Oxidative stress in non-small cell lung cancer: Role of nicotinamide adenine dinucleotide phosphate oxidase and glutathione

Abstract

Background. Cigarette smoke is strongly associated with NSCLC, but the carcinogenesis of NSCLC is poorly understood. Methods. To discover the role of oxidative stress and anti-oxidative defense in NSCLC, we measured NADPH oxidase (NOX) activity, myeloperoxidase activity, 8-OHdG, and glutathione content from lung specimens. These came from 32 patients: 22 NSCLC patients and ten controls without cancer. Results. In NSCLC patients, NOX activity was significantly higher both in the malignant (p = 0.001) and non-malignant (p = 0.044) samples from NSCLC patients, than in the control specimens. Myeloperoxidase activity was lower (p = 0.001) and glutathione content (p = 0.009) higher in malignant tissue. No significant difference was observable in 8-OHdG content between patient groups. Conclusions. Increase in NOX activity in the malignant tissues was independent of smoking history and myeloperoxidase activity, suggesting its independent role in NSCLC pathogenesis.

The main cause of non-small cell lung cancer (NSCLC) is cigarette smoke (CS). It has been hypothesized that CS induces carcinogenesis mainly by an oxidant-mediated mechanism. CS contains high levels of oxidants and oxidant-generating compounds. It increases oxidant generation by inflammatory and non-phagocytic cells [1], it contributes to the persistence of chronic inflammation, and it modulates tissue antioxidant capacity [2]. An excess of reactive oxygen species (ROS) can cause multiple alterations in cellular molecules [3], which in turn alter cells’ permanent growth arrest, apoptosis [4], and angiogenesis [5]. ROS have been connected with carcinogenesis in a number of different malignancies [3].

Endogenous ROS are mostly produced by activated inflammatory cells such as polymorphonuclear neutrophilic leukocytes (PMNs) [6]. For example, the number of PMNs in the pulmonary circulation is 35 – 100 times as high as in the systemic circulation. The most abundant protein in PMNs is myeloperoxidase (MPO), a lysosomal enzyme producing hydrochlorous acid (HOCl) for defense against various micro-organisms [7]. Increased MPO activity in the lung or lung secretions has been identified in smokers and in chronic obstructive pulmonary disease (COPD) [8].

Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) is the initial enzyme in ROS production, transporting electrons to reduce oxygen to superoxide. Besides the classical NOX present in the phagocytes, to date seven isozymes have been detected in structural non-phagocytic mammalian cells [9]. Initially NOX was considered an inflammatory component of host defense, but the NOX family of enzymes has also been shown to be an integral part of cell signaling, apoptosis, angiogenesis, and differentiation [9]. The NOX isoforms expressed in lung tissue include NOX2, DUOX1, and NOX4 [9], [10].

The redox state in the lung is controlled by complex and cell-specific antioxidative mechanisms including primary antioxidant enzymes, thiol-containing redox modulatory proteins, detoxification enzymes, and small molecular-weight antioxidants [3], [6]. Among the latter, in the human lung, one of the most interesting is glutathione (GSH) (L-γ-glutamyl-L-cysteinyl-glycine), an intracellular free thiol containing tripeptide; its levels in airway secretions are even 140-fold higher in the lungs than in the circulating blood [11]. GSH participates in numerous antioxidant and detoxification reactions [12]. Glutamate cysteine ligase is the rate-limiting enzyme of GSH synthesis, which has been shown to be down-regulated in COPD [13] but to be highly elevated in lung cancer [14]. Studies strongly suggest that increased intracellular concentrations of GSH and the levels of other thiol-containing antioxidants are linked both to apoptosis and to tumor-cell chemoresistance [15], [16].

In many malignancies, footprints of oxidative damage, the markers of the imbalance between oxidants and antioxidants, have been detectable [17]. Among such markers, 8-oxo-hydroxydeoxyguanosine (8-OHdG), the most common form of DNA oxidation product produces G to T tranversions [18]. It induces spontaneous mutations [19], thus also proving a significant factor in carcinogenesis [20] and a predictor of relative oxidative exposure [21]. Chronic inflammation is linked to carcinogenesis, and activated neutrophils induce 8-OHdG formation in pulmonary epithelial cells [22].

