Prabhat C. Goswami, PhD, Free Radical & Radiation Biology Program, Department of Radiation Oncology, University of Iowa, B180 Medical Laboratories, Iowa City, IA 52242, USA. Tel.: (319) 384-4666; fax: (319) 335-8039; e-mail: firstname.lastname@example.org
In recent years, the intracellular reactive oxygen species (ROS) levels have gained increasing attention as a critical regulator of cellular proliferation. We investigated the hypothesis that manganese superoxide dismutase (MnSOD) activity regulates proliferative and quiescent growth by modulating cellular ROS levels. Decreasing MnSOD activity favored proliferation in mouse embryonic fibroblasts (MEF), while increasing MnSOD activity facilitated proliferating cells’ transitions into quiescence. MnSOD (+/–) and (–/–) MEFs demonstrated increased superoxide steady-state levels; these fibroblasts failed to exit from the proliferative cycle, and showed increasing cyclin D1 and cyclin B1 protein levels. MnSOD (+/–) MEFs exhibited an increase in the percentage of G2 cells compared to MnSOD (+/+) MEFs. Overexpression of MnSOD in MnSOD (+/–) MEFs suppressed superoxide levels and G2 accumulation, decreased cyclin B1 protein levels, and facilitated cells’ transit into quiescence. While ROS are known to regulate differentiation and cell death pathways, both of which are irreversible processes, our results show MnSOD activity and, therefore, mitochondria-derived ROS levels regulate cellular proliferation and quiescence, which are reversible processes essential to prevent aberrant proliferation and subsequent exhaustion of normal cell proliferative capacity. These results support the hypothesis that MnSOD activity regulates a mitochondrial ‘ROS-switch’ favoring a superoxide-signaling regulating proliferation and a hydrogen peroxide-signaling supporting quiescence.
Cellular redox environment is influenced by the production of reactive oxygen species (ROS) and their removal by antioxidants. ROS (superoxide, hydrogen peroxide, hydroxyl radical, singlet molecular oxygen, and organic hydroperoxides) are primarily produced intracellularly by two metabolic sources: the mitochondrial electron transport chain and enzymatic reactions. ROS levels are tightly regulated by antioxidant enzymes and low molecular weight antioxidants (McCord & Fridovich, 1969; Oberley, 1982; Fernandes & Holmgren, 2004). Manganese superoxide dismutase (MnSOD) is a nuclear-encoded and mitochondrial matrix-localized antioxidant enzyme that catalyzes the conversion of superoxide to hydrogen peroxide, which is further converted to water by glutathione peroxidase and catalase. Traditionally, ROS were thought to be toxic by-products of living in an aerobic environment, causing damage to cellular macromolecules. However, in recent years, it is increasingly evident that ROS levels regulate multiple cellular processes, including proliferation (Burdon & Gill, 1993; Sundaresan et al., 1995; Bae et al., 1997; Irani et al., 1997).
Previous studies reported that ROS signalling regulates cellular proliferation in response to mitogenic stimuli including platelet-derived growth factor, epidermal growth factor, cytokine, antigen receptors, and oncogenic Ras (Sundaresan et al., 1995; Bae et al., 1997; Irani et al., 1997). Results from these previous studies suggested membrane-bound nicotinamide adenine dinucleotide phosphate (NADPH) oxidase as the source of ROS signalling. We have shown previously that the cellular redox environment could regulate cell cycle phase transitions (Goswami et al., 2000; Menon et al., 2003; Sarsour et al., 2005; Menon & Goswami, 2007). Mitochondria being the major source of cellular ROS, it is important to determine whether MnSOD activity and mitochondria-generated ROS regulate normal cell proliferation.
Normal cell proliferation has two distinctive stages: the proliferative cycle and quiescent growth. The proliferative cycle includes entry into, and progression through the cell cycle. It also includes transitioning into the quiescent growth, which is essential to prevent aberrant proliferation as well as protecting cellular lifespan. Transition from quiescence (G0) to proliferative growth (G1 to S to G2 and M) involves the sequential activation of cyclin-dependent kinase activities (Sherr, 1993; Grana & Reddy, 1995). Progression from G0/G1 to S is largely regulated by the D-type cyclins in association with the cyclin-dependent kinases, CDK4/6. Cyclin D1/CDK4-6 kinase complex partially phosphorylates the retinoblastoma protein, causing the release of the E2F family of proteins, which transcriptionally activates expression of multiple S-phase-specific genes required for DNA replication (Nevins, 1992). Cyclin B1/CDK1 kinase complex, along with Cdc25C phosphatase, regulates progression from G2 to M phase. Because cyclins are the positive regulators of cellular proliferation, cyclin D1 and cyclin B1 protein levels are often used as indicators of cells entering into and exiting from the proliferative cycle.
