Stress granules and hormetic adaptation of cancer

Abstract

Cell stress is inherent to cancer and a key driver of tumorigenesis. Recent studies have proposed that cell stress promotes tumorigenesis through the non-membranous organelles, stress granules (SGs). While SG biology is an emerging field, all studies to date point to an enhanced ability of cancer cells to form SGs compared to normal cells, a heightened dependence on SGs for survival under adverse conditions and for chemotherapy resistance, and dependence of tumors on SGs for growth. Why cancer cells become dependent on SGs, or how SGs promote tumorigenesis, remains to be elucidated. Here, we attempt to provide a framework for answering these questions by framing SGs as a hormetic response to tumor-associated stress stimuli.

Tumor-associated stress stimuli and hormesis-based adaptations

In 1943, Southam and Ehrlich made the surprising observation that extracts from red cedar trees enhanced the metabolism of fungi at low concentrations but had an inhibitory effect at higher concentrations [1]. This biphasic response in which low doses of a stressor provide a beneficial effect on the organism, but high doses are detrimental, was named “hormesis.” This concept can be exemplified in a variety of models, all of which converge on the conclusion that moderate doses of stress stimuli protect against higher levels of both acute and long-term stresses [2]. For example, short episodes of mild ischemia have been reported to shield the brain and heart from severe depletion of blood [3, 4]. Moderate exercise can promote positive physiological adaptation in active skeletal muscles, such as mitochondrial biogenesis and synthesis of antioxidant enzymes [5–7]. However, oxidative stress from high-intensity exercise can result in damage to macromolecular structures, such as proteins, lipids, and DNA. Extreme calorie restriction (CR) can lead to malnutrition and associated pathologies, but moderate CR can protect against ageing, cancer, and neurodegenerative diseases [8–11]. In the context of organismal biology, hormesis is seen as an evolutionarily conserved response that is beneficial to the organism; however, the principle itself applies to any living system, including those that are ultimately detrimental, such as cancerous cells and cells in the tumor microenvironment. Cancer occurrence and progression are tightly linked to stress stimuli, and the stress phenotype of cancer is a well-recognized hallmark. These include hypoxia, oxidative stress, ER-stress, nutrient deprivation, DNA damage, acidosis, and biomechanical stress, which are triggered by cancer cell intrinsic factors like oncogenic mutations, as well as cancer cell extrinsic factors like extracellular matrix, poor/faulty vascularization, and chemotherapy (Figure 1) [12, 13]. Tumor-associated stress stimuli are not just byproducts of cancer progression, but are also key mediators of cancer initiation, progression, metastasis, and resistance to therapy [14, 15]. We propose that tumor-associated stressors operate as hormetic stimuli to elicit concerted changes in cellular behavior to benefit the growth and survival of cancer cells (Figure 2). The mechanisms underlying hormesis are poorly understood but are largely attributed to increases in stress adaptive responses, including the endoplasmic reticulum (ER)-stress response, autophagy, and DNA damage response [16–22]. Coincidentally, these adaptive responses have also been shown to play a critical role in cancer progression and therapy efficacy [23, 24]. Evidence emerging over the past few years has identified a newcomer to the stress adaptive responses implicated in tumorigenesis, namely the non-membranous organelles termed stress granules (SGs) (Box 1). [25–29]. SGs are evolutionarily conserved and are distinct from the cancer-associated hormetic responses mentioned above, in that they are triggered by a broad number of stress stimuli (Box 2). A distinguishing feature of SGs is that exposure to one type of stress can enhance SG formation and protection/resistance from another type of the multitude of stresses that are present in cancer (Figure 2A–B). As such we propose that SGs act as a unique platform for cross-stress hormesis. Repeated cycles of this response and the integration of adaptive responses to various stress stimuli may be a major contributor to the supercharged fitness, resistance, and plasticity that are prominent features of tumorigenesis, as well as key challenges to cancer therapeutics.

Figure 1. Tumor-associated stress stimuli.

(A) During the trajectory of a normal cell’s development into a hyperproliferative tumor, cellular stress cooperates with the accumulation of oncogenic drivers, contributing to cellular transformation. The particular stresses that are experienced by the cell during this stage can shape its future dependence on certain stress response pathways for survival. (B) A variety of different stress stimuli are experienced by a cancer cell and a developing tumor, including hypoxia, oxidative stress, ER stress, DNA damage, acidosis, biomechanical stress, and metabolic stress. The intensity, duration, and combination of cellular stresses experienced by a cancer cell can either lead to cell death, or to stress tolerance and the development of evolutionary mechanisms.

