Loss of SOD3 (EcSOD) Expression Promotes an Aggressive Phenotype in Human Pancreatic Ductal Adenocarcinoma

Extracellular superoxide dismutase (EcSOD; SOD3) is a glycosylated enzyme that binds to heparin sulfate moieties in the extracellular environment and catalyzes the dismutation of superoxide (O₂^·−) to hydrogen peroxide (H₂O₂) with a rate constant of >10⁹/(mol/L)·s. By limiting O₂^·− availability and producing H₂O₂, EcSOD is thought to regulate the redox state of the extracellular environment (1). By virtue of removing O₂^·−, EcSOD has also been hypothesized to decrease the formation of highly reactive peroxynitrite (ONOO⁻) from nitric oxide (·NO) in the extracellular environment and exert a protective role over ·NO bioavailability (2). Compared with SOD1 and SOD2 (CuZnSOD and MnSOD, respectively) expression of EcSOD is highly cell-type specific and can be found most abundantly in the pancreas, lung, kidney, and vasculature (3).

A common feature in human cancers is an imbalance in the production and removal of reactive oxygen species (ROS), such as O₂^·− and H₂O₂, due to both increased fluxes of these ROS as well as alterations in antioxidant capacity. These redox imbalances in both the intracellular as well as extracellular environment have been hypothesized to significantly contribute to cancer cell proliferation and progression to malignancy (4–13). However, historically, many more studies have focused on imbalances in the intracellular redox environment, relative to the extracellular environment (4–13). In addition, many studies have not differentiated the effects of specific ROS.

Increased intracellular or extracellular steady-state levels of specific ROS can occur because of a decrease in the activity of antioxidant enzymes, such as SODs, catalase, and glutathione peroxidases (5, 14–16). The loss of EcSOD expression has been reported in several cancers (17–19), suggesting that EcSOD may play an essential role in maintaining an extracellular redox balance that opposes tumor growth and metastasis. This concept was highlighted in previous work showing that human pancreatic ductal adenocarcinoma (PDA) cell lines transduced with adenoviral vectors, enforcing the transient overexpression of EcSOD, demonstrated decreased cell growth and plating efficiency in vitro, as well as slower tumor growth and increased survival in a xenograft model (17, 19). Furthermore, similar results were obtained in vitro and in vivo with human breast and lung carcinoma cells where EcSOD overexpression not only decreased growth rate and clonogenic survival but also the invasive phenotype of these cancer cells (18, 20, 21).

Building on the previous findings that EcSOD is lost in PDA cell lines and overexpression inhibits growth and cloning efficiency in vitro and in xenograft models (17), we hypothesized that decreased levels of EcSOD in human PDA would correlate with poorer outcomes as well as increased aggressiveness of clinical disease progression. We also performed preclinical studies to evaluate the impact of overexpression of EcSOD on PDA tumor cell survival in cell culture and animal models, as well as determine the potential contribution of H₂O₂ and ·NO to any of the observed biologic effects of EcSOD on PDA progression. Results from our studies reveal that loss of EcSOD expression is a common event in patients with PDA, and that loss of EcSOD predicts a more aggressive tumor phenotype with respect to invasive and metastatic behavior. These findings highlight the important role of specific extracellular ROS in PDA biology and suggest potential utility for increasing EcSOD activity as a therapeutic target for slowing disease progression in patients with PDA.

Human cell lines (BxPC3, MiaPaCa2, and HPNE) were obtained directly from the ATCC and were passaged actively for less than 6 months from the time of receiving them from the ATCC and were not further authenticated and maintained in the recommended medium.

Luciferase-expressing BxPC3 cells were generated by subcloning the firefly luciferase gene, Luc2 (Promega) into a pQCXIP vector. Transfection of the construct was carried out using 3 μg of DNA and Lipofectamine 2000 into Phoenix packaging cells. Retroviral supernatants were collected 48 hours after transfection, spun down, and filtered through a 0.45-μm filter. Filtered retroviral supernatant was then incubated with cells for 24 hours before changing to complete medium. Transduced cells were selected 24 hours later with G418 (Geneticin). EcSOD-expressing BxPC3 and MiaPaCa-2 cells were generated as previously described (26).

