p53 Activation by Cr(VI): A Transcriptionally Limited Response Induced by ATR Kinase in S-Phase

Abstract

Cellular reduction of carcinogenic chromium(VI) causes several forms of Cr-DNA damage with different genotoxic properties. Chromate-treated cultured cells have shown a strong proapoptotic activity of the DNA damage-sensitive transcription factor p53. However, induction of p53 transcriptional targets by Cr(VI) in rodent lungs was weak or undetectable. We examined Cr(VI) effects on the p53 pathway in human cells with restored levels of ascorbate that acts as a principal reducer of Cr(VI) in vivo but is nearly absent in standard cell cultures. Ascorbate-restored H460 and primary human cells treated with Cr(VI) contained higher levels of p53 and its Ser15 phosphorylation, which were induced by ATR kinase. Cr(VI)-stimulated p53 phosphorylation occurred in S-phase by a diffusible pool of ATR that was separate from the chromatin-bound pool targeting DNA repair substrates at the sites of toxic mismatch repair (MMR) of Cr-DNA adducts. Even when more abundantly present than after exposure to the radiomimetic bleomycin, Cr(VI)-stabilized p53 showed a much more limited activation of its target genes in two types of primary human cells. No increases in mRNA were found for nucleotide excision repair factors and a majority of proapoptotic genes. A weak transcription activity of Cr(VI)-upregulated p53 was associated with its low lysine acetylation in the regulatory C-terminal domain, resulting from the inability of Cr(VI) to activate ATM in ascorbate-restored cells. Thus, p53 activation by ascorbate-metabolized Cr(VI) represents a limited genome-protective response that is defective in upregulation of DNA repair genes and proapoptotic transcripts for elimination of damaged cells.

Epidemiological studies have firmly established chromium(VI)-containing compounds as human lung carcinogens by inhalation (Langard, 1990; Salnikow and Zhitkovich, 2008). The most extensive characterization of Cr(VI)-related incidence of lung cancers is available for large cohorts of workers who were exposed to soluble chromates (Gibb et al., 2000; Luippold et al., 2003). Data from these cohorts have been used for a risk assessment modeling and establishment of a new permissible exposure limit by the Occupational Health and Safety Administration (OSHA, 2006). Rodents are resistant to lung carcinogenicity by soluble Cr(VI) due to its very rapid extracellular inactivation (Krawic et al., 2017). DNA damage by Cr(VI) occurs as a result of its intracellular reduction to Cr(III) by small-sized reducers ascorbate (Asc), glutathione or cysteine (Salnikow and Zhitkovich, 2008). Cr(VI) reactions with Asc differ from those with thiols by the absence of Cr(V) intermediate (DeLoughery et al., 2014; Stearns and Wetterhahn, 1994). Cr(V) is more reactive than Cr(IV) (Sugden and Stearns, 2000; Sugden et al., 2001) and it can also participate in redox reactions with cellular oxidants such as H₂O₂, producing more toxic reactive oxidants damaging DNA (Messer et al., 2006). Asc is a dominant reducer of Cr(VI) in the lung and other tissues in vivo (Standeven and Wetterhahn, 1991, 1992; Suzuki and Fukuda, 1990) but not in standard cell cultures that are all severely deficient in vitamin C (Blaszczak et al., 2019). Reactions of Cr(III) with DNA leads to the production of several forms of small Cr-DNA adducts (Quievryn et al., 2002; Voitkun et al., 1994). Cr(VI) also causes the formation of DNA-protein crosslinks (DPCs) that were characterized biochemically in Cr(VI)-treated cells (Costa, 1991) and detected in lymphocytes of Cr(VI)-exposed individuals (Zhitkovich et al., 1996). DNA-protein crosslinks arise from binary Cr-DNA adducts through a capture and attachment of protein (Macfie et al., 2010). Among small DNA adducts, ternary complexes showed the highest mutagenicity during replication in human cells (Quievryn et al., 2003; Voitkun et al., 1998). Toxicity and genotoxicity of ternary DNA adducts were linked to their aberrant processing by DNA MMR, which results in the formation of DNA double-strand breaks (DSBs) (Peterson-Roth et al., 2005; Reynolds et al., 2007). The genotoxic significance of DPCs formed in Cr(VI)-treated cells is less clear as these potentially toxic lesions are much less abundant than small Cr-DNA adducts (Zhitkovich et al., 1996). Mismatch repair did not play any detectable role in cytotoxicity or activation of genotoxic stress signaling in response to DPCs formed by formaldehyde (Peterson-Roth et al., 2005; Wong et al., 2012). Thus, it is possible that the presence of DPCs in Cr(VI)-treated cells causes additional genotoxic stress that is unrelated to MMR and small ternary Cr-DNA adducts.

The tumor suppressor p53 is a main activator of gene expression after DNA damage (Ellinger-Ziegelbauer et al., 2004; Kruse and Gu, 2009; Nikulenkov et al., 2012). This property of p53 led to numerous efforts to use either its transcriptomic signature or specific p53-responsive reporters for the detection of genotoxic responses. Studies in standard Asc-deficient cell cultures have found a robust p53 activation and p53-dependent apoptosis after large doses of Cr(VI) (Carlisle et al. 2000; Hill et al., 2008; Russo et al., 2005; Son et al., 2010). However, restoration of Asc in cells prior to treatments with Cr(VI) diminished p53 accumulation and its Ser15 phosphorylation (Luczak et al., 2016). Asc-replete cells have also lost Cr(VI)-induced activation of ATM kinase that upregulates p53 in response to oxidative stress. The p53-related transcriptome changes were borderline significant or undetectable in the lungs of rodents after intratracheal instillations of Cr(VI) despite the use of genotoxic doses and a higher abundance of p53 protein (D’Agostini et al., 2002; Izzotti et al., 2004). Thus, the standard cell cultures do not recapitulate p53 activity of Cr(VI) in vivo. It is also unclear what DNA damage-responsive sensors are involved in the p53 protein upregulation after Cr(VI) treatments.

Here, we examined how Cr(VI) affects the p53 pathway in human cells with physiological levels of its principal reducer Asc. We found that Cr(VI)-stabilized p53 produced limited transcriptomic changes, which were weak or undetectable for several targets. We further determined that Cr(VI) caused p53 accumulation and its Ser15 phosphorylation by ATR in the S-phase of the cell cycle. Activation of the ATR-p53 axis was independent on DNA MMR of Cr-DNA adducts and may represent a genotoxic response to DPCs.

