Oral cancers represent a significant health problem, with over 270,000 new cases reported annually worldwide.1 Several risk factors have been identified for this form of cancer including, tobacco and alcohol use, advanced age, and decreased intake of fruits and vegetables.2, 3, 4, 5 Although the mechanisms by which these factors enhance cancer risk remain largely unclear, early oxidative damage is believed to play a key role.6, 7, 8 Free radicals generated in tobacco smoke have been implicated in tobacco-induced carcinogenesis at numerous sites within the oral cavity9 and, in similar fashion, ethanol-induced oxidative damage has been linked with oral carcinogenesis among chronic alcohol abusers.10 An accumulation of cellular insults resulting from oxidative damage is a hallmark of the aging process11 and may underlie the close relationship between aging and cancer. Finally, many of the nutritional factors linked to the development of oral cancers are associated with maintaining redox status within oral mucosa.12
The major intracellular antioxidant glutathione (GSH) represents a critical line of defense against oxidative stress.13 Numerous epidemiological and laboratory findings suggest that an inverse relationship exists between the levels of GSH and initial risk factors for cancer.14 In addition to protecting cells against oxidative damage, GSH, which is normally in high concentrations in cells and tissues, also detoxifies carcinogens by phase II conjugation15 and maintains immune function by regulating mitogenic response and lymphocytic proliferation.16 Furthermore, systemic depletion of GSH has been observed in tissues and organisms as they age and has been proposed as a potential mechanistic factor for enhanced susceptibility to carcinogenesis.17
To date, no epidemiological studies have examined the relationship between endogenous GSH and oral cancer. However, in studies of diet and oral cancer, an inverse association has been observed between increased dietary intake of GSH from fruits and vegetables and a decreased risk for cancer.18 In addition, dietary supplementation of GSH has been shown to reduce epidermoid carcinoma formation in the 7,12-dimethylbenz[a]anthracene (DMBA)-hamster cheek pouch model19 while simultaneously inhibiting angiogenesis and increasing expression of wild-type p53 protein.20 GSH supplementation acts synergistically with other antioxidants to reduce tumorigenesis in the same model.21
While high levels of GSH appear to be protective against cancer development, high levels of GSH have also been found in various tumor types.22 In clinical studies, GSH levels were approximately 4-fold greater in oral tumor tissue than that in the surrounding “normal” epithelium.23, 24, 25, 26 In one study, increased tumor GSH levels correlated with the stage of the disease.25 Also, increased GSH levels have been reported in cheek pouch tumors in the DMBA-induced hamster model.27 It has been proposed that the elevated levels of GSH in tumors may facilitate the propagation and protection of tumor cells.28, 29, 30, 31, 32 Indeed, overexpression of GSH in tumors has been related to increased resistance to chemotherapeutic drugs,15 protection against radiation,33 enhanced cell proliferation34, 35 and decreased levels of apoptosis.36
Protein glutathiolation, a major product of GSH oxidation in vivo, represents a potential mechanism by which oxidative stress can impact carcinogenesis. This process, which results in the formation of mixed disulfides of GSH with cysteinyl moieties on proteins, is enhanced during oxidative stress and has recently been recognized as a means by which to posttranslationally control the structure and function of redox-sensitive proteins associated with the regulation of diverse cellular and metabolic processes including cell proliferation and apoptosis.37, 38, 39, 40 Indeed, progressive glutathiolation of key proteins may serve as molecular switches by which cells respond to varying degrees of oxidative stresses in an immediate and reversible manner. To date, glutathiolation has been found to regulate numerous carcinogenesis-related proteins, including transcription factors, protein kinases, and other cell signaling proteins40 such as protein kinase C isozymes,41H-ras,42 annexin II,43 ubiquitin conjugating enzymes,44 thioredoxin,45 c-jun,46 NF-kB47 Pax-8,48 cAMP-dependant protein kinase (PKA),49 MEKK150 and mitochondrial complex I.51
To our knowledge, there have to date been no studies examining the levels of glutathiolated proteins in tumor tissue or during the course of carcinogenesis in the oral cavity, or at any site. In the present study, we have examined the levels of GSH, glutathiolated proteins and related thiols and disulfides in oral tumors and in oral epithelium during the course of carcinogenesis in a relevant oral cancer model. We have used the well-established 4-nitroquinoline-N-oxide (4-NQO)-induced F-344 rat tongue cancer model52 that has numerous advantages including the site of cancer induction at the dorsal end of the posterior tongue, a common site in human disease, and the observed spectra of preneoplastic and neoplastic lesions similar to those observed clinically.