Figure 2. Stress Granules and Hormetic Adaptation of Cancer.

(A) When exposed to stressful stimuli, such as 1) oxidative stress, 2) acidosis, 3) ER stress, 4) DNA damage, 5) nutrient deprivation, 6) hypoxia, or 7) biomechanical stress, cancer cells respond by stimulating the formation of SGs. SGs function as a stress adaptive program that promotes cell survival under stress. Mild but repeated episodes of stress may lead to hormetic SG induction and consequently a pre-conditioned or adapted cellular state that is more favorable in dealing with a new/high stress compared to cells with a low stress history (pink box). (B) As a hormetic adaptive response, SGs may influence several cellular processes relevant to tumorigenesis. These processes include proliferation, migration, avoiding cell death, and specialized functions of cells in the tumor microenvironment, as blocking SG formation has been implicated in altering these processes in cancer cells. Mechanistically, SGs can affect these processes through the context specific recruitment and modifications of proteins and mRNA transcripts. SG formation and function is regulated by SG composition (i), as well as by post-translational modifications of the proteins and transcripts involved in the SG network. Post-translational modifications likely occur at the SG site (ii), and SGs have been shown to influence other cellular processes such as RNA stability (iii) and protein trafficking (iv). (C) A hormetic window signifies the amount or duration of a particular stress setting that can provide cellular or organismal benefit (dotted lines). Since SGs can alter cellular state and increase stress tolerance, these organelles are working to expand this hormetic window (blue) from its original boundary (black) and bestow cellular fitness during chronic stress encounters.

Box 1. SG formation.

SGs were initially discovered as cytoplasmic foci in tomato cells undergoing heat shock [30]. Subsequently, they have been observed to be induced by various stress conditions, including oxidative stress, ER stress, hypoxia, UV irradiation, energy depletion, osmotic stress, acidosis, viral and bacterial infections, across a wide range of organisms such as protozoans, fungi, Caenorhabditis elegans, and mammals [11, 31–39] [40]. The primary role of SGs reported to date is to safeguard cells from stress stimuli, and the inhibition of SG formation in cells under stress has been shown to lead to cell death [29]. On a structural level SGs are thought to form through liquid–liquid phase separation (LLPS). LLPS refers to the stable de-mixing of a liquid into two distinct liquid states, driven by multivalent macromolecular interactions and the formation of biochemically distinct condensates [41]. Studies aimed at the characterization of the composition of SGs have revealed that they consist of over 450 proteins and over 11,000 mRNA transcripts [42–44]. About half of the proteins contained within the SG core are RNA binding proteins (RBPs), whereas the others are thought to be pulled in through protein-protein interactions [65]. Recent studies demonstrated that SG formation is mediated by the interactions of RNA and a network of ~36 proteins [45, 46]. These studies have set forth a mechanism whereby SG formation is determined by the capacity of G3BP1 to LLPS with protein-free RNA molecules that are released from polysomes during stress [45]. As such, the increase in valency that is required for phase separation is achieved predominantly through RNA-driven conformational rearrangements of G3BP1, through an interplay between three intrinsically disordered regions (IDR) and dimerization domain, and further tuned by phosphorylation within the IDR. Interactions with additional components of the SG core network, including extrinsic G3BP1-binding factors and regulators, as well as post-translational modifications, can further regulate SG formation in both positive and negative manners [45].

Box 2. SGs and the cellular response to stress.