MATERIALS AND METHODS

Chemicals

Dehydro-l-(+)-ascorbic acid (DHA) dimer, potassium chromate (99% pure), cisplatin, and hydroxyurea (HU) were purchased from Sigma-Aldrich. KU60019, VE821, and NU7441 were obtained from SelleckChem. NU7026 and ETP46464 were from Calbiochem; 1, 2-diamino-4, 5-dimethoxybenzene and dihydrorhodamine 123 (DHR123) from Molecular Probes; bleomycin from LKT Labs and KU55933 from BioVision.

Cells and treatments

All cells were obtained from the American Type Culture Collection. Human H460 lung epithelial cells were cultured under 95% air/5% CO₂ using RPMI-1640 medium containing 10% fetal bovine serum and antibiotics. IMR90 primary human lung fibroblasts were propagated in DMEM medium containing 10% serum and antibiotics in the atmosphere containing 5% O₂ and 5% CO₂. Growth of primary human fibroblasts at 5% O₂ improves their proliferative properties. Normal human neonatal keratinocytes were grown in 95% air/5% CO₂ using vendor-recommended media supplemented with growth factors. The source of Cr(VI) was K₂CrO₄ dissolved in deionized water and filter-sterilized before addition to cells. Asc-restored cells were treated with Cr(VI) in the complete growth media. Inhibitors of DNA damage-responsive kinases were added together with Cr(VI).

Restoration of Asc levels

Cells were incubated for 90 min with DHA in Krebs-HEPES buffer [30 mM HEPES (pH 7.5), 130 mM NaCl, 4 mM KH₂PO₄, 1 mM MgSO₄, 1 mM CaCl₂] containing 0.5 mM d-glucose and 5% fetal bovine serum. Following the initial dose testing, we used 0.2 mM DHA for H460 and keratinocytes and 1 mM DHA for IMR90 cells to restore physiological concentrations of Asc prior to the addition of Cr(VI). Cellular Asc was extracted and measured as a fluorescent product formed after reaction with 1, 2-diamino-4, 5-dimethoxybenzene (Reynolds et al., 2012). Cell volumes were measured by forward scattering analyses using a flow cytometer (FACSCalibur, BD Biosciences).

Western blotting

Total protein extracts were prepared by boiling cells for 10 min in a 2% SDS buffer (2% SDS, 50 mM Tris-HCl pH 6.8, 10% glycerol) containing Halt Protease and Phosphatase Inhibitors (Thermo Scientific). Chromatin and soluble protein fractions were isolated as recently described (Rubis et al., 2019) except that NP40 was replaced with Triton X-100. For detection of proteins smaller than 100 kDa, samples were run on 12% SDS-PAGE gels and then electrotransferred onto polyvinylidene difluoride (PVDF) membranes using PierceG2 Fast Blotter and 12% ethanol-supplemented buffer (Luczak and Zhitkovich, 2018). Larger proteins were separated on 6% or 10% gels and transferred onto PVDF membranes overnight by a wet transfer at 4°C. Primary antibodies were incubated with membranes overnight at 4°C. The following primary antibodies were used: anti-phospho DNAPK (S2056) (Abcam, ab18192), anti-MSH6 (BD Biosciences, 610919), anti-γ-tubulin (Sigma-Aldrich, T6557), anti-phospho-CHK1 (Ser317) (Cell Signaling, 2344), anti-phospho-H2AX (Ser139) (Cell Signaling, 2577), anti-p53 (Santa Cruz, sc-126), anti-GAPDH (Cell Signaling, 3683), anti-fibrillarin (Abcam, ab5821), anti-phospho-ATM (S1981) (Cell Signaling, 13050), anti-phospho-RPA32 (S33) (Bethyl, A300-246A), anti-MSH2 (BD Biosciences, 556349), anti-phospho-53 (S15) (Cell Signaling, 9284), anti-histone H3 (Cell Signaling, 9715), anti-MDM4 (Bethyl, A300-287A), anti-acetylated p53 (Lys382) (Cell Signaling, 2525), anti-p21 (Cell Signaling, 2947), anti-XPC (Cell Signaling technology, 12701), anti-phospho-CHK2 (Thr68) (Cell Signaling, 2661), anti-ribosomal protein L7A (Cell Signaling, 2415), anti-ATR (Santa Cruz, sc-1887), anti-ATM (Santa Cruz, sc-23921), anti-MDM2 (Cell Signaling, 86934), anti-p300 (Millipore, 05-257), anti-CBP (Cell Signaling, 7389), and anti-PCAF (Cell Signaling, 3378). Horseradish peroxidase-conjugated goat anti-mouse IgG (Millipore) and goat anti-rabbit IgG (Cell Signaling) were used as secondary antibodies.

Knockdowns with shRNA

The pSUPER.retro.puro vector (OligoEngine) was used for intracellular delivery and stable expression of shRNA in IMR90 primary human cells. The targeting sequences were 5′-GACTCCAGTGGTAATCTA-3′ for p53, 5′-GAGTGTTGTGCTTAGTAAA-3′ for MSH2, and 5′-GGTGATCCCTCTGAGAACT-3′ for MSH6. A nonspecific control was purchased from the supplier of the vector. Packaging of retroviral particles, infection of cells, and selection conditions were as described previously (Reynolds et al., 2004).

Quantitative reverse transcription PCR (RT-qPCR)

DHA-preincubated-preincubated IMR90 and keratinocytes were treated with Cr(VI) for 3 h followed by 3 h incubation in Cr-free media prior to isolation of RNA. Total cellular RNA purified with TRIzol Reagent (Ambion) was reverse transcribed with the RT First Strand Kit from Qiagen. Real-time PCR reactions were run on the ViiA7 Real-Time PCR System (Applied Biosystems) using the RT SYBR Green ROX qPCR Mastermix and gene-specific primers from Qiagen. Expression of target genes was determined by the ΔΔCt method and normalized to the mRNA levels of three housekeeping genes (B2M, GAPDH, and TBP).

