada et al., 2011).Tioconazole, originally used to treat vaginal yeast infections, was reported to enhance chemotherapy efficacy in colorectal tumor xenografts (Liu et al., 2018).Table 1 provides a comprehensive summary of various repurposed drugs in CRC.

Pancreatic cancer
In the field of PC treatment research, recent developments in DR have yielded promising strategies and novel candidates for therapeutic intervention.Pham et al. (2022) have recently introduced a deep learning framework that employs various genome-wide chemical-induced gene expression datasets to predict gene rankings in expression profiles induced by de novo chemicals, based on their chemical structures.They used this model for DR to identify potential treatments for PC from all existing drugs in DrugBank, and proposed candidates including dipyridamole, AZD-8055, linagliptin, and preladenant, which were subsequently validated in vitro (Pham et al., 2022).A recent study established a biobank of over 30 genetically distinct human pancreatic ductal adenocarcinoma (PDAC) organoid lines, demonstrating their correlation with the molecular and phenotypic heterogeneity observed in primary PDAC tissue and in vivo drug responses.Using a fully automated screening platform, this study conducted a DR analysis covering 1,172 FDA-approved drugs.Among the in vivo validated hits were several drugs currently approved for noncancer indications, including emetine and ouabain.These drugs were found to specifically target PDAC organoids by disrupting their response to hypoxia.Notably, a dose of 0.56 mg/kg/d of ouabain significantly reduced PDAC xenograft growth in mice (Hirt et al., 2022).Mugiyanto et al. (2022) used genomic data from the cBio Cancer Genomics Portal to identify PC-associated and drug target genes.Through functional annotations, they prioritized 318 PC risk genes, of which 216 were druggable according to DrugBank.The Connectivity Map (CMap) Touchstone analysis revealed 13 potential PC drugs, including midostaurin and fulvestrant, which target PRKA and ESR1 respectively, as promising candidates for PC treatment (Mugiyanto et al., 2022).Another study investigated whether aspirin (ASA) and oseltamivir phosphate (OP) treatment could enhance PC cell sensitivity to gemcitabine-induced cytotoxicity and hinder chemoresistance development.The combination of ASA and OP with gemcitabine significantly disrupted PC cell viability, clonogenicity, ECM protein expression, migration, and induced apoptosis in MiaPaCa-2 and PANC-1 cells (Qorri et al., 2022).Several studies have recently identified Meflin, a glycosylphosphatidylinositol-anchored membrane molecule, as a functional maker of cancer-restraining CAFs (rCAFs) in PDAC (Kobayashi et al., 2019(Kobayashi et al., , 2021;;Mizutani et al., 2019;Takahashi et al., 2021).Lida et al. conducted a screening of nuclear receptor ligands and identified Am580, a synthetic retinoid and RARα-selective agonist, as a compound that upregulates Meflin expression in both human pancreatic stellate cells (PSCs) and mouse mesenchymal stem cells (MSCs).Furthermore, Increasing Meflin enhances tumor sensitivity to chemotherapy in a PDAC xenograft model (Iida et al., 2022).The phase I and II clinical trials are conducting to investigate the efficacy of a combination of AM80 and gemcitabine and nabpaclitaxel in patients with advanced PC (Mizutani et al., 2022).Molecular docking technique confirmed that ZINC000001612996, ZINC000052955754, ZINC000003978005, and ZINC000006716957 could potentially act as small molecule drugs and co-ligands for TRPC3 and TRPC7, both of which are part of the Transient Receptor Potential Channels (TRPs)-related gene signature in PDAC (Shi W. et al., 2022).Chen et al. (2020) reported dose-dependent negative effects of albendazole, an anthelmintic drug, on proliferation, migration and viability of the human PC cell lines SW1990 and PANC-1.The cytotoxicity of albendazole was further confirmed in vivo using a nude mouse xenograft model, showing a significant reduction in tumor growth (Chen H. et al., 2020).Table 2 presents details on various repurposed drugs for PC.

Liver cancer
Recently, several innovative transcriptomics-based DR methods were employed to uncover novel therapeutic candidates and combinations for the treatment of hepatocellular carcinoma (HCC or LC).Regan-Fendt K et al. (2020) showed that fostamatinib and dasatinib could be effective for sorafenibresistant HCC (Regan-Fendt et al., 2020).Additionally, an mRNA expression profile-based DR method revealed that TOP2A has consistently unfavorable association with HCC patient survival and successfully repositioned withaferin-a (WFA) and mitoxantrone (MTX) through molecular docking studies as potential inhibitors of HepG2 cell proliferation for HCC treatment (Yuan et al., 2022).Tan et al. (2023) have screened a compound library containing 419 FDA-approved drugs and discovered that desloratadine, an antiallergic drug, inhibits proliferation in HCC cell lines as well as CDX, patient-derived organoid (PDO) and PDX.The study also identified N-myristoyl transferase 1 (NMT1) as its target, linking high NMT1 and VILIP3 expression to advanced HCC stages and poor survival (Tan et al., 2023).Furthermore, the combined effects of sorafenib, raloxifene, and loratadine on LC cells were assessed, finding that these two-or three-drug combinations significantly reduced metabolic activity, increased apoptosis, and decreased colony formation compared to single-drug treatments, suggesting the potential of the triple combination as a promising approach for LC treatment (Villarruel-Melquiades et al., 2023).We have summarized several repurposed drugs for HCC in Table 3.