The present study was undertaken to investigate in NSCLC two potential oxidant-producing enzymes, NOX and MPO, a major small molecular weight antioxidant of the lung, GSH, and a marker of oxidative damage, 8-OHdG, by comparing the activity or level of these components between the malignant and non-malignant areas of the same lung and between subjects with or without NSCLC. Additionally these markers were compared between squamous cell carcinoma (SCC) and adenocarcinoma (AC).

Patients and methods

The study material included lung specimens from 32 Caucasian patients (Table I), of whom 22 underwent lobectomy via thoracotomy because of NSCLC between January 2003 and May 2006 at the Helsinki University Central Hospital. None of these patients received prior neoadjuvant chemo- or radiotherapy. All NSCLC patients, with the exception of one AC patient, had smoked over 25 pack years. This value is calculated by multiplying the number of packs of cigarettes smoked per day by the number of years the person has smoked. Five AC and three SCC patients had quit smoking at least a year prior to the operation. Co-morbidities included hypertension in nine, hypercholesterolemia in eight, COPD in six, and diabetes in three patients. Patients with hypercholesterolemia had been prescribed daily statin medication, and COPD patients used inhalable corticosteroids. Five patients took a daily non-steroidal anti-inflammatory drug (NSAID), 100 mg acetylsalicylic acid.

Table I.  Patient groups.

		Control	NSCLC	Adenocarcinoma	Squamous cell carcinoma
n		10	22	11	11
Age, years,		38.5 (16–71)	62.5 (40–82)	67.0 (56–82)	60.0 (40–73)
median (range)
Male/Female		7/3	16/6	7/4	9/2
Pack years smoked, median (range)		0 (0–45)	40 (0–100)	30 (0–50)	50 (37–100)
Stage	IA		5	2	3
	IB		8	3	5
	IIA		0	0	0
	IIB		5	2	3
	IIIA		4	4	0

NSCLC = Non-small cell lung cancer.

Ten patients were included as controls; they had undergone surgery for benign lung disease: seven had recurrent pneumothorax, two had a lung hamartoma, and one, vascular anomaly. Six of the control patients were never-smokers. None of the control patients had any daily medication before the surgery. To none of the pneumothorax patients had been applied talc mixture or doxycycline before surgery. All controls underwent wedge resection by video-assisted thoracosopy.

The Ethics Committee of Helsinki University Central Hospital approved this protocol.

Samples

Samples of lung carcinomas and adjacent non-malignant lung tissue were obtained during surgery. Each non-malignant lung sample was taken 50 millimeters distal from the malignant tumor in the resected lobe. In the control group, samples came from the most normal-appearing part of the resected lung tissue. Specimens for all analyses were immediately frozen in liquid nitrogen and stored at −80°C. Histopathological specimens were taken adjacent to the same tumor as were fresh-frozen samples. Histopathological specimens were stored in paraffin. These were examined by one of the authors (K.S.) by standard histology and immunohistochemistry.

Analysis of NOX and MPO activities

Total NOX activity was measured by lucigenin-enhanced chemiluminescent detection of superoxide anion essentially as described [23]. In brief, the homogenate was diluted in 50 mM potassium phosphate buffer (pH 7.4). To this solution was added 25 µM dark-adapted lucigenin. The reaction was initiated by 200 µM NADPH. Chemiluminescence in duplicate cuvettes was recorded for 20 min in 1-minute cycles by a Bio-Orbit 1251 Luminometer. The mean area under the curve was automatically counted to give an integral value. Results are expressed as mV/mg protein.

MPO activity was determined by modification of the method of Suzuki, in which the enzyme catalyzes the oxidation of 3,3′, 5,5′-tetramethylbenzidine by H₂O₂ to yield a blue chromogen with a maximum wavelength of 655 nm [24]. MPO activity is expressed as units/milligram protein (U/mg protein).

Free glutathione

GSH content, expressed as nmol/mg protein, was estimated by Saville's method [25].