Immunofluorescence

IMR90 cells were seeded at 40%–50% confluence on human fibronectin-coated coverslips (Corning, catalogue #356088). After overnight attachment, cells were preincubated with 1 mM DHA and then treated with Cr(VI) for 3 h. Cells were incubated in Cr-free media for 3 h before fixation. Replicating cells were labeled by the addition of 10 µM 5-ethynil-2′-deoxyuridine (EdU) for 40 min before fixation. Cells were directly fixed to the slides by incubation with 3.7% paraformaldehyde for 15 min at room temperature. Cells were permeabilized in PBS/0.5% Triton X-100 for 15 min at room temperature and then blocked with 2% fetal bovine serum in PBS at 37°C for 30 min. DNA-incorporated EdU was fluorescently labeled using Click-iT EdU-Alexa Fluor 488 Imaging kit (Invitrogen, C10337). Primary antibodies [anti-phospho-53 (S15), Cell Signaling Technology, 9284] were diluted in 1% BSA, 0.5% Tween-20 in PBS, and added to cells for 2 h at 37°C. Secondary antibodies (Alexa Fluor 568 goat anti-rabbit, Life Technologies, A11036) were added to the slides for 1 h at room temperature in the dark. Coverslips were then mounted on glass slides with Vectashield fluorescence mounting media containing DAPI (H-1200). Cells were imaged using the Nikon E-800 Eclipse fluorescent scope (×200–400 magnification, wide-field images).

Cell viability

Cytotoxic effects of Cr(VI) were measured using the CellTiter-Glo luminescent cell viability assay (Promega, G7571). This assay quantifies the metabolic activity of cell populations, making it sensitive to all forms of cytotoxicity including anti-proliferative effects. Cells were seeded at 1000 cells/well into black 96-well optical bottom cell culture plates (Thermo Scientific, 165305) one day before the addition of Cr(VI). Cell viability measurements were taken at 72 h post-Cr.

Detection of oxidants during Cr(VI) reduction

Formation of oxidants in Cr(VI) reduction reactions was monitored by the redox-sensitive fluorescent probe DHR123, which responds to reactive oxygen species as well as free and glutathione-bound Cr(V) (DeLoughery et al., 2014). Reaction mixtures contained 0 or 40 µM Cr(VI), 100 mM NaCl, 50 mM MOPS (pH 7.0), 10 µM DHR123, 3 mM glutathione, 0.2 mM cysteine, and various concentrations of Asc. Samples were incubated for 30 min at 37°C followed by recording DHR123 fluorescence (excitation = 500 nm, emission = 535 nm). The extent of Cr(VI) reduction was measured by chromate absorbance at 372 nm.

Cr(VI) uptake

Asc-restored H460 and IMR90 cells were treated with 0 or 10 µM Cr(VI) for 3 h, followed by nitric acid extraction of cellular Cr and its measurements by graphite furnace atomic absorption spectroscopy (Messer et al., 2006).

Statistics

Statistical significance was evaluated by two-tailed, unpaired t-test.

RESULTS

Functionally Limited Activation of p53 by Cr(VI)

Lung is the main target of carcinogenic effects of Cr(VI) in occupationally exposed populations. Our previous studies have extensively characterized cytotoxic and genotoxic effects of Cr(VI) in H460 lung epithelial cells and IMR90 primary lung fibroblasts (Luczak et al., 2016; Reynolds et al., 2009, 2012). In this work, we continued to use these two human cell lines as our main biological models. Although H460 are transformed cells, they retain wild-type p53 gene and are frequently used for studies of p53-dependent responses. In our hands, H460 cells have shown a robust activation of the p53 pathway in response to the DNA-protein crosslinker formaldehyde (Wong et al., 2012) and the hypoxia mimetic nickel (Wong et al., 2013). Asc is the primary reducer of Cr(VI) in the cells in vivo but it is present in very low or undetectable concentrations in cultured cells. In the freshly fed cultures, we found that H460 cells had only 11.1±3 µM Asc, which is less than 1% of its physiological concentration in the human lung (Slade et al., 1985). A preincubation with the oxidized form of vitamin C, DHA, resulted in a very efficient restoration of Asc levels in H460 cells (Figure 1A). Similar to the majority of other cells in culture, H460 gradually lose Asc through its leakage into the media, with approximately 40% decrease in cellular Asc following 3 h post-DHA incubations. In all subsequent experiments, we used 0.2 mM DHA for the delivery of Asc into H460 cells, which created the initial concentration of 2.4 mM. Treatment of Asc-restored H460 cells with 5 µM Cr(VI), which corresponds to approximately LD₃₅ (Figure 1B), produced very clear increases in p53 protein and its Ser15 phosphorylation levels at 0–4 h recovery times (Figure 1C). The protein abundance of the p53-inducible CDK inhibitor p21 (CDNK1A1) also increased at 2 and 4 h post-Cr times although not as dramatically as did p53 or its phosphorylation. Cellular responses at 0–4 h post-Cr were not affected by cell death, as a Western blot for a sensitive apoptotic marker PARP showed only its intact form. Next, we tested how p53 upregulation by Cr(VI) compared with that by the radiomimetic drug bleomycin. We found that relative to bleomycin, Cr(VI) was more effective in increasing amounts of p53 but less potent in the stimulation of p53-Ser15 phosphorylation and upregulation of the p53 target, p21 protein (Figure 1D). In agreement with the earlier study using a slightly different (serum-free) DHA incubation protocol (Luczak et al., 2016), the restoration of Asc levels in H460 cells dramatically suppressed p53-Ser15 phosphorylation by Cr(VI) with approximately a 15-fold decrease for 5 µM Cr (Figure 1E). For a more toxic 10 µM Cr, the loss of p53 phosphorylation in Asc-restored cells was moderately lower but still very large (6.7-fold). In contrast, the inhibitory effects of cellular Asc on p53 protein increases were small. Similar to its effects on p53-Ser15 phosphorylation, cellular Asc was also a very potent suppressor of CHK2 phosphorylation at Thr68 (Figure 1F), which is a target site for the DNA damage-responsive kinase ATM. Protein levels of CHK2, ATM, and another DNA damage-sensitive kinase ATR were not altered by Cr(VI) or cellular Asc. In contrast to Cr(VI), restoration of cellular Asc had no appreciable effects on phosphorylation of p53 or CHK2 by the DNA adduct-forming but redox-inactive metal compound cisplatin (Figure 1G). Asc also showed no significant impact on the DNA damage signaling in response to topoisomerase I-induced DNA breaks (Luczak et al., 2016). Thus, Asc effects on the DNA damage responses appeared to be specific to Cr(VI). ATM is a redox-sensitive kinase that can be activated directly by oxidants in the absence of DNA damage (Guo et al., 2010). We found that introduction of Asc into the thiols-based Cr(VI) reduction reactions in the amount equal to as little as 5% of the glutathione concentration completely eliminated the production of oxidants despite a larger amount of metabolized Cr(VI) (Figs. 1H and 1I). Thus, the absence of oxidants such as Cr(V) intermediate (Zhitkovich, 2005) during Asc-driven Cr(VI) reduction probably was responsible for the loss of ATM activation in Asc-restored cells, strongly diminishing p53-Ser15 and CHK2 phosphorylation.