A PDA tissue microarray (TMA) was obtained from US Biomax, Inc. (#PA2081), and IHC was performed (EcSOD; AbFontier; #LF-MA1021). Sections were deparaffinized, rehydrated, and subjected to heat-induced epitope retrieval in citrate buffer (pH6). Endogenous peroxidase activity was blocked with 3% H₂O₂ followed by serum block. After incubation with primary antibody at a dilution of 1:400 for 30 minutes, the Dako Mouse Envision System was used for detection. Dako DAB Plus was applied, and slides were lightly counterstained with hematoxylin.

The approval by the University of Iowa Institutional Review Board was obtained to retrospectively evaluate tissue from patients with PDA. Tissue blocks from 17 patients treated surgically for PDA were retrieved, and IHC for EcSOD and 3-nitrotyrosine (3-NT; Millipore; #06284; 1:2,000) was performed as above.

Oncomine is a cancer microarray database (22) that is publicly available at www.oncomine.org. The database was accessed, and a search for SOD3 and PDA was performed. Seven datasets were reported as published on the website.

EcSOD expression in human normal and PDA samples was examined using a TissueScan Real-Time PDA qPCR Panel (I) from ORIGENE Technologies (#PNRT301) and from tissue specimens obtained from patients with PDA at the University of Iowa (Iowa City, IA). EcSOD expression was measured using a TaqMan primer/probe set (Hs00162090_m1) from Applied Biosystems and normalized against β-actin control primers.

Case identification and selection methods of Surveillance, Epidemiology and End Results (SEER) PDA TMA have been described previously (23). This study includes PDA specimens from 114 patients with stage I–IV disease. Ethical and transfer of material approvals were obtained from originating sites and the National Cancer Institute (Bethesda, MD). The association of EcSOD expression in the 114 patients with disease stage and pancreatic resection status was evaluated with χ² analysis. Survival was defined as months from diagnosis to death and Kaplan–Meier survival plots were constructed to estimate survival as a function of time and log-rank was used to compare survival between patients with and without EcSOD expression. The Cox proportional hazards regression was used in multivariate analyses to estimate the impact of loss of EcSOD on survival (SAS 9.3, Cary, NC).

Cells were plated at a density of 1 × 10⁶ in 60-mm dishes in media containing 10% FBS and allowed to attach for 24 hours. Cells were rinsed and incubated with adenoviral vectors (Ad. EcSOD (24) or Ad. Empty) as previously described (21).

Cells were plated in triplicate into 60-mm dishes at a density of 1 × 10³ (Bx-PC3) or 1 × 10² (Mia-PaCa-2) in complete medium. Dishes were incubated for 2 weeks to allow for colony formation. The colonies were then fixed in ice-cold methanol and stained with 0.5% w/v crystal violet in 25% methanol. Colonies containing >50 cells were scored as survivors.

Cells were seeded in duplicate at a density of 1 × 10⁵ in 6-well plates with complete medium and were counted daily for 96 hours. Cell population doubling time in hours was determined as previously described (25).