Analysis of 8-oxo-hydroxydeoxyguanosine (8-OHdG)

DNA was isolated by a non-enzymatic method: pure DNA was dissolved in high performance liquid chromatography (HPLC)-grade water, with deferoxamine mesylate added to reduce artificial oxidation. DNA was hydrolyzed to nucleotides on incubation with nuclease P1, further hydrolyzed to nucleosides with alkaline phosphatase, and the nucleosides were separated by C18 reverse-phase column. The amount of 8-oxo-dG was determined with HPLC equipped with an electrochemical detector; deoxyguanosine (dG) was determined with a UV detector, with 8-OHdG concentration is expressed as the ratio of 8-oxo-dG per 10⁵ dG (8-oxo-dG/10⁵ dG).

Statistical analyses

Values, unless otherwise stated, are expressed as median and range. Comparison between independent groups was performed by the Mann-Whitney U-test. Comparison between malignant and non-malignant lung samples was done by Wilcoxon rank sum tests. Spearman's correlation coefficient was used to examine the relationship between variables. Non-parametric methods were applied, due to non-normalities in the data. All reported p-values were based on two-tailed tests. Statistical calculations were carried out with SPSS 16.0.1 software. A p-value of < 0.05 was considered significant.

Results

NOX activity

The lowest NOX activity occurred in the control samples, the value being significantly lower than the value of the non-malignant areas in NSCLC (p = 0.044) (Table II). NOX activity was still higher in the malignant NSCLC specimens when both SCC and AC were combined, compared to activity in the same patients’ non-malignant lung specimens (p = 0.001) or compared to control subjects’ lung tissue (p = 0.001). When the cancer subtypes were compared to the non-malignant lung tissue of the same patients, NOX activity appeared to be highest in AC (p = 0.005) (Table III). This activity did not significantly differ, however, between AC and SCC. Overall these results suggest that a considerable part of the NOX activity is related to the malignant tissues in NSCLC.

Table II.  Descriptive statistics with p-values by parameter and group

		NSCLC 22
Parameter n	Control 10	non-malignant	malignant
NADPH oxidase (mV/mg)
Median	0.94	1.83	7.30
(range)	(0.28–2.69)	(0.26–19.67)	(0.21–47.20)
P-value^1		0.044	0.001
P-value^2			0.001
Myeloperoxidase (U/mg)
Median	4.60	7.60	1.60
(range)	(0.40–50.50)	(0.00–23.30)	(0.00–12.50)
P-value^1		NS	NS
P-value^2			0.001
Glutathione (nmol/mg)
Median	8.25	7.40	15.35
(range)	(5.40–16.10)	(0.20–36.90)	(0.10–43.70)
P-value^1		NS	NS
P-value^2			0.009
8-OHdG (8-oxo-dG/10^5 dg)
Median	1.59	1.69	1.13
(range)	(0.80–4.67)	(0.00–72.90)	(0.00–188.30)
P-value^1		NS	NS
P-value^2			NS

¹P-values based on Mann-Whitney U-test (pairwise comparisons, group and control).

²P-values based on Wilcoxon Signed rank test (pairwise comparisons, malignant and non-malignant lung).

NSCLC = Non-small cell lung cancer.

NADPH = Nicotinamide adenine dinucleotide phosphate.

8-OHdG = 8-oxo-hydroxydeoxyguanosine.

NS = Not significant.

Table III.  Descriptive statistics with p-values by parameter and histology in non-small cell lung cancer patients.

