Glutathione S-transferase π expression in a pathological section of nasopharyngeal cancer by immunohistochemical staining. The glutathione S-transferase π–positive cells in this tumor section exhibit more intense nuclear than cytoplasmic staining (hematoxylin-eosin counterstain, original magnification ×540).
Glutathione S-transferase π expression in a normal nasopharyngeal tissue section by immunohistochemical staining. Glutathione S-transferase π immunopositivity was observed in basal cells and cilia (hematoxylin-eosin counterstain, original magnification ×540).
Relationship between glutathione S-transferase π (GST-π) immunohistochemical expression and total tissue iron levels in nasopharyngeal cancer tissues (Pearson r= 0.85) (A) and nuclear iron levels in nasopharyngeal cancer tissues (Pearson r= 0.76) (B).
Jayasurya A, Yap W, Tan N, Tan BK, Bay B. Glutathione S-Transferase π Expression in Nasopharyngeal Cancer. Arch Otolaryngol Head Neck Surg. 2002;128(12):1396-1399. doi:10.1001/archotol.128.12.1396
Glutathione S-transferase π (GST-π) is an enzyme that catalyzes the conjugation of electrophilic substrates and prevents oxidative damage. Although GST-π expression has been analyzed in many cancers, the significance of GST-π expression in nasopharyngeal cancer (NPC), a tumor with a high treatment failure rate, is still unclear.
To elucidate the significance of GST-π expression in NPC.
Evaluation of GST-π expression in NPC tissue specimens and determination of its relationship with tissue iron (a pro-oxidant) and clinicopathological factors in NPC.
Materials and Methods
Immunohistochemical expression of GST-π was carried out in 55 NPC and 4 normal nasopharyngeal tissue sections. Eleven nasopharyngeal biopsy specimens (4 normal and 7 NPC) were analyzed for tissue iron levels. The expression of GST-π in NPC was correlated with corresponding tissue iron levels. The relationships between GST-π expression with sex, race, tumor stage, cervical nodal status, and clinical staging were also analyzed.
Glutathione S-transferase π immunoreactivity was observed in all NPC sections, with the percentage of immunopositive cells ranging from 1.0% to 72.0%. Tissue iron levels were significantly higher in the NPC tissues compared with normal tissues (P = .001). A direct correlation was observed between GST-π expression and total and nuclear iron levels in NPC (P = .01 and P = .047, respectively). A significant association was also observed between GST-π expression and cervical nodal disease (P = .007).
Nasopharyngeal tumor cells may respond to pro-oxidant conditions by modulating intracellular antioxidant defense. Glutathione S-transferase π expression appears to be associated with lymphogenous metastasis in NPC.
GLUTATHIONE S-transferases (GSTs), a multigene family of enzymes, are known to play key roles in protecting cells from cytotoxic agents and in catalyzing the conjugation of electrophilic substrates with the tripeptide glutathione, thus preventing oxidative damage by an intrinsic organic peroxidation activity.1 As GSTs act as intracellular binding proteins for a variety of compounds, expression of this enzyme has been widely studied in both normal and tumor tissue. The 5 cytoplasmic forms of GST (ie, alpha, mu, sigma, pi, and theta) are categorized on the basis of their structural and functional characteristics.2 Glutathione S-transferase π is the most ubiquitous of the human GST enzymes,3 and it is believed that GST-π has an important constitutive function in tissue because of its widespread presence. Glutathione S-transferase π is the primary GST isozyme expressed in tumors.4 Overexpression of GST-π due to transcriptional activation or stabilization of the protein or messenger RNA may be a factor contributing to antitumor drug resistance.2
Iron, a transition metal, has been shown to promote cancer formation in experimental animal models.5 Increased iron stores in the body are known to elevate the risk of cancer because free iron has been reported to be mutagenic and carcinogenic.6 Dietary iron deficiency, on the other hand, is known to reduce tumor formation in murine models. Iron facilitates tumor growth, while redox cycling of iron produces reactive oxygen species,7 which in turn can induce DNA damage and initiate tumor formation.5 However, reactive oxygen species is also known to mediate oxidative stress-induced apoptosis and affects many functions integral to cellular homeostasis.8,9
The above findings prompted us to examine the relationship between the expression of GST-π with tissue iron levels and clinicopathologic characteristics in undifferentiated nasopharyngeal cancer (NPC), a tumor of epidermoid origin with the highest incidence in Southern China and Asia but uncommon in the Western population.10
A total of 59 patients who underwent postnasal biopsies were included in the present study. Fifty-five specimens were histologically confirmed to be undifferentiated NPC (World Health Organization type 3), while 4 showed normal histologic characteristics of the nasopharynx. All the biopsy specimens were fixed in 10% formalin before they were embedded in paraffin. Seven NPC and 4 normal biopsy specimens were also collected in 1.15% potassium chloride for elemental analysis. The tissues were relatively free of hemorrhage. The ages of the patients with NPC ranged from 26 to 81 years. There were 8 female NPC patients in the cohort. Thirty NPC patients had positive cervical lymph node enlargement, which were confirmed by computed tomographic scanning to be metastatic disease. The tumors were also clinically staged by the criteria of the American Joint Committee on Cancer (Table 1).
