Upregulation of fibronectin and its integrin receptors – an adaptation to isolation stress that facilitates tumor initiation Free

Tumor initiation is an inefficient biological process in which malignant cells must fulfill a series of conditions, such as accumulating gene mutations, losing tumor suppressor genes, evading immune surveillance and overcoming microenvironment-imposed stresses (Welch and Hurst, 2019; de Visser and Joyce, 2023; Wu et al., 2023b). Only a few tumor cells harbor the ability to initiate the formation of a tumor at primary and metastatic sites; these tumor cells are called tumor-initiating cells (TICs) or cancer stem cells (CSCs). CSCs possess intrinsic properties that allow them to overcome microenvironmental stressors including hypoxia, nutrient deprivation, loss of appropriate matrix and chemotherapy (Huang et al., 2020). However, recent evidence suggests that tumor-initiating capacity is a plastic feature that non-CSCs can acquire as an adaptive response to overcome stress (Nallasamy et al., 2022).

Emerging questions in cancer research concern the role that CSCs play in solid tumor development. Do all tumors require CSCs? While CSCs are linked to tumor initiation, do they also play a role in tumor progression and metastasis? There is evidence that CSC properties are plastic and can be induced by stresses within the tumor microenvironment (Nallasamy et al., 2022). Understanding how the tumor microenvironment regulates cancer stemness at the molecular level could lead to a new class of therapies to prevent or reverse cancer stemness and tumor progression. Tumor cells that can adapt to various microenvironmental stresses, such as hypoxia, oxidative stress, nutrient stress, and stress during treatment with chemotherapy or radiation, represent cells with stem-like properties (Ferguson et al., 2021; Nallasamy et al., 2022).

The immediate tumor cell niche, including the extracellular matrix (ECM), represents a significant and powerful tool utilized by tumor cells to overcome microenvironment-imposed stresses and to gain tumor-initiating potential (Winkler et al., 2020). The ECM interacts with tumor cell surface receptors, primarily integrins, to initiate signaling cascades that promote cell survival and proliferation in an otherwise hostile microenvironment (Winkler et al., 2020). Integrins function as cell surface sensors and drivers of tumor cell survival within various microenvironments, including those that result in cellular stress (Cooper and Giancotti, 2019). In addition, non-integrin ECM receptors play important roles in stress tolerance and cancer stemness; for example, discoidin domain receptor 1 (DDR1), a collagen receptor and tyrosine kinase, is involved in creating an immune-exclusive microenvironment, conferring therapy resistance and maintaining tumor dormancy at metastatic sites (Sun et al., 2021; Sirvent et al., 2022).

Stromal cells, such as cancer-associated fibroblasts (CAFs) and tumor-infiltrating immune cells, are thought to largely account for the production of ECM in established tumors (Yuan et al., 2023). However, it is also interesting to consider how individual tumor cells must overcome hurdles to form colonies within non-permissive tissues prior to the appearance of stromal cells, blood vessels and inflammatory mediators that are enriched within a pro-tumor microenvironment. Accumulating evidence suggests that certain tumor cells can remodel their surrounding ECM, allowing cells to survive cellular stress and thereby promote tumor initiation and/or metastasis (de Visser and Joyce, 2023). Here, we further consider whether a single tumor cell that finds itself embedded within a normal (i.e. tumor-suppressive) tissue environment can autonomously produce and organize an ECM niche to support the initiation and growth of a new tumor colony.

One key tumor-promoting ECM protein is fibronectin (FN), a provisional matrix protein that serves as a scaffold for the deposition of a range of ECM components, including collagens, fibrin and proteoglycans, as well as growth factors and cytokines (Efthymiou et al., 2020). FN expression is highly elevated in various types of cancer, and high FN expression correlates with poor clinical outcomes (Spada et al., 2021). Importantly, FN expression is also enriched in a tumor-initiating or metastatic niche, indicating that FN could play a role in tumor initiation and metastasis (Efthymiou et al., 2020). Whereas previous studies have established that stromal cell-derived FN promotes tumor initiation and metastasis (Efthymiou et al., 2020; Erdogan et al., 2017; Rick et al., 2019; Spada et al., 2021), the role of tumor cell-derived FN remains poorly understood.

