The different toxicities of Cr(III) and Cr(VI) compounds can be rationalized in terms of their physico-chemical properties. In particular, their ability to cross biological membranes and, ultimately, induce intracellular damage is determined by their sizes, structures and charges. Cr(VI) crosses the cell membrane as a tetrahedral CrO42– anion, due to its structural analogy to the sulfate and phosphate anions, whereas the bulkier Cr(III) ions do not.
2. Cancer Initiation, Promotion and Progression: The Critical Importance of Oxidative, Proteotoxic and Genotoxic Stresses
Carcinogenesis is driven by stepwise genetic mutations and concomitantly enhanced and uncontrolled cell proliferation. DNA damage is believed to be in the genesis of this process by creating a transformed cell, which over the course of additional genomic and cellular insults becomes a fully malignant and metastatic cancer cell
[10]. Traditionally, carcinogenesis has been divided into three phases: initiation, promotion and progression
[11] (
Figure 1). Initiation entails the acquisition of mutations in proto-oncogenes and tumor suppressor genes. Significantly, incipient cancer cells feature a deranged metabolism, leading to high levels of reactive oxygen species (ROS) and, consequently, oxidative stress
[12][13][14][15] (
Figure 1). ROS damage not only DNA, but also proteins and membrane lipids. Yet, ROS also play a role in cellular signaling, promoting cell proliferation and adaptation to the hypoxic conditions often found in the tumor microenvironment
[16]. In particular, mitochondrial ROS inactivate inhibitory phosphatases (e.g., PTEN), unleashing the PI3K/AKT cell survival and growth pathway, and prolyl hydroxylases (e.g., PHD2). In turn, this inactivation stabilizes hypoxia inducible factors (HIF), concomitantly triggering angiogenesis. Next, cancer cells enter a promotion phase, where mutations in oncoproteins such as growth factor receptors and kinases gradually lead to independence from extracellular growth factors
[17]. As mutations often disrupt a protein’s ability to fold
[18], accumulation of increasingly larger amounts of mutated proteins represents yet another type of cell intrinsic stress—proteotoxic stress
[19] (
Figure 1). This type of stress can be created by any structural alteration that may lead to protein misfolding and aggregation. Ultimately, incipient cancer cells form a solid tumor mass, creating a tumor microenvironment (TME). Here, cancer cells reprogram stromal cells to produce tumorigenic cytokines, chemokines and tissue-remodeling metalloproteinases
[20], inhibit anticancer immune responses
[21] and recruit blood vessels via angiogenesis to sustain their continued growth
[22]. The TME also creates a host of cell extrinsic stressors, including hypoxia, acidosis and nutrient deprivation
[23][24][25][26]. Malignant tumors are also characterized by rampant chromosomal instability and aneuploidy, caused by chromosome segregation errors during mitosis. Such extensive damage leads to genotoxic stress. While genotoxic stress leads to p53-induced apoptosis in normal cells, in malignant cells it is tolerated and subverted, giving rise to a mosaic of genomic mutations and karyotypic abnormalities in solid tumors
[27][28][29].
Figure 1. The different types of stress associated with the three stages of carcinogenesis. Carcinogenesis has been traditionally divided in three stages: initiation, promotion and progression. Different types of cellular stress have been implicated in these stages. Reactive oxygen species (ROS) damage proteins, lipids and DNA, inducing mutations in the latter. Incipient cancer cells at the promotion stage harbor an increasing number of DNA mutations, resulting in dramatically higher levels of mutant proteins and, ultimately, proteotoxic stress. Transition to a fully malignant phenotype, i.e., progression, is thought to require chromosomal instability and resulting karyotypic abnormalities, causing genotoxic stress. Of note, all types of stress indicated (oxidative, proteotoxic and genotoxic) play roles in all three stages of carcinogenesis described above; their relative importance likely differs amongst different types of cancer.
3. Links between the Cellular Response to Stress and Carcinogenesis
Carcinogenesis entails the acquisition of a growing ability to survive in the face of cellular stress levels that normal cells are unable to withstand. There is now a growing perception that this ability results, at least in part, from a subversion of the cellular systems that evolved to protect normal cells against stress. One such system is the so-called cellular stress response, a homeostatic system to combat proteotoxic stress that is found across all three domains of life.