BxPC3/MiaPaCa2 cells were seeded at a density of 2 × 10⁵ in serum-free medium into Matrigel chambers (8-μm pore size) from the BD Bio-Coat Matrigel Invasion Assay System (BD Biosciences; #354480). The effects of various interventions on invasion were evaluated, including 20 mmol/L N-acetylcysteine (NAC; Cumberland Pharmaceuticals; #66220-207-30), 10 U/mL catalase (Sigma-Aldrich; #C30), and 100 μmol/L N^G-nitro-l-arginine (L-NNA; Calbiochem; #483120). To evaluate the effects of the SOD mimetic (Galera Therapeutics; #GC4419) on invasion, GC4419 was placed on the cells for 24 hours before, as well as in the invasion chambers during the invasion assay. To evaluate the effects of conditioned media on invasion, media (2% FBS) was harvested from Bx-EcSOD or Mia-EcSOD cells and placed in the invasion chambers. In all cases, the Transwell chambers were placed into 24-well plates containing medium with 10% FBS and incubated for 22 to 24 hours at 37°C, 5% CO₂. Noninvading cells from the upper chamber were removed. Invading BxPC3 cells were stained by placing Transwell chambers into 0.5 mg/mL of a Thiazolyl Blue Tetrazolium Bromide (MTT) solution (Sigma-Aldrich) in RPMI-1640 medium without FBS or phenol red. Inserts were incubated at 37°C for 1 hour before membranes were removed and incubated in 150 μL of DMSO. DMSO solution was transferred to a 96-well plate, and the absorption at 550 nm was measured immediately and compared with known cell concentrations from a standard curve. Invading MiaPaCa2 cells were fixed in 100% methanol and stained with Giemsa and the number of invading cells counted from five random fields. Invasion data are represented as relative invasion except for data containing L-NNA and catalase, which are shown as percent invasion (invasion/migration × 100) due to the impact of these chemicals on migration. For Transwell migration, cells were seeded at a density of 2 × 10⁵ in serum-free medium into BD Falcon PET track-etched membrane chambers (8-μm pore size; BD Biosciences).

Total RNA for all in vitro studies was extracted using TRizol and the PureLink RNA Mini Kit (Life Technologies; #12183018A). Each sample was reverse transcribed and gene expression was measured using Power SYBR PCR master mix (Applied Biosystems; #4367659). Samples were normalized to 18 s control using the ΔΔC_t method.

Protein expression of intracellular SODs was assayed as previously described (26). Antibodies to MnSOD and CuZnSOD were prepared as previously characterized (17). The EcSOD antibody was a kind gift from Drs. James Crapo and Rebecca Oberley-Deegan, National Jewish Medical and Research Center, Denver, Co (27).

The concentration of EcSOD that accumulated in cell culture media was determined by a kinetic analysis of the competition of the reaction of O₂^•− with 5,5-dimethylpyrroline-N-oxide (DMPO), k_DMPO = 19/(mol/L)·s (28), and EcSOD, k_EcSOD = 4.0 × 10⁹/(mol/L)·s (29) using xanthine oxidase/hypoxanthine as a source of O₂^•−. Samples were examined by electron paramagnetic resonance (EPR) using a Bruker EMX spectrometer with an HS cavity and AquaX 4-bore sample cell (Bruker Biospin). Bovine CnZnSOD was used as a verified standard (30); 1 U of activity (50% inhibition of the DMPO-OOH signal) corresponds to 7.3 × 10⁻¹⁰ mol/L of bovine CuZnSOD as dimer. One unit of activity for EcSOD corresponds to a concentration of 2.4 × 10⁻¹⁰ mol/L of fully active EcSOD tetramers.

To quantify EcSOD activity, the WST-1 SOD assay kit (Dojindo Molecular Technologies; #S311) was used as previously described (24). To verify L-NNA–inhibited NO-synthase (NOS) activity, a nitrite assay was performed on media samples using a kit as per the manufacturer's instructions (Griess Reagent System; Promega; #TB229). Media from BxPC3-control cells incubated with or without 100 μmol/L L-NNA for 24 hours were assayed for nitrite concentration using media without FBS or phenol red and compared with a nitrite standard curve for each assay. The data were normalized per cell.

Female 6- to 10-week-old athymic-nu/nu mice were purchased from Harlan Laboratories and handled in accordance with the University of Iowa Animal Care and Use Committee (IACUC). BxPC3 cells (2.5 × 10⁶) stably expressing either EcSOD or empty vector were injected subcutaneously and tumor volumes were determined using calipers applying the equation: volume = length (longest diameter) × width (shortest diameter)² × 0.5. Animals were sacrificed when the tumors reached 15 mm in diameter.