	Adenocarcinoma 11		Squamous Cell Carcinoma 11
Parameter n	non-malignant	malignant	non-malignant	malignant
NADPH oxidase (mV/mg)
Median	1.53	8.43	2.51	3.65
(range)	(0.45–19.67)	(2.44–19.67)	(0.26–7.36)	(0.21–47.20)
P-value^1			NS	NS
P-value^2		0.005		NS
Myeloperoxidase (U/mg)
Median	9.90	2.60	5.90	1.40
(range)	(0.70–23.30)	(0.50–10.40)	(0.00–15.80)	(0.0–12.50)
P-value^1			NS	NS
P-value^2		0.007		0.007
Glutathione (nmol/mg)
Median	11.60	20.8	4.90	6.90
(range)	(3.10–36.90)	(5.80–36.90)	(0.20–14.0)	(0.10–43.70)
P-value^1			0.008	NS
P-value^2		0.022		NS
8-OHdG (8-oxo-dG/10^5 dg)
Median	0.93	0.92	4.26	4.24
(range)	(0.00–6.95)	(0.00–2.46)	(0.54–72.90)	(0.80–188.30)
P-value^1			0.017	0.001
P-value^2		NS		NS

¹P-values based on Mann-Whitney U-test (pairwise comparisons, adenocarcinoma and squamous cell carcinoma, malignant and non-malignant lung, respectively).

²Values based on Wilcoxon Signed rank test (pairwise comparisons, adenocarcinoma vs. non-malignant lung or squamous cell carcinoma vs. non-malignant lung, respectively).

NADPH = Nicotinamide adenine dinucleotide phosphate.

8-OHdG = 8-oxo-hydroxydeoxyguanosine.

NS = Not significant.

MPO activity

Between the control patients’ lung tissue and non-malignant tissue of cancer patients, lung MPO activity did not differ. When MPO activity was compared between the combined group with NSCLC and the non-malignant areas of the same patients or to the control patients’ lungs, the levels for the malignant specimens were significantly lower (p = 0.001) (Table II). Significance was greatest between the malignant and non-malignant specimens in the same patients. MPO activity did not differ between AC and SCC, but in both cases the activity was lower in the malignant tissue than in the non-malignant area (Table III).

Free GSH content

The lowest GSH levels were in control and non-malignant tissues, and did not differ (Table II). GSH was significantly higher in the NSCLC than in the non-malignant tissues of the same patients (p = 0.009). When GSH was compared between cancer subtypes, GSH was higher in AC than in the corresponding non-malignant tissues (p = 0.022), the corresponding values for SCC were non-significant (Table III).

Footprint of oxidative stress, 8-OHdG

No difference emerged for this marker between either control group, nor the combined group with NSCLC (Table II). However, 8-OHdG content was significantly higher in SCC than in the control patient lung, both in malignant (p = 0.001) and non-malignant (p = 0.017) areas (Table III).

Correlations

Relationship to smoking history

SCC patients had, as expected, a significantly higher number of pack years smoked than did AC patients (p = 0.003). A positive correlation emerged between pack years and 8-OHdG content of the malignant samples (0.49, p = 0.020) (Table IV), and a negative correlation between higher number of pack years and GSH content of NSCLC patients’ non-malignant samples (−0.47, p = 0.029).

Table IV. Correlations by parameter and sample group

		NADPH oxidase	Myelo-peroxidase	Glutathione	Pack years
NADPH oxidase	Control (n = 10)	–	0.21	0.81	0.46
	P-value		NS	0.005	NS
	Non-Malignant (n = 11)	–	−0.49	0.13	0.22
	P-value		0.023	NS	NS
	Malignant (n = 11)	–	−0.41	0.27	−0.01
	P-value		NS	NS	NS
Myeloperoxidase	Control (n = 10)	0.21	–	0.27	−0.06
	P-value	NS		NS	NS
	Non-malignant (n = 11)	−0.49	–	0.53	−0.33
	P-value	0.023		0.013	NS
	Malignant (n = 11)	−0.41	–	0.38	−0.37
	P-value	NS		NS	NS
Glutathione	Control (n = 10)	0.81	0.27	–	0.15
	P-value	0.005	NS		NS
	Non-malignant (n = 11)	0.13	0.53	–	−0.47
	P-value	NS	0.013		0.029
	Malignant (n = 11)	0.27	0.384	–	−0.24
	P-value	NS	NS		NS
8-OHdG	Control (n = 10)	−0.41	−0.15	−0.69	0.35
	P-value	NS	NS	0.029	NS
	Non-malignant (n = 11)	−0.13	−0.35	−0.75	0.34
	P-value	NS	NS	0.001	NS
	Malignant (n = 11)	−0.10	−0.35	0.56	0.49
	P-value	NS	NS	0.007	0.020

P − values based on Sperman's rank correlation coefficient by rank.

