NADPH oxidase 4 is an oncoprotein localized to mitochondria

Abstract

Reactive oxygen species (ROS) are known to be involved in many physiological and pathological processes. Initially ROS-producing NADPH oxidase (NOX) proteins were thought to be present in phagocytes. However, recent studies have demonstrated that NOX proteins are expressed in many other cell types and tissues. NOX family members' expression and function seems to vary from tissue to tissue. We determined the expression of the NOX family of proteins (NOX1-5) in normal breast tissue and breast tumors. Our study revealed that normal breast tissues express NOX1, 4 and 5 genes. Similar pattern of expression was revealed in a breast epithelial cell line. We found that NOX4 was overexpressed in the majority of breast cancer cell lines and primary breast tumors. NOX4 was also overexpressed in ovarian tumors. Overexpression of NOX4 in normal breast epithelial cells resulted in cellular senescence, resistance to apoptosis, and tumorigenic transformation. Overexpression of NOX4 in already transformed breast tumor cells also showed increased tumorigenicity. Strong evidence suggests that regulation of these processes occurs through NOX4 generation of ROS in the mitochondria. We demonstrate that the NOX4 protein contains a 73 amino acid long mitochondrial localization signal at the N-terminus that is capable of transporting a passenger protein GFP into the mitochondria. Treatment of NOX4 overexpressing cells with catalase resulted in decreased tumorigenic characteristics. Together, this study provides evidence for an oncogenic function for NOX4 protein localized to mitochondria and suggests that NOX4 is a novel source of ROS produced in the mitochondria. This study also identifies a possible treatment of NOX4-induced breast cancer by antioxidant treatment.

Introduction

NADPH oxidases and the mitochondria are a major source of cellular reactive oxygen species (ROS).¹ There are seven identified family members in the NADPH family: five NADPH Oxidases (NOX), NOX1–5, and two NOX homologues, DUOX1 and DUOX2.² NOX enzymes play a fundamental role in many cell functions including signal transduction, differentiation, proliferation and cell death.¹ NADPH oxidases are differentially activated by a whole host of binding partners, including p22^phox, p40^phox, p47^phox/NOXO1, p67^phox/NOXA1 and Rac. Activation of NOX enzymes 1–3 are dependent on different combinations of these binding partners, while NOX5 has only been shown to be activated by Ca²⁺ ions³ (reviewed in ref. 4).

NADPH Oxidase 4 (NOX4) is an active component of the NOX complex. The NOX4 enzyme has been shown to be activated solely by the p22^phox binding partner.^5,6 NOX4 has been correlated with the activation of the p38^MAPK pathway in endothelial cells.⁷ NOX4 expression plays a role in regulation of the other NOX family members; a dominant negative NOX4 phenotype results in a lack of NOX2 expression.⁸ Overexpression of NOX4 in fibroblasts and kidney cells resulted in an increase in the generation of ROS.^9–11 NOX4 has been demonstrated to localize to the endoplasmic reticulum and the nucleus,^12,13 and it has been identified in many cell types, including endothelial cells, adipocytes and fibroblasts.^9,14,15 Splice variants of NOX4 have been identified in lung cells.¹⁶ It has been suggested that NOX4 may be an active oxygen sensor, relying on the p22^phox subunit to function.^6,17 NOX4 functions have been well-characterized in cardiac diseases and diabetes.^5,18 However NOX4 function in carcinogenesis is unclear.

In this study, we aimed to identify a role for NOX4 and other NOX family members in breast carcinogenesis. Our study reveals that NOX4 is overexpressed in breast tumors when compared to normal breast tissues. We conducted heterologous NOX4 expression in normal breast epithelial cells. These studies suggest that NOX4 expression in normal cells results in resistance to apoptosis and increased tumorigencity.

Results

NOX4 is overexpressed in malignant breast cell lines and breast tumors.