Bx-EcSOD and Bx-Control cells (1 × 10⁷) were administered via intraperitoneal (i.p.) injection into athymic nude mice and tumor formation was monitored by bioluminescent imaging. VivoGloTM i.p. luciferin (15 mg/mL; Promega) was used, and after 5 minutes, mice were anesthetized by isoflurane inhalation maintained at 3% through a nose and imaged on the AMI-1000 (Spectral Instruments Imaging).

Statistical analyses for the in vitro and in vivo studies were examined and analyzed primarily by the Student t and Mann–Whitney tests. The resulting data from the studies with more than two groups were analyzed using the ANOVA or nonparametric ANOVA (Kruskal–Wallis) test, followed by a Tukey or Dunn multiple comparisons. Nonparametric tests were chosen when the sample size was small or the data were not normally distributed. All statistical tests were two-sided and assessed for significance at the 0.05 level (SAS 9.2) and GraphPad Prism 6.0.

EcSOD has been shown to be highly expressed in several normal human tissues including lung, breast, and pancreas (17, 18, 20, 21). However, in cancers of the breast, prostate, and lung, EcSOD expression, protein levels, and/or activity are frequently downregulated compared with nonmalignant tissues (18, 20, 31, 32). To test the hypothesis that EcSOD expression is similarly altered in patients with PDA, a TMA consisting of normal and malignant pancreatic ductal epithelium was evaluated. As shown in Fig. 1A, all core biopsies from normal pancreatic ductal epithelium had expression of EcSOD (28 of 28; 100%), while only 26 of 44 core biopsies from PDA had EcSOD expression (59%; P < 0.001). Because the core biopsies in the TMA examined were not from matched tumor-normal biopsies, specimens from patients treated with resection were evaluated to determine how EcSOD expression varied between malignant tissue and adjacent, nondiseased pancreatic ductal epithelium. As shown in Fig. 1B and C, a, EcSOD expression was reduced in tumor tissue compared with adjacent nondiseased pancreas in 11 of 16 specimens evaluated (69%; P = 0.0007). A decrease in EcSOD expression was also noted in areas of premalignant pancreatic ductal epithelium (P; PanIN-3) and chronic pancreatitis (CP; Fig. 1C, b).

Loss of EcSOD expression is expected to result in higher steady-state superoxide (O₂^•−) levels that can be evaluated via biomarkers indicative of oxidative damage to proteins. To evaluate the presence of a protein damage marker associated with the reactions of O₂^•− with nitric oxide (^•NO) to form peroxynitrite (ONOO⁻; ref. 33) in samples of human PDAs, immunoreactive 3-NT was evaluated in resected specimens of PDA (Fig. 1C, c and d). None of the nondisease adjacent normal pancreatic ductal epithelium (D) demonstrated significant staining for 3-NT (Fig. 1C and D). However, 12 of 17 (71%) PDA (T) stained positively for 3-NT (P < 0.0001; Fig. 1C, c). From these data, we conclude that the presence of 3-NT is a common finding in PDA, highlighting the reactions of O₂^•− with ^•NO that are ongoing as this disease progresses.

To evaluate changes in EcSOD mRNA expression, we evaluated specimens obtained from patients with PDA (n = 19) and compared the expression of EcSOD with specimens obtained from normal pancreas (n = 9). As shown in Fig. 1D, there was a significant reduction in mRNA expression in PDA tissues relative to normal pancreas (P = 0.001). An evaluation of Oncomine data revealed a similar pattern where EcSOD expression was reduced in five of seven studies evaluated, two of which were highly significant (Fig. 1E). From these data, we conclude that in addition to protein expression, EcSOD mRNA expression is also reduced in malignant pancreatic tumors. The decrease or absence of EcSOD expression in PDA relative to normal pancreas could explain, at least in part, the elevation of 3-NT seen in these tumor specimens.