3.1. Note on Nomenclature
Several of the studies discussed in this research were carried out at a time when very little was known regarding heat shock proteins (HSP) and their role in the cellular response to stress. Back then, HSP were named based on their approximate subunit molecular weights, as determined by polyacrylamide gel electrophoresis. For instance, the designations Hsp90 and HSP90 were used interchangeably to describe any protein with an approximate subunit molecular weight of 90 kDa whose expression was rapidly and strongly induced by stress. Since then, the number of known stress-responsive proteins, some of which constitutively expressed, has expanded enormously. Many of the now known isoforms share identical subunit molecular weights and it is often not possible to retrospectively identify the specific isoforms being described in the earlier studies.
To complicate matters further, some stress-responsive proteins were not initially classified as HSP and were given unrelated names. Currently, up to ten different names can be found in the literature for the same gene product
[30]. In 2009, aiming at reducing inconsistencies and increase clarity, Kampinga and collaborators put forward new guidelines for the nomenclature of the human HSP
[30]. Unfortunately, this nomenclature has not yet been widely adopted, remaining unfamiliar to most readers.
In this Encyclopedia entry, the abbreviation Hsp is used when referring to a clearly identified isoform (e.g., Hsp72), whilst HSP abbreviates either one or more unspecified isoforms of a given family or the family as a whole (e.g., HSP90 is used to describe an unidentified isoform of an approximate subunit molecular weight of 90 kDa or the HSP90 family as a whole).
3.2. The Cellular Stress Response: Basic Concepts
The cytoprotective effects of the cellular stress response are mediated by HSP. These molecular chaperones promote proper protein folding, translocation and degradation, as well as the assembly and disassembly of protein complexes
[31][32]. In mammals, heat shock factor 1 (HSF1) is the main transcriptional regulator of the cellular stress response
[33][34].
In eukaryotic cells, the cellular stress response comprises different sub-systems, which fulfil organelle-specific functions, such as the unfolded protein response (UPR), which operates in the endoplasmic reticulum (ER)
[35], and the mitochondrial unfolded protein response (UPRmt). The ER is a major site for the synthesis, folding, modification and transport of secretory and transmembrane proteins, as well as for the assembly of protein complexes
[36][37]. Incorrect protein maturation can occur even under physiological conditions, due to, among other causes, the very high protein concentrations normally found in the ER (~100 mg/mL
[38][39]). ER stress, i.e., the incapacity of this organelle to manage its load of client proteins, is further aggravated under conditions often found in the TME, such as nutrient deprivation, hypoxia, augmented ROS levels and low pH
[40]. Certain cancers, such as B-cell-derived multiple myeloma, produce extremely high levels of immunoglobulins, which translates into protein overload and further ER stress
[41].
Accumulation of unfolded or misfolded proteins triggers the UPR, which signals transient attenuation of protein translation, while increasing the ER capacity of protein folding and degradation of misfolded proteins
[38][39][42]. Amongst the molecular chaperones involved in the re-establishment of protein homeostasis (i.e., proteostasis) are numerous glucose-regulated proteins (induced by glucose starvation), including Grp78, which is the most abundant ER-resident chaperone, and Grp94
[38][39][42][43][44]. Grp78 and Grp94 are the ER members of the HSP70 and HSP90 families, respectively. After a certain time, proteins that remain aggregated, misfolded and/or unassembled are targeted for ER-associated degradation (ERAD), leading to their translocation from the ER to the cytosol to be degraded by the ubiquitin-proteasome machinery
[45]. If ER stress becomes chronic, abnormal calcium signaling from ER to mitochondria and apoptotic pathways can be activated
[46].
In eukaryotes, the metabolic energy required to sustain cellular processes, including stress-induced adaptations, is generated mostly in the mitochondria. Interestingly, mitochondria are closely connected to the ER through mitochondria-associated membranes (MAMs), which allow the exchange between these two organelles of lipids, calcium ions (Ca
2+) and, possibly, ROS. It has also been suggested that MAMs are involved in glucose homeostasis
[47]. ER and mitochondrial stress pathways seem to be interconnected, as a mitochondria-resident HSP90, tumor necrosis factor receptor-associated protein 1 (TRAP1), has been associated with UPR in the ER
[48][49]. In addition, p53-upregulated PUMA and NOXA
[50], as well as Lon protease
[51], which is also a chaperone
[52], seem to be part of a signaling pathway that transmits ER dysfunction to the mitochondria. ER stress, amino acid depletion, excessive ROS levels, oxidative phosphorylation (OXPHOS) perturbation, impaired complex assembly (mitonuclear protein imbalance) and the accumulation of misfolded proteins all impair mitochondrial protein import efficiency and lead to nuclear translocation of the activating transcription factor associated with stress (ATF) and subsequent activation of the UPRmt
[53][54][55]. In the nucleus, ATF mediates the transcription of genes involved in the re-establishment of mitochondrial function, mitochondrial proteostasis and protein import efficiency
[56][57]. Resistance to ER and mitochondrial stresses can contribute to carcinogenesis
[58][59].