SGs are transient; they only form in response to cellular stress and dissipate once the stress has subsided [29]. SGs formation and the activation of the Integrated Stress Response (ISR) are tightly linked as ISR is essential for the rise in free RNA [47]. Different types of stress will activate the ISR through different kinases, such as PERK (endoplasmic reticulum stress), HRI (heat shock or osmotic stress), GCN2 (nutrient deprivation), and PKR (viral infection), which converge onto the phosphorylation of eIF2α [48]. This event inhibits Cap-dependent mRNA translation, leading to polysome disassembly and an accumulation of free RNA in the cytoplasm, which acts as a platform for SG formation. Concomitant with inhibition of Cap-dependent translation, ISR activation favors Cap-independent translation of transcripts like ATF4, which go on to activate the expression of stress-adaptive genes [49]. The early stall in translation following eIF2α phosphorylation is an important first step in cell survival under stress, as this begins to conserve amino acids and ER chaperones. ATF4 activity extends this period of conservation by promoting the expression of genes involved in amino acid transport, metabolism, redox balance, and protein homeostasis [48]. While ISR activation can promote SG formation, and both the ISR and SGs are working concomitantly during stress, little is known as to whether this relationship is bidirectional and interactive. The mammalian proteome database and SG transcriptome do not report the association of key ISR effectors like PERK, GCN2, or PKR proteins or transcripts with SGs [42, 43]. However, the transcripts of three major negative regulators of the ISR, namely PPPIR15A/B and TRIB3, have been reported in the SG transcriptome dataset [43]. The proteins encoded by these transcripts function to turn off the ISR and restore normal cellular signaling, through directly de-phosphorylating eIF2α and blocking ATF4 activity, respectively [50]. One potential reason for this localization could be that SGs are protecting the stability of these transcripts during stress, so that when the stress subsides and SGs disassemble, these transcripts can be released, translated, and contribute to the inhibition of any residual ISR signaling. ATF4 can activate the expression of apoptotic proteins, like BCL-2 and NOXA/PUMA, to induce cell death during intense or longer periods of stress. Antagonistically, SGs have been shown to sequester pro-apoptotic proteins. If ATF4 is activating the transcription of apoptotic gene products, but SGs are soaking up these products or other apoptotic proteins that interact with these products, this could help skew cellular signaling towards survival. Interestingly, it has also been reported that phosphorylated eIF2α is recruited solely to disassembling SGs [51]. Perhaps as the stress subsides and SGs begin to disassemble, this potent ISR inducer is relocated and modified at the SG site. These data suggest that SGs may have the potential to regulate ISR signaling, and this effect may be strongest during the disassembly stage.

SGs as a hormetic adaptive response in cancer

Although SGs are necessary for normal cell function, their role in stress conditioning and adaptation has been hijacked in various disease states. These include neurodegenerative disorders, such as Amyotrophic Lateral Sclerosis (ALS) and Alzheimer’s Disease (AD), viral infections, such as Rabies and Herpes, and a variety of cancers, including pancreatic, colorectal, osteosarcoma, and head and neck cancers [25, 52–57]. SGs are a prominent feature of cancer cells under stress and are detected in tissues of pancreatic cancer and osteosarcoma [25, 27]. Many oncogenic proteins including KRAS, Musashi-1, BCR-ABL, mTOR, and DDX3X have been found to impact the formation of SGs, whereas BCR-ABL, DDX3X, and Rbfox2 have also been shown to localize to SGs [25, 26, 58–63]. Additionally, many SG components of the SG core network including G3BP1, YB1, RBFOX2, and HDAC 6 are frequently overexpressed in cancer and as such, support a notion of altered SG dynamics in tumorigenesis [27, 64–66]. To date, the majority of the studies linking SGs to tumorigenesis have relied on genetically or pharmacologically impairing either SG regulators or central nodes in the SG core network, and have implicated SGs in tumor resistance to chemotherapy, growth, and metastasis. To discern whether these outputs are a direct consequence of SGs, or simply a non-SG related function of these proteins, one study utilized a synthetic protein that consisted of domains heterologous to the domains of G3BP1 that determine its SG-formation capabilities. These domains included a dimerization domain, IDR regions, and RNA binding regions [25, 45]. When the synthetic construct was expressed in the background of endogenous G3BP1 knockdown, it successfully rescued both SG formation and pancreatic tumor progression. On the other hand, the expression of a truncated G3BP1 construct which lacks the N-terminal dimerization domain and is deficient in SG formation failed to achieve the same effect [25]. As such, the recent advances in the understanding of the structural determinants of SG formation have provided invaluable tools and knowledge about the function of SGs and have paved the way for similar approaches in gaining a causal understanding of the role of SGs in cancer.

Multiple lines of evidence support the notion that SGs can function as a hormetic adaptive response in cancer. For instance, Arimoto et al demonstrated that SGs helped promote cancer cell survival in hypoxic conditions by sequestering the pro-apoptotic protein RACK1 and suppressing apoptotic signaling [31]. Notably, this resulted in a resistance to the genotoxic agent etoposide, as the level of etoposide-induced cell death was lower in hypoxic cancer cells than in normoxic ones. Mechanistically, this priming for survival was a result of the inhibition of p38 that occurred due to RACK1 sequestration in SGs in the previous stress event [31]. These observations show that stress conditioning, or an enhanced ability to mitigate future stress encounters, is mediated through SGs, which is consistent with a hormetic adaptive response.