To test the hypothesis that loss of EcSOD expression is predictive for a worse outcome in patients with PDA, a TMA composed of SEER patients with stage I–IV PDA was evaluated and clinical parameters and outcome were correlated with EcSOD expression. Among the 114 patients with PDA, EcSOD expression was absent in 41 (36%). Patients with complete loss of EcSOD were more often diagnosed with stage IV disease (54% vs. 30%; P = 0.019) and less often had surgical resection (39% vs. 67%; P = 0.004) compared with patients with intact EcSOD. In an adjusted proportional hazard model with the 114 patients with PDA, female gender and tumor resection were strongly associated with reduced risk for death. Factors independently predictive of increased risk of death included age >65 years, year of diagnosis, registry, distant disease, and loss of EcSOD expression [HR, 1.6; 95% confidence interval (CI), 1.0–2.6; P = 0.04]. The Kaplan–Meier analysis was performed to compare the overall survival of patients with intact (2+) and with complete loss (0+) of EcSOD expression. The median survival of patients with intact EcSOD expression was 11.0 months compared with 6.5 months for patients with loss of expression. As shown in Fig. 2, loss of EcSOD expression was associated with a nonsignificant trend toward decreased overall survival. From these data, we conclude that EcSOD expression is decreased in patients with PDA and that loss of EcSOD expression correlates with worse clinical and pathologic variables and may predict outcome.

To determine the expression of EcSOD in human pancreas and PDA cell lines, EcSOD mRNA was evaluated in three human PDA cell lines relative to HPNE, an immortalized but nontumorigenic human pancreatic ductal epithelial cell line. Similar to what was shown in clinical specimens, there was a significant reduction in EcSOD mRNA expression in PDA cells relative to HPNE (Supplementary Fig. S1A). To determine the impact of EcSOD overexpression on PDA cells, EcSOD was stably overexpressed in two PDA cell lines (Supplementary Fig. S1B and S1C) and subsequently characterized. Overexpression of EcSOD resulted in the detection of EcSOD protein in cell lysates as well as in the media, but did not impact expression of either MnSOD or CuZnSOD (Supplementary Fig. S1C). SOD activity in the culture media from transduced cells was increased as determined by an indirect colorimetric assay as well as by a direct, quantitative EPR (Supplementary Fig. S2A and S2B). These results show that active EcSOD protein was normally processed and secreted into the media.

These stably expressing cells were then used to test the hypothesis that overexpression of EcSOD would impact PDA tumor cell growth in culture. As shown in Fig. 3A and B, stable overexpression of EcSOD resulted in a significant reduction in clonogenic potential in BxPC3 and MiaPaCa2 PDA cell lines. EcSOD stably expressing cells also displayed a significantly increased doubling time compared with controls (Fig. 3C).

We next evaluated the impact of EcSOD overexpression on tumor cell invasion. BxPC3 and MiaPaCa2 cell lines stably overexpressing EcSOD and empty vector showed no difference in migration capacity at 24 hours (data not shown). EcSOD-overexpressing BxPC3 and MiaPaCa2 cells showed a significant reduction in invasion at 24 hours compared with controls (Fig. 4A and B). Parent BxPC3 cells were then transiently transduced with Ad.EcSOD or Ad.Empty for 72 hours, harvested, and placed on the Matrigel. As shown in Fig. 4C, cells transduced with EcSOD adenovirus also showed significantly less invasion compared with controls. To determine whether EcSOD would have a cell-autonomous effect on tumor cell invasion, conditioned media from stably expressing BxPC3 or MiaPaCa2 cells were harvested after 48 hours and placed onto the respective parent cells in the Matrigel. Conditioned media from the EcSOD-overexpressing cells, but not the parent cells, significantly reduced tumor cell invasion (Fig. 4D and E). These findings demonstrate the potent effect of EcSOD on tumor cell invasion both when expressed by cells and when enriched in the tumor microenvironment. These findings support the hypothesis that the more aggressive disease progression seen in patients who have lost EcSOD expression in their tumors may directly relate to less EcSOD activity in the extracellular environment.