DMBA, 7,12-dimethylbenz[a]anthracene; DTNB, 5,5′-dithiobis-(2-nitrobenzoic acid); γ-GCL, γ-glutamylcysteine ligase; GGT, γ-glutatmyltranspeptidase; γ-Glu-Cys, γ-glutamylcysteine; GSH, glutathione; GSSG, glutathione disulfide; GSSP, protein glutathiolation; MPA, metaphosphoric acid; 4-NQO, 4-nitroquinoline-N-oxide. The first two authors contributed equally to this paper.
Male F344 rats were obtained at 5 weeks of age (Charles River labs, Wilmington, MA). At 6 weeks of age (Week 1 of the experiment), rats were randomly divided into 2 groups. The experimental group (N = 16) was administered the carcinogen 4-NQO at a concentration of 20 ppm continuously in drinking water. 4-NQO treatment continued for a period of 8 weeks, after which time rats received tap water only. Control animals (N = 12) received tap water throughout the study. Rats were sacrificed at Weeks 16 and 32 (Fig. 1).
Sacrifice and tissue harvesting
Animals were sacrificed by CO2 asphyxiation between 8:00 am and 9:00 am to prevent interference due to circadian fluctuations. Tongues were immediately excised, rinsed in ice-cold saline, cut longitudinally into 2 sections and weighed. One section from each rat was fixed in 10% formalin and was paraffin embedded. Serial sections were cut for histopathology. From the other portions of the tongue, surface epithelia were removed by scraping with a microtome knife, and removal of only epithelial tissue was confirmed by histopathological examination of the remaining tongue tissue. Epithelial cells were immediately processed for GSH and protein glutathiolation measurements (as described later). When distinct tumors were present, they were removed and processed as described below. All procedures were performed at 0–4°C and tissues were processed immediately.
Tissues were weighed and homogenates (10% w/v) were prepared in ice-cold 5% (w/v) metaphosphoric acid (MPA) using a Ten-Broeck all glass homogenizer, kept on ice for 15 min, and centrifuged at 13,000g for 5 min. After removal of the resulting supernatant fraction, pellets were stored at −80°C until analysis of protein bound GSH. Supernatant fractions were stored at −80°C until analysis for free thiols and disulfides.
Measurement of GSH, GSSG and other thiols and disulfides
Total free GSH was determined using our modification53 of a 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB; Elman's Reagent) enzymatic recycling method originally described by Tietze.54 To measure GSH, glutathione disulfide (GSSG) and other relevant thiols and disulfides, we used our HPLC-dual electrochemical detection method.55
Protein bound GSH was determined in the acid-insoluble pellets derived from MPA-extracts.56 After washing 3 times by resuspension in 5% MPA and centrifugation, the pellets were resuspended in 8 M urea/1 mM EDTA and incubated 10 min at 40°C. Potassium borohydride was added to a final concentration of 35 mg/ml and the solution was incubated 45 min at 40°C. A few drops of octanol were added prior to the addition of potassium borohydride to reduce foaming. The solution was precipitated by 20% MPA for 15 min on ice. The mixture was then centrifuged at 13,000g for 15 min and the supernatant was stored at −80°C until analysis for GSH or other related thiols as described earlier.
All results were expressed as the mean ± SEM and were analyzed either by Student's t-test or one-way ANOVA where appropriate. The differences were considered statistically significant at p < 0.05.
At Week 16, in the rats treated with 4-NQO for 8 weeks, no tumors were observed, but preneoplastic lesions ranging from dysplasia to carcinoma in situ were observed (Table I and Fig. 2). Tongue hyperplasia was present in all 4-NQO-treated rats, whereas significant dysplasia was observed in 25% and carcinoma in situ in 12.5% of the rat tongues from 4-NQO-treated rats (Table I). At Week 32, all animals treated with 4-NQO developed large tumors in the posterior tongue (dorsal region). Histologically, they were identified as well-differentiated squamous cell carcinoma. Squamous cell papillomas as well as dysplastic and hyperplastic lesions were also present in these tongues (Data not shown). No lesions or regions of hyperplasia or dysplasia were observed in control rats after either 16 or 32 weeks.
Table I. Histopathological Lesions in Rat Tongues From Animals Sacrificed at Week 16
Percentage of animals with lesions (n)
Squamous cell carcinoma
At Week 16, total free GSH levels (GSH + GSSG) in the tongue epithelia of carcinogen-treated rats were >50% higher than those in control animals (p < 0.0001) (Table II). These differences in GSH levels were the same whether expressed per gram tissue or milligram protein. No significant differences in GSH levels were observed in liver or blood from these animals (Table III), consistent with the specificity of 4-NQO for oral epithelium. After 32 weeks, in tumors derived from rats given 4-NQO, GSH levels were 2.5-fold higher than those observed in the epithelia from control animals (p < 0.0001) (Table II). Also, in 4-NQO-treated rats, 2-fold higher levels of GSH were observed in tongue epithelium adjacent to regions expressing tumor (p < 0.0001). While the levels of GSH in adjacent tissues were approximately 20% lower that in the tumor itself, these differences did not reach a level of statistical significance.