We investigated the hypothesis that MnSOD activity regulates transitions between proliferative and quiescent growth by modulating cellular ROS (superoxide and hydrogen peroxide) levels. Our results show that a decrease in MnSOD activity and, therefore, an increase in superoxide steady-state levels, promotes proliferation. In contrast, an increase in MnSOD activity and the resultant decrease in superoxide levels support quiescence. Decreasing MnSOD activity was associated with an increase in cyclin D1 and cyclin B1 protein levels supporting proliferative growth, while increasing MnSOD activity decreased cyclin B1 and facilitated fibroblasts’ transition into quiescent growth.
MnSOD activity regulates transitions between quiescent and proliferative growth
Initially, we determined whether changes in MnSOD activity affect cellular superoxide steady-state levels and activities of other antioxidant enzymes. Exponentially growing asynchronous cultures of mouse embryonic fibroblasts (MEFs) were harvested and total cellular protein extracts were used for measurement of antioxidant enzyme activity. MnSOD activity was higher in wild-type (+/+) compared to the heterozygous (+/–) cells and absent in homozygous knockout (–/–) MEFs (Fig. 1A). Changes in MnSOD activity did not significantly alter CuZnSOD and catalase activities. Cells from replicate dishes were collected for electron paramagnetic resonance (EPR) spectroscopy measurements of superoxide steady-state levels (Fig. 1A, bottom panel). The steady-state levels of superoxide was low in MnSOD (+/+) MEFs and increased approximately fivefold in MnSOD (+/–) MEFs. MnSOD (–/–) MEFs showed an approximately ninefold higher level of superoxide compared to MnSOD (+/+) MEFs.
To determine whether MnSOD activity regulates cellular proliferation, cell growth was measured in MnSOD (+/+), MnSOD (+/–) and MnSOD (–/–) MEFs. Results from the cell growth assay showed all three cell types had an initial lag period followed by exponential growth (Fig. 1B). MnSOD (+/+) fibroblasts reached confluence (quiescence; < 5% S-phase) between 6 and 9 days postplating. MnSOD (+/–) fibroblasts continued to grow and failed to exit from the proliferative cycle. MnSOD (–/–) had a much longer lag period and cell proliferation was significantly suppressed. Cell population doubling time for the three cell types were as follows: MnSOD (+/+), 41 h; MnSOD (+/–), 47 h; MnSOD (–/–), 110 h.
MnSOD activity was lower in wild-type exponential cultures and increased as cells transitioned into quiescence (Fig. 1B inset). The growth state-dependent changes in MnSOD activity were independent of MnSOD protein levels and mRNA, which did not show any difference in quiescent growth (9 days postplating) compared to exponential growth (4 days postplating, Fig. 1B inset and data not shown). MnSOD activity remained low in MnSOD (+/–) MEFs at 6 and 9 days postplating, which is consistent with these cells’ inability to exit from the proliferative cycle.
To determine whether MnSOD activity regulates progression through the proliferation cycle, fibroblasts from replicate dishes were trypsinized and fixed in ethanol. Ethanol-fixed cells were analyzed for DNA content by flow cytometry. Representative DNA histograms are shown (Fig. 1C, upper panels) and quantitation of results is presented (Fig. 1C, lower panels). The percentage of S-phase cells was higher (23%) at 2 days postplating in MnSOD (+/+) fibroblasts, gradually decreasing to less than 4% at 12 days postplating as cells transitioned into quiescence. These results suggest that MnSOD (+/+) MEFs retained the capacity to successfully transit into quiescence from the proliferative growth state.
The extended lag period in MnSOD (–/–) fibroblasts’ growth (Fig. 1B) was associated with a large increase in the percentage of cells with sub-G1 DNA content (Fig. 1C); indicating an absence of MnSOD activity resulted in extensive cell loss. However, it appears that MnSOD (–/–) fibroblasts must have adapted to proliferation because the percentage of cells with sub-G1 DNA content decreased at 12 days postplating, coinciding with an increase in the percentage of S-phase (14–16%) as well as an increase in cell number.