NOX proteins are variably expressed in different tissues.¹ To identify NOX expression in human breast tissue, we performed RT-PCR analyses of NOX1, 2, 3, 4 and 5. Figure 1 shows NOX expression in normal human breast tissue and a normal human breast cell line. Our analyses of the five NADPH oxidases demonstrates variable expression of the individual family members in the normal breast sample, and identifies consistent expression of only NOX1, 4 and 5 in human breast tissue and normal human breast epithelial cell line (MCF12A). RT-PCR analyses of NOX2 yielded varying results ranging from low to no expression. NOX3 expression was undetectable. Interestingly, expression levels of NOX1 and 4 appear similar in both normal breast tissue and normal human breast epithelial cell line. However, NOX5 expression was higher when compared to NOX1 and 4 expression levels (Fig. 1A). We conclude that primary breast tissue and breast epithelial cell line MCF12A express NOX1, 4 and 5 members of NOX family.

Figure 1.

(A) NOX expression in normal breast cell line and normal breast tissue. RT-PCR was used to analyze the gene expression of the NADPH family members in normal breast tissue (I) and the MCF12A normal breast cell line (II). (B) NOX4 is overexpressed in malignant breast cell lines and breast tumors. RT-PCR was conducted to determine expression of the NOX4 gene in cell lines and breast tissue and tumor samples. The NOX4 gene is overexpressed in most breast cancer cell lines and breast tumors (T) when compared to normal breast tissue (N). (C) NOX4 is overexpressed in breast tumors. IHC analysis was done on tissue array (TARP5) containing breast carcinomas of different grades. (I) Representative positive breast carcinoma case. (II) Representative negative breast carcinoma case. (III) Graph representing NOX4 expression in all breast tumors assayed. (IV) Graph representing NOX4 expression in breast tumors stratified by grade.

After identifying expression of NOX genes in normal human breast tissue we analyzed human breast cancer cell lines to determine expression of all NOX proteins (data not shown). We determined NOX4 was overexpressed in breast cancer cell lines of varying invasiveness (Fig. 1B). Our analysis of NOX4 expression in primary breast tumors also demonstrated a higher expression when compared to normal breast tissue samples (Fig. 1B). A TARP5 slide with several breast carcinoma sections from different cases helped us to screen a relatively large number of cases under the same experimental conditions (Fig. 1C). Examination of the breast tumor sections revealed that 73% of breast carcinoma cases as a whole were positive for NOX4 expression (Fig. 1C-III). Negative control sections incubated with secondary antibody showed no reaction (Fig. 1C-II). We also determined whether or not NOX4 expression correlated with tumor grade. Our analysis revealed no correlation between tumor grade and NOX4 expression (r = 0.04). Forty three percent of grade 1, 75% of grade 2 and 80% of grade 3 breast carcinomas were positive for NOX4 (Fig. 1C-IV). We conclude that NOX4 is overexpressed in breast tumors and that its expression does not correlate with tumor grade.

NOX4 is overexpressed in primary ovarian tumors.

To extend our findings to other tumor types, we conducted immunohistochemical analysis of NOX4 protein expression in 63 human ovarian tumors. We found 42/63 tumors (67%) showed positive staining for NOX4 expression (Fig. 2A and C). An analysis of NOX4 expression in ovarian tumors stratified by grade demonstrated no correlation between tumor grade and positive staining for NOX4, with all percentages falling between 60–70% (Fig. 2D). We conclude that NOX4 is overexpressed in ovarian carcinomas and that no apparent correlation exists between the NOX4 expression and tumor grades.

Figure 2.

NOX4 is overexpressed in ovarian tumors. IHC analysis was performed on tissue array (TARP5) containing ovarian carcinomas of different grades. (A) Representative positive ovarian carcinoma case. (B) Representative negative ovarian carcinoma case. (C) Graph representing NOX4 expression in all ovarian tumors assayed. (D) Graph representing NOX4 expression in ovarian tumors stratified by grade.

NOX4 protein localizes into the mitochondria.