Our group recently reported that cancer cells undergoing tumor initiation can produce and deposit FN into their surrounding ECM, and that this skill is enabled by the stress-induced upregulation of the cell surface molecule lysophosphatidic acid receptor 4 (LPAR4), a G-protein-coupled receptor (GPCR) for lysophosphatidic acid (LPA) (Wu et al., 2023a). We observed that LPAR4 activates the AKT–cAMP response element-binding protein (CREB) pathway, directly leading to the expression and deposition of a FN-containing matrix. Cells that engage this matrix can overcome isolation stress in vitro and more readily accomplish tumor initiation in vivo, and we found the interaction between FN and the integrins αvβ3 and α5β1 to be indispensable for this niche-induced cancer stemness (Wu et al., 2023a). In fact, other groups have reported that various environmental stresses can promote the expression of the FN-binding integrins αvβ3 and α5β1 (Ju et al., 2017; Cooper and Giancotti, 2019). In this Opinion article, we will define isolation stress and discuss its role in adaptive stress-responsive mechanisms utilized by cancer cells, with a particular focus on how FN–integrin interactions mediate pro-tumor functions. We will also discuss potential therapeutic approaches to target adaptive stress-responsive pathways to sensitize cancer cells to the effects of therapy.

During tumor initiation at primary or metastatic sites, solitary tumor cells must overcome challenges without support from pro-tumor blood vessels, stroma or immune cells (Bakhshandeh et al., 2022). Although a solitary cancer cell has gained various oncogenic gene mutations, it has not yet lost interactions with surrounding normal cells or with the normal ECM at the early stages of tumor initiation. These interactions, in fact, have a suppressive effect on tumor initiation (Joyce and Pollard, 2009; Quail and Joyce, 2013). Consequently, solitary tumor cells experience a series of growth hurdles, including a lack of oxygen and nutrients, a lack of cell–cell contact with other tumor cells, limited access to appropriate ECM, and deprivation of growth factors and cytokines supplied by pro-tumor stroma and immune cells that are present in established tumors (Bakhshandeh et al., 2022). During metastasis, solitary tumor cells that have successfully escaped from the primary tumor site and entered the bloodstream (i.e. circulating tumor cells) would encounter new microenvironment-imposed stressors, such as oxidative stress (Tasdogan et al., 2021), loss of cell–cell contact and detachment of cell–matrix interactions (Shen and Kang, 2020). Additionally, chemotherapy treatments designed to target proliferating tumor cells create excessive oxidative stress in tumor cells (Jiang et al., 2023). We collectively define such stressors faced by solitary tumor cells during tumor initiation, progression and metastasis as ‘isolation stress’ (Wu et al., 2023a,b) (Fig. 1).

Stress tolerance is not only an inherent CSC trait; this feature can be acquired by non-CSCs as they adapt to microenvironmental cues or sublethal stressors (Nallasamy et al., 2022). Tumor cells that successfully adapt to isolation stress tend to be more cross-stress tolerant and aggressive. As an example of this reprogramming cascade, hypoxia-inducible factor 1α (HIF1α) represses the expression of miR-338-5p in colon cancer cells exposed to hypoxia, and the downregulated miR-338-5p leads to increased expression of interleukin-6 (IL-6), which in turn activates signal transducer and activator of transcription 3 (STAT3) and upregulates the STAT3 target B-cell lymphoma 2 (BCL2). These molecular changes consequently contribute to cell resistance to oxaliplatin, a standard-of-care treatment for colon cancer patients (Xu et al., 2019). T-cell lymphoma and esophageal squamous cancer cells treated with sublethal doses of hydrogen peroxide (an oxidative stress inducer) exhibit a more aggressive tumor phenotype with increased therapy resistance and clonogenicity (Wu et al., 2018, 2016; Zhang et al., 2016). Pancreatic cancer cells exposed to nutrient deprivation become more migratory and metastatic through the upregulation of Slug (also known as SNAI2) (Recouvreux et al., 2020). Furthermore, our recent findings indicate that pancreatic cancer cells subjected to isolation stress, such as hypoxia, oxidative stress, nutrient stress or chemotherapy, develop increased therapy resistance and gain tumor-initiating capacity (Wu et al., 2023a). Taken together, these studies illustrate how isolation stress can contribute to the development of a more aggressive tumor phenotype.