The percentage of S-phase in MnSOD (+/–) fibroblasts was approximately 15% at day 2 and remained relatively constant during the 12 days of culture, indicating a decrease in MnSOD activity inhibited MEFs exit from the proliferative cycle (Fig. 1C). The percentages of G2 + M-phases in MnSOD (+/+) and MnSOD (–/–) MEFs were comparable. However, MnSOD (+/–) fibroblasts exhibited a significantly higher percentage (> 40%) of cells in G2 + M (Fig. 1C).
A bromodeoxyuridine (BrdU) pulse chase assay was used to determine whether the increase in percent G2 in MnSOD (+/–) MEFs could be due to a delay in G2 transit (Fig. 2). Exponentially growing asynchronous cultures were pulse-labelled with BrdU and continued in culture in BrdU-free media. MEFs were harvested at various times post-BrdU labelling, and ethanol-fixed cells were assayed for BrdU-positive and -negative population by flow cytometry. Three distinct cell populations, BrdU-positive S-phase and BrdU-negative G1 and G2, were detected at the end of 30-min BrdU labelling for both cell types (Fig. 2A). Representative histograms demonstrating cell cycle progression at 2, 4, and 8 h post-BrdU labelling are presented in Fig. 2(A). Results showed cells in G2 progressed normally to M-phase with a median transit duration of approximately 3 h in MnSOD (+/+) fibroblasts (Fig. 2B, left panel). However, the median G2 transit time was higher (> 8 h) in MnSOD (+/–) fibroblasts. Likewise, the fraction of BrdU-positive cells completing cell division (cells in box marked ) was higher in MnSOD (+/+) vs. MnSOD (+/–) fibroblasts (Fig. 2B, right panel) suggesting the duration of transit through G2 and M phases was longer in MnSOD (+/–) compared to MnSOD (+/+) MEFs.
Next we determined whether the delay in G2 + M transit could be due to MnSOD (+/–) MEFs’ inability to maintain an appropriate superoxide threshold level. MnSOD (+/–) fibroblasts were treated with a superoxide scavenging agent Tiron and cells were harvested for EPR and flow cytometry measurements. Tiron decreased the superoxide steady-state levels by approximately 50%; these cells also exhibited a decrease in the percentage of cells in G2-phase (25% in control vs. 10% in Tiron-treated cells; Fig. 2D, left panel). In a separate experiment, asynchronously growing exponential cultures of MnSOD (+/+) fibroblasts were infected with adenovirus carrying a dominant negative mutant form of MnSOD, and cells were harvested for measurements of superoxide steady-state levels and percent G2 + M. Overexpression of the mutant MnSOD showed a significant decrease in MnSOD activity (Fig. 2E, inset). The decrease in MnSOD activity was associated with a fourfold increase in the percentage of cells in G2 + M phase (Fig. 2E). These results were comparable with results obtained from the MnSOD (+/–) MEFs suggesting that the cells’ failure to maintain an appropriate superoxide threshold level could affect G2 + M transit and subsequent transition into quiescent growth in daughter generations.
To determine whether MnSOD activity regulates quiescent fibroblasts’ entry into the proliferative cycle, 10- and 30-day quiescent MEFs were replated at a lower cell density and cell growth measured for 6 days (Fig. 3). Cell cycle phase distributions prior to replating were assayed by flow cytometry analysis of DNA content. While the percentage of G1 cells was higher in 30-day compared to 10-day MnSOD (+/+) MEFs, the percentage of G1 cells did not differ in 10- and 30-day cultures of MnSOD (+/–) and (–/–) MEFs (Fig. 3A). The percentage of S-phase cells was higher in MnSOD (+/–) and (–/–) MEFs compared to wild-type cells suggesting that these cells might not have exited from the proliferative cycle. MEFs subcultured from 10-day confluent cultures showed comparable growth characteristics among the three cell types (Fig. 3B, left panel). Interestingly, while cell growth was comparable in fibroblasts subcultured from 30-day MnSOD (+/–) and (–/–) MEFs, it was significantly inhibited in MnSOD (+/+) MEFs (Fig. 3B, right panel).