3.3. Cancer and the Cellular Stress Response
Most types of tumors display augmented HSP levels
[60]. Increased HSF1 activity likely contributes to the augmented HSP levels, yet it has been reported that HSP gene promoters can also be activated by the oncogenic transcription factor c-MYC, as well as by loss of the tumor suppressor protein p53
[61]. Strikingly, deletion of HSF1 in mice bearing mutations in the
Ras oncogene and
Tp53 tumor suppressor gene protected them from tumor formation
[62].
Specific HSP have been directly implicated in p53 inactivation and malignant transformation
[63], as well as in cancer invasiveness and resistance to chemotherapy
[64]. For instance, HSP90 overexpression, which was observed in a broad spectrum of cancers, correlated with tumor growth, metastatic potential and resistance to chemotherapy
[60][65][66]. This observation led to the proposal that tumors develop a dependency on HSP90
[67][68]. It is noteworthy that, unlike other HSP, HSP90 proteins are not necessary for the correct folding of newly synthesized proteins. Instead, their main role is to stabilize meta-stable proteins, ultimately suppressing the formation of protein aggregates. Importantly, numerous oncoproteins are HSP90 clients
[69]. Chief among these are several receptor tyrosine kinases and steroid hormone receptors, such as the human epidermal growth factor 2 (HER2), associated with uncontrolled cellular proliferation
[66][70], telomerase, an enzyme required for immortalization
[71], AKT, involved in apoptosis deregulation
[72], hypoxia-inducible factor 1-alpha (HIF-1α), essential for angiogenesis
[73] and the metabolic shift observed in tumors
[12][66][74], and matrix metalloproteinases (MMPs), crucial for successful tissue invasion and metastasis
[75]. According to the "HSP90 addiction hypothesis", cancer cells depend on an increased pool of HSP90 to retrieve essential proteins that became misfolded due to extensive proteotoxic stress and to allow increasingly mutated oncoproteins and tumor suppressor proteins to function, by preventing their misfolding and degradation.
Remarkably, HSP90 proteins have also been found in the extracellular milieu, where they act as potent stimulators of immune responses
[76]. Unsurprisingly, HSP90 is currently being explored as a target for cancer therapy. There are currently over 70 clinical trials employing HSP90 inhibitors registered in ClinicalTrials.gov, but none has been approved for cancer treatment yet
[77].
Altogether, the cellular stress response emerges as a double-edged sword: evolved to protect cells from menaces to homeostasis, it might constitute, in its extreme, one of the main mechanisms behind cancer cells’ formidable resilience.
4. The Molecular Mechanisms of Hexavalent Chromium Carcinogenicity: A Brief State of the Art
Genetic mechanisms likely play a critical role Cr(VI) carcinogenesis, as supported by the observation of genetic lesions in both the lung cells of chromate workers and in cultured cells exposed to different Cr(VI) concentrations [78][79][80][81][82][83][84][85][86][87]. Thus, the initial observation, in test tube experiments, that Cr(VI) is mostly unreactive towards DNA (and most other biomolecules) puzzled researchers. However, it is now known that, following its rapid cellular uptake, Cr(VI) undergoes a multi-step reduction that generates a variety of species that react extensively with biomolecules, namely Cr(III), which is the final reduction species, and the unstable intermediates Cr(IV) and Cr(V) [88][89]. Under physiological conditions, ascorbate accounts for about 90% of Cr(VI) reduction, but non-protein thiols, such as glutathione and cysteine, also contribute significantly to its reduction [90]. Thus, Cr(VI) reduction generates additional reactive species, such as carbon-based radicals from ascorbate, thiyl radicals from glutathione and cysteine and, possibly, ROS. Among the Cr(III)-DNA complexes formed are Cr(III)-DNA adducts, DNA-protein crosslinks and DNA interstrand crosslinks [78][91].