Cancer progression is accompanied by increasing levels of stress in type, frequency, and intensity (Figure 1). As a hormetic adaptive response, SG formation would be expected to increase in tumor vs. normal tissue. In fact, cancer cells have a several-fold higher capability of forming SGs compared to normal cells [26]. This capability also appears to be specific to cancer type and mutational status. Highly aggressive tumors driven by oncogenes that induce cellular stress, such as mutant KRAS-driven pancreatic cancers, have a higher SG formation capability compared to tumors lacking this mutation [26]. Additionally, suppression of SGs impairs the growth of KRAS mutant pancreatic tumors, suggesting that SGs facilitate the maintenance of pancreatic cancer cells in the ‘hormetic window’ that supports survival and growth (Figure 2) [25]. In support of this notion, higher SG formation capability has also been linked to resistance to chemotherapy, which often relies on cell stress for its cytotoxic activity [26]. The role of SGs in tumors that arise in pre-existing stress conditions provides another line of evidence. Epidemiological evidence and studies in animal models have clearly demonstrated that conditions where tissues are repeatedly exposed to stress such as obesity, inflammation, alcohol, smoking, and exposure to environmental toxins can increase cancer occurrence and accelerate progression [67]. One possible explanation for this association is that such stress conditions may operate along the same axis as oncogenic-induced stress, and fuel growth and survival by increasing hormetic adaptive strategies. In fact, pancreatic tumors that arise in the setting of obesity have markedly higher (5–8-fold) levels of SGs compared to those in standard weight animal models and patients [25]. The upregulation of SGs in obesity-associated pancreatic cancer is specifically regulated by the IGF1-S6K1-SRPK2 signaling axis. Inhibition of this pathway by pharmacological inhibition of S6K1 reduced SGs to their levels in standard weight tumors but had drastic effects in tumor growth. The duality of hormesis (low dose favorable/high dose toxic) may be a potential explanation for this effect, whereby partial/incomplete inhibition of SGs is sufficient to push cancer cells out of the hormetic window of stress that is favorable (Figure 2). Supporting a role for SGs in maintaining this favorable hormetic window, extensive inhibition (~90%) of SGs in pancreatic tumors in standard weight and obese mouse models, had a far more drastic effect in the obese setting; an approximately doubling of survival (90 days) was observed in standard weight mice whereas in the obese setting ~40% of the mice were tumor free even at 300 days [25].

Is the role of SGs in tumorigenesis limited to cancer cell survival? In general terms, hormesis involves the simultaneous stimulation of various pathways and biological processes. Thus, while the key end-product of hormesis is survival, the effects of hormetic responses may extend to encompass biological processes beyond simply cell survival. A prime example of this is mitohormesis, a term used to describe hormesis elicited by mitochondrial reactive oxygen species (ROS) [68]. In mitohormesis, mitochondrial ROS activate mitoprotective mechanisms which persist even after the stress subsides. In non-pathological settings this is associated with health and longevity, but cancer cells can utilize mitohormesis to promote tumorigenesis and pro-metastatic programs [68]. The potential role of SGs in cellular functions beyond cell survival has yet to be elucidated. The challenge of purifying SGs due to their non-membranous property has been a major limitation in this direction, as a complete understanding of their composition is lacking. Nonetheless, biolabeling and crosslinking approaches have revealed a broad network of protein and RNA constituents [43, 44, 69–71]. As expected, SGs consist of many stress-sensing molecules, but also contain a large portion of proteins and transcripts that are unrelated to the stress response. Thus, the impact of SG formation may extend beyond cell survival under stress. Below, we look deeper into how SGs are interwoven with key themes of cancer evolution, such as proliferation, protein and gene regulation, and interactions with the tumor microenvironment.