Cancer cells are thought to generate ROS which are known to promote tumorigenesis via cell proliferation, cell motility and invasion. Interestingly ROS are produced by both the NOX proteins and by mitochondria. Our previous study suggested a cross talk between the ROS generating NOX and mitochondria.¹⁹ We analyzed whether NOX4 localizes to mitochondria. We analyzed a putative mitochondrial localization signal (MLS) in NOX4 protein. Our analyses revealed that the N-terminus of NOX4 proteins contains a probable MLS. Using MITOPROTII (http://ihg2.helmholtz-muenchen.de/ihg/mitoprot.html), we found a 98% probability of the NOX4 being localized to the mitochondria. We PCR amplified the first 73 amino acids and cloned in frame with GFP in pEGFP-N2 vector. Figure 3A shows that MLS derived from NOX4 protein target GFP to mitochondria. In addition, western blot analyses of mitochondrial proteins isolated from pEGFP-N2-NOX4MLS transfected cells demonstrates mitochondrial localization (Fig. 3B). We conclude that human NOX4 contains an authentic MLS and NOX4 localizes into the mitochondria.

Figure 3.

NOX4 localizes to the mitochondria. pEGFP-N2-NOX4MLS construct was used to transfect NIH3T3 cells. (A) MitoTracker staining determined mitochondrial location, DAPI staining determined nuclear location, GFP staining determined location of the NOX4 protein. (B) Western blot analyses of pEGFP-N2-NOX4MLS expressing cells showing NOX4 import into the mitochondria.

NOX4 overexpression leads to H₂O₂ production and cellular senescence.

The above study showed overexpression of NOX4 in breast tumors, therefore we created a breast epithelial cell line that stably overexpresses NOX4. The MCF12A breast cell line is a nontumorigenic cell line derived from reduction mammoplasty tissue. We determined that NOX4 expression in this cell line is lower than in breast cancer cell lines (Fig. 1B-I). We transfected human NOX4 into the MCF12A cell line to overexpress the NOX4 gene. We used RT-PCR and western blot to verify that NOX4 is overexpressed in the transfected cell lines (Fig. 4A-I and II), respectively). NOX4 requires phox22 for its actvity.^6,17 We measured phox22 expression in NOX4 overexpressing cells. Figure 4A-III shows that phox22 expression remained unaffected. We confirmed that NOX4 overexpression results in increased activity that is induced by NADPH and suppressed by DPI (Fig. 4B). As expected overexpression of NOX4 led to a detectible increase in H₂O₂ production (and not superoxide) (Fig. 4C).

Figure 4.

Generation of an MCF12A cell line with stable overexpression of NOX4. A(I) RT-PCR analysis demonstrates a higher expression of the NOX4 gene when compared to the empty vector control. A(II) Western blot analysis was performed using the membrane fraction isolated from wild-type and NOX4 overexpressing MCF12A cells. Amido Black was used as a loading control. (B) Measurement of NOX4 activity using a nitroblue terazolum (NBT) assay. NBT reduction was normalized to cell number. (C) DHE and DCF assays were conducted to analyze the amount of ROS produced by NOX4 overexpressing MCF12A cells when compared to the empty vector (Mock) control. (D) Beta-galactosidase assays were conducted to determine cellular senescence in MCF12A (I) and MDA-MB435 (II) cell lines.

Since cellular senescence is described as a barrier to tumorigenesis,^20–22 we performed a SA beta-galactosidase assay on the empty vector and NOX4 overexpressing cells, and determined a higher instance of senescence in the NOX4 overexpressing cell lines when compared to the empty vector controls (Fig. 5D). An increase in senescence was observed in both breast epithelial and cancer cell lines when NOX4 was overexpressed. These results demonstrate NOX4 overexpression leads to cellular senescence.

Figure 5.

NOX4 overexpression induces resistance to cell death. Apoptosis induced by intrinsic (TRAIL) and extrinsic (Etoposide) factors was measured by quantification of Annexin staining. Values represent fold differences in apoptosis normalized to the untreated control.