Figure 1.

Cytotoxicity and upregulation of p53 by Cr(VI) in H460 cells. Cells were incubated with DHA for 90 min in a glucose-supplemented Krebs buffer containing 5% FBS. Cells were treated in complete growth media with Cr or bleomycin for 3 h in 0.2 mM DHA-preincubated cells. A, Cellular levels of Asc immediately (0 h) and 3 h later after incubations with DHA. Data are means±SD, n = 3. B, Cytotoxicity of Cr(VI) as measured by the CellTiter-Glow assay at 72 h post-Cr. Data are means±SD, n = 3. C, Westerns for Asc-restored cells treated with Cr(VI) and collected after different recovery (post-Cr) times. GAPDH was used as a loading control. D, Comparison of the p53 pathway activation by Cr(VI) and bleomycin (bleo) in Asc-restored cells. E, Effects of Asc restoration on p53 upregulation by Cr(VI). Cells were preincubated with 0 or 0.2 mM DHA and then treated with Cr(VI) for 3 h and collected for westerns. Longer and shorter film exposures for phospho-p53 are shown. F, Cellular Asc inhibits CHK2-T68 phosphorylation by Cr(VI) but does not alter protein levels of CHK2, ATR, or ATM. G, Asc restoration has no effect on p53 and CHK2 phosphorylation by cisplatin. H, Completion of Cr(VI) reduction in glutathione/Cys mixtures supplemented with various concentrations of Asc. Data are means±SD, n = 3. I, Elimination of oxidants during Cr(VI) reduction by the addition of Asc to thiol-containing reactions (3 mM glutathione, 0.2 mM Cys). Data are means±SD, n = 3. Abbreviations: DHA, Dehydro-l-(+)-ascorbic acid.

Next, we examined effects of Cr(VI) on p53 in IMR90 primary human cells in which we restored physiological levels of Asc by incubations with 1 mM DHA (Figure 2A). A restoration of physiological concentrations of Asc in IMR90 primary cells required a preincubation with 1 mM DHA versus 0.2 mM DHA for transformed H460 cells. Dehydro-l-(+)-ascorbic acid enters cells through glucose transporters that are overexpressed in transformed cells, making them more efficient in taking up oxidized vitamin C (Blaszczak et al., 2019). IMR90 showed a higher viability than H460 cells after Cr(VI) treatments (Figure 2BvsFigure 1B), which was associated with their lower uptake of Cr (2.1-fold for 10 µM Cr). Using Cr(VI) concentrations that corresponded to approximately LD₂₅ and LD₅₀, we also found a clear upregulation of both p53 protein and its Ser15 phosphorylation in IMR90 primary cells (Figure 2C). Both responses peaked at 3 h posttreatment for the lower dose of Cr and plateaued for the higher dose at this time. To examine the cause of increased p53 protein abundance after Cr(VI), we measured stability of p53 and its gene expression. These analyses found that Cr(VI) caused stabilization of total p53 protein and especially, its Ser15-phosphorylated form (Figure 2D). No significant changes were detected in the levels of p53 mRNA (Figure 2E). Gene expression effects of any transcription factor are preceded by its binding to chromatin. Our examination of IMR90 primary cells showed that chromatin binding of Cr-upregulated p53 was similar to that of other DNA-damaging agents such as bleomycin and cisplatin (Figure 2F). For all three treatments, the majority of p53 protein was bound to chromatin. Transcription activity of p53 is repressed by its inhibitor MDM4 (Kruse and Gu, 2009). No changes in the protein abundance of this inhibitor were found in Cr(VI)-treated primary cells (Figure 2G). At the time of the maximal increases in p53 abundance, Cr(VI)-treated IMR90 cells did not show marked changes in protein expression of three transcriptional targets of p53 such as p21 and XPC (Figure 2H). Next, we compared p53 upregulation by Cr(VI) versus oxidatively induced DNA breaks, which is a canonical genotoxic activator of p53. We found that two doses of the radiomimetic bleomycin that were less potent than Cr(VI) in increasing p53 protein abundance, caused a much stronger upregulation of p21 (Figure 2I). Despite a higher p53 abundance, Cr(VI) was a less potent inducer of Ser15 phosphorylation in comparison to bleomycin. This disparity in Ser15 phosphorylation was not caused by different levels of p53-targeting ATM or ATR kinases (Figure 2J). A lower protein abundance of p53 for bleomycin relative to Cr(VI) was associated with higher levels of MDM2, the main E3 ubiquitin ligase targeting p53 degradation (Figure 2J). Transactivation activity of p53 is strongly promoted by acetylation of a cluster of lysines in its C-terminal domain (Wang et al., 2016). Judging by the levels of acetyl-K382, Cr(VI) was a very weak inducer of acetylation at this regulatory region as compared with bleomycin (Figure 2K). Overall, our studies in H460 and primary lung cells found a clear ability of Cr(VI) to induce classic markers of the genotoxic stress-activated p53 such as its protein stabilization and Ser15 phosphorylation. However, protein expression of p53 transcriptional targets was only modestly increased or unchanged.

Figure 2.

Cr(VI) effects on p53 in IMR90 primary human cells. Treatments with Cr(VI) and DNA-damaging drugs were for 3 h in cells preincubated with 1 mM DHA. Cells were collected at different postexposure time (recovery time). A, Cellular concentrations of Asc in cells after incubations with 1 mM DHA. Data are means±SD, n = 3. B, Cytotoxicity of Cr(VI) in cells as determined by the CellTiter-Glow assay at 72 h post-Cr. Data are means±SD, n = 3. C, Westerns for cells collected at 0, 3 and 6 h recovery (post-Cr) times. GAPDH was used as a loading control. D, Stability of p53 in Cr(VI)-treated cells. The protein synthesis inhibitor cycloheximide (CHX, 0.1 mg/ml) was added for 1 h after Cr removal. E, Levels of p53 mRNA at 3 h postexposure to 20 µM Cr or 10 µM bleomycin (bleo). Cells expressing nonspecific (sh-ns) or p53-targeting shRNA (sh-p53) were used for RT-qPCR analyses. Data are means±SD, n = 3, ***p < .001 relative to sh-ns controls. F, Chromatin binding of p53 in cells treated with 20 µM Cr(VI), 10 µM bleomycin (bleo), or 20 µM cisplatin (Pt) and collected at 3 h postexposure. GAPDH and histone H3 served as loading controls and markers of soluble and chromatin fractions, respectively. G, MDM4 levels in cells at 0 and 3 h recovery times post-Cr. H, Protein expression of p53-regulated genes in Cr(VI)-treated cells at 3 h postexposure. I, Comparison of the p53 pathway activation by Cr(VI) and bleomycin (bleo). J, ATM, ATR, and MDM2 levels in Cr(VI)- and bleomycin-treated cells. K, p53-Lys382 acetylation by Cr(VI) and bleomycin. Abbreviations: RT-qPCR, quantitative reverse transcription PCR; DHA, Dehydro-l-(+)-ascorbic acid.