SGs in Proliferation and Senescence

Cancer cells are highly proliferative cells that are defined by their dysregulation of cell cycle signaling. However, cancer cells can also enter into senescence, which is a non-proliferative stage that is largely unresponsive to growth or death signals [72]. This often poses a problem for cancer reoccurrence following therapy, as these cells can remain viable under treatment, and later exit senescence to return to a proliferative state [72]. In the context of proliferative cells, it has been shown that blocking SG formation inhibits cancer cell proliferation [25]. In turn, the proliferative state of the cell may also dictate the composition and function of SGs. For example, the Transport Protein Particle (TRAPP) complex, which is essential to the anti-apoptotic function of SGs, localizes to these organelles only in cycling cells [73]. This accumulation of the TRAPP complex in SGs was shown to be dependent on CDK1/2, which are often hyperactivated in cancer [73]. These data suggests that there may be cell state-dependent specificity in SG composition, which could lead to unique approaches to target SGs across the proliferative state spectrum. Interestingly, mitotic cells do not form SGs, suggesting that this cell state may be more prone to cell death, and may not become “primed” by a stress encounter in the way that other cell types would [74]. In terms of senescent cells, it has been shown that SGs sequester a senescent-promoting factor, PAI-1, potentially influencing the transition from a proliferative state to a senescent state [75]. Another study revealed that the loss of SG formation through knockdown of G3BP1 blocks the senescent-associated secretory phenotype (SASP) in primary human lung fibroblasts [76]. This loss of G3BP1 did not impact commitment to senescence, but rather impacted the RNA levels of common SASP transcripts, including matrix metalloproteases and promoters of inflammation, after senescence induction [76]. This effect on the SASP is important given its role in promoting cell migration, invasion, and pro-inflammatory signaling in cancer [77]. In line with this, lung adenocarcinoma cancer cells co-cultured with senescent fibroblasts deficient in G3BP1 showed reduced migratory capacity as compared to those co-cultured with G3BP1-expressing senescent fibroblasts. This effect translated in vivo, as co-implantation of cancer cells with senescent fibroblasts deficient in G3BP1 gave rise to smaller tumors [76]. Taken together, these studies indicate a role for SGs across the proliferative spectrum, however the specific functional outputs may differ with cell state (Figure 2).

SGs in Gene Expression, Protein Modification, and Protein Trafficking

Several studies have shown that SGs can elicit broad changes in cellular signaling by modifying the stability of transcripts or by altering protein function through sequestration (Figure 2). For example, SGs have been shown to enhance the mRNA stability of the prominent tumor promoter MYC in sarcoma cells, which could imply a novel route of enhancement of MYC activity [78]. SGs have been shown to recruit the mRNA of multiple transcription factors, including NFKB, STAT3, and HIF1-alpha, which have been implicated in cancer cell signaling [43]. Alterations in the stability of these transcripts could lead to changes in gene expression and regulation. Interestingly, G3BP1 has been shown to impact the stability of mRNAs through binding to RNA guanine quadruplexes (rG4) [79]. rG4s are located primarily in the 3’ and 5’ UTR regions and play important roles in mRNA translation and degradation. G3BP1 binding in these regions was shown to stabilize these mRNA transcripts [79]. It would be interesting to see how G3BP1 impacts mRNA stability in stressed and non-stressed conditions, as mRNAs containing rG4 have been shown to accumulate in SGs [80]. Additionally, analysis of the SG transcriptome has revealed that the mRNA encoding multiple DNA Damage Response proteins, including major players like RAD50, ATM, and ATR, are recruited to SGs [43]. The sequestration of such transcripts could regulate the apoptotic response to DNA damage and impact DNA quality control. SGs have also been implicated in RNA degradation via crosstalk with other organelles in the cytoplasm [81]. One study suggests that SGs and processing bodies physically interact, and that mRNA is remodeled within SGs, sorted, and specific transcripts are then delivered to processing bodies for degradation [81]. The particular transcripts that are shuffled between SGs and processing bodies have not been identified, but one study that compared the processing body transcriptome between stressed and non-stressed conditions reported an elevated content in certain transcripts, such as AHNAK and DYNC1H1, which were also identified in the SG transcriptome [43, 82]. Interestingly, AHNAK has been shown to suppress tumor proliferation in triple-negative breast cancer [83]. Based upon this data, if SGs contribute to AHNAK transcript degradation through association with processing bodies, an increased formation of SGs could enhance cancer cell fitness under stress.