By restraining the normal DNA replication and transcription processes, these adducts activate the various cellular DNA repair systems in a lesion-dependent manner. Mutations in key proteins involved in these DNA repair systems have been described both in Cr(VI)-induced lung cancer patients and in cultured cells exposed to Cr(VI) compounds, impairing their ability to remove chromium-DNA adducts [92]. Importantly, double-strand break (DSB) induction by the mismatch repair (MMR) system may drive genomic instability [3][93]. Hirose and co-workers reported a high incidence of microsatellite instability (MSI), a particular type of genomic instability that specifically affects the microsatellites, in lung cancers from chromate-exposed workers [3][94][95]. However, MSI was not be observed upon in vitro exposure of human lung epithelial cells to Cr(VI) [96].
Cr(VI) exposure can also generate DNA damage by indirect mechanisms. For instance, Cr(VI) exposure may lead to loss of thiol redox control through interference with antioxidant defense systems [97]. This and additional lines of evidence, namely the observation of 8-hydroxy-2’-deoxyguanosine formation in rat lungs following intratracheal administration of Cr(VI) [86], suggest that Cr(VI) exposure can damage DNA through the generation of oxidative stress [98][99][100]. Additionally, altered ROS levels affect gene expression [101].
5. The Impact of Hexavalent Chromium on the Cellular Stress Response
Cr(VI) may promote proteotoxic stress and, ultimately, activate the cellular stress response through various mechanisms. For instance, changes in protein conformation may result from their direct interaction with Cr(III). Conformational changes may also be a consequence of oxidative stress, as it may originate incorrect disulfide bonds and other forms of protein modification
[97]. Gene mutations induced by Cr(VI), as observed in vivo and in vitro
[102], can also compromise correct folding of the affected proteins
[18].
The number of published studies on the impact of Cr(VI) on the cellular stress response is still small. In addition, most of these studies did not specifically address the role of this response on carcinogenesis. Namely, some of the earlier studies were exploiting the then recent gene array technology to identify gene pathways that might be affected by Cr(VI) exposure
[103][104].
The first observation of an effect of Cr(VI) on the cellular stress response was made in 1998, on a molecular toxicology study aimed at developing a sensitive biological system for the rapid detection of low levels of environmental pollutants
[105]. Using a radiolabeled antisense RNA probe, the authors found that, at mildly cytotoxic concentrations, a 6 h exposure to Cr(VI) increased Hsp72 mRNA levels in HepG2 and HT29 cells. These results confirmed that HSP activation is exquisitely sensitive to Cr(VI) exposure, as changes in Hsp72 transcript levels could be detected for Cr(VI) concentrations as low as 0.5 μM. Of note, mRNA levels were determined 3 h after the stressing exposure, as it was observed that, after heat shock, transcript levels strongly increased in the first 3 h, then decreasing to nearly basal levels 6 h after shock. In an independent study, protein levels peaked instead at 6 h after exposure
[106], stressing the importance of conducting adequate time courses.
The second report of Cr(VI) impacting the cellular stress response came from a study aimed at identifying metal-responsive promoters and, ultimately, new signal transduction pathways that might be modulated by exposure to this and other environmental pollutants
[98]. To this end, 13 recombinant HepG2 cell lines, each of which stably transfected with a specific stress-responsive promoter regulating the expression of the chloramphenicol acetyl transferase (CAT) reporter gene, was exposed, for 48 h, to different Cr(VI) concentrations. Intracellular levels of CAT protein were determined immediately after exposure. A subcytotoxic Cr(VI) concentration induced CAT upregulation in the cell lines transfected with the HSP70 and Gpr78 promoters, but statistical significance was only reached for HSP70. At a higher Cr(VI) concentration, CAT protein levels were further augmented, yet this was accompanied by a dramatic decrease in cell viability. The results of this study highlighted the different susceptibilities of these two HSP to Cr(VI).
Cr(VI) is a lung carcinogen and, as such, studies conducted on its main targets (
i.e., human epithelial lung cells) should be particularly informative. In the A549 cell line, established from a human lung adenocarcinoma, a 2 h Cr(VI) exposure upregulated the transcript levels of TRAP1, the mitochondrial homologue of Hsp90
[104]. However, the Cr(VI) concentration used in this study was extremely high (600 µM) and would likely cause massive cell death for longer exposures. Therefore, the results of this study must be interpreted with caution. Nonetheless, it was recently reported, also in the A549 cell line, that a much lower Cr(VI) concentration (0.5 µM) upregulated Grp78 protein levels, again peaking at 6 h of Cr(VI) exposure
[107]. The exquisite sensitivity of Grp78 to Cr(VI) is noteworthy. In L-02 hepatocytes, Grp78 mRNA levels were increased after a 24 h exposure to Cr(VI) in the low micromolar range
[108]. In the same cell line, a similar exposure regimen, which was found to induce significant cytotoxicity, decreased the protein levels of both HSP70 and HSP90
[109].