To elucidate the mechanism of action of EcSOD on PDA cell invasion, BxPC3 cells were incubated with 20 mmol/L NAC, which is a nonspecific thiol antioxidant that scavenges a wide variety of ROS and reactive nitrogen species (RNS) containing species as well as reducing protein thiols (34). As shown in Fig. 5A, NAC significantly reduced invasion in control cells after 24 hours without a significant impact on the invasion of EcSOD-overexpressing PDA cells. These results support the hypothesis that EcSOD overexpression alters the redox balance leading to inhibition of PDA cell invasion. When the experiment was repeated using 10 U/mL catalase, a specific enzymatic scavenger of H₂O₂ (Fig. 5B), catalase did not significantly decrease invasion of either Bx-Control or Bx-EcSOD cells after 24 hours. These results support the hypothesis that extracellular H₂O₂ (which is the product of EcSOD reacting with superoxide) was not responsible for the inhibition of invasion caused by EcSOD over expression.

By catalyzing the dismutation of O₂^•−, EcSOD inhibits the reaction of nitric oxide (·NO) with O₂^•− to form RNS such as ONOO⁻ (35), and therefore the potential roles of RNS derived from reactions of ^•NO with O₂^•− were examined. Bx-EcSOD and Bx-Control cells were analyzed in a Matrigel assay that included the arginine analog L-NNA (100 μmol/L), a competitive inhibitor of NOS activity. L-NNA was found to significantly inhibit the accumulation of nitrite (which is derived from the reaction of ^•NO with O₂) in the media of BxPC3 cells supporting the conclusion that NOS activity was inhibited by L-NNA (Supplementary Fig. S3). Similar to what was seen with NAC, L-NNA did not significantly impact the invasive capacity of the Bx-EcSOD cells, while causing a significant inhibition of the invasive capacity of the Bx-Control cell line (Fig. 5C). These results support the hypothesis that the decrease in tumor cell invasion seen with the overexpression of EcSOD was due to the removal of O₂^•−, reducing the reactions of ^•NO with O₂^•−, and limiting the formation of ONOO⁻.

Superoxide-specific SOD mimetics have also been shown to limit the formation of ONOO⁻-induced damage to proteins, as assayed by the formation of immunoreactive 3-NT in the livers from irradiated animals (36). Therefore, the effects of the small-molecule superoxide dismutase mimetic, GC4419, on PDA cell invasion were determined. GC4419 is reported to be highly selective in its dismutation of O₂^•− (and its protonated form, hydroperoxyl radical, HOO^•), while not catalyzing reactions with other ROS under physiologic conditions, making it a suitable chemical probe to look at the specific role of O₂^•− in this biology. As shown in Fig. 5D, 50 μmol/L of GC4419 resulted in a significant reduction of tumor cell invasion. These results further support the hypothesis that excess extracellular O₂^•− significantly accelerates PDA cell invasion. GC4419 is currently being evaluated in a phase Ib clinical trial (NCT01921426) in patients with head and neck cancer receiving chemoradiation therapy, highlighting the potential translational impact of these findings.

To test the hypothesis that stable EcSOD expression would reduce tumor growth in vivo, Bx-EcSOD and Bx-Control xenografts were generated in athymic nude mice. Overexpression of EcSOD effectively suppressed tumor growth compared with controls (Fig. 6A). Bx-EcSOD tumors also took significantly longer time to double in size and mice bearing EcSOD-expressing BxPC3 tumors had prolonged median survival compared with mice bearing empty vector tumors (Fig. 6B and C). To examine the impact of EcSOD overexpression on peritoneal tumor formation, Bx-EcSOD and Bx-Control cells were delivered via i.p. injection into athymic nude mice and monitored as a function of time with bioluminescence. The peritoneal cavity provides an environment more similar to clinical PDA than the xenograft tumor model. The bioluminescence signal was equivalent during the first 5 days after injection but at 16 and 28 days after injection, the animals bearing EcSOD stably overexpressing cells formed less peritoneal tumor growth as compared with controls (Fig. 6D and E).