Multiple lines of evidence suggest that solitary tumor cells must overcome isolation stress to initiate tumor formation at both primary and remote sites (Shen and Kang, 2020; Insua-Rodriguez et al., 2018). As noted above, tumor cells that have overcome isolation stress show signs of increased cross-stress tolerance and cancer stemness features, allowing such cells to contribute to therapy resistance, tumor initiation and metastasis. To design new strategies to prevent tumors from overcoming isolation stress, it will be important to understand the molecular events that govern this process, since the adaptive mechanisms developed by tumor cells in response to isolation stress are often plastic and reversible.

Over the past decade, our group has focused on integrin αvβ3, an isolation stress-inducible cell surface molecule that is involved in cancer stemness, tumor initiation and metastasis. We and others have reported that αvβ3 expression is enriched on therapy-resistant tumor cells (Fox et al., 2021; Seguin et al., 2015; Haeger et al., 2020). For instance, we have shown that αvβ3 expression is upregulated on lung adenocarcinoma cells that gain resistance to the tyrosine kinase inhibitor erlotinib (Seguin et al., 2014). Integrins, such as αvβ3 and some β1 integrins, can function independently of their ECM ligands, which include vitronectin, FN, fibrinogen and osteopontin, among others (Xiong et al., 2021; Haake et al., 2022; Seguin et al., 2015; Desgrosellier et al., 2009) (Fig. 2). For example, αvβ3 enhances anchorage-independent growth by activating Src, independently of cell adhesion-mediated focal adhesion kinase (FAK, also known as PTK2) activation, and the activated Src is required for αvβ3-induced tumor cell survival and metastasis in vivo (Desgrosellier et al., 2009). Additionally, αvβ3 recruits Kirsten rat sarcoma viral oncogene homolog (KRAS) GTPase and v-ral simian leukemia oncogene homolog B (RalB) GTPase to the tumor cell plasma membrane, leading to activation of the IκB kinase (IKK)-related kinase TBK1 and nuclear factor κB (NF-κB) signaling, ultimately driving cancer stemness and therapy resistance (Seguin et al., 2014). Further evidence suggests that αvβ3–KRAS–RalB–NF-κB signaling is necessary and sufficient for tumor initiation, anchorage independence and drug resistance (Seguin et al., 2014). Interestingly, αvβ3 is a biomarker for mutant KRAS-addicted tumor cells (Seguin et al., 2017). Specifically, we have found that the carbohydrate-binding protein galectin-3 binds to and triggers clustering of αvβ3 on the cell surface, which subsequently results in the recruitment of KRAS to the tumor cell plasma membrane and activation of AKT kinases (herein referring to AKT1, AKT2 and AKT3). These events lead to two major KRAS-induced effects: macropinosome formation and redox balancing. Ultimately, these changes result in mutant KRAS-expressing tumor cells becoming KRAS addicted (Seguin et al., 2017). Importantly, αvβ3 expression is enriched on circulating drug-resistant non-small cell lung tumor cells in a nude mouse model, and the population of circulating tumor cells dramatically decreases in these mice following treatment with LM609, an antibody that targets αvβ3 (Wettersten et al., 2019). Furthermore, αvβ3 expression is enriched on mammary stem cells, where it is required for initiation of alveologenesis and remodeling during mouse pregnancy (Desgrosellier et al., 2014). In breast cancer, αvβ3, in the absence of its ligands, activates Src and Slug, and represses p53-upregulated modulator of apoptosis (PUMA, also known as BBC3), an important tumor suppressor involved in breast cancer metastasis, thereby leading to enhanced cancer stemness and metastasis (Desgrosellier et al., 2014; Sun et al., 2018). In addition, αvβ3 is involved in cancer glucose metabolism, as it upregulates the expression of glucose transporter 3 (Glut3, also known as SLC2A3), a glucose transporter with high affinity, through PAK4–YAP1–TAZ (WWTR1) signaling in a subset of classical and proneural glioblastomas (Cosset et al., 2017). This αvβ3 dependency is particularly significant for this subset of glioblastoma cells, since αvβ3-induced Glut3 facilitates uptake of extracellular glucose more efficiently to overcome the brain microenvironment, which is particularly deficient in glucose (Cosset et al., 2017). Collectively, these studies suggest that even in the absence of its ligands, integrin αvβ3 can function as an isolation stress-induced cell surface molecule that drives stress tolerance, tumor initiation and metastasis. As such, αvβ3 is an interesting cell surface receptor to target therapeutically for multiple types of cancer.