Two studies have been conducted in the BEAS-2B cell line, established from normal human bronchial epithelium. Both studies used Cr(VI) concentrations that did not cause overt cytotoxicity. The first study aimed at identifying specific and sensitive biomarkers of toxic metal exposure [110]. One significant finding was the extreme specificity of the Cr(VI) effects: of the 1200 gene transcripts analyzed, only 44 had their expression altered after a 4 h Cr(VI) exposure. Of the 44 genes affected, 3 encoded HSP (HSP40, HSP60 and HSP90A) and were all down-regulated. The transcript levels of all other HSP analyzed (HSP27, HSP-70, HSP70.1, HSP-71) remained unchanged, giving further support to the perception that the impact of Cr(VI) is isoform-specific.
The second study employing BEAS-2B cells investigated the impact of Cr(VI) on the expression of the Hsp72 and Hsp90α isoforms at both the transcript and protein levels [111]. Importantly, this study unveiled decoupling of mRNA and protein levels for both Hsp72 and Hsp90α. After a 48 h incubation with Cr(VI), Hsp72 mRNA levels were decreased, whereas Hsp72 protein levels remained unchanged. For Hsp90α, mRNA levels were unaltered, whereas protein levels were decreased. This decoupling is likely multifactorial, potentially involving critical post-transcriptional regulators, such as RNA binding proteins and microRNAs [112][113]. Protein stability and turnover may also have to be considered [114][115]. Thus, in future studies, it will be important to conduct detailed time-courses of the effects of Cr(VI) on gene expression at both levels.
There are two other published cellular studies on the impact of Cr(VI) on HSP70 [116][117] and one on the impact of this carcinogen on HSP90 [117]. Altogether, these studies clearly show that this impact is dependent on both the cellular model employed and on the experimental design.
Another study, conducted in rat lung epithelial cells, showed the impact of Cr(VI) on additional HSP isoforms, namely Hsp10 and Hsp105, whose protein levels were increased after a 24 h incubation, which was shown to produce significant cytotoxicity [118]. Two other studies, one employing HaCaT [119] cells and the other employing human primary skin fibroblasts [120], unveiled the ability of Cr(VI) to alter the phosphorylation state of HSP27. Of note, aberrant phosphorylation of HSP27 has been associated with cancer [121]. In HaCaT cells, HSP27 expression was upregulated by Cr(VI) at both transcript and protein levels, but the phosphorylation of this HSP was decreased [119]. On the contrary, levels of phosphorylated HSP27 were found to be increased in Cr(VI)-exposed in human primary skin fibroblasts [120]. This apparent contradiction might be explained by differences in cell model, Cr(VI) concentration and duration of exposure.
In the only two in vivo studies conducted to date, one employing Institute of Cancer Research (ICR) mice [122] and the other one Sprague-Dawley rats [103], Cr(VI) administration induced HSP expression. In ICR mice, Cr(VI) intraperitoneal injection increased liver HSP27 and HSP70 protein levels. In Sprague-Dawley rats, Cr(VI) intratracheal instillation increased HSP70 mRNA levels in the lungs, but not in the liver. HSP60, Grp75 and Grp94 mRNA levels, on the other hand, were unaffected in both lung and liver. In fact, none of the 216 genes assessed had their liver mRNA levels altered, whereas changes in lung mRNA levels were observed for 52 genes. The observed lack of effect in the liver was ascribed to the upstream reduction and consequent detoxification of Cr(VI), firstly in the lung, then in the blood of the general circulation and finally in the liver itself.
While the results obtained in the studies published thus far do not constitute a direct proof of a link between the cellular stress response and Cr(VI)-induced carcinogenesis, they do show the ability of this carcinogen to modulate the expression of several components of this response under conditions of biological relevance. It has also become clear that the observed effects are dependent on tissue, cell type, Cr(VI) concentration, duration of exposure and HSP isoform. Future studies must address the issue of biological relevance and should also include adequate time courses, as HSP transcript and protein levels change over time during the recovery period.