Protein expression is not always a good measure of gene expression, as some p53-regulated proteins such as p21 can also become unstable in response to DNA damage (Hall et al., 2014; Jascur et al., 2011). Therefore, we measured mRNA levels of 10 genes that are known as direct targets of the transcription factor p53. The p53-dependence of induction of these genes in IMR90 primary cells was verified by the loss of their upregulation in cells with stable shRNA knockdown of p53, which produced more than a 20-fold decrease in p53 mRNA (Figure 2E). Using doses that gave a stronger p53 accumulation for Cr(VI) than bleomycin (Figure 2I), we found that among four proapoptotic genes, all of which were strongly upregulated by bleomycin, only BBC3 (also known as PUMA) was modestly induced by Cr(VI) (Figure 3A). Two growth arrest-related genes p21 (CDKN1A1) and BTG2 were induced by Cr(VI) but to a lesser extent than by bleomycin (Figure 3B). Expression of DNA repair genes XPC and DDB2 was not changed by Cr(VI) in contrast to their significant induction by bleomycin (Figure 3C). Among the negative feedback targets, mRNA levels of Ser/Thr protein phosphatase PPM1D and the p53 E3 ubiquitin ligase MDM2 were only modestly elevated by Cr(VI) (Figure 3D). In contrast, bleomycin produced a strong induction of both genes, especially MDM2. To exclude the possibility that stable expression of our nonspecific shRNA in IMR90 cells somehow interfered with normal gene induction by p53, we tested gene expression in response to Cr(VI) in early passage IMR90 cells without any genetic and other manipulations. These cells showed the same pattern of Cr(VI)-stimulated gene expression, namely, clear increases in mRNA levels for p21 and BBC3, a modest induction of MDM2 and no upregulation of proapoptotic PIDD1 and DNA repair factors DDB2 and XPC (Figure 3E).

Figure 3.

mRNA levels of p53-regulated genes in IMR90 primary cells. Cells expressing nontargeting (sh-ns) and p53-targeting (sh-p53) shRNA were Asc-restored, treated with a solvent (H₂O), 20 μM Cr(VI), or 10 μM bleomycin for 3 h and collected 3 h later. mRNA levels were measured by RT-qPCR. Data are means±SD, n = 3, *p < .05, **p < .01, ***p < .001 relative to mock-treated cells. A, Expression of proapoptotic genes, (B) growth arrest-related, (C) DNA repair and (D) negative feedback response genes. E, Dose-dependent changes in gene expression by Cr(VI) in early passage IMR90 cells (passage 7). Cells were collected at 3 h post-Cr. Abbreviations: RT-qPCR, quantitative reverse transcription PCR.

Finally, we evaluated the effects of Cr(VI) on p53 in Asc-restored primary human keratinocytes. Although Cr(VI) is not known as a skin carcinogen, it may act as a cocarcinogen with UV (Davidson et al., 2004; Salnikow and Zhitkovich, 2008; Uddin et al., 2007). Our other reason for exploring responses in keratinocytes was related to their evolutionary selected robustness of UV-DNA damage removal via nucleotide excision repair, which also operates on Cr-DNA adducts (O’Brien et al., 2005; Reynolds et al., 2004). Similar to primary lung cells, Cr(VI)-treated primary keratinocytes also showed a clear upregulation of p53 and its Ser15 phosphorylation, which was accompanied by only modest increases in the p53-dependent transcriptional target, p21 (Figure 4A). However, when compared with the radiomimetic bleomycin at doses that produced a comparable protein abundance of p53, Ser15 phosphorylation by Cr(VI) was much weaker (Figure 4B). A near absence of phospho-Ser15 in untreated keratinocytes makes Cr(VI) appear as a strong inducer of this modification (Figure 4A), however, this increase was small when compared with the bleomycin-induced response. The presence of Lys382 acetylation was undetectable for Cr(VI) but this transcriptionally important modification was abundantly increased by bleomycin (Figure 4B). Similar to IMR90 cells, we found a strong induction of mRNA for all six studied genes in bleomycin-treated cells whereas only BBC3, BTG2, and MDM2 showed modest (<twofold) increases by Cr(VI) (Figure 4C). Again, gene expression of PIDD1 and DNA repair proteins XPC and DDB2 was not elevated by Cr(VI) despite their very strong upregulation by the DNA-breaking bleomycin. Thus, Cr(VI)-upregulated p53 is functionally limited in its ability to transactivate expression of a full spectrum of target genes.

Figure 4.

Activation of the p53 pathway in primary human keratinocytes. Asc-restored keratinocytes (2.8±0.3 mM, n = 3) were treated with a solvent (H₂O), 10 μM Cr(VI), or 10 μM bleomycin (bleo) for 3 h and collected after 3 h recovery in the regular media. mRNA levels were measured by RT-qPCR. Data are means±SD, n = 3, *p < .05, **p < .01, ***p < .001 relative to mock-treated cells. A, Westerns for readouts of activated p53 in primary human keratinocytes. B, Westerns for p53 forms in cells collected at 3 h recovery after Cr(VI) and bleomycin treatments (both at 10 μM). GAPDH was used as a loading control. C, mRNA levels of p53-regulated genes. Abbreviations: RT-qPCR, quantitative reverse transcription PCR.