Four-micrometer sections were cut from the paraffin-embedded tissue specimens and immunostained, as previously described.11 The sections were dewaxed and rehydrated in graded series of alcohols. Endogenous peroxidase was quenched with hydrogen peroxide, and nonspecific binding of the antibodies were blocked by normal goat serum. Glutathione S-transferase π immunostaining was detected by incubating overnight at 4°C with primary GST-π antibody (Dako, Copenhagen, Denmark) at a dilution of 1:200. After incubation with avidin-biotin complex (Dako), the immunostains were developed in diaminobenzidine. The sections were then counterstained with hematoxylin-eosin for 20 seconds. Negative controls were obtained by omitting the primary antibody. Known GST-π positive breast cancer slides were used as positive controls. To assess the percentage of GST-π immunopositive cells, sections were analyzed and scored by one researcher (A.J.) under a light microscope (Zeiss Axioplan; Carl Zeiss, Oberkochen, Germany) using a ×40 objective linked to a monitor. A total of 10 high-power fields (with a microscopic field area of 8 µm2) were randomly chosen, and 100 to 150 cells were counted per 10 high-power fields.
For estimation of iron levels, the nasopharyngeal tissues (which were not contaminated by blood) were initially homogenized in 1.15% potassium chloride before overnight freeze-drying. Two and a half milliliters of 10mM nitric acid was then added to the homogenate, mixed thoroughly, and allowed to stand for 5 hours. After centrifugation at 38 000g for 30 minutes, the supernatant was then collected for estimation of total tissue iron levels by flame atomic absorption spectrometry.12 For subcellular fractionation, the tissues were cut and processed in 4 mL of cold homogenizing medium (0.25M sucrose, 5mM magnesium chloride, and 10mM Tris–hydrochloric acid buffer [pH 7.4]) as described by David et al.13 The nuclear fraction was collected by centrifuging the homogenate at 1000g for 10 minutes. The resulting supernatant was further centrifuged at 10 000g for 1 hour to pellet plasma membranes (representing crude lysosomal, mitochondrial, and microsomal fractions), with the final supernatant corresponding to the cytosol fraction. The nuclear pellet was purified by centrifuging at 60 000g for 1 hour through 2.2M sucrose, 1mM magnesium chloride, and 10mM Tris–hydrochloric acid buffer (pH 7.4). The subcellular fractions were then analyzed for iron after digestion with 10mM nitric acid for 5 hours. Protein level was estimated using the Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, Calif).
The commercially available GraphPad Prism software (GraphPad Software Inc, San Diego, Calif) was used for data analysis. Results are expressed as mean ± SEM. The t test was used to compare means and Pearson correlation for analysis of the relationship between variables. A P value less than .05 was considered statistically significant.
All 55 nasopharyngeal tumor tissues showed positive GST-π immunoreactivity. Glutathione S-transferase π staining was present in both the nucleus and the cytoplasm of the tumor cells (Figure 1). Infiltrating lymphocytes in the tumor sections exhibited negative immunoreactivity for GST-π. In the normal tissues, GST-π immunopositivity was observed in the basal cells and the cilia (Figure 2). Weak staining was also seen in the germinal centers within the normal nasal tissues. The percentage of GST-π immunopositive cells in NPC sections ranged from 1.0% to 72.0%, with a mean of 29.4% ± 3.1%. The association of GST-π expression with clinicopathological data is given in Table 2. A significant correlation was found between GST-π and cervical nodal metastasis (P = .007).
Total iron levels in the tumor tissues ranged from 1.5 to 8.88 µg/g of tissue. The total tissue iron level in cancer tissues (n = 7) was significantly higher than in the normal nasopharyngeal tissues (n = 4) (4.73 ± 1.11 vs 0.56 ± 0.12 µg/g of tissue; P = .001). The mean iron levels for the nuclear, membrane, and cytosolic fractions in the nasopharyngeal tumor tissues were 0.88 ± 0.02, 0.51 ± 0.08, and 1.69 ± 0.48 µg/g of protein, respectively. A positive correlation was noted between GST-π expression and total tissue iron levels in tumors (P = .01) (Figure 3A). A direct correlation between GST-π expression in NPC and nuclear iron levels was also observed (P = .047) (Figure 3B).