Oxidative stress as a contributor to malignant disease progression has been extensively studied and proposed to play a role in several solid tumors refractory to conventional therapies (13, 18, 20, 31, 37). PDA is currently the fourth leading cause of cancer-related deaths in the United States and remains highly lethal due to a poor understanding of the biology of the disease and resistance to available, conventional therapies (38). Prior evaluations of SODs in PDA have demonstrated potential for antioxidant therapy as a treatment for this disease; however, the mechanism underlying the anticancer effects of this treatment and translational avenues to the clinic remain relatively unexplored (17, 19).

The current studies describe both analyses of clinical specimens and preclinical in vitro and in vivo studies, testing the hypothesis that the loss of EcSOD leads to an increase in the progression and invasiveness of PDA. These investigations are also the first to provide an in-depth investigation correlating EcSOD with clinical outcomes in patients with PDA. In addition, these studies examined the effect of EcSOD activity on invasion and metastasis in in vitro and in vivo models of PDA using multiple methods of EcSOD overexpression.

The results using patient samples indicate that loss of EcSOD expression is a common feature of PDAs that is accompanied by an increase in 3-NT formation (39), which may indicate increased protein nitration mediated by ONOO⁻. The loss of EcSOD expression in patient samples correlated with more advanced disease and a trend toward shorter overall survival. These results support the hypothesis that loss of EcSOD expression in PDAs may lead to an increase in ONOO⁻ in the extracellular environment that could contribute to a more aggressive disease phenotype. Because the primary route of ONOO⁻ formation is via reaction of O₂^•− with ^•NO, either increases in O₂^•− or ^•NO would be expected to give rise to increase ONOO⁻ formation, which we propose would be exacerbated by loss of EcSOD expression.

In support of the aforementioned hypothesis, we also found in preclinical cell culture models that both the stable and transient overexpression of EcSOD in PDA cells decreased cloning efficiency and increased the doubling time of PDA cells. This was true for both cell-directed expression and using conditioned media. In addition, in vitro overexpression of EcSOD or treatment with the O₂^•−-specific small-molecule dismutase mimetic (GC4419) also caused significant decreases in PDA cell invasion.

Interestingly, in parent BxPC3 cells, but not BxPC3 cells overexpressing EcSOD, in vitro invasion was inhibited by a thiol antioxidant (NAC) and a ^•NOS inhibitor (L-NNA). NAC is a nonspecific thiol antioxidant that can react directly with several ROS/RNS, react to keep protein thiols reduced, as well as be taken up by cells to increase levels of small molecular weight thiols, such as cysteine, γ-glutamylcysteine and glutathione (34). Metabolic oxidative stress activates signal transduction and gene expression during glucose deprivation in human tumor cells (40). Because NAC only had an effect on invasion in the absence of EcSOD overexpression, these results support the conclusion that redox imbalances reversed by NAC in the absence of EcSOD are contributing to the invasive properties of BxPC3 cells. Furthermore, the finding that L-NNA also only inhibited invasion in the absence of EcSOD overexpression, supports the hypothesis that reactions of ^•NO with O₂^•− only contribute to invasion of BxPC3 cells when EcSOD is minimally expressed. Taken together, these results support the hypothesis that loss of EcSOD in PDA leads to increased reactions of ^•NO with O₂^•− that contribute to alterations in the extracellular redox balance leading to the invasive phenotype. This idea is further supported by the increases in 3-NT seen in human PDA specimens. Finally, these results support the speculation that the formation of ONOO⁻ may contribute to the aggressive progression of clinical PDAs.

Although we have not addressed the precise mechanism of how ONOO⁻ could contribute to invasiveness and progression of PDA in this study, a previous report found that extracellular lecithinized superoxide dismutase (PC-SOD) inhibited metastasis in a lung cancer animal model by a mechanism that might involve the inhibition of ONOO⁻ formation leading to less damage the extracellular matrix and reduced tumor metastasis (41). Future studies will seek to identify the targets of RNS such as ONOO⁻ that may contribute to increase disease progression in PDA.