MnSOD activity regulated transition from quiescent to the proliferative cycle was further evident from the results presented in Fig. 3(C). Quiescent cultures of MnSOD (+/+) MEFs were infected with adenovirus carrying a dominant negative mutant form of MnSOD (AdmMnSOD) cDNA and assayed for MnSOD activity and superoxide levels. Overexpression of the dominant negative mutant MnSOD significantly increased superoxide levels (Fig. 3C). Control and adenovirus-infected cells were replated at lower cell density and harvested at 48 h postreplating for flow cytometry analysis of cell cycle phase distributions. The percentage of G1 cells decreased in control fibroblasts and fibroblasts infected with adenovirus containing an empty vector (AdBgl), suggesting that these cells entered into S-phase. The percentage of G1 cells was higher in AdmMnSOD overexpressing MEFs indicating that the appropriate amount of MnSOD activity during quiescence is required to maintain MEFs capacity to re-enter the proliferative cycle. Similar results were obtained from quiescent normal human skin fibroblasts (NHFs) that were infected with AdmMnSOD (Supplementary Fig. S1). Quiescent NHFs were infected with adenoviruses carrying wild-type and mutant form of human MnSOD cDNAs. Control and adenovirus-infected cells were replated at a lower cell density and cell numbers counted at 2, 4, and 6 days postplating. Cell growth was significantly suppressed in AdmMnSOD-infected cells compared to control cells and cells infected with adenovirus containing human MnSOD cDNA (AdhMnSOD) (Supplementary Fig. S1).
MnSOD activity-regulated transitions between quiescence and proliferative growth are associated with changes in cyclin D1 and cyclin B1 protein levels
To determine whether MnSOD activity could influence cell cycle regulatory proteins, immunoblotting assays were performed using total cellular protein extracts isolated from MnSOD (+/+) and MnSOD (+/–) fibroblasts harvested at indicated times. These measurements were not made in MnSOD (–/–) MEFs because of extensive cell death and subsequent aberrant proliferation indicating that these cells might have undergone transformation (see next section below). Both cyclin D1 and cyclin B1 protein levels decreased in MnSOD (+/+) fibroblasts as cells exited from the proliferative cycle. Cyclin D1 protein levels decreased to 17%, and cyclin B1 protein levels decreased to 14% at 9 days postplating (Fig. 4A). The decrease in cyclin D1 and B1 protein levels corresponded to a decrease in S-phase from 25 to < 5%. However, cyclin D1 and B1 protein levels showed minimal changes in MnSOD (+/–) fibroblasts during 4–9 days postplating (Fig. 4A), correlating with cells’ inability to exit from the proliferative cycle.
In a separate series of experiments, it was determined whether MnSOD overexpression could facilitate MnSOD (+/–) fibroblasts’ exit from the proliferative cycle. MnSOD (+/–) MEFs were infected with adenovirus carrying mouse wild-type MnSOD cDNA (AdMnSOD) and then harvested for measurements of MnSOD activity, superoxide steady-state levels, and cyclin B1 protein levels. MnSOD overexpression exhibited a dose-dependent increase in MnSOD activity (Fig. 4B). This increase corresponded to a decrease in superoxide steady-state levels (Supplementary Fig. S3). Furthermore, MnSOD (+/–) fibroblasts infected with 50 and 100 MOI AdMnSOD significantly decreased cyclin B1 protein levels (Fig. 4B). At 3 days postinfection, cell numbers were similar in MnSOD (+/–) control, AdBgl II, and AdMnSOD-infected cells (Fig. 4C). The cell numbers in MnSOD (+/–) fibroblasts control, AdBgl II, and 30 MOI AdMnSOD-infected cells continued to increase at 6 days postinfection. However, MnSOD (+/–) fibroblasts infected with 100 MOI AdMnSOD showed no change in cell numbers at 6 vs. 3 days postinfection (Fig. 4C). These results suggest MnSOD overexpression facilitated MnSOD (+/–) MEFs’ transition from the proliferative to quiescent growth.