NOX4 overexpression leads to resistance to apoptosis.

We determined apoptotic cell death in the NOX4 overexpressing breast epithelial cells. We found no difference in cell death induced by TRAIL. However, when treated with etoposide, NOX4 expressing cells showed apoptosis resistance when compared to mock transfected cells (Fig. 5). This result demonstrates that NOX4 expression protects cells from cell death induced by anticancer agents.

NOX4 overexpression induces tumorigenic phenotypes which are inhibited by catalase.

We investigated the effects of overexpressing NOX4 in the MCF12A normal breast cell line, and the MDA-MB435 malignant cell line. Using in vitro assays, we measured anchorage independent growth and cell invasion. We found that the NOX4-overexpressing cells showed a two fold higher increase in anchorage-independent growth when compared to the empty vector controls. This observation was true for both the normal and malignant cell lines (Fig. 6A). We also conducted a matrigel invasion analysis. We determined that normal cells that overexpress NOX4 were much more invasive than the empty vector control (Fig. 6B). These results demonstrate that overexpression of NOX4 causes cells to take on a tumorigenic phenotype. As another measure of tumorigenicity, we performed a proliferation assay. We determined that normal cells that overexpress the NOX4 gene proliferate slower than empty vector controls and wild-type cells. However, the malignant cell line overexpressing NOX4 proliferated faster than their empty vector control (Fig. 6C). These studies suggest that overexpression of NOX4 induces tumorigenic chararacteristics.

Figure 6.

Tumorigenic transformation of breast epithelial cells overexpressing NOX4. Parallel analyses were conducted in MCF12A and MDA-MB435 cells overexpressing NOX4. (A) Anchorage-independent cell growth was assayed. Colonies represents an average value for assays repeated in triplicate with analysis of six fields of view per assay. (B) Invasion was assayed using a matrigel boyden chamber. Matrigel invasion assays for MCF12A cells were completed in duplicate with three chambers per assay. Matrigel invasion assays for MDA-MB435 cells were completed once with three chambers per assay. For each analysis, 10 fields of view were captured and the number of cells per field of view was quantified. Mean number of cells were determined by taking the average of the ten fields of view. The average of the two analyses were calculated and plotted as mean cells/field. (C) Cell doubling time was analyzed for NOX4 and Mock MCF12A and MDA-MB435 cell lines. Cell doubling time was calculated as described in the methods. (D) Catalase treatment inhibits Matrigel invasion induced by NOX4 overexpression. Matrigel invasion assays were completed once for the MCF12A cells and twice for the MDA-MB435 cells with three chambers per cell type per assay. Invasion was quantified as outlined in (B). Stars denote statistically significant differences between treated and untreated NOX4 overexpressing cell lines.

Because hydrogen peroxide is one of the downstream effects of NOX4 overexpression, we tested the efficacy of catalase treatment on abating the tumorigenic phenotype. Catalase is an enzyme that breaks hydrogen peroxide down into water and oxygen. Cells were treated with catalase and seeded in matrigel invasion chambers. Catalase treatment resulted in a significant decrease in invasion of NOX4 overexpressing cells (Fig. 6D). These results demonstrate that catalase inhibits the tumorigenic effects of NOX4 overexpression.

Discussion

NOX proteins represent the major non-mitochondrial sources of ROS. The NOX enzyme complex was first described in neutrophils where it is normally quiescent but generates a large quantity of ROS upon activation during phagocytosis and plays a vital role in non-specific host defense against ingested pathogens.^1,23 In the past few years many non-phagocytic cells have been found to contain NADPH oxidases (reviewed in ref. 24).