As mentioned above, we recently discovered another isolation stress-inducible cell surface molecule, LPAR4, and described its contribution to pancreatic cancer (Wu et al., 2023a). We reported that LPAR4-expressing pancreatic cancer cells create a pro-tumor FN-containing ECM niche that not only provides growth support to LPAR4-expressing cells, but also to neighboring LPAR4-negative cells (Wu et al., 2023a). The FN within this ECM niche was shown to be required to promote tumor initiation in vivo. Interestingly, we found that LPAR4 activates the AKT–CREB pathway, leading to the expression of FN even in the absence of exogenous LPA, hinting that LPAR4 might have ligand-independent functions (Wu et al., 2023a). In support of this notion, there is additional evidence that many GPCRs display ligand-independent functions, referred to as basal (or constitutive) GPCR activation (Yao et al., 2009; Gavriilidou et al., 2019; Chen et al., 2022a).

We found the induction of LPAR4 in response to isolation stress to be transient and plastic in nature (Wu et al., 2023a). In general, evidence suggests that other cellular responses to isolation stress are also plastic, and that stress-adaptive responses are reversible upon the alleviation of stress (Frank et al., 2020; Sánchez-Navarro et al., 2022). In the case of LPAR4, expression is upregulated only during the early stages of tumor initiation, after which it decreases significantly in well-established tumors in vivo (Wu et al., 2023a). We identified miR-139-5p as a tumor suppressor that represses LPAR4 expression in pancreatic cancer and observed that tumor cells suppress miR-139-5p expression in response to isolation stress to release the brake on LPAR4 expression (Wu et al., 2023a). In well-established tumors, individual tumor cells are less likely to experience isolation stress, since they receive growth support from stromal and immune cells, as well as neighboring tumor cells. In the absence of isolation stress, these tumor cells retain a high level of miR-139-5p expression that triggers the brake on LPAR4 expression. This might help explain why only moderate LPAR4 expression is detected in well-established pancreatic tumors. Echoing our study, Lee et al. have reported that LPAR4 expression reaches its peak after 3–7 days of in vitro differentiation in cardiac progenitor cells, after which it subsequently declines (Lee et al., 2021). The authors also found that LPAR4 expression is only transiently expressed in the early stage of embryonal heart development in vivo (Lee et al., 2021), suggesting the presence of a dominant-negative regulator of LPAR4 expression in mature mouse cardiac cells.

Taken together, these examples illustrate how cells mitigate isolation stress by employing various biological strategies that involve collaboration between cell surface molecules and ECM proteins. Since gaining this stress-tolerant state can contribute to tumor initiation and metastasis, targeting these adaptive mechanisms might provide significant therapeutic benefit.

FN is a critical component of the ECM that has been well established to function as a driver of stress tolerance, tumor initiation and metastasis (Rick et al., 2019; Efthymiou et al., 2020). Tumor-associated stromal cells, such as CAFs and immune cells, are the major source of FN in the tumor microenvironment (Spada et al., 2021). However, tumor cells are also believed to produce FN, and this autonomous effect is critical for their survival under isolation stress and facilitates metastasis to distant organs. For example, dormant breast cancer cells deposit FN to help overcome nutrient-deprivation stress (Barney et al., 2020), whereas metastatic breast cancer cells produce FN to support invasion and survival under nutrient-stress conditions (Jun et al., 2020). In pancreatic cancer, tumor cells upregulate FN expression in response to isolation stress, thereby leading to enhanced stress tolerance and tumor initiation (Wu et al., 2023a). We speculate that pancreatic cancer cells may reactivate this mechanism during metastasis, particularly during circulation and tumor cell seeding in new distant organs, when solitary tumor cells are once again exposed to isolation stress. Chronic inflammation has been linked to cancer initiation and stress tolerance (Greten and Grivennikov, 2019; Grivennikov et al., 2010). One of the underlying mechanisms involves the production of oxidizing compounds upon chronic inflammation, resulting in a tumor microenvironment that has elevated levels of reactive oxygen species (ROS) (Grivennikov et al., 2010). This stressful microenvironment is expected to upregulate FN expression in tumor cells as well. In summary, this stress-inducible cell-autonomous mechanism exploited by cancer cells is important for their survival in an adverse microenvironment, especially where they cannot gain access to appropriate ECM, growth factors and cytokines normally provided by pro-tumor stromal or immune cells.