NADPH = Nicotinamide adenine dinucleotide phosphate.

8-OHdG = 8-oxo-hydroxydeoxyguanosine.

NS = Not significant.

GSH, NOX, and 8-OHdG in control patients

In the control patients, a significant positive correlation appeared between GSH content and NOX activity (0.81, p = 0.005), and a negative one between GSH and 8-OHdG (−0.69, p = 0.029).

GSH, MPO activity and 8-OHdG in cancer and non-malignant tissues

The correlation was positive for non-malignant lung samples versus GSH content and MPO activity (0.53, p = 0.013), but a negative one between GSH and 8-OHdG (−0.75, p = 0.001) in the non-malignant lung, whereas the correlation between these indices was positive in malignant tissues (0.56, p = 0.007).

Discussion

The present study, to our knowledge, may be among the first studies of NOX in human lung diseases, and it shows modestly elevated NOX activity in the non-malignant lung of NSCLC patients, this elevation being, however, much greater in the malignant tissues, especially in AC. The strength of this investigation lies in the functional NOX activity in two separate cancer subtypes combined with two separate control groups. We are aware of no other studies on NOXs and lung cancer, but the present results strongly suggest that most of the NOX activity is related to cancer cells–not to the interstitium or to inflammatory cells of the non-malignant tissues. Our findings are in full agreement with earlier ones revealing NOX expression in cancer cells, in those studies prostate and colon cancers [26], [27]. More studies will be necessary to investigate the role of various NOX isoforms in lung cancer and the possible role of NOXs in tumor invasion, since the oxidant milieu in tumor cells has multiple effects on tumor behavior, one of these being cell invasion [28].

In contrast to increased NOX activity, MPO was significantly lower in the malignant tissues, both in AC and SCC, than in the non-malignant lung of these same patients. Moreover, a trend appeared toward elevated MPO in the non-malignant lung compared to that in control samples. These results suggest that MPO, being a marker of PMNs, reflects the oxidant burden of the lung interstitium, but not of the cancer cell. What needs emphasis is that these control lung specimens were not totally normal, either. That these patients had undergone lung surgery for other reasons was the reason that their neutrophilic inflammation was probably higher than in lungs of totally healthy individuals. Moreover no systematic histological analyses were done on grade of inflammation or of neutrophil accumulation. However, our results are in line with other findings showing elevated levels of oxidants related to inflammatory cells in malignancies and in non-malignant lung tissue of patients with lung cancer [29], so this may also be the case in our patients. The relationship between MPO and other oxidant-producing enzymes and redox imbalance in the parenchyma of various cancers including lung cancer is a generally unresolved area of research, and further research is warranted.

Our GSH content was higher in the malignant than in the non-malignant lung, with the highest GSH levels detectable in AC. This is in agreement with Blair et al., who revealed that NSCLC contained higher GSH levels than did adjacent normal lung tissue [30]. Our lung tissues contained lung homogenates, whereas Cook et al. have demonstrated that subpopulations of cancer cells may differ in their content of GSH from that of cells from non-malignant tissue [31]. Other significant effects on GSH levels in the lung stem from smoking history, other exposures, diseases, medications, age, and sex. Our SCC patients, for example, had smoked for a significantly higher number of pack years than our AC patients.

GSH levels of malignant tissues can be expected to reflect the antioxidant capacity of the tumor cells. Glutamate cysteine ligase, the rate-limiting enzyme in GSH synthesis, is elevated in lung cancer [14]; these alterations in turn may contribute both to tumor invasion and to increased drug resistance. Only part of the GSH is free, a significant portion being bound to proteins such as those in many drug-transport mechanisms and proteins that regulate the cellular redox state, including glutathione-S-transferases and glutaredoxins [32]. Protein-bound GSH is difficult to measure. Elevated free GSH levels in malignant lung tissues, especially in AC, are, however, in agreement with those in other investigations of GSH-related enzymes in lung and pleural malignancies.