Here, we have shown that loss of EcSOD expression in human PDA may be associated with more aggressive disease progression and poorer patient outcomes; however, the mechanism underlying silencing of EcSOD in PDA remains unknown. In human lung and breast cancers, EcSOD is lost at the transcriptional level in association with a distinct gain of aberrant cytosine methylation in the EcSOD promoter region (18, 20), so this mechanism may also extend to PDA and will be the subject of future inquiry. Overall, our results support the hypothesis that increasing SOD activity in the extracellular space could provide a useful target for limiting disease progression in PDA. Furthermore, these results continue to support the hypothesis that small molecule SOD mimetics could be useful adjuvants to therapies designed to limit disease progression.

R.A. Beardsley is an employee of Euclise Pharmaceuticals; has ownership interest (including patents) in Confluence Life Sciences, Euclises Pharmaceuticals, and Galera Therapeutics; and is a consultant/advisory board member for Confluence Life Sciences. J. Keene has ownership interest (including patents) in Galera Therapeutics. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J.E. Hrabe, B.G. Allen, A.J. Case, G.R. Buettner, F.E. Domann, D.R. Spitz, J.J. Mezhir

Development of methodology: B.R. O'Leary, M.A. Fath, A.M. Bellizzi, J.E. Hrabe, B.G. Allen, A.J. Case, B.A. Wagner, G.R. Buettner, F.E. Domann, J.J. Mezhir

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B.R. O'Leary, M.A. Fath, A.M. Bellizzi, A.J. Case, S. Altekruse, B.A. Wagner, G.R. Buettner, C.F. Lynch, B.Y. Hernandez, W. Cozen, R.A. Beardsley, J. Keene, M.D. Henry, D.R. Spitz, J.J. Mezhir

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B.R. O'Leary, M.A. Fath, A.M. Bellizzi, J.E. Hrabe, A.M. Button, A.J. Case, S. Altekruse, B.A. Wagner, G.R. Buettner, R.A. Beardsley, J. Keene, M.D. Henry, D.R. Spitz, J.J. Mezhir

Writing, review, and/or revision of the manuscript: B.R. O'Leary, M.A. Fath, A.M. Bellizzi, J.E. Hrabe, A.M. Button, S. Altekruse, B.A. Wagner, G.R. Buettner, C.F. Lynch, B.Y. Hernandez, W. Cozen, R.A. Beardsley, J. Keene, M.D. Henry, F.E. Domann, D.R. Spitz, J.J. Mezhir

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B.R. O'Leary, J.E. Hrabe, S. Altekruse, B.Y. Hernandez, J.J. Mezhir

Study supervision: J.J. Mezhir

The authors thank Dr. James Crapo and Dr. Rebecca Oberley-Deegan (National Jewish Medical and Research Center, Denver, CO) for providing antibodies to EcSOD. The authors also thank Dr. Apollina Goel (Free Radical and Radiation Biology Program, The University of Iowa, Iowa City, IA) for the use of a Tecan plate reader and Dr. Nalin Goonesekere for assistance with Oncomine data gathering and interpretation.

This work was supported by the American Surgical Association Foundation Fellowship (to J.J. Mezhir); This work was supported by Grant IRG-77-004-31 from the American Cancer Society, administered through The Holden Comprehensive Cancer Center at The University of Iowa (to J.J. Mezhir); NIH grant T32CA148062 (to J.J. Mezhir); University of Iowa Carver College of Medicine Collaborative Pilot Grant (J.J. Mezhir); generous gifts from the families of John Van Sickel and Jack Kloet and grateful patient donations to the Holden Comprehensive Cancer Center for pancreatic cancer research. Radiation and Free Radical Research Core in the Holden Comprehensive Cancer Center at The University of Iowa core laboratory support through NIH P30 CA086862. D.R. Spitz and M.A. Fath were supported by R01 CA182804, R01CA133114, and R21 CA161182; G.R. Buettner and B.A. Wagner were supported by R01 CA169046; F.E. Domann was supported by RO1 CA115438; and M.D. Henry was supported by RO1 CA130916. Construction of the pancreatic TMA was supported by NCI contract numbers N01-PC-35143 (Iowa), N01-PC-35137 (Hawaii), and N01-PC-35139 (Los Angeles).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.