Due to splicing events, FN has over 19 different isoforms, which are presented in soluble plasma form (pFN) or as an insoluble cellular mesh (cFN) (Spada et al., 2021). cFN forms homodimers and assembles into fibrillar structures in a process mediated by cell–matrix interactions involving integrins α5β1 and αvβ3 (Spada et al., 2021). The interaction between FN and integrins induces integrin clustering, formation of focal adhesion complexes, actin filament polymerization and cell cytoskeleton contractility. This, in turn, leads to conformational changes in FN-binding domains that are required for FN–FN molecular interactions and FN fibrillogenesis (Spada et al., 2021). CAFs are known to efficiently assemble and deposit FN, which enhances tumor initiation and progression (Erdogan et al., 2017; Zeltz et al., 2020). However, tumor cells might lack the capacity to assemble FN as efficiently as CAFs do. We have observed that FN is not assembled in LPAR4-expressing pancreatic cancer cells (Wu et al., 2023a), which is consistent with a recent study showing that breast cancer cells expressing high levels of FN are also unable to assemble FN (Libring et al., 2020). Although LPAR4-expressing pancreatic cancer cells deposit non-fibrillar FN, LPAR4-negative pancreatic tumor cells can still benefit from this form of FN, indicating that non-fibrillar FN is capable of binding to receptors and driving pro-tumor functions.

FN interacts with multiple receptors, including integrins α5β1, αvβ3, α4β1 and α3β1, as well as syndecans, and these interactions not only integrate cell–matrix signaling but also influence FN assembly (Efthymiou et al., 2020; Stepp et al., 2010; Danen et al., 2002). Among the FN-interacting integrins, α5β1 and αvβ3 are both isolation stress inducible and play important roles in stress tolerance and tumor progression. In breast cancer and melanoma cells, hypoxia stress induces expression of both α5β1 and αvβ3, which in turn promotes cell adhesion, cell migration or metastasis (Ju et al., 2017; Cowden Dahl et al., 2005; Sesé et al., 2017). Mesenchymal renal cell carcinoma and soft tissue sarcoma cells express higher levels of integrin αvβ3 and FN, and the interaction between FN and αvβ3 upregulates Slug expression, thereafter promoting epithelial-to-mesenchymal transition and lung metastasis (Knowles et al., 2013). Isolation stress induces pancreatic cancer cells to deposit a FN-rich ECM that can confer cancer stemness to non-CSCs in a manner that depends on FN, α5β1 and αvβ3, since blockage of FN–integrin interactions using anti-integrin α5β1 or αvβ3 antibodies prevents the ability of the FN-rich ECM to induce cancer stemness (Wu et al., 2023a). Given that FN, α5β1 and αvβ3 are all isolation stress inducible, solitary tumor cells are expected to utilize these molecules cooperatively to trigger survival mechanisms to overcome hostile microenvironments. The increased levels of pericellular FN allow binding of FN to the two enriched integrins on tumor cells to promote stress tolerance, cancer stemness, tumor initiation and metastasis (Fig. 3).

In conclusion, isolation stress upregulates the expression of both FN and FN-binding integrins, especially α5β1 and αvβ3. Tumor cells utilize the interaction between this matrix protein and its relevant cell surface receptors to collaboratively promote stress tolerance and cancer stemness. Hence, targeting FN–integrin interactions could represent a promising strategy to counteract cellular stress-induced cancer progression, as discussed below.

Tumor cells have evolved various strategies to overcome isolation stress and initiate tumor formation in primary and metastatic sites. Targeting these stress-adaptive mechanisms could provide therapeutic opportunities to prevent tumor initiation and metastasis and to sensitize tumor cells to existing standard-of-care treatments.

Our group and others have demonstrated that LPAR4, FN, αvβ3 and α5β1 are induced by isolation stress and drive stress tolerance and tumor initiation in cancer cells. Therefore, function-blocking antibodies targeting these surface molecules or strategies to block their downstream signaling cascades could be promising therapeutic approaches for cancer. For instance, AKT kinases, which act downstream of LPAR4, are promising therapeutic targets in multiple types of cancer (Hua et al., 2021), and several AKT inhibitors, such as ipatasertib and capivasertib, have shown good safety profiles and have advanced into Phase II or III clinical trials for solid tumors (https://clinicaltrials.gov/; NCT05276973, NCT05172245 and NCT04862663). Alternatively, strategies to prevent tumor cells from gaining expression of these cell surface receptors should also be evaluated.