Several excellent reviews have detailed the recruitment of individual proteins to SGs and how it can impact protein function [29, 71, 84]. One potential mechanism of altered protein function would be post-translational modifications that occur at the SG site. SG formation, maintenance, and disassembly programs greatly rely on the post-translational modification of various SG core network proteins [29]. Acetylation of G3BP1 and Pab1 have been shown to inhibit their SG forming capabilities, each through a decrease in RNA binding potential [85, 86]. Similarly, the methylation of G3BP1 and UBAP2L in their RGG domains inhibits SG formation, but in this case by blocking their association with various RBPs and ribosomal subunits. Various enzymes involved in conferring O-linked N-acetylglucosamine modifications have proven critical in promoting SG formation, and the proteins that these enzymes modify are shown to localize within the SG [87]. These proteins include multiple ribosomal proteins, as well as the apoptotic factor RACK-1, which serves as a primary route of anti-apoptotic activity conferred by granule sequestration. Interestingly, a majority of ubiquitination machinery is recruited to the SG site, and ubiquitination has been shown to play a role in SG disassembly. For example, G3BP1 becomes polyubiquitinated following heat stress, is subsequently extracted from SGs, and SG disassembly promptly follows [88]. ADP-ribosylation is another protein modification that has been documented in both SG formation and disassembly [89]. Several PARPs are localized to SGs, and their overexpression has been shown to induce SG assembly. PARP1 and PARP5 in particular have been shown to bind SG proteins like G3BP1, FUS, and Ago2, with this binding initiating LLPS or SG formation [89]. Interestingly, since PAR chain length can impact protein binding partners, and since PAR levels increase upon stress induction, it is plausible that PARylation can enhance protein-protein interactions that support phase separation, and that the PAR chain itself may impact the composition of condensates through these binding partners [89]. One important area of future investigation is whether post-translational modifications are occurring at the SG site, as various protein-modifying enzymes are recruited to SGs. Identifying the particular modifications that may occur to proteins or transcripts in SGs is an interesting area of investigation.

In addition to potential protein or transcript modifications, SGs have also been shown to directly affect protein transport processes (Figure 2). For example, the COPII coat protein which is essential to the formation of anterograde transport vesicles and cargo selection has been shown to be recruited to SGs [73]. Sequestration of COPII to SGs was shown to disrupt anterograde transport. While the interplay between SGs and anterograde transport remains to be elucidated, it may represent a unique platform through which SGs could impact several cellular processes. Overall, these findings suggest that the capacity of SGs to recruit various types of transcripts and proteins, thus altering their stability and activity, can lead to widespread changes in cell signaling and gene expression under stress.

SGs in the Tumor Microenvironment

SG formation has been shown to be both a cell autonomous and cell non-autonomous process, as the expression of mutant KRAS can lead to enhanced SG formation intracellularly and in surrounding wildtype cells [26]. This non-autonomous mechanism is driven by the secretion of a SG-promoting molecule, namely the inflammatory prostaglandin 15-Deoxy-Delta-12,14-prostaglandin J2 (15-dPGJ2) [26]. Enhancement of SG formation in surrounding wildtype cells furthered their hormetic window for stress tolerance and enhanced stress resistance [26]. Therefore, cancer cells can impact SG formation and stress adaptation within the microenvironment. Interestingly, SG formation has been shown to influence immune cell function in the tumor microenvironment. One study utilized a truncated form of TIA-1 to block SG formation in T cells and investigated how SG deficiency impacted T cell activation based on their cytokine profiles. When T cells expressed this mutant form of TIA-1, there was an increase in IL-4 release during T cell priming, suggesting that SGs may impact the cytokine profile of T cells during activation [90]. In support of this, another study showed that T cell stimulation results in the expression of SG markers and SG formation, suggesting a role for SGs during immune cell activation [91]. SGs may also be involved in the expression of the immune checkpoint protein Programmed Cell Death-1 (PD-1) expression in T cells by recruiting and stabilizing PD-1 mRNA [92]. Altogether these studies indicate that the capacity of cells in the tumor microenvironment to form SGs may impact their specialized functions and contribution to tumor progression (Figure 2).