GSH prevents 8-OHdG formation [33], and in our material, GSH correlated negatively with 8-OHdG both in the controls and in the non-malignant tissues. However, this correlation was positive in malignant tissues, representing an uncontrolled redox state. Our findings are consistent with those of Yang et al. showing down-regulation of GSH pathway genes as associated with lung cancer pathogenesis [34], but for lung tumors, upregulation of this system induces drug resistance.

We found no significant changes in 8-OHdG content between NSCLCs and non-malignant tissues and control samples. Content was significantly higher both in malignant and non-malignant tissue of SCC than in AC patients with a significant correlation with smoking. Neither NOX activity nor MPO activity correlated with 8-OHdG content in NSCLC patients’ non-malignant or malignant samples. It has been established that 8-OHdG content in the normal lung is highly related to current smoking [35]. Although 8-OHdG has been a good marker of oxidative stress in rodents, this may not be the case for NSCLC in humans. In addition, no relation emerged for the 8-OHdG and neutrophil contents in the nasal lavage [36], indicating that antioxidative mechanisms differ significantly from those of rodents. It is important to note as well that many of our correlations were calculated in relatively small sample sizes, as was also the case here, and therefore need to be interpreted cautiously.

This study is a reflection of oxidative production, neutrophilic infiltration, antioxidative capacity, and cellular damage in NSCLC patients’ established malignant tumor and non-malignant lung samples. The main finding was the increased NOX activity in malignant tumor and non-malignant lung samples, which was independent of neutrophil infiltration. The role of NOX in the carcinogenesis of NSCLC is unknown, needing further research.

Limitations of the study

This study has several weaknesses. SCC patients had smoked for significantly more pack years and most were current smokers, compared to AC patients and to the controls. Therefore, the results are mainly descriptive for established tumors, and relation to CS exposure should be interpreted cautiously. Between SCC and AC patients, clinical stage also varied. This is mainly because pathogenesis and cancer biology differ between SCC and AC, as SCC is more closely associated with CS and has a more central tumor presentation than in AC. Because there we had only six COPD patients, no study of the difference between non-COPD and COPD NSCLC patients could be established. The lung tissue specimens gathered from control patients represented a population which was younger and had smoked less. This mismatch has practical reasons; for these kinds of studies, age- and smoking-matched control material with normal lung histopathology is very difficult to obtain. Some studies suggest that oxidative stress increases with age, but based on 8-OHdG, this difference was not evident in our two groups of controls. Moreover, our lung specimen from the normal-appearing area in cancer patients served as an additional control.

Conclusion

The main finding of this study was increased NOX activity in the malignant and non-malignant lung samples of NSCLC patients, when compared to control samples. In contrast, MPO appeared to be localized especially in the non-malignant areas in the cancer patients. GSH was elevated in malignant tissues, in line with many previous findings on the GSH-related pathways in human malignancies.

Abbreviations

8-OHdG, 8-oxo-hydroxydeoxyguanosine; AC, adenocarcinoma; CS, cigarette smoke; COPD, chronic obstructive pulmonary disease; dG, deoxyguanosine; DUOX1, Dual oxidase 1; GSH, glutathione, HOCl, hydrochlorous acid; HPLC, High-performance liquid chromatography; MPO, myeloperoxidase; mV/mg, millivolt/milligram; NADPH, nicotinamide adenine dinucleotide phosphate; nmol/mg, nanomole/milligram; NSAID, non-steroidal anti-inflammatory drug; NSCLC, non-small cell lung cancer; NOX, Nicotinamide adenine dinucleotide phosphate oxidase; NOXNOX2/4, NADPH oxidase 2/4; PMN, polymorphonuclear neutrophilic leukocytes; ROS, reactive oxygen species; SCC, squamous cell carcinoma; U/mg, units/milligram; UV, ultraviolet.

Acknowledgements

This study has been supported by a special governmental grant for health sciences research. The authors express their gratitude for the skillful technical and secretarial assistance of Yvonne Sundström. Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.