Gastric cancer
Zhang et al. (2022) identified 6-Thioguanine (6-TG) as a potential therapeutic agent for GC by inducing ferroptosis through inactivation of system xc − , inhibition of glutathione (GSH) production, downregulation of GPX4, and elevation of lipid reactive oxygen species (ROS) levels in MGC-803 and AGS cell lines, with in vivo data supporting its anti-tumor activity (Zhang et al., 2022a).Furthermore, it has been shown that HC-056456, a CatSper channel blocker, as a novel ferroptosis-inducing compound inhibits GC cell growth by reducing GSH through the p53/ SLC7A11 pathway, leading to increased Fe2+ and lipid peroxides in vitro and in vivo (Zhang et al., 2022b).In a comprehensive bioinformatic analysis of highly differentially expressed genes in GC, Hossain et al. (2023) found that CDH2, COL4A1, and COL5A are associated with the survival of GC patients.They also identified docetaxel, lanreotide, venetoclax, temsirolimus, and nilotinib as the top six candidate drugs, targeting aforementioned proteins, for the treatment of GC patients (Hossain et al., 2023).Nitazoxanide also yielded favorable outcomes, as it exhibited activity across GC cell lines tested (Ribeiro et al., 2023).In addition, Rabben et al. (2021a) computationally predicted the repositioning of ivermectin for the treatment of GC based on gene expression profiles of both human and mouse models of GC.They further validated their in silico prediction used human GC cell lines MKN74 and KATO-III in vitro.Transgenic insulin-gastrin (INS-GAS) mice were employed for experimental validation of ivermectin in GC treatment (Rabben et al., 2021a).Furthermore, in vivo and in vitro anti-tumor and growth suppression effects of ivermectin were demonstrated on GC, showing that ivermectin suppressed MKN1 cells growth through yes-associated protein 1 (YAP1) downregulation (Nambara et al., 2017).To identify drugs capable of inhibiting the DNA-binding activity of the helicobacter pylori transcription factor HP104, a combination of computational and in vitro methods led to the discovery of three promising drugs, including temoporfin, trientine, and tetraethylenepentamine, for potential antibacterial applications (Antoniciello et al., 2022).Table 4 provides a summary of available repurposed drugs in GC.