In non-phagocytic cells NOX family member expression varies in a cell specific manner (reviewed in ref. 5). In this paper we analyzed the expression of NOX1 to 5 in breast tissues. Our study identified that normal breast tissues and epithelial cell line express NOX1, 4 and 5. Previous studies demonstrate organ specific expression of NOX genes.^23,24 Interestingly NOX1 is expressed in epithelial cells (e.g., colon) as well as in vascular smooth muscle cells (VSMC). NOX2 is expressed in endothelial cells, cardiomyocytes and fibroblasts. NOX3 is primarily expressed in fetal tissues and adult inner ear. NOX4 appears to be expressed in many tissues including kidney, placenta, endothelial cells, VSMC, cardiomyocytes, fibroblasts, ovary, testis and skeletal muscle. NOX5 is expressed in fetal tissues and adult testis, spleen, ovary, placenta and pancreas.^23,24

Our analyses revealed that NOX1, 4 and 5 were coexpressed in both breast epithelial cell line and primary breast tissues. Such a coexpression of NOX has also been reported in other tissues. These include VSMC which coexpress NOX1 and 4,²⁵ and endothelial cells¹⁴ and cardiomyocytes which coexpress NOX2 and 4.^23,26 Notably, expression levels of NOX1 and 4 were equal in primary breast tissues. However, NOX5 expression in comparison to NOX1 and 4 was higher. A similar pattern of NOX coexpression was also revealed in breast epithelial cell line.

Our previous study revealed an important role for NOX1 in breast (and ovarian) tumorigenesis.¹⁹ In this paper we aimed to identify a role for NOX4 in breast tumorigenesis. We determined the NOX4 expression in breast (and ovarian) tumor tissue via RT-PCR, and found that NOX4 was overexpressed in both breast cancer cell lines and primary tumors. We found that 4 out of 6 breast cancer cell lines (66.6%) overexpressed NOX4. Similar analyses in primary breast tumors revealed an overexpression of NOX4 in approximately 87% primary breast tumors at the mRNA level (See Fig. 1C). When we analyzed a larger number of primary breast tumors by histochemical analyses we found that approximately 70% overexpressed NOX4. Since ovarian cells also express NOX4,²³ we determined its expression in primary tumors. This analysis revealed that approximately 65% of ovarian tumors overexpressed NOX4. We did not find a correlation with tumor grade either in breast or ovarian tumors. Interestingly NOX4 overexpression is reported in prostate, melanoma and glioblastoma cell lines.^27–29 Our study together with studies from other laboratories suggest that NOX4 may function as an oncogene involved in tumorigenesis related to different organ sites.

NOX4 localizes to the nucleus.¹² We asked whether NOX4 localizes into the mitochondria. Both NOX protein and mitochondria are known to be major sources of ROS in the cell but the identity of proteins involved in ROS production within mitochondria is not clear. It is hypothesized that mitochondria produce ROS as a by-product of oxidative phosporylation involving various oxidative-reductive reactions which ultimately produce ATP.^30–32 Using cell biology and biochemical approaches we experimentally tested whether NOX4 is a mitochondrial protein. Our study revealed that the first 73 amino acids at the protein's N-terminus contains a mitochondrial targeting signal and this signal is sufficient to transport a passenger protein (GFP) into the mitochondria. A complimentary biochemical approach also demonstrated that NOX4 is a mitochondrial protein. Since other studies show that NOX4 also localizes to the nucleus it is likely this protein functions in both compartments. NOX4 was recently shown to physically interact with PDIP38 (also known as POLDIP2, proliferating cell nuclear antigen–and DNA polymerase–interacting protein).³³ The interaction with PDIP38 suggests a possible role for NOX4 in nuclear DNA replication and/or repair. In this context it is important to note that PDIP38 also localizes to mitochondria and in particular associates with mitochondrial nucleoids.³⁴ It is conceivable that NOX4 localization to mitochondria regulate mtDNA replication, repair and nucleoid structure via PDIP38 interaction or intra mitochondrial production of ROS which modify/regulate proteins involved in mtDNA maintenance. Indeed our preliminary studies demonstrate that NOX4 is involved in mtDNA homeostasis (unpublished data).