Table II. Levels of Free Glutathione in Rat Tongue Epithelium During 4-NQO-Induced Carcinogenesis
Mean ± SEM
% of control
Mean ± SD
% of control
Significantly different from respective controls, p < 0.0001.
The level of GSH covalently bound to protein was assessed by measuring the levels of GSH released after potassium borohydride reduction of acid-insoluble pellets. At Week 16, the mean protein bound GSH content of tongue epithelia in control rats was 36.7 nmol/g tissue (0.189 nmol/mg protein) (Table IV) and represented about 3% of the total cellular GSH pool documented in Table II. The levels of protein bound GSH in the tongue epithelia from carcinogen-treated rats were 53.7 nmol/g tissue or 0.274 nmol/mg protein, representing increases of 46% and 31% compared with control rats, (p < 0.02). In tumors derived from rats given 4-NQO for 32 weeks, protein glutathiolation levels were 2-fold greater than those observed in tongue epithelia from control rats (p < 0.02) (Table IV). Comparable to that of unbound GSH, protein bound GSH levels were ∼50% greater than those observed in control animals. Differences in protein bound GSH were similar regardless of whether expressed on a tissue weight or protein weight basis.
Table IV. Enhanced Protein Glutathiolation in Rat Tongue Epithelium During 4-NQO-Induced Carcinogenesis
Mean ± SD
% of control
Mean ± SD
% of control
Significantly different from respective controls, p < 0.02.
In addition to GSH, the levels of other biologically relevant thiols and disulfides in tongue epithelia from all animals sacrificed after 16 weeks were analyzed by HPLC with dual electrochemical detection.55 Three major peaks were apparent in tongue tissues corresponding to cysteine, GSSG and γ-glutamylcysteine (γ-Glu-Cys). After 16 weeks, the levels of cysteine, the second most abundant low molecular weight thiol in cells and the limiting amino acid in GSH synthesis were increased significantly by 62% in carcinogen-treated animals compared with those in control animals (p < 0.05) (Table V). The levels of GSSG were increased 2-fold in carcinogen-treated animals compared with those in control animals after 16 weeks (p < 0.05). Since the levels of GSH and GSSG were similarly increased in 4-NQO treated rats, GSSG/GSH ratios were similar between carcinogen treated and control rats. There was a strong correlation between the levels of GSSG and glutathiolated protein in tongue tissues from rats sacrificed after 16 weeks (r = 0.88, p < 0.0001) (Fig. 3). No differences were observed in the levels of γ-Glu-Cys.
Table V. Changes in Glutathione–Related Thiols and Disulfides in Tongue Epithelium from 4-NQO-Treated Rats
We observed that GSH levels, as well as the levels of protein bound GSH, were enhanced in tongue tumors in the 4-NQO model. Although it is well known that GSH levels are upregulated in various tumor types, to our knowledge, this is the first report of enhanced protein glutathiolation (GSSP) during carcinogenesis. Further, we report that the upregulation of GSH and GSSP occurs during preneoplastic stages of carcinogenesis (hyperplasia, dysplasia) and long before the presence of clinically observable lesions.
While the mechanism(s) responsible for the observed increases in glutathiolation is unknown, it may be explained, in part, by the observed increase in GSSG levels. Indeed after 16 weeks, levels of GSSP were strongly correlated with those for GSSG in the tongue epithelium (Fig. 3). Thiol disulfide exchange between GSSG and free thiol groups on proteins is thought to be an important route of formation for GSSP. While increases in GSSG normally occur during periods of oxidative stress, it is unclear whether the increased levels of GSSG and GSSP observed during 4-NQO-induced tongue carcinogenesis are a result of continuous development of a pro-oxidative environment within these cells. No changes were observed in GSSG/GSH ratios, an often-used indicator of redox status, primarily due to the occurrence of a concomitant increase in both GSH and GSSG. Decreases in the activity of the thioredoxin/thioredoxin reductase system may also be involved due to its role in reducing protein mixed disulfides.57 However, in most tumor tissues examined to date, enhanced thioredoxin/thioredoxin reductase activity has been observed.58 Since acute administration of 4-NQO induces high levels of oxidative stress, it is possible that this can lead to increases in the levels of GSSP. However, acute administration of 4-NQO depletes GSH levels by conjugation.59 Moreover, enhanced levels of GSH and GSSP are observed 8 and 24 weeks after cessation of 4-NQO administration, long after the acute effects of 4-NQO would have occurred.