SGs in Cancer Therapy and Resistance

DNA damaging chemotherapies, such as doxorubicin, etoposide, cisplatin, and oxaliplatin have been shown to promote SG formation through phosphorylation of eIF2a [93]. The microtubule-stabilizing drug paclitaxel has also been shown to promote SG formation, mainly through mechanisms of microtubule assembly [93]. However, many of these studies have been conducted with drug concentrations outside of a therapeutically relevant scope, ranging from 50μM-600μM [93]. Whether or not these therapies can induce SGs at physiological doses remains to be addressed. Among studies that have observed SG formation in response to anti-cancer inhibitors within the clinical range, the proteosome inhibitor Bortezomib, which is an effective treatment for myelomas, was shown to induce SG formation at 1μM (3hr), mainly through its activation of heme-regulated inhibitor kinase (HRI) and phosphorylation of eIF2α. This concentration is approximately 3 times the maximum used clinically but is not largely outside of therapeutic range [94]. Fournier et al. investigated whether SGs are a source of Bortezomib resistance in cancer cells and found that depletion of HRI through siRNA inhibited SG formation, reduced cancer cell growth, and increased apoptosis in HeLa cells that were exposed to Bortezomib. This data suggests that SGs may play a role in cancer cell resistance to Bortezomib, and that blocking SG formation through HRI inhibition may sensitize cancer cells to this proteosome inhibitor [95].

Treatment with the tyrosine kinase inhibitor Sorafenib has also been shown to promote SG formation, mainly through the activation of PERK, which is another avenue of eIF2α phosphorylation. Adjibade et al. treated multiple cancer cell lines with a Sorafenib concentration ranging from 5–25μM, which falls in line with the maximum clinical concentration of 20 μM [96]. Chemical inhibition or siRNA knockdown of PERK was able to reduce p-eIF2α levels and inhibit SG formation following Sorafenib treatment in Sorafenib-resistant HCC [96]. These studies point to specific regulators that are activated in response to therapy-induced stress and identify the potential for these regulators to induce SG formation. SGs are likely contributing resistance mechanisms in the presence of these therapeutic inhibitors, and further studies directly linking SGs to such mechanisms of resistance may aid in identifying appropriate combinatorial therapies to restore drug sensitivity in non-responsive tumors.

Concluding remarks

Several lines of evidence indicate that SGs may function in a hormetic manner in tumorigenesis by pre-conditioning cancer cells to future stress stimuli and enhancing their stress adaptive responses. Evidence suggests that in addition to their cytoprotective function, SGs may interact with multiple processes involved in tumor cell signaling and evolution, including the cell cycle, protein signaling, DNA damage response, microenvironment crosstalk, and immune cell function. The identification of some of the major structural determinates of SG formation will allow for more causative studies to be performed, linking SGs to these cellular processes. Overall, studies suggest that both SG levels and composition likely regulate their signaling output and function. Further investigation is necessary to identify any specific changes in SG composition and function during tumorigenesis, as well as decipher how SGs function to retain a memory of previous stress events to promote future stress tolerance (see Outstanding questions). Such studies may provide important insight into the regulation and function of SGs in cancer and identify key therapeutic targets for cancer treatment.

Outstanding Questions.

Is the role of SGs in tumorigenesis limited to stress resistance or do they impact additional aspects such as proliferation, metastasis, and tumor microenvironment regulation. What is the mechanistic basis behind the role of SGs in tumorigenesis?

How does SG composition change during cellular transformation and across cancer progression? Can SG composition be used to assess a particular stress history of a tumor, to pinpoint cancer cell dependencies?

How long does a pre-conditioned, adapted cellular state last after SG formation? How do SGs impact other programs, such as epigenetics and gene expression, to coordinate a memory of previous stress encounters?

How does SG composition impact SG function, and do distinct stress stimuli or proliferative states influence SG composition?

Do protein modifications occur at the SG site? How are particular proteins and transcripts modified within the SG, thus changing their activity pre- and post-exit from SGs?

Highlights.

Tumor growth relies on an increased capacity for cellular stress adaptation due to the stress stimuli that accompany tumor progression. Recent evidence indicates that stress granules (SGs) are upregulated in response to several tumor-associated stress stimuli, determine the cellular capacity to cope with stress, and are critical to tumorigenesis. Nonetheless, how SGs promote tumorigenesis is yet to be understood.

Advances in the understanding of the composition of SGs have furthered our knowledge of how SGs impinge on the apoptotic machinery to enhance stress resistance. In addition, these studies have identified a large proteome and transcriptome that are constituents of SGs, suggesting that they may function beyond blocking stress-induced cell death. These data, combined with the elucidation of the molecular and structural determinants of SGs formation, set the stage for deciphering the mechanisms and processes through which SGs promote tumorigenesis.

We propose a framework whereby SGs function as a hormetic adaptive response in cancer. As such, mild repeated tumor-associated cellular stressors act as hormetic stimuli that tune the formation of SGs to regulate several cellular processes, in a way that culminates as beneficial to the survival of cancer cells.