Decrease in MnSOD activity was associated with aberrant proliferation
The hypothesis that MnSOD activity regulates transitions between proliferative and quiescent growth was further evident in long-term cultures. MEFs were maintained in cultures with regular change in growth medium and cell numbers were measured at 6 and 30 days postplating (Fig. 4D). Cell numbers in MnSOD (+/+) fibroblasts were comparable at 6 and 30 days postplating, suggesting these cells exited from the proliferative cycle and maintained quiescent growth. However, cell numbers in MnSOD (+/–) and MnSOD (–/–) fibroblasts continued to increase, suggesting these cells failed to exit from the proliferative cycle (Fig. 4D, left panel). The aberrant proliferation in MnSOD (+/–) and (–/–) MEFs were also evident in long-term culture dishes stained with Coomassie dye (Fig. 4D, right panel). MnSOD (+/+) fibroblasts exhibited contact inhibited monolayer cultures characteristic of normal cell growth. In contrast, cell densities were significantly higher in MnSOD (+/–) fibroblasts and isolated foci with intense cell growth were observed in MnSOD (–/–) fibroblasts.
MnSOD activity regulates an ROS switch during transitions between proliferative and quiescent growth
MnSOD activity converts superoxide produced in the mitochondria to hydrogen peroxide, thereby regulating cellular ROS steady-state levels. To determine whether changes in ROS levels are associated with transitions between proliferative and quiescent growth, dihydroethidium (DHE; indicative of superoxide) and 5-(and 6)-chloromethyl-2′,7′-dichlorodihydro fluorescein diacetate (DCFH; indicative of hydrogen peroxide) fluorescence were measured by flow cytometry. The specificity of DHE and DCFH assays for measurements of superoxide and hydrogen peroxide levels is shown in results presented in Fig. 3(C) and supplementary Fig. S2. Results (Fig. 5A) showed that steady-state levels of superoxide were higher in MnSOD (+/+) exponential cultures (2 days postplating, 20–25% S-phase) compared to quiescent growth (20 days postplating, < 5% S-phase). Hydrogen peroxide levels were lower in exponential cultures (2 days postplating) compared to quiescent cultures (20 days postplating, Fig. 5B). The steady-state levels of superoxide in MnSOD (+/–) and MnSOD (–/–) fibroblasts were higher both in 2- and 20 days cultures. The increase in superoxide and decrease in hydrogen peroxide levels in MnSOD (+/–) and (–/–) fibroblasts in 20-day cultures were associated with cells’ inability to exit from the proliferative cycle. Consistent with these results, the percentage of S-phase cells remained higher in MnSOD (+/–) and (–/–) MEFs (15 and 25%, respectively). Interestingly, a direct correlation was observed between superoxide levels and percent S-phase (R2 = 0.95, Fig. 5C).
The concept of mitochondria-derived ROS regulating quiescent and proliferative growth was further supported by results presented in Fig. 6. Exponential and quiescent cultures of MnSOD (+/+) MEFs were assayed for extracellular hydrogen peroxide levels using the horseradish peroxidase–p-hydroxyphenyl acetic acid (HRA-pHPA) method (Fig. 6A). Hydrogen peroxide levels were low in exponential cultures and increased approximately fourfold in quiescent cultures. The specificity of the assay for hydrogen peroxide measurements was determined using catalase in the reaction buffer. Catalase treatment neutralized hydrogen peroxide both in exponential and quiescent cultures (Fig. 6A, upper panel). The increase in hydrogen peroxide levels was associated with a decrease in the percentage of S-phase cells (Fig. 6A, bottom panel). In a separate set of experiments, exponential and quiescent cultures of MnSOD (+/+) MEFs were incubated with MitoSox and MitoTracker dyes and fluorescence measured by flow cytometry. MitoTracker fluorescence was comparable in exponential and quiescent cultures. However, exponential cultures showed significantly higher (approximately four- to fivefold) MitoSox fluorescence compared to quiescent cultures (Fig. 6B). These results were consistent with results presented in Fig. 5 supporting the hypothesis that mitochondria-generated ROS could regulate transitions between quiescent and proliferative growth.
The concept of cellular redox environment regulating proliferation can be traced back to 1931 when Rapkine first demonstrated the presence of soluble thiols that fluctuated cyclically during cell division in sea urchin eggs. Later, in 1960, Kawamura demonstrated that protein thiol staining increased as the mitotic spindle was assembling in sea urchin eggs. Subsequently, Mauro et al. (1969) reported that the protein-bound and nonprotein sulfhydryl (-SH) and disulfide (-SS-) groups oscillate during the mammalian cell cycle. Consistent with these earlier observations, we reported previously that the fluorescence of a pro-oxidant sensitive dye increased two- to threefold in G2 compared to G1, a pro-oxidant event is necessary prior to S-entry, and pharmacological manipulations of cellular redox environment resulted in G1 delay (Goswami et al., 2000; Menon et al., 2003, 2005; Sarsour et al., 2005). While these previous reports support the hypothesis of cellular ROS regulating proliferation, a direct relationship between antioxidant enzyme, ROS, and cell proliferation is currently lacking.