Mechanisms of p53 Upregulation

To further understand mechanisms of p53 upregulation by Cr(VI), we examined cell cycle specificity of this response. Immunofluorescence imaging of Cr(VI)-treated IMR90 primary cells showed that Ser15-phosphorylated p53 was present only in S-phase that was identified by incorporation of the thymidine analog EdU (Figure 5A). For cells treated in this experiment with 20 µM Cr(VI) and analyzed after 3 h recovery, 40.8±5% and 35.5±5.4% (n = 3) of all cells were stained positive for phospho-p53 and double positive for phospho-p53/EdU, respectively. Exposure to a lower concentration of 10 µM Cr(VI) also produced S-phase-specific p53 phosphorylation at different post-Cr times (Figure 5B). Stabilization of p53 in response to genotoxic agents is usually a result of its Ser15 phosphorylation by one of the apical DNA damage-responsive kinases such as ATM, ATR, or DNAPK (Kruse and Gu, 2009). To identify which of these kinases is responsible for targeting p53 in response to Cr-DNA damage, we employed a set of highly selective inhibitors, which allows a rapid inactivation of kinase activity without the loss of protein-protein interactions. Validation studies in cells treated with the known activators of ATM and DNAPK (DNA-breaking bleomycin) or ATR (dNTP depletion by HU) showed that the selected concentrations of the inhibitors were very effective in the suppression of activity of specific kinases (Figure 6A). Both inhibitors of ATM eliminated its activating autophosphorylation at Ser1981 whereas two tested inhibitors of DNAPK blocked its activating autophosphorylation at Ser2056. Inhibitors of ATR kinase suppressed phosphorylation of its target protein CHK1 at Ser317. In Cr(VI)-treated H460 cells, the use of two sets of inhibitors for each kinase showed that only the inactivation of ATR blocked phosphorylation and stabilization of p53 (Figs. 6B and 6C). The same result was observed in IMR90 primary cells, which showed the loss of p53 upregulation in the presence of inhibitors of ATR but not two other DNA damage-responsive kinases (Figs. 6D and 6E). The lack of ATM involvement in p53 upregulation is consistent with the absence of the activated (autophosphorylated) form of this kinase in Cr(VI)-treated cells (Figure 6C).

Figure 5.

Cell cycle specificity of p53 phosphorylation by Cr(VI). Asc-restored IMR90 cells were treated with Cr(VI) for 3 h and processed for immunostaining at the indicated recovery times. Replicating cells were labeled by the addition of 10 μM EdU for 40 min before fixation of slides. Nuclei were identified by DNA staining with DAPI. A, Representative images of mock- and 20 μM Cr(VI)-treated IMR90 cells fixed at 3 h post-Cr. B, Percentage of cells stained for phospho-p53 (pS15+) and for both phospho-p53 and EdU at 0–6 h recovery after exposure to 0 or 10 μM Cr(VI). Data are means±SD, n = 3.

Figure 6.

Role of DNA damage-responsive kinases in p53 phosphorylation. Inhibitors of ATM (ATM-i), ATR (ATR-i), or DNAPK (DNAPK-i) kinases were present during Cr(VI) treatments and added to the media during post-Cr incubations. ATM-i1—1 μM KU60019, ATM-i2—10 μM KU55933, ATR-i1—3 μM ETP46464, ATR-i2—5 μM VE821, DNAPK-i1—30 μM NU7026, and DNAPK-i2—3 μM NU7441. A, Validation of the selected kinase inhibitors. Asc-restored H460 cells were treated with 5 μM bleomycin (bleo) for 1 h or HU for 3 h. Inhibitors were present during Asc loading and drug treatments. B, Phosphorylation and protein levels of p53 in H460 cells after 2 h recovery post-Cr. C, Phosphorylation of p53 and ATM in H460 cells after 4 h recovery post-Cr. D and E, Westerns for p53 and phospho-p53 in IMR90 cells after 3 h recovery post-Cr. Abbreviations: HU, hydroxyurea.

We have previously found that ATR was responsible for Ser139-phosphorylation of histone H2AX (γ−H2AX formation) in Cr(VI)-treated H460 and IMR90 cells (DeLoughery et al., 2015). γ−H2AX is the biochemical marker of DNA DSBs, whose induction by Cr(VI) required active DNA MMR and it was initiated in S-phase (Peterson-Roth et al., 2005; Reynolds et al., 2009; Zecevic et al., 2009). Therefore, we next examined whether p53 upregulation by Cr(VI) was also dependent on processing of Cr-DNA damage by MMR. Cr(VI) treatments of IMR90 primary cells with stable knockdowns of MSH2 or MSH6, which are Cr-DNA damage-sensing MMR components (Reynolds et al., 2009), showed the expected loss of the DSB marker γ−H2AX (Figure 7A). Phosphorylation of two other targets of ATR kinase, CHK1 at Ser317 and RPA32 at Ser33, was also abrogated by inactivation of MMR (Figs. 7A and 7C). Despite their established above ATR dependence, p53 accumulation and p53-Ser15 phosphorylation by Cr(VI) were unaffected by the loss of MSH2 and MSH6 proteins (Figs. 7B and 7C). MSH2 and MSH6 operate as a heterodimer and the loss of one component causes destabilization of its binding partner, which explains depletion of both proteins in single knockdowns (Figure 7B). Thus, p53 upregulation represents an ATR-dependent signaling branch that is activated by different DNA lesions than other genotoxic responses mediated by this kinase such as checkpoint signaling (CHK1 phosphorylation) and DNA repair (H2AX and RPA32 phosphorylation).

Figure 7.

Genotoxic stress responses to Cr(VI) in primary human cells with inactivated MMR. Asc-restored IMR90 cells with chromosome-integrated vectors expressing nonspecific, MSH2- or MSH6-targeting shRNA were treated with Cr(VI) for 3 h and collected immediately or after 4 h recovery in regular media. A, Mismatch repair-dependence of Cr(VI)-induced phosphorylation of CHK1, RPA32, and histone H2AX (γ-H2AX—Ser139 phosphorylated histone H2AX, γ-H2AX-ub₁—monoubiquitinated form of γ-H2AX). B, p53 upregulation after 0 h post-Cr or (C) 4 h recovery post-Cr in control and mismatch repair knockdown cells. Abbreviations: MMR, mismatch repair.