Given the importance of protein glutathiolation as a post-translational regulator of protein activity, it is possible that the regulation of key proteins by glutathiolation may be playing a role in the carcinogenic process. However, no information is yet available on the specific proteins that are glutathiolated or which proteins are more susceptible to glutathiolation during tumor development. While many studies have examined the glutathiolation of proteins in vitro, few studies have examined the specific proteins that are modified in vivo. Further, published reports on intact cells have focused primarily on cells in culture and few have examined protein thiolation in actual tissues.
A variety of findings support the direct involvement of GSH in inhibiting tumor initiation, promotion and progression. The induction of GSH-related detoxification is thought to be a primary mechanism by which numerous antioxidant-based chemopreventive agents work,60 including selenium,61N-acetylcysteine62, 63 and 2-oxothiazolidine-4-carboxylic acid64 and thiol-related compounds found in garlic such as S-allylcysteine.65 Additionally, topical application of GSH inhibited tumor progression in the murine skin multistage carcinogenesis model66 and dietary GSH produced a significant regression of aflatoxin B1-induced hepatocellular carcinoma in rats.67
The mechanisms by which GSH levels are upregulated during carcinogenesis are unknown. The overall regulation of GSH levels in cells can depend upon numerous factors including the availability of a critical precursor, cysteine, activity of a rate limiting de novo biosynthetic enzyme, γ-glutamylcysteine ligase (γ-GCL), redox cycling of GSH with GSSG disulfide and the transport of GSH or GSSG from cells via its numerous transporters. Our finding that cysteine levels are enhanced during carcinogenesis supports the possibility that increased synthesis may be responsible for the increase in GSH levels. High levels of cysteine have also been previously reported in human esophageal tumors.68 Previous studies indicated that the cystine transport system Xc− is involved in upregulation of intracellular GSH levels and drug resistance in ovarian cancer cell lines.69 γ-GCS has been reported to be upregulated in numerous cancers, including prostate cancer,70 multiple myeloma,71 nonsmall cell lung carcinoma,72 colorectal carcinoma73 and mesothelioma.74 In the present study, no changes in expression levels of the catalytic subunit of this enzyme were observed during carcinogenesis (data not shown); however, activity levels were not measured. γ-Glutatmyltranspeptidase (GGT), another enzyme involved in the metabolism of GSH, is also upregulated in various tumor types. This enzyme, which is the only known enzyme that can break the γ-glutamyl bond in GSH, is located on the external side of the plasma membrane and transfers the glutamyl residue from extracellular GSH to other amino acids, yielding cysteinylglycine which can subsequently be degraded to cysteine and glycine by a dipeptidase on the cell surface. These breakdown products can then be transported into the cells and reassembled into GSH.75 Many tumors express high levels of GGT,76 including those from the hamster cheek pouch oral cancer model77 as well as human oral squamous cell carcinoma.78
Increased GSH levels in tumor cells may provide these cells with a distinct growth advantage due to the role GSH plays in defending cells from exogenous and endogenous insults. The development of multidrug resistance is associated with an enhancement in GSH levels, as well as GSH conjugation and transport systems.79 Apoptosis, is another critical pathway involved in carcinogenesis which is dependent upon GSH levels in many cell types.36 In most cases, high GSH levels are thought to inhibit apoptosis, whereas GSH depletion has been linked with increases in apoptosis. Thus, it is possible that high levels of GSH in tumor tissues also serve to promote tumor growth by inhibiting apoptotic mechanisms.
The increases in GSH, GSSG and glutathiolated proteins during carcinogenesis were observed in the target tissue, tongue, but not in liver and in blood. This is consistent with the site specificity of 4-NQO in this model. It also suggests that blood levels of GSH or GSSP are not suitable as biomarkers of oral tumor development. However, it is possible that tongue or buccal scrapings may be more appropriate for this purpose.
Since GSH is the target for many putative clinically relevant chemopreventive agents as well as tumor-specific sensitizing drugs, information regarding the mechanisms and timing of GSH enhancement during tumorigenesis is of importance in the design of intervention strategies. The present results suggest that GSH depleting strategies that target preneoplastic cells may be beneficial in delaying the disease progression. Alternatively, caution may be warranted when using GSH enhancing agents in individuals with more advanced neoplastic disease. Further studies will be required to examine the possibility that GSH and GSSP induction is initiated even earlier during the process of carcinogenesis.