Our analyses of NOX proteins revealed that NOX4 is overexpressed in the majority of breast cancer cell lines and primary breast tumors. In order to provide direct evidence and cause and effect relationship we overexpressed NOX4 in a normal breast epithelial cell line. We found that overexpression of NOX4 in normal breast epithelial cell line induces tumorigenic characteristics including cellular senescence, anchorage-independent cell growth, matrigel invasion and resistance to apoptotic cell death. Additionally overexpression of NOX4 in a malignant cell line further enhanced these tumorigenic characteristics of the cell line. These results suggest that NOX4 overexpression plays an oncogenic role in breast tumorigenesis.

Overexpression of NOX4 resulted in increased H₂O₂ production but not the superoxide radical production. This observation is consistent with recent studies suggesting NOX4 produces H₂O₂ and not the superoxide.^35–38 NOX4 activity was inhibitable by DPI treatment (Fig. 4B). We demonstrated a higher concentration of H₂O₂ in the NOX4 overexpressing breast epithelial cell line than the wild type line. Consistent with our studies other laboratories have also reported increased H₂O₂ production with NOX4 overexpression.^39–41 It is likely that increased production of H₂O₂ due to overexpression of NOX4 mediates breast tumorigenesis. Indeed, our studies demonstrate that NOX4 induced tumorigenicity was reduced when cells were treated with catalase. Reduction in NOX4 induced tumorigenicity with antioxidant agents such as catalase provide a new avenue for development of treatment for breast cancers.

In summary, this paper describes that NOX4 is an oncoprotein that localizes to mitochondria. Our studies also establish that NOX4 is a novel source of ROS produced in mitochondria.

Materials and Methods

Cell line and culture conditions.

Wild-type MCF12A human mammary epithelial cells were cultured in a 1:1 mixture of Ham's F-12 medium and DMEM with 10% horse serum, supplemented with 2 mM glutamine, 0.1 µg/ml cholera enterotoxin, 10 µg/ml insulin, 0.5 µg/ml hydrocortisone, and 20 ng/ml epidermal growth factor. MDA-MB435 cells were cultured in 1x DMEM with 10% FBS supplemented with 0.01% penicillin/streptomycin. Cell lines with vectors were maintained in their respective wild-type media supplemented with G418. All cell lines were maintained under an atmosphere of 95% air-5% CO₂ at 37°C and subcultured weekly using 0.02% EDTA and 0.05% trypsin. Media were changed every 2–3 days.

NOX gene expression analyses.

NOX1 to 5 expression analyses was determined by RT-PCR using total RNA as a template. Total RNA was extracted from MCF12A cells using the TRIzol reagent (GIBCO-BRL, Grand Island, NY), and cDNA was synthesized by reverse transcription using the Invitrogen cDNA synthesis kit. The NADPH Oxidase transcripts were amplified using the following primers: NOX1 forward: 5′-ACA AAT TCC AGT GTG CAG ACC-3′, NOX1 reverse: 5′-AGA CTG GAA TAT CGG TGA CAG-3′; NOX2 forward: 5′-TCA CAC CCT TCG CAT CCA-3′, NOX2 reverse: 5′-ATC ATG GTG CAC AGC AAA GTG-3′; NOX3 forward: 5′-GAG TTC ATC AGA CAG GCC TCC-3′, NOX3 reverse: 5′-ACC ACA GGG CCT AAA ATC CAT-3′; NOX4 forward: 5′-CTC AGC GGA ATC AAT CAG CTG-3′; NOX4 reverse: 5′-AGA GGA ACA CGA CAA TCA GCC-3′; NOX5 forward: 5′-GGG TCT GAT GCC TTG AAG GA-3, NOX5 reverse: 5′-GCA GCC GTG TGC ATC ATG-3′. 5′-ATG GGT CAG AAG GAT TCC TAT GT-3′ (forward primer) and 5′-AAG GTC TCA AAC ATG ATC TGG G-3′ (reverse primer) were used for the actin cDNA. NOX4 expression levels were determined in breast cancer cell lines by RT-PCR using cDNA extracted from MCF7, T47D, SKBR3, MDA-MB231, BT474 and MDA-MB435 cells prepared as outlined above, with the same primers as outlined above.