The present study was designed to determine whether MnSOD activity and, therefore, mitochondria-derived ROS regulate transitions between quiescent and proliferative growth. Mouse embryonic fibroblasts with MnSOD wild-type (+/+) (+/–) and (–/–) genotypes were used to demonstrate that MnSOD activity regulates cellular ROS levels favouring a superoxide-signalling regulating proliferation and hydrogen peroxide-signalling supporting quiescent growth. For many years, ROS were only thought to be toxic unwanted by-products of living in an aerobic environment. It was suggested that ROS cause random damage to cellular macromolecules and this cumulative damage of critical cellular macromolecules is believed to be the cause of slow degeneration of biological systems during the aging process. This idea was originally proposed in the Free Radical Theory of Aging and Oxygen Toxicity (Harman, 1956). However, numerous recent reports, including our own research, suggest ROS are essential physiological requirements for maintaining cellular redox-homeostasis, and that lower levels of ROS, in fact, act as secondary messengers regulating multiple cellular processes including proliferation.
Our results show mouse embryonic fibroblasts with different MnSOD genotypes exhibited discrete differences in cellular proliferation. MnSOD (+/+) fibroblasts demonstrated an initial lag period followed by exponential and quiescent growth, characteristics of normal cellular proliferation (Fig. 1). An additional characteristic of normal cell proliferation relates to the gradual loss in proliferation capacity of quiescent cells. We have defined previously this property of normal cells as the telomerase-independent nonreplicative senescence (Sarsour et al., 2005). MnSOD (+/+) fibroblasts subcultured from 10- and 30-day quiescent growth showed significant differences in their proliferative capacity. The percentage of S-phase decreased to approximately 5% at 48 h of replating of 30-day quiescent cultures, which correlated with an inhibition in cell growth.
Although MnSOD (+/–) and (–/–) fibroblasts exhibited an initial lag period followed by exponential growth, these cells were unable to exit from the proliferation cycle and failed to establish quiescent growth (Fig. 1). Likewise, both cell types maintained their proliferative capacities upon subculturing from 10- and 30-day cultures (Fig. 3). Additional differences were also observed in exponential cultures of MnSOD (+/–) and MnSOD (–/–) fibroblasts compared to wild-type cells. MnSOD (+/–) fibroblasts showed a significant increase in percentage of G2 cells. Results from the BrdU-pulse-chase assay demonstrated that the increase in the percentage of G2 cells in MnSOD (+/–) fibroblasts was primarily due to a delayed transit through G2, indicating that an appropriate amount of MnSOD activity is required for progression from G2 to M and subsequent exit from the proliferative cycle in daughter generations (Fig. 2). MnSOD (–/–) fibroblasts maintained a higher percentage of cells with sub-G1 DNA content until 9 days postplating (Fig. 1). This increase in the percentage of cells with sub-G1 DNA content suggested the lack of MnSOD activity is detrimental to cells. However, it appears that MnSOD (–/–) fibroblasts overcame this initial wave of cell loss during 9–12 days postplating and adapted to proliferation, which is indicated by a corresponding increase in S-phase percentage. These results clearly suggest MnSOD activity is required for proper transitions between quiescent and proliferative growth.
A key concept emerged from our results, which indicates MnSOD activity regulates a ‘ROS-switch’ favouring a superoxide-signal regulating proliferation and a hydrogen peroxide-signal supporting quiescent growth. MnSOD is a nuclear encoded mitochondrial localized antioxidant enzyme known to convert mitochondrial-generated superoxide to hydrogen peroxide. Therefore, changes in MnSOD activity are anticipated to influence cellular ROS steady-state levels. Consistent with this hypothesis our results show increased superoxide steady-state levels in exponential cultures of all three MnSOD genotype fibroblasts (Fig. 5). In fact, a direct correlation was observed between superoxide steady-state levels and percent S-phase (R2 = 0.95, Fig. 5