Other examples of drug repurposing in gastrointestinal cancers
Surveillance for individuals at risk of GI cancers is essential for early diagnosis and prognosis improvement, and in choosing longterm chemoprotective drugs, approved molecules with well-known long-term effects are preferred.Aspirin (acetylsalicylic acid), introduced at the end of the 19th century, has been proposed for various diseases, such as cardiovascular diseases, strokes (Brighton et al., 2012;Dimitriadis et al., 2022;Gdovinova et al., 2022) and the chronic treatment of Fabry Disease (Monticelli et al., 2022).Specifically, aspirin has been shown to prevent CRC (Baron et al., 2003;Benamouzig et al., 2012;Guirguis-Blake et al., 2022) and PC (Streicher et al., 2014).Other non-steroid anti-inflammatory molecules like celecoxib (Arber et al., 2006;Bertagnolli et al., 2006) and sulindac (Long et al., 2020) have also been suggested for CRC.
Brefeldin A, originally used as a macrolide antibiotic, has shown significant induction of autophagy in CRC cells both in vitro and in vivo (Bei et al., 2022).It functions by provoking endoplasmic reticulum stress (ER-stress) and upregulating Bip which decreases Akt phosphorylation through increased Bip/Akt interaction leading to autophagy induction in CRC cells (Zhou L. et al., 2019).In addition, antifungal drug ketoconazole has been reported to induce PINK1/Parkin-mediated mitophagy and accelerate apoptosis in HCC cells via COX-2 downregulation (Chen Y. et al., 2019).
Genistein, originally prescribed for reducing symptoms of menopause, osteoporosis, and obesity, has shown promising effects in cancer therapy.Genistein has been reported to inhibit proliferation, induce apoptosis and cell cycle arrest in G2 by inhibiting the Wnt/β-catenin signaling pathway in CRC cells (Oliveira et al., 2022).It also promotes apoptosis in HT29 CRC cells by modulating the caspase-3 and p38 MAPK signaling pathways (Shafiee et al., 2016).Furthermore, genistein inhibits glycolysis and induces mitochondrial apoptosis through downregulating of HIF-1α which leads to GLUT1 and HK2 inactivation and apoptosis induction in drug-resistant HCC cells (Li J. et al., 2017).Genistein modulates telomerase activity and reduces tumorigenesis by hTERT downregulation as well as modulation of Gli1 gene expression to weaken cancer stem-like properties in GC cells (Jian- Hui et al., 2016).Genistein, which shows promise as an anticancer drug candidate, is currently in phase II of clinical trials (Cao et al., 2022;Chu et al., 2023).
Metformin is another successful repositioned drug for GI cancers which is currently in phase II of the clinical trial.Metformin reduces cell survival and tumorigenesis by lowering serum insulin levels and downregulation of IGF-1 (Sarfstein et al., 2013).Additionally, it induces G1-arrest via AMPK activation and cyclin D1 downregulation (Wang Y. et al., 2018), inhibits proliferation through mTOR signaling pathway regulation regardless of AMPK dependency (Demaré et al., 2021).It has been shown that metformin inhibits the progression of GC through the inhabitation of HIF1α/PKM2 signaling (Chen G. et al., 2015).Furthermore, it inactivates RAS/ERK and AKT/mTOR signaling pathways and reduces proliferation in KRAS-derived tumors.Metformin has been reported to selectively inhibit KRAS-driven metastatic CRC by silencing MATE1 (Xie et al., 2020).
Several Studies have demonstrated that the combination of repurposed drugs with cytotoxic drugs, radiotherapy or even the combination of multiple repositioned drugs can exhibit synergistic antitumor effects.A recent meta-analysis by Heer et al. (2022) revealed that aspirin when co-administered with sulindac and difluoromethylornithine (DFMO), an inhibitor of ornithine decarboxylase used to treat facial hirsutism, showed significantly more effective results in protecting against CRC adenomas (Heer et al., 2022).The combination of bortezomib and chloroquine has been shown to suppress proliferation and induce apoptosis in human liver tumors, whether orthotopically or subcutaneously xenografted in mice (Hui et al., 2012).In clinical research, combining nelfinavir with a short course of hypofractionated radiotherapy (SCHRT) showed increased sensitivity of CRC tumors to radiotherapy (Meyn et al., 2016).In addition, FOLFIRINOX, a chemotherapy regimen comprising leucovorin calcium (folinic acid), fluorouracil, irinotecan hydrochloride and oxaliplatin, is used for the treatment of advanced PC.In a phase II clinical trial, FOLFIRINOX is combined with losartan as neoadjuvant therapy, followed by chemoradiotherapy, for locally advanced PC (Murphy et al., 2019).Another successful example of DR is a combination of metformin, digoxin and somatostatin, which has shown significant suppression of PC cell proliferation in clinically relevant animal models.Currently, this combination is being evaluated in a clinical trial (Liu S.-H. et al., 2020).Tables 1-4 provide comprehensive information on various drugs that have been successfully repurposed to treat GI cancers and approved for clinical use by the FDA.Furthermore, we have prepared a Supplementary Table S2, which presents a list of repurposed unapproved or withdrawn drugs/natural components targeting GI-related cancers.
6 Challenges and future perspectives Drug repurposing for cancer therapy is widely used to discover new indications for existing compounds; However, only a few repurposed drugs have been formally subjected to the clinical treatment guidelines.Despite advantages such as anticancer pharmacokinetic parameters, acceptable safety and tolerability in humans, there is still a risk of late-phase clinical trial failure due to competition with new drug development.The legal and regulatory barriers such as patents issues and prescription charges must also be addressed (Vickers, 2017;Breckenridge and Jacob, 2019;Pushpakom et al., 2019).
The financial factors significantly influence the clinical development and approval of repurposing drugs, as private sector organizations prioritize higher returns due to intellectual property rights (Vickers, 2017).Most registered clinical trials listed in the Repurposing Drugs in Oncology (ReDo) project are sponsored by Universities or Hospitals (67%), research institutes or non-profit organizations (28%), and only a small percentage by pharmaceutical companies (Pantziarka et al., 2018).Patent considerations for off-patent drugs can pose significant barriers to DR, requiring credible and statistically strong evidence for a new indication and legislative efforts to address this issue (Pushpakom et al., 2019).
Physicians prescribe drugs based on scientific evidence from clinical trials, and both generic and repurposed drugs should be used when suitable.Nevertheless, the pharmaceutical industry can exert substantial investments on drug promotion, physician marketing and consumer advertising.A clear example is thalidomide, originally used as a sedative or antiemetic, which has been repurposed for multiple myeloma.Despite phase III clinical trials showing no survival advantage for the combination of melphalan-prednisonelenalidomide over melphalan-prednisone-thalidomide (Stewart et al., 2015;Zweegman et al., 2016) lenalidomide became the standard treatment approach, even with a higher estimated cost than thalidomide.Thus, underfunding for clinical trials in oncology and prescribing rejection bias remain significant challenges for cancer drug repurposing.
Another challenge lies in the genetic diversity of individuals and the complex nature of disease (Pritchard et al., 2017).While there are many common pathways involved in cancer development, differences in pathways and genes exist among subgroups and individuals.This genetic diversity results in varying side effects and treatment responses to routine drugs and therapies.
Moreover, current GI cancer therapies may be ineffective for specific patients or cancer types due to inadequate drug targeting or inefficient drug interactions (Apicella et al., 2017).In rectal cancer, personalized DR based on gene expression signatures and reverse drug-induced gene expression profiles has shown promising results.For instance, Carvalho et al. (2021) identified potential topoisomerase II inhibitors like doxorubicin, teniposide, idarubicin, mitoxantrone and epirubicin for CRC therapy, leading to a significant reversal of rectal cancer gene expression signatures (Carvalho et al., 2021).Given the significant differences in drug efficacy among individuals due to gene profiles and tumor heterogeneity, it becomes crucial to focus on DR based on tumor/subject molecular profiles to reduce inefficiencies in cancer treatment (Li and Jones, 2012).
While drug repurposing offers various benefits compared to the conventional de novo approach, it may not always lead to success due to lack of efficacy or toxicity issues.Bevacizumab, initially developed to treat CRC, was determined to be a strong candidate to treat other kinds of cancer such as colon, rectal, brain, lung, and kidney through drug repositioning.However, it failed in phase III trials despite positive results (Kim and Oh, 2018).These failed drug candidates in clinical trials still represent an affluent resource for repositioning, as they are well studied pharmacokinetically and clinically.Personalized genomics studies focusing on patient and disease heterogeneities may reveal that many of these failures were tested in inappropriate subject groups, making them practical options for future personalized medicine approaches, particularly for subjects with limited treatment options.