To further explore potential causes for a low p53 acetylation by Cr(VI), we examined protein levels of the main acetyltransferases targeting lysines in the C-terminal domain (Reed and Quelle, 2014). We found that the protein abundance of p300, CBP, and PCAF was not significantly affected by Cr(VI) in Asc-restored H460 or primary human cells (Figure 8A). Next, we turned our attention to the possibility that the suppression of ATM activation by Cr(VI) in cells with the physiological Asc (Figure 1F) could be a factor in the poor p53 acetylation. Similar to the observed dramatic decrease in Ser15 phosphorylation (Figure 1E), normalization of cellular Asc also caused the loss of K382 acetylation despite only a modestly lower p53 abundance (Figure 8B). To assess the contribution of ATM and ATR to p53 activation, including its K382 acetylation, we examined responses of Asc-restored cells to bleomycin and the ribonucleotide reductase inhibitor HU. The DNA damage response to bleomycin is caused by DSBs triggering activation of ATM. ATM signaling stimulates a nucleolytic processing of DSB ends to produce a 3′single-stranded DNA tail, which acts as a potent activator of ATR. Hydroxyurea causes a loss of dNTPs and stalling of DNA polymerases, resulting in the accumulation of ssDNA due to a continuous unwinding of the DNA duplex by the replicative helicase (Saldivar et al., 2017; Yazinski and Zou, 2016). As expected from its DNA damage characteristics, bleomycin-induced activation of ATR (assessed by CHK1 phosphorylation) in the ATM-dependent manner (Figure 8C). As a result of the accumulation of ssDNA, HU triggered a string activation of ATR resulting in a massive formation of phospho-CHK1. Strikingly, ssDNA-activated ATR in HU-treated cells did not induce Ser15 phosphorylation or increase p53 protein levels. In bleomycin-treated cells, Ser15-p53 phosphorylation was completely dependent on ATM activity, which in part occurred indirectly through activation of ATR, as evidenced by a decreased phosphorylation in the presence of the ATR inhibitor. The inhibition of ATM only partially suppressed the p53 accumulation by bleomycin. Similar to Ser15 phosphorylation, the induction of p53-K382 acetylation and MDM2 by bleomycin required ATM activity (Figure 8D). Thus, in bleomycin-treated cells, ATM was a much more potent inducer of Ser15 phosphorylation and especially, K382 acetylation than ATR. However, ATR still makes a significant contribution to the accumulation of p53, probably through inactivating phosphorylation of the MDM2 complex (Fu et al., 2010). Hydroxyurea-treated cells showed no appreciable changes in p53-K382 acetylation or MDM2 accumulation (Figure 8D). The absence of Ser15 phosphorylation and p53 stabilization by the ssDNA-producing HU supports our findings in Cr(VI)-treated cells, namely, that the ssDNA-bound/activated ATR does not act on p53, which is targeted by a diffusible pool of active ATR that can reach a globally distributed p53 in either soluble or chromatin-bound forms.

Figure 8.

Assessment of p53 acetylation-related factors. Cells with restored Asc levels were used for 3 h treatments with Cr(VI) and other genotoxic stressors. ATM (10 µM KU55933) and ATR (0.5 µM AZD6738) inhibitors were added to cells for 30 min before treatments with bleomycin and HU in panels C and D. A, Normal protein expression of p53-targeting lysine acetyltransferases p300, CBP, and PCAF in H460 (5 µM Cr) and IMR90 cells. MSH2 and MSH6 were used as loading controls. B, Cr(VI) effects on K382 acetylation and p53 protein levels in Asc-deficient and Asc-restored H460 cells. C, Westerns for phospho-CHK1 (Ser317), phospho-p53 (Ser15), and p53 in IMR90 cells treated with 10 µM bleomycin and 2 mM HU. D, Westerns for acetylated p53 (K382) and MDM2 protein in IMR90 cells treated as in panel C. Images of controls in panels C and D are from the same membranes/film exposures as bleomycin and HU samples. Unrelated lanes between controls and bleomycin samples were cropped out. Abbreviations: HU, hydroxyurea.

DISCUSSION

Functionally Limited Activation of p53 by Cr(VI)

We found that treatments of Asc-restored primary and H460 human cells with mild-to-moderate cytotoxic doses of Cr(VI) produced clear increases in protein and Ser15 phosphorylation levels of p53, which are classic markers of the DNA damage-triggered stress response (Kruse and Gu, 2009; Siliciano et al., 1997). In our previous studies, we have found that these doses induced DSBs and Cr-DNA adducts (DeLoughery et al., 2015; Reynolds et al., 2009, 2012). The accumulation of p53 after Cr(VI) occurred as a result of protein stabilization without changes in mRNA levels. Similar to genotoxic drugs bleomycin and cisplatin, Cr(VI)-stabilized p53 was bound to chromatin, indicating that initial steps in p53 activation by Cr(VI) were apparently normal. However, RT-qPCR measurements of mRNA showed only upregulation of a subset of p53-regulated genes by Cr(VI) in two primary cell lines of different histological origin. mRNA increases by Cr(VI)-stabilized p53 was moderate for cell cycle inhibitors, weak for negative feedback genes, weak or absent for proapoptotic genes, and completely absent for DNA repair genes. A weaker upregulation of the E3 ubiquitin ligase MDM2 confirmed by westerns (Figure 2J) probably contributed to a strong p53 accumulation by Cr(VI) despite a more modest increase in Ser15 phosphorylation when compared with the radiomimetic bleomycin. Additionally, ATR may act on the MDM2 complex to inhibit its activity toward p53 (Fu et al., 2010).

Acetylation of p53 at a cluster of six C-terminal Lys residues is important for expression of the majority of p53-regulated genes (Kruse and Gu, 2009; Reed and Quelle, 2014). The C-terminal cluster of unacetylated lysines acts a docking site for the acidic domain reader SET, which prevents a histone acetylation activity but not promoter recruitment of p300 acetyltransferase (Wang et al., 2016). Consistent with its limited transactivation activity, we found that p53 increases by Cr(VI) were not accompanied by parallel changes in its C-terminal Lys382 acetylation. The degree of the nonresponsiveness to unacetylated p53 varies among different promoters, which is in agreement with our findings on the upregulation of some p53-regulated genes (p21) but not others (FAS and DNA repair genes). A higher occupancy by RNA polymerase II and the absence of inhibitory factors at the p21 promoter were linked to its better responsiveness to p53 in comparison to FAS (Sullivan et al., 2018). These additional levels of promoter regulation can explain a differential gene activation by underacetylated p53 in Cr(VI)-treated cells.