Antibodies provide one opportunity to prevent FN from interacting with its integrin receptors. Integrin α5β1 and/or αvβ3 might be ideal therapeutic targets for cancer, since these receptors are usually expressed at low or undetectable levels in normal tissues but become highly expressed in cancer or during wound repair (Li et al., 2021; Desgrosellier and Cheresh, 2010). Considering that FN binds to at least two major integrins, α5β1 and αvβ3, targeting only one of them might not be sufficient to hinder FN-induced stress tolerance and tumor progression. Although demonstrating some activity in patients and exhibiting no serious side effects, individual antibodies targeting α5β1 or αvβ3 have not yet proceeded to Phase III trials (Alday-Parejo et al., 2019; Li et al., 2021; Chen et al., 2022b; Gutheil et al., 2000). This might be due to (1) the potential functional redundancy between integrins; (2) the complicated role integrins play in different stages of tumor development; for instance, α5β1 or αvβ3 might be necessary for the early stages of tumor development but not the late stages; or (3) a lack of studies combining integrin inhibitors with the standard-of-care treatment (Alday-Parejo et al., 2019; Chen et al., 2022b). The fact that both α5β1 and αvβ3 appear to be expressed in response to stress and/or drug treatments and recognize FN might make it difficult to achieve clinical success by targeting either integrin alone. Alternatively, targeting the cell-binding domain of FN itself might represent an approach to prevent the pro-tumor effects of FN more effectively. Given the contributions of CAF-deposited FN to tumor progression, blocking FN fibrillogenesis or assembly by CAFs via targeting of FN–integrin interactions warrants further consideration.

In summary, targeting stress-adaptive mechanisms such as LPAR4-induced AKT signaling or FN–integrin interactions might be promising therapeutic strategies for cancer treatment, given their already proven safety profiles. These approaches have the potential to prevent tumor initiation and metastasis, and to prevent drug resistance or enhance the efficacy of existing therapies for cancer patients. Although it could prove impossible to prevent tumor initiation, such treatments might inhibit the initiation of new metastatic lesions where tumor cells are attempting to overcome isolation stress.

Tumor initiation in primary and metastatic sites is a challenging and inefficient process that requires tumor cells to overcome isolation stress to establish the formation of a tumor. In the tumor microenvironment, tumor cells receive growth support from ancillary cells that secrete cytokines, growth factors and ECM proteins. Furthermore, tumor cells develop cell-autonomous programs to sustain themselves under specific circumstances, such as when they have limited access to blood vessels, immune cells or stromal cells.

We have uncovered mechanisms employed by tumor cells to overcome isolation stress and promote tumor initiation that involve stress-induced FN expression, as well as the FN-binding integrin αvβ3 (Wu et al., 2023a; Seguin et al., 2014). Tumor cells collaboratively utilize isolation stress-induced FN, α5β1 and αvβ3 to mitigate isolation stress and gain stress tolerance, cancer stemness and drug resistance. Furthermore, the FN-rich ECM produced by LPAR4-expressing pancreatic cancer cells transfers cancer stemness to neighboring LPAR4-negative cells (Wu et al., 2023a). This indicates that not all tumor cells need to upregulate LPAR4 or FN in response to isolation stress, as a small population of LPAR4-expressing tumor cells is sufficient to provide stemness features to neighboring cells by creating a FN-rich pro-tumor niche. Importantly, it is necessary to explore whether isolation stress-induced LPAR4, FN and FN-binding integrins represent generic mechanisms exploited by epithelial cancer cells, such as those from the lung, prostate and breast. Additionally, testing the therapeutic potential of blocking FN–integrin interactions in preclinical models and clinical trials is urgently needed.

In conclusion, we have recently unraveled several unique stress-responsive mechanisms involving the interaction between FN and integrins that enable tumor cells to overcome inhospitable microenvironments. As such, identifying additional stress-inducible surface receptors might reveal a new class of therapeutic targets to prevent drug resistance and enhance the effectiveness of current cancer treatments.