Glutathione S-transferase π expression has been reported to be present in many tumors such as primary breast carcinomas11 and cervical cancer.14 Glutathione S-transferase-π–positive staining has been demonstrated in squamous cell carcinoma of the pharynx and larynx.15 Similarly, GSTP1-subtype immunopositivity has also been observed in primary oral, oropharyngeal, hypopharyngeal, and laryngeal cancers, with differential expression of GSTP1 genotype frequencies.16 Recently, Chen and coworkers17 analyzed GST-π expression in primary and metastatic NPC.
The presence of cytoplasmic and nuclear GST-π immunostaining have been observed in tumors like primary breast carcinomas11 and sarcomas.4 In human uterine cervical lesions, GST-π has been shown to be expressed mainly in the cytoplasm of mildly dysplastic cells, whereas in the severely dysplastic cells, both nuclei and cytoplasm were strongly stained.14 In our NPC study, we demonstrated the presence of nuclear and cytoplasmic GST-π immunostaining. The presence of nuclear localization of GST-π, a cytosolic enzyme, is unlikely to be artifactual because (1) similar immunohistochemical findings were obtained using pure GST-π antibody prepared by GST-π antibody–bound affinity column chromatography,14 (2) immunoblotting of different subcellular fractions support the finding that GST-π is present within the nucleus,18 and (3) GST-π (with a molecular weight of 22 kd) has the capacity to gain access to the nucleus through the nuclear pore complex. Although the biological significance of nuclear localization has hitherto not been ascertained,4 nuclear localization of GST-π in undifferentiated NPC cells could be attributed to the presence of nuclear iron as evidenced by the direct correlation between nuclear iron levels and GST-π expression.
Elemental analyses of trace metals in neoplastic breast, prostate, esophageal, and nasopharyngeal tissues have revealed an association between malignancy and raised levels of elements such as zinc, cadmium, and copper.9,12,19 Earlier studies have also shown that the presence of excess iron increases mammary carcinogenesis.20 Normal and neoplastic cells have a qualitative requirement for iron, but an increased supply of iron is needed for the continuous proliferation of tumor cells.7 In the present study, the NPC tissues had mean tissue iron levels that were 8-fold higher compared with the normal nasopharyngeal tissues.
However, tumoricidal activity of the immune effector cells has also been reported to be markedly suppressed by iron salts, carbonyl iron, and iron dextran,21 and excessive iron is known to inhibit the activity of CD4 lymphocytes.22 This is especially relevant in undifferentiated NPC, a tumor characterized by a heavy T-cell lymphocytic infiltrate, which is composed of nonneoplastic elements and appears to have beneficial effects.23 In addition, free and catalytic forms of iron can mediate the production of reactive oxygen species, which produces tissue damage via the Fenton and Haber-Weiss reactions.8 Generation of reactive oxygen species can lead to lipid peroxidation and DNA fragmentation, the hallmark of apoptosis.
It has been previously observed by Misra and colleagues24 that nickel-induced lipid peroxidation in rat kidneys resulted in an increased renal iron level and enhanced GST activity. A direct correlation between GST-π and tissue iron levels in NPC observed in our study may represent an adaptive response of the tumor cells to the toxic metabolites produced by oxidative stress. This is supported by the fact that aldehydic products of lipid peroxidation are known to be effective inducers of an efficient detoxification system involving GST and glutathione.25 Although the significant correlation between GST-π expression with lymph node status in NPC observed in our study could be due to several factors, one such possibility is enhanced survival.
In conclusion, results of the present study clearly show the relationship between GST-π, an antioxidant, and iron, a pro-oxidant. The direct correlation of iron levels with GST-π expression could reflect the modulation of antioxidant defense in tumor cells as a protective mechanism against the deleterious effects of pro-oxidants. It also appears that GST-π expression has a direct correlation with lymphogenous metastasis in NPC.
Accepted for publication June 13, 2002.
This study was supported by grant 0244/1997 from the National Medical Research Council, Singapore.
We are grateful to Ms A. Hsu for technical assistance.
Corresponding author and reprints: Boon-Huat Bay, MBBS, PhD, Department of Anatomy, National University of Singapore, 4 Medical Dr, Blk MD 10, S117 597 Singapore (e-mail: email@example.com).