Conclusion
Drug repurposing is increasingly considered by both academia and the pharmaceutical industry as a cost and timesaving alternative to de novo drug development.Repurposing non-oncology drugs in cancer therapy provides a promising therapeutic opportunity, especially for patients with rare cancers, advanced diseases, or chemo-resistant tumors.In the present review, we have explored the potential of DR approaches, with a particular focus on their application in GI cancers.Repurposed drugs can target known pathways and key molecular targets in cancer biology due to their established functional mechanisms.DR has received attention owing to its potential to enhance treatment effectiveness and ability to overwhelm resistance to standard chemotherapy as well as improve therapy outcomes in tumors with limited response to conventional treatments.Additionally, when repurposed drugs are used in combination with routine oncology drugs, they offer a unique opportunity to target multiple pathways and molecular targets in cancer cells, going beyond the scope of traditional chemotherapy drugs and modulating diverse cancer-relevant pathways.However, it should be noted that the interaction of repurposed drugs with standard cancer drugs may pose challenges during the clinical trials.
The repositioning of drugs to treat GI cancers presents an attractive option given the increasing number of new cases, annual deaths, and the challenges in treating certain tumors.This review outlines various DR approaches that can be used to improve the efficiency of existing GI therapies.However, further clinical studies are needed to determine their potential for clinical adoption.
FIGURE 1Repurposed drugs currently being evaluated against gastrointestinal cancers in various clinical phases.Metformin, simvastatin, and propranolol are entering phases 2 and 3 to treat more than 2 gastrointestinal cancers.Two drugs indomethacin and azithromycin for colorectal cancer, and gemcitabine for pancreatic cancer have reached phase IV.

TABLE 1
List of repurposed drugs proposed for targeting colorectal carcinoma.

TABLE 1 (
Continued) List of repurposed drugs proposed for targeting colorectal carcinoma.

TABLE 1 (
Continued) List of repurposed drugs proposed for targeting colorectal carcinoma.

TABLE 2
List of repurposed drugs proposed for targeting pancreatic cancer.

TABLE 2 (
Continued) List of repurposed drugs proposed for targeting pancreatic cancer.

TABLE 2 (
Continued) List of repurposed drugs proposed for targeting pancreatic cancer.

TABLE 2 (
Continued) List of repurposed drugs proposed for targeting pancreatic cancer.

TABLE 3 (
Continued) List of repurposed drugs proposed for targeting hepatocellular carcinoma.TABLE 4 List of repurposed drugs proposed for targeting Gastric cancer.