Immunohistochemical (IHC) analyses.

A tissue array slide from the Cooperative Human Tissue Network (CHTN) and Tissue Array Research Program (TARP5) of the National Cancer Institute, National Institutes of Health, Bethesda, MD was used in the present study. The slide contains breast and ovarian carcinomas as well as multiple benign tissues from different organs. One section from formalin-fixed, paraffin-embedded benign liver tissue slide was used as a positive control. NOX4 (N-15) pre-adsorbed blocking peptide (Santa Cruz Biotech, Santa Cruz, CA) was used as negative control. Characterization of the lesions and grading of the tumors was previously done by a pathologist (Mohamed Mokhtar Desouki) and published.¹⁹ The immunohistochemistry protocol as described by Desouki et al.¹⁹ was applied with modifications. The slides were deparaffinized by incubation in xylene and ascending grades of alcohol. Antigen retrieval was done by heating in citrate-based, antigen unmasking solution (Vector Laboratories, Burlingame, CA) for 30 minutes at 98°C, incubated in 3% hydrogen peroxide (H₂O₂) for 10 minutes, blocked with blocking peptide for 30 minutes, incubated with 4 ug/ml anti-NOX4 (Santa Cruz Biotech, Santa Cruz, CA) antibody for one hour at room temperature, followed by incubation with biotinylated secondary anti-goat solution for 30 min and another 30 minutes with Vectastatin ABC kit (Vector Laboratories, Burlingame, CA). Color was developed by incubating slides with peroxidase substrate solution followed by counterstaining with Hematoxylin. Sections were incubated with secondary antibody to verify no nonspecific binding occurred. All sections were examined with an Olympus BX50 microscope. The pictures were taken with an Olympus DP 70 connected to DP Controller software (Olympus, Center Valley, PA). Statistical analyses were performed by using Epi-Info software program, version 3.5.1.⁴² A linear correlation test to determine correlation between tumor grades and NOX4 immunoreactivity was performed. Scoring of immunoreactivity was considered to be negative or positive, with the same parameters as previously described (score + <10% positive, score ++ 10–50% positive and score +++ >50% positive).¹⁹

Cloning of NOX4 mitochondrial localization sequence (MLS) into a GFP vector.

The NOX4 MLS fragment was produced from normal breast tissue cDNA by PCR using NOX4MLS forward primer (5′-TAA TGA ATT CAT-GGC TGT GTC CTG GAG GAG C-3′, which contains EcoR1site) and reverse primer (5′-ACC TGG ATC CGT AAA AGG ATA AGG CTG-3′, which contains BamH1 site). The NOX4 MLS fragment was cloned in frame with GFP into the pEGFP-N2 expression vector.

Mitochondrial localization of NOX4.

NIH3T3 cells were seeded on glass slides and cultured. pEGFP-N2-NOX4MLS construct was transfected into the cells the following day with a transfection reagent (Fugene HD, Roche). Cell staining was performed 48 hours post-transfection using MitoTracker Red (Invitrogen) and DAPI. The images were obtained by confocal fluorescence microscopy.

Transfection of MCF12A cells and single clone selection.

A NOX4 construct containing human NADPH oxidase 4 (NOX4) accession number NM_016931 in pReceiver-M10 expression vector was purchased from GeneCopoeia (Rockville, MD). Transfection of MCF12A cells was performed using FuGENE6HD following manufacturer's instructions. MCF12A cells were seeded at a density of 10⁶ cells/10-cm dish and incubated overnight. Cells were transfected the following day with 5 µg of plasmid DNA using 30 µl of Fugene-6 transfection reagent (Roche, Nutley, NJ) and 1 ml of serum-free culture medium for 2 h at 37°C, after which time culture medium containing 5% horse serum was added. Cells were harvested 18 to 24 h later with trypsin-EDTA and seeded at a concentration of 10⁴ cells/10-cm dish in medium containing 500 µg of Geneticin G418 (Roche, Nutley, NJ) per ml. Individual colonies were isolated after 2 weeks of selection, using cloning rings and then were passaged continuously in a selection medium.