Ser15 phosphorylation is typically associated with the stabilization of p53 as a result of MDM2 dissociation. However, a low level of this modification was also required for transactivation of p21 promoter by p53, which was linked to a greater degree of histone acetylation and chromatin relaxation (Loughery et al., 2014). Ser15 phosphorylation alone had much weaker effects on expression of proapoptotic genes, which is consistent with the need for additional events for their full activation and the initiation of the terminal cell fate decision. A priming phosphorylation at Ser15 and subsequent phosphorylation events at other N-terminal Ser/Thr residues were found to be important for the recruitment of p300 and CBP acetyltransferases and the ensuing lysine acetylation of the transactivation-promoting C-terminal domain (Lee et al., 2010; Reed and Quelle, 2014; Teufel et al., 2009). Phosphorylation at Ser15 and Ser20 also promoted recruitment and p53 acetylation by the MYST family lysine acetyltransferase MOZ (Rokudai et al., 2009). Thus, it is possible that a relatively low Ser15 phosphorylation by Cr(VI)-induced ATR was a cause of a low lysine acetylation of p53, which limited its transactivation activity to a subset of target genes, such as cell cycle checkpoint-promoting CDK inhibitor p21. A weak transactivation activity of p53 somewhat resembles the inability of Cr(VI)-treated cells to upregulate expression of inducible genes by other stimuli (Hamilton and Wetterhahn, 1989; Wei et al., 2004). Thus, it is possible that there is a general cause for impairment of inducible gene expression. In addition to defective protein acetylation, a global chromatin remodeling by Cr(VI) (VonHandorf et al., 2018) could exert inhibitory effects on gene induction.

Mechanism of p53 Activation by Cr(VI)

We found that ATR kinase was responsible for the upregulation of p53 abundance by Cr(VI). ATR is an apical DNA damage-responsive kinase that is most commonly activated in response to replication abnormalities (Saldivar et al., 2017; Yazinski and Zou, 2016). Cr(VI)-induced p53 phosphorylation was found exclusively in S-phase, pointing to replication events as a trigger in activation of the ATR-p53 axis. Replication was also necessary for the formation of DSBs by MMR processing of Cr-DNA adducts (Peterson-Roth et al., 2005; Reynolds et al., 2007). The formation of a biochemical marker of DSBs, Ser139-phosphorylated histone H2AX, was mediated by ATR kinase (DeLoughery et al., 2015). Despite this dependence on the same kinase, the type of DNA damage triggering p53 phosphorylation by ATR were different from DNA breaks produced by MMR of Cr-DNA adducts. In addition to H2AX phosphorylation, we found that MMR-generated secondary DNA lesions also induced phosphorylation of other targets of ATR such as checkpoint kinase CHK1 and the ssDNA-binding protein RPA. The most common activator of ATR is single-stranded DNA that becomes coated with RPA, which stimulates binding of ATR-ATRIP complex and activation of ATR kinase activity toward specific substrates (Saldivar et al., 2017; Yazinski and Zou, 2016). The ssDNA dependence is firmly established for CHK1 phosphorylation by ATR, including in our studies with the same cell lines used in the present work (Wong et al., 2012). We observed a disappearance of phosphorylated CHK1, RPA32, and H2AX in Cr(VI)-treated cells with MMR knockdowns, indicating that these cells lost DNA breaks and the source of ATR-activating ssDNA. Another mode of ATR activation, which does not involve the presence of ssDNA, is the collision of replication complexes with DPCs (Wong et al., 2012). These very bulky lesions block the progression of replicative helicase complexes, which prevents DNA unwinding and the appearance of ssDNA (Duxin et al., 2014). Based on the apparent ssDNA independence of p53 phosphorylation and the well-established ability of Cr(VI) to cause these large DNA adducts (Costa, 1991; Zhitkovich, 2005), DPCs could be an important trigger for activated ATR targeting p53 in Cr(VI)-treated cells. In contrast to the limited functionality of p53 in Cr(VI)-treated cells, the DNA-protein crosslinker formaldehyde produced a fully functional p53 despite its activation by ATR in S-phase (Wong et al., 2012). Additional stress signaling branches likely contribute to the full activation of the p53 pathway, as formaldehyde also caused a destabilization of the p53 repressor MDM4 (Wong et al., 2012) and stimulated another genome stress-sensitive apical kinase ATM (Ortega-Atienza et al., 2016), which did not occur in Cr(VI)-treated cells. Results of our studies with other genotoxic stressors (bleomycin and HU) indicate that ATM is a much more potent activator of p53 acetylation than ATR and support the model of p53 activation by a pool of a diffusible active ATR that is separate from ATR activated by binding to ssDNA. Figure 9 graphically summarizes our main results and activation mechanisms for two separate pools of ATR.

Figure 9.

Summary of the main findings. Cr-DNA damage formed as a result of Cr(VI) reduction by ascorbate triggers activation of ATR kinase through two replication-dependent mechanisms. Formation of ssDNA at the sites of DNA DSBs produced by toxic mismatch repair (MMR) of Cr-DNA adducts activates ATR via the ssDNA-binding mechanism and leads to the phosphorylation of adjacent chromatin substrates and ssDNA-recruited proteins. The second pool of activated ATR is diffusible and targets p53, resulting in selected transcription effects. Collision of replication complexes with DPCs inducing release of the MCM helicase-bound ATR is a plausible source of the active diffusible kinase. Abbreviations: DPC, DNA-protein crosslink; DSB, double-strand breaks.

Summary and Toxicological Implications

Cells with restored physiological levels of Asc recapitulated a limited transcriptional activity of p53 observed in response Cr(VI) in vivo (D’Agostini et al., 2002; Izzotti et al., 2004). Cr(VI)-stabilized p53 was impaired in upregulation of its proapoptotic targets and completely inactive in the induction of nucleotide excision repair genes XPC and DDB2. A lack of XPC upregulation is detrimental for genome stability of Cr(VI)-exposed cells as this factor acts as a sensor of DNA damage in global nucleotide excision repair, which removes Cr-DNA adducts (Reynolds et al., 2004). ATR, one of the apical DNA damage-sensitive kinases, was responsible for p53 activation, which occurred specifically in S-phase and was independent on the formation of secondary DNA lesions by MMR and phosphorylation of other ATR targets. Cr(VI) exposures are associated with the development of squamous cell lung carcinomas (Ishikawa et al., 1994; Satoh et al., 1997), which unlike tobacco smoke-induced or spontaneous cancers of this type (Kandoth et al., 2013), retained normal p53 tumor-suppressor (Kondo et al., 1997) despite their very high mutability (Hirose et al., 2002; Takahashi et al., 2005). The presence of wild-type p53 in Cr(VI)-induced lung cancers can be related to its weak proapoptotic transcriptional activity in response to Cr-DNA damage, which avoids a strong selection for cells with mutated p53 during chronic exposures.

DECLARATION OF CONFLICTING INTERESTS

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

FUNDING

This work was supported by grants ES008786 and ES028072 from the National Institute of Environmental Health Sciences.