Research ArticleCancer

Antioxidants Accelerate Lung Cancer Progression in Mice

See allHide authors and affiliations

Science Translational Medicine  29 Jan 2014:
Vol. 6, Issue 221, pp. 221ra15
DOI: 10.1126/scitranslmed.3007653

Abstract

Antioxidants are widely used to protect cells from damage induced by reactive oxygen species (ROS). The concept that antioxidants can help fight cancer is deeply rooted in the general population, promoted by the food supplement industry, and supported by some scientific studies. However, clinical trials have reported inconsistent results. We show that supplementing the diet with the antioxidants N-acetylcysteine (NAC) and vitamin E markedly increases tumor progression and reduces survival in mouse models of B-RAF– and K-RAS–induced lung cancer. RNA sequencing revealed that NAC and vitamin E, which are structurally unrelated, produce highly coordinated changes in tumor transcriptome profiles, dominated by reduced expression of endogenous antioxidant genes. NAC and vitamin E increase tumor cell proliferation by reducing ROS, DNA damage, and p53 expression in mouse and human lung tumor cells. Inactivation of p53 increases tumor growth to a similar degree as antioxidants and abolishes the antioxidant effect. Thus, antioxidants accelerate tumor growth by disrupting the ROS-p53 axis. Because somatic mutations in p53 occur late in tumor progression, antioxidants may accelerate the growth of early tumors or precancerous lesions in high-risk populations such as smokers and patients with chronic obstructive pulmonary disease who receive NAC to relieve mucus production.

INTRODUCTION

Antioxidants including vitamins, carotenes, and minerals are found naturally in the diet and are added to food, cosmetic products, and pharmaceuticals. Antioxidants act as electron donors that neutralize reactive oxygen species (ROS) and other free radicals that may otherwise damage DNA and promote tumorigenesis (1). Consequently, popular wisdom—supported by numerous cellular and preclinical studies—holds that antioxidants protect against cancer (24). However, large randomized clinical trials have produced inconsistent results, and some studies indicate that antioxidants may even increase cancer risk (510). Moreover, recent genomic analyses of lung cancers have shown a high frequency of mutations in genes that activate an endogenous antioxidant program, suggesting that decreasing the amounts of ROS promotes tumor growth (11, 12). Consistent with this notion, experimental studies show that oncogenes such as K-RAS and B-RAF promote tumor growth by stimulating NRF2-mediated transcription of endogenous antioxidant genes (13, 14). Despite the striking discordance between the use of antioxidants and the lack of experimental support for their anticancer properties, no studies have yet examined their impact on tumor growth in state-of-the-art mouse models of cancer, including lung cancer—the most common form in humans (15).

RESULTS

Antioxidants accelerate tumor development in endogenous mouse models of lung cancer

To define the impact of antioxidants in lung tumorigenesis, we administered N-acetylcysteine (NAC) in the drinking water to mice harboring a Cre-inducible endogenous oncogenic Kras2LSL allele, 1 week after they inhaled a Cre adenovirus to activate K-RASG12D expression in lung epithelial cells. K-RASG12D mice develop multifocal tumors that vary in grade from epithelial hyperplasia and adenomatous hyperplasia to adenoma and adenocarcinoma (16). The mice were sacrificed 10 weeks after tumor initiation and were found to have 2.8-fold higher tumor burden than control mice (Fig. 1A and table S1). We next tested the effect of vitamin E, a structurally unrelated antioxidant. Vitamin E supplementation of the diet increased tumor burden in a dose-dependent fashion (Fig. 1B and table S1). To determine whether the effect of antioxidants on lung tumor growth occurs with a different oncogene and mouse line, we administered NAC and vitamin E to mice harboring a Cre-inducible endogenous oncogenic BrafCA allele (17). After Cre adenovirus inhalation, these mice express B-RAFV600E in lung epithelial cells and develop more tumors than K-RASG12D mice, but with a lower histological grade (17, 18). NAC increased the B-RAFV600E–induced tumor burden by 3.4-fold (Fig. 1C and table S1); the effect was similar in mice given the antioxidant starting 1 week before Cre adenovirus (Fig. 1C and table S1). In B-RAFV600E mice, vitamin E produced a dose-dependent effect similar to that in K-RASG12D mice (Fig. 1D and table S1). Moreover, NAC and vitamin E reduced the median and maximal survival of B-RAFV600E mice by 60 and 50%, respectively (Fig. 1E). Histological analyses revealed that the tumors of antioxidant-treated mice had a more advanced histological grade than tumors from control mice (Fig. 1F and fig. S1, A and B).

Fig. 1. NAC and vitamin E increase tumor growth in mice with K-RASG12D– and B-RAFV600E–induced lung cancer.

(A and B) Tumor burden (percent tumor area per lung area) in lungs from NAC-treated (A), vitamin E (Vit E)–treated (B), and littermate control K-RASG12D mice 10 weeks after inhalation of Cre adenovirus. NAC (1 g/liter) was administered in the drinking water and vitamin E (0.1 and 0.5 g/kg) in the chow diet, starting 1 week after Cre adenovirus inhalation. (C and D) Tumor burden in lungs from NAC-treated (C), vitamin E–treated (D), and littermate control B-RAFV600E mice 6 weeks after inhalation of Cre adenovirus. NAC (1 g/liter) was administered 1 week after or 1 week before Cre adenovirus. (E) Kaplan-Meier plot showing survival of NAC-treated, vitamin E–treated, and littermate control B-RAFV600E mice (n = 20 to 23 per group). (F) Left: Tumor stage (stages 1 to 4) in lungs from mice with K-RASG12D–induced lung cancer (n = 649 to 894 tumors in lungs from 17 to 33 mice per group). P values are for comparisons of the percentage of stage 3 and 4 tumors in antioxidant-treated and control mice. Right: Representative tumors in hematoxylin and eosin (H&E)–stained lung sections. (G) Ratio-ratio plot of K-RASG12D tumor transcriptome sequencing data showing overlapping regulation of genes by NAC and vitamin E (n = 10 tumors from five mice per group). Statistical analyses identified 310 and 905 differentially expressed transcripts (false discovery rate, q, <0.05) in the NAC- and vitamin E–treated tumors, respectively. The direction of expression change (up or down) was consistent for most genes that were significantly regulated by either antioxidant (gray data points) and for all genes that were significantly regulated by both (blue data points). Unbiased pathway analysis of genes down-regulated by both antioxidants identified the “Metabolism of xenobiotics by cytochrome P450” [Kyoto Encyclopedia of Genes and Genomes (KEGG), P = 1.8 × 10−6; red data points] category and the “Glutathione transferase activity” (GO; P = 2.4 × 10−4) category, which is largely a subset of the former. (H) Expression of eight endogenous antioxidant genes identified in the RNAseq analyses. Numbers in bars = n. Scale bar, 500 μm. *P < 0.05, ***P < 0.001, ****P < 10−30. Exact P values are provided in table S2. Graphical data are presented as means ± SEM.

Although both NAC and vitamin E are antioxidants, they have distinct molecular properties. Vitamin E is fat-soluble, regulates enzymatic activities, and is used as a dietary supplement, whereas NAC is water-soluble, participates in glutathione metabolism, and is used as a mucolytic agent (19, 20). We hypothesized that the marked overlap in the effects of NAC and vitamin E on tumor growth and survival reflects their common antioxidant properties. If so, NAC and vitamin E supplementation should affect similar molecular pathways. Indeed, transcriptome sequencing (RNAseq) of K-RASG12D tumors revealed that the transcriptional changes induced by NAC and vitamin E were highly overlapping (Fig. 1G and fig. S2, A and B). Pathway analyses revealed that the antioxidants suppressed the expression of genes that participate in the endogenous ROS defense system (Fig. 1, G and H, fig. S2C, and table S1). Those genes were also suppressed by antioxidants in normal lung tissue (fig. S2D and table S1). Verified target genes for the transcription factor NRF2 (21) were enriched among the suppressed genes (fig. S2E), but the percentage of NRF2 targets was low, suggesting that the transcriptional response to antioxidants is also mediated by other factors. Verified p53 target genes were not enriched among the suppressed genes (fig. S2F).

Antioxidants reduce ROS and DNA damage and increase tumor cell proliferation

To determine whether NAC and vitamin E suppress the amounts of ROS in tumors, we quantified fluorescence in lung sections stained with a redox-sensitive probe. In untreated mice, ROS was lower in lung tumors than in the surrounding normal tissue (Fig. 2A, fig. S3A, and table S1). This result is consistent with previous studies showing that oncogenes reduce ROS in tumor cells by activating endogenous antioxidants (14). ROS in tumors were further reduced by NAC and vitamin E (Fig. 2A, fig. S3A, and table S1). The reduced ROS was accompanied by increased ratios of reduced (GSH) to oxidized (GSSG) forms of glutathione (fig. S3B and table S1). Consistent with the reduced ROS, NAC and vitamin E reduced the amounts of ROS-induced DNA damage in lung tumors, as judged by immunohistochemical analyses with antibodies against 8-oxoguanine (Fig. 2B, fig. S3C, and table S1). Further immunohistochemical analyses revealed that tumors from antioxidant-treated mice contained more proliferating cells than did tumors from controls, as judged by the incorporation of 5′-bromo-2′-deoxyuridine (BrdU) and staining with antibodies against phosphorylated histone 3 (pH3) (Fig. 2C, fig. S3D, and table S1). The increased proliferation was not associated with increased signaling of the RAS–RAF–extracellular signal–regulated kinase (ERK) pathway, because the percentages of tumors with high amounts of phosphorylated ERK1/2 did not differ between groups (fig. S3E). Apoptotic cells, evaluated with TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling) staining and antibodies against cleaved caspase-3, were essentially undetectable in antioxidant-treated and control tumors (fig. S4, A and B). The amounts of senescent or quiescent cells, evaluated with antibodies against p16, p19ARF, and p21CIP1, were largely unaffected (fig. S4, C to E), although p19ARF in B-RAFV600E tumors was reduced by antioxidants (fig. S4E). When antioxidant supplementation was stopped 1 week before the mice were sacrificed, the numbers of proliferating cells in tumors were reduced compared to tumors of mice receiving antioxidants throughout the experiment (Fig. 2D and table S1). Thus, antioxidants increase tumor growth by reducing ROS and DNA damage and by promoting tumor cell proliferation.

Fig. 2. NAC and vitamin E reduce ROS and DNA damage and increase tumor cell proliferation.

(A) Left: Quantification of ROS in lung sections of antioxidant-treated and control K-RASG12D mice as judged by dichlorofluorescein (DCF) fluorescence with the redox-sensitive probe CM-H2DCFDA [5-(and-6-)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, acetyl ester] [n = 5 fields of view per lung; five lungs per group for normal lung tissue (NLT); n = 25 tumors from five mice for Ctrl, NAC, and vitamin E (Vit E)]. a.u., arbitraty units. Right: Representative micrographs showing DCF fluorescence. (B) Left: Quantification of 8-oxoguanine–positive cells in lungs of antioxidant-treated and control K-RASG12D mice (n = 2 to 3 tumors per lung; 10 lungs per group). Right: Representative immunohistochemistry micrographs showing 8-oxoguanine staining. (C) Left: Quantification of BrdU-positive cells in tumors from antioxidant-treated and control B-RAFV600E mice (n = 3 tumors per lung; five lungs per group). Right: Representative immunofluorescence micrographs showing BrdU-positive tumor cells. BrdU was injected into the peritoneal cavity 3 hours before the mice were sacrificed. (D) Quantification of BrdU-positive cells in lung tumors of K-RASG12D mice treated with NAC for 9 weeks and sacrificed at 10 weeks (NAC Stop). Control mice received NAC for 10 weeks. Scale bars, 100 μm. **P < 0.01, ***P < 0.001. Exact P values are provided in table S2. Graphical data are presented as means ± SEM.

NAC and a soluble vitamin E analog increase the proliferation of oncogene-expressing mouse fibroblasts

To determine whether NAC and vitamin E increase the proliferation of cultured cells, we incubated Kras2LSL/+ and BrafCA/+ primary mouse fibroblasts with a Cre adenovirus to activate the expression of K-RASG12D and B-RAFV600E, and then added NAC or the soluble vitamin E analog Trolox to the culture medium. The antioxidants did not affect the proliferation of the parental Kras2LSL/+ and BrafCA/+ fibroblasts (fig. S5A), but the proliferation of K-RASG12D and B-RAFV600E cells increased in a dose-dependent fashion (Fig. 3A). The antioxidants increased BrdU incorporation and the percentage of cells in the S phase of the cell cycle, but did not affect apoptosis (fig. S5, B to D, and table S1). The proliferation of fibroblasts transformed by c-MYC also increased in response to NAC and Trolox (fig. S6 and table S1). ROS analyses revealed that whereas oncogene expression transiently suppressed the amounts of ROS, NAC and Trolox caused sustained suppression of ROS (Fig. 3B and table S1). Thus, the response to antioxidants is similar in oncogene-expressing fibroblasts and lung tumor cells.

Fig. 3. Antioxidants increase tumor cell proliferation in vitro and in vivo by reducing p53 activation.

(A) Proliferation of primary K-RASG12D– (left) and B-RAFV600E–expressing (right) fibroblasts in medium supplemented with 250 μM (low) and 1 mM (high) NAC or 25 μM (low) and 100 μM (high) Trolox. Values are the mean proliferation of fibroblasts from three embryos per genotype assayed in triplicate. (B) The amounts of ROS in fibroblasts incubated for 4 and 12 days in medium supplemented with antioxidants, as judged by fluorescence-activated cell sorting (FACS) analyses of DCF fluorescence (n = 3 cell lines per genotype). (C and D) Western blots of fibroblast (C) and tumor (D) lysates with antibodies against total and phosphorylated forms of p53 and phosphorylated H2AX (γH2AX) and ATMSer1981. Actin and β-tubulin were used as loading controls. (E) Proliferation of primary p53-deficient K-RASG12D– (left) and B-RAFV600E–expressing (right) fibroblasts in medium supplemented with 1 mM NAC or 100 μM Trolox. The cells were generated by incubating Kras2LSL/+Trp53fl/fl and BrafCA/+Trp53fl/fl fibroblasts with Cre adenovirus. Control cells, designated as WT, were the parental cells incubated with a βgal adenovirus. Values are the mean proliferation of fibroblasts from three embryos per genotype assayed in triplicate. (F) Tumor burden in lungs of NAC-treated and littermate control conditional p53-deficient K-RASG12D (left) and B-RAFV600E (right) mice 10 weeks after inhalation of Cre adenovirus. Numbers in bars = n. *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in table S2. Graphical data are presented as means ± SEM.

Antioxidants reduce the amounts of p53 and thereby stimulate tumor cell proliferation

The tumor suppressor p53 regulates cell proliferation and is activated by ROS and DNA damage. We tested the possibility that reduced ROS and DNA damage in response to antioxidants would affect p53. Indeed, NAC and vitamin E markedly reduced the amounts of p53, as judged by Western blots of fibroblast and tumor lysates (Fig. 3, C and D). The amounts of phosphorylated H2AXSer139 and ATMSer1981, markers of DNA damage, were reduced in lysates of antioxidant-treated cells and tumors (Fig. 3, C and D), consistent with the reduced amounts of 8-oxoguanine in tumors (Fig. 2B, fig. S3B, and table S1). To determine whether p53 is required for the antioxidant effect, we bred Kras2LSL/+ and BrafCA/+ mice on a background of a conditional p53 knockout allele (Trp53fl/fl) and used Cre to simultaneously activate oncogene expression and inactivate p53. Inactivation of p53 abolished the ability of antioxidants to increase the proliferation of oncogene-expressing fibroblasts in vitro and tumor cells in vivo (Fig. 3, E and F, table S1, and fig. S6D).

The impact of antioxidants is similar in human and mouse tumor cells

We next tested the impact of antioxidants on human lung cancer cell lines. NAC and Trolox increased the proliferation of cells expressing wild-type p53, but not that of cell lines with p53 mutations (Fig. 4, A and B). In wild-type p53 cell lines, the antioxidants reduced ROS, increased BrdU incorporation and the percentage of cells in S phase, had no impact on apoptosis, and reduced DNA damage and p53 expression (Fig. 4, C and D, fig. S7, A to C, and table S1). p53 in a cell line expressing a mutant form of the protein was not reduced by antioxidants (Fig. 4E). Knockdown of p53 expression with lentiviral short hairpin RNAs (shRNAs) in wild-type p53 cell lines abolished the antioxidant effect (Fig. 4F, fig. S8, and table S1). Thus, human and mouse tumor cells respond similarly to antioxidants.

Fig. 4. Antioxidants reduce ROS, DNA damage, and the amounts of p53 and increase the proliferation of human lung cancer cell lines.

(A and B) Proliferation of human lung cancer cell lines expressing wild-type (A) or mutant (B) p53 in medium supplemented with 1 mM NAC or 100 μM Trolox. Values are the means of triplicate analyses per cell line. (C) Amounts of ROS in human lung cancer cell lines incubated for 4 days in medium supplemented with antioxidants, as judged by FACS analyses of DCF fluorescence (n = 4 analyses per cell line). (D) Western blots of lysates of cancer cell lines (wild-type p53) with antibodies against total and phosphorylated forms of p53 and phosphorylated H2AX (γH2AX). β-Tubulin was used as a loading control. (E) Western blots of lysates of cancer cell lines (mutant p53) with antibodies against p53. β-Tubulin was used as a loading control. Note that H358 and H1299 are p53-deficient and that the amount of p53 in H23 is not reduced by antioxidant treatment. (F) Proliferation of cell lines from the experiment in (A), incubated with lentiviruses expressing shRNAs targeting TP53 or containing a scrambled (SCR) sequence, in medium supplemented with NAC or Trolox. Values are the means of triplicate analyses per cell line and experiments with two different shTP53 lentiviral clones. *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in table S2. Graphical data are presented as means ± SEM.

DISCUSSION

This study demonstrates that antioxidant supplementation of the diet reduces ROS and DNA damage, prevents p53 activation, and markedly increases tumor cell proliferation and tumor growth in mice. The data demonstrate that tumor cells proliferate faster when oxidative stress is suppressed. This reasoning is consistent with previous studies showing that oncogenes stimulate NRF2-mediated expression of endogenous antioxidants, reduce ROS, and thereby increase tumor cell proliferation (13, 14, 22).

The antioxidants reduced the expression of genes involved in the endogenous ROS defense system. This result is consistent with the reduced amounts of ROS and oxidative DNA damage and the increased GSH/GSSG ratio. The simplest explanation for this result is that a feedback mechanism in lung cells down-regulates the endogenous ROS defense system when the amounts of ROS are suppressed by NAC or vitamin E.

Our data do not support a direct role for the reduced expression of endogenous antioxidant genes in the increased tumor growth. Instead, several lines of evidence suggest that reduced amounts of p53 mediate the antioxidant-induced increase in tumor growth. First, NAC and vitamin E reduced p53 in tumors and cultured mouse and human tumor cells. Second, antioxidants increased the proliferation of human lung cancer cells with wild-type, but not mutant, p53. Third, the ability of antioxidants to increase tumor cell proliferation was abolished when p53 was inactivated or suppressed by shRNAs. One potential explanation for the reduced amounts of p53 is that the antioxidants reduced oxidative DNA damage, γH2AX, and phospho-ATM and thereby removed potent stimuli for p53 activation and stabilization. However, we cannot rule out the possibility that additional factors are involved.

One limitation of the study is that the K-RAS and B-RAF models only allow us to study the impact of antioxidants on tumor progression, and not tumor initiation or prevention. In previous studies, antioxidants were protective against chemically induced lung cancer, and it is possible that high amounts of ROS are required for tumor development in that setting (23, 24). However, experimental studies and large clinical trials quite convincingly suggest that antioxidants, including isoflavones, carotenes, vitamins, and NAC, should not be recommended for the prevention of lung cancer and that their use may promote tumor growth (10, 2527).

Another limitation is that although antioxidants accelerated the proliferation of human lung cancer cell lines via similar mechanisms as in mouse tumor cells, the precise clinical relevance of our findings is not yet clear. We speculate that because the antioxidant effect was dependent on p53, and TP53 mutations in humans are believed to occur late in tumor progression (28), antioxidants may accelerate the progression of early tumors and precancerous lesions. This would suggest that antioxidants are unsafe in patients with early stages of lung cancer and in people at risk of developing the disease. For example, this may be relevant to patients with chronic obstructive pulmonary disease, who are often smokers with an increased risk of developing lung cancer and ingest high amounts of NAC to relieve mucus production.

MATERIALS AND METHODS

Study design

The objective of the study was to define the impact of antioxidant supplementation on tumor progression, severity, and lethality in mouse models of endogenous lung cancer. Littermate mice were randomized to antioxidant and control groups. Antioxidants were administered 1 week after the induction of lung cancer, and mice were euthanized 8 to 10 weeks later. For survival experiments, mice were sacrificed when they were moribund, defined as when they became listless and exhibited ruffled fur and shortness of breath. The impact of antioxidants on tumor growth, amounts of ROS, DNA damage, redox state of glutathione, cellular proliferation, apoptosis, and signaling was monitored by routine histology, immunohistochemical methods, and enzyme-linked immunosorbent assays. The impact of antioxidants on downstream signaling and protein and gene expression was further evaluated by Western blotting, RNA sequencing, and TaqMan analyses in tumor biopsies, cultured mouse fibroblasts, and human lung cancer cell lines. Tumor analyses, RNAseq, Western blots, immunohistochemistry quantifications, and cell experiments were performed in a blinded fashion.

Mice and tumor induction

Six- to 8-week-old Kras2LSL/+ (16) and BrafCA/+ (17) mice on Trp53+/+ or Trp53fl/fl (29) backgrounds were allowed to inhale a Cre adenovirus [5 × 107 plaque-forming units (PFU) for experiments on the Trp53+/+ background and 5 × 106 PFU for the Trp53fl/fl background] under general anesthesia, as described (16). NAC (616-91-1, ≥99% purity, Sigma) was administered in the drinking water (1 g/liter). Vitamin E (dl-α-tocopheryl acetate; 7695-91-2, Zhejiang Medicine Co.) was administered in chow pellets (Lantmännen) at doses of 0.1 and 0.5 g/kg chow (12.5 and 61.5 mg/kg body weight), calculated on the basis of observed daily food intake. Control mice were always littermates. NAC and vitamin E supplementation did not alter body weight or food and water intake. Mice had a mixed genetic background (129SV and C57BL/6). Animal experiments were approved by the Research Animal Ethics Committee in Gothenburg.

Histology and immunohistochemical analyses

For routine histology, 4-μm sections of paraformaldehyde inflation–fixed, paraffin-embedded lungs were stained with H&E. Immunofluorescence and immunohistochemical analyses were performed as described (30). The sections were incubated with antibodies recognizing 8-oxoguanine (1:1000; 45.1, Abcam), pH3 (1:400; 9714), phosphorylated ERK1/2 (pERK, 1:100; 9106, Cell Signaling Technology), p16 (1:500; sc-1661), p19ARF (1:100; sc-32748), and p21CIP1 (1:50; sc-397, Santa Cruz Biotechnology), and then processed with the Vectastain Elite ABC Kit (PK6101) and the DAB Peroxidase Substrate Kit (SK4100, Vector Laboratories). Cells positive for BrdU were visualized as described (31). Apoptosis was detected by TUNEL staining (ApopTag Fluorescein In Situ Apoptosis Detection Kit, s7110, Millipore) and by immunohistochemistry with an antibody to cleaved caspase-3 (1:200; 9661, Cell Signaling Technology). Sections were examined on a Leica TCS SP5 microscope with LAS Advanced Fluorescence software (version 2.0.2, Leica Microsystems) and on a Zeiss Axio-Imager M1 with AxioVision software (version 4.6, Carl Zeiss).

Quantification of tumor burden and immunostaining

Tumor burden (percent tumor area per lung area) in H&E-stained sections of all five lung lobes on three levels separated by 2 mm was quantified with Biopix iQ software (version 2.3.1., Biopix). Tumor stage was quantified in the same sections. K-RASG12D tumors were classified into four stages (1, epithelial hyperplasia; 2, atypical adenomatous hyperplasia; 3, adenoma; and 4, adenocarcinoma), and B-RAFV600E tumors into three (stages 1 to 3), as described (18, 32). For quantification of BrdU and pH3 staining, the total numbers of positively stained cells in tumors were counted and normalized to tumor area; for 8-oxoguanine staining, tumors were photographed, the images were converted to grayscale, and the total numbers of cells with positive nuclear staining were counted and normalized to tumor area [adapted from (14)]; for pERK staining, the numbers of tumors with low and high staining were scored as described (33).

Transcriptome sequencing and TaqMan

RNA from 10 NAC-treated, 10 vitamin E–treated, and 10 untreated tumors was isolated with the RNeasy Plus Mini kit (Qiagen). The tumors were isolated from 15 mice (2 tumors per mouse). The polyadenylated RNA fraction was sequenced with an Illumina HiScanSQ instrument using the TruSeq RNA Sample Preparation v2 kit. Sequencing reads were mapped to the NCBI37/mm9 mouse genome assembly with TopHat aligner (34), and genes in the University of California Santa Cruz annotation (35) were quantified with the HTSeq-count tool (http://www-huber.embl.de/users/anders/HTSeq). For TaqMan analyses, complementary DNA (cDNA) was synthesized from tumor RNA with the iScript cDNA synthesis kit (170-889, Bio-Rad), and expression of Aldh3a1, Aldh3b1, Cyp2s1, Gsta2, Gsta3, Gsta4, Gstm2, Gpx2, Myc, Gapdh, and Actb was analyzed by quantitative reverse transcription polymerase chain reaction on a ABI 7900HT (Life Technologies) using the probe sets Mm00839312_m1, Mm00550698_m1, Mm00512037_m1, Mm00833353_mH, Mm01233706_m1, Mm00494803_m1, Mm01199654_gH, Mm00850074_g1, Mm00487804_m1, Mm99999915_g1, and Mm00607939_s1, respectively.

ROS and glutathione

For analyses of tissue ROS, lungs were frozen in optimum cutting temperature compound, cryosectioned, and incubated with 5 μM CM-H2DCFDA (C6827, Life Technologies), a DCF-labeled redox-sensitive probe, for 90 min at 37°C. Randomly selected areas were photographed with the same exposure time. Mean signal intensity normalized to tissue area was quantified after conversion to grayscale images with ImageJ 1.48b [National Institutes of Health; adapted from (14)]. ROS in cultured cells were measured by incubating 106 cells with 5 μM CM-H2DCFDA for 30 min at 37°C. DCF fluorescence was then analyzed by FACS (Accuri C6, BD Biosciences). Glutathione in reduced (GSH) and oxidized (GSSG) forms was measured in tumor lysates with the Bioxytech GSH/GSSG-412 kit (OxisResearch).

Cell culture and cell proliferation

Primary mouse fibroblasts were isolated from E13.5 to E14.5 embryos, and experiments were done with cells at passages 2 to 3. For adenovirus transduction, 106 cells were seeded in 100-mm dishes and incubated for 24 hours with 10 multiplicity of infection (MOI) adenoviruses expressing Cre recombinase or β-galactosidase (Ad5CMVCre and Ad5CMVntLacZ, Gene Transfer Vector Core, University of Iowa) as described (36). Human lung cancer cell lines were from the American Type Culture Collection. Proliferation assays were carried out by plating 104 cells per well on 12-well plates. For each data point, the cells were trypsinized and counted. NAC and Trolox (6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid; cat. no. 53101-49-8, Sigma) were added to the culture medium 6 hours after plating. MYC-transformed primary mouse fibroblasts were generated with a retrovirus encoding c-Myc (MSCV-MYC-IRES-GFP), as described (37). For lentiviral experiments, 106 cells were incubated for 72 hours with 20 MOI lentiviruses expressing shRNAs targeting TP53 (clone 1: TP53 TRCN0000003756; clone 2: TP53 TRCN0000003753; Sigma-Aldrich) or a scrambled (SCR) sequence (SHC002V; Sigma-Aldrich). BrdU incorporation and analyses of percentage of cells in the S phase of the cell cycle were performed as described (31).

Western blots

Lysates (10 to 20 μg) from cells and tumor pieces were resolved on 4 to 12% bis-tris gels (Bolt, Life Technologies), transferred onto nitrocellulose membranes, and incubated with antibodies as described (38). Primary antibodies were total p53 (fl393, Santa Cruz Biotechnology), phospho-p53Ser15 (9284, Cell Signaling Technology), p21CIP1 (F-5, Santa Cruz Biotechnology), γH2AX (05-636, Millipore), phospho-ATMSer1981 (NB100-306, Novus Biologicals), c-MYC (9402, Cell Signaling Technology), actin (2228, Sigma-Aldrich), and β-tubulin (926-68072, Li-Cor). Secondary antibodies were anti-mouse IRDye 680RD (926-68072) and anti-rabbit 680RD (926-68071, Li-Cor). Protein bands were detected on a Li-Cor Odyssey Imager with Odyssey software (version 3.0, Li-Cor).

Statistics

Values are presented as means ± SEM. For statistical analyses, we used GraphPad Prism software v.6.03: one-way analysis of variance (ANOVA) with Tukey’s post hoc test for tumor burden and the amounts of ROS in tissues; Fisher’s exact test for tumor grade and enrichment of NRF2 and p53 target genes; the log-rank test for survival; χ2 test for ERK and p19ARF staining in tumor sections; one-way ANOVA with Dunnett’s post hoc test for 8-oxoguanine, BrdU, and pH3 staining of lungs, TaqMan analyses, cells in S-phase, BrdU incorporation, and the amounts of ROS and the redox state of glutathione in cells; and two-way ANOVA for cell proliferation. Tests for differential expression of genes identified in RNAseq analyses were performed with DESeq (39) with a q value cutoff of 0.05 and excluding low-abundance genes (lowest 60% quantile). To test for enriched pathways among 144 genes that were significantly repressed by both antioxidants, we used the online version of GSEA (gene set enrichment analysis) (40).

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/6/221/221ra15/DC1

Fig. S1. Antioxidant supplementation increases tumor stage in mice with B-RAFV600E–induced lung cancer.

Fig. S2. Administration of antioxidants to mice with K-RASG12D–induced lung cancer reduces tumor expression of endogenous antioxidant genes.

Fig. S3. NAC and vitamin E reduce ROS and DNA damage and increase tumor cell proliferation.

Fig. S4. Antioxidants do not affect the amounts of apoptotic and senescent cells in tumors of mice with K-RASG12D– and B-RAFV600E–induced lung cancer.

Fig. S5. NAC and Trolox increase the proliferation of fibroblasts expressing oncogenic, but not wild-type, K-RAS and B-RAF.

Fig. S6. Antioxidants increase the proliferation of MYC-transformed fibroblasts.

Fig. S7. Antioxidants increase the proliferation of human lung cancer cell lines expressing wild-type p53.

Fig. S8. TP53 is required for antioxidants to increase the proliferation of human lung cancer cell lines.

Table S1. Original data (Excel spreadsheet).

Table S2. Exact P values (Excel spreadsheet).

Reference (41)

REFERENCES AND NOTES

  1. Acknowledgments: We thank L. Pehlivanoglu for technical assistance, S. Ordway for editorial assistance, O. Hammarsten for materials and discussions, D. Dankort and M. McMahon for providing BrafCA/+ mice, and Histo-Center AB for histology. Funding: This study was supported by grants from the European Research Council, the Göran Gustafsson Foundation, the Swedish Children’s Cancer Fund, BioCARE—a National Strategic Research Program at the University of Gothenburg, and Ingabritt and Arne Lundberg’s Research Foundation (to M.O.B); the Swedish Cancer Society and the Swedish Research Council (to E.L., J.A.N., P.L., and M.O.B.); Västra Götalandsregionen (to J.A.N., P.L., and M.O.B.); the Heart and Lung Foundation and Polysackaridforskning AB (to P.L.); Assar Gabrielsson Foundation (to V.I.S. and E.L.); and the Magnus Bergvall, Åke Wiberg, and Lars Hierta Memorial Foundations (to E.L.). Author contributions: V.I.S.: experimental design, performance, evaluation, and manuscript preparation; M.X.I.: performance, evaluation, and figure preparation; E.L.: experimental design, evaluation, and bioinformatics; J.A.N.: experimental design and MYC experiments; P.L.: project development and coordination, experimental design, and evaluation; M.O.B.: project development and coordination, experimental design, evaluation, and manuscript preparation. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The RNAseq data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE52594.
View Abstract

Navigate This Article