Cell fractionation.

Mitochondrial and membrane fractions were isolated according to Paterson and Gottesman.⁴³ Cells were trypsinized, collected from 5x T175 flasks and centrifuged at 500 xg for 5 min following two washes in PBS. Cells were centrifuged at 1,000 xg for 15 min at 4°C, resuspended in ice cold cell homogenization buffer (150 mM MgCl₂, 10 mM KCl, 10 mM Tris-Cl) and incubated at −80°C for a minimum of 3 hours. Next cells were dounce homogenized and resuspended in homogenization buffer + sucrose. Nuclei were pelleted by centrifugation at 1,000 xg for 5 min at 4°C in a swinging bucket rotor. Supernatant was centrifuged at 5,000 xg for 10 min at 4°C in a fixed angle rotor to separate the membrane fraction from the pelleted mitochondria. The pelleted mitochondria fraction was purified in OptiPrep Solution containing 500 mM Hepes, 100 mM EDTA and iodixanol fractions. Mitochondria were collected from the 30%-10% iodixanol fractions, washed to remove iodixanol and resuspend in mitochondria suspension buffer containing 0.25 M sucrose and 10 mM Tris-Cl. Protein content in the nuclear, membrane and mitochondrial fractions were measured using Bradford assay.

NADPH oxidase activity assay and ROS measurement.

Cells were plated in a 24 well plate 48 hr prior to measuring NADPH oxidase activity by nitroblue tetrazolum (NBT) reduction as described in Serrander et al.³⁵ NOX activity was measured in the presence and absence of 500 µM NADPH. To inhibit NOX activity, cells were preincubated with 10 µM diphenyleneiodonium chloride (DPI) for 30 min. A duplicate plate was used to count cells and NBT reduction was normalized to cell number. ROS measurments was performed as described previously.¹⁹

Cell staining for senescence-associated β-galactosidase (SA-β-Gal) activity.

Senescence associated cell staining was done according to Dimri et al.⁴⁴ Culture media was removed from cells growing on 10 cm dishes and cells were washed with PBS. The fixative solution (2% (v/v) Formaldehyde 0.2% (w/v) Glutaraldehyde) was added to the cells for 5 min. Next cells were stained with staining solution (5 mM K₃Fe(CN)₆, 5 mM K₄Fe(CN)₆, 2 mM MgCl₂, 150 mM NaCl and 1 mg/ml X-Gal in 30 mM) Citric Acid/Phosphate Buffer and incubated for 16 hr at 37°C.

Detection of apoptosis by Annexin V-PI binding.

Detection of apoptosis was determined by staining cells with annexin V-FITC and PI as described in Kulawiec et al.⁴⁵

Analysis of in vitro tumorigenic phenotype.

Soft agar and matrigel invasion assays were performed as described previously.^19,46 Thirty minutes prior to cell seeding, cells were treated with catalase (750 U/ml) and seeded in media containing the same concentration of catalase. Matrigel invasion assays were performed as described previously.^19,46

Analysis of proliferation and cell cycle.

The MTT assay was performed by routine method. For each cell line, 12 wells of the 96-well microplate were used. Cells were plated at a concentration of 5,000 cells/well in 200 µl of media. After 24 hr and at subsequent time points 10 µl of MTT stock solution (5 mg of MTT/ml of PBS) were placed in each well. After 3 hr incubation, DMSO was added in order to dissolve formazan crystals. The optical density of samples was measured immediately using a Microplate Spectrophotometer (Bio-Tek instruments, USA) at a wavelength of 560 nm. MTT reduction rate (optical density) for each sample was normalized to the untreated control. Cell doubling times were determined during the Log Phase period using free software available on www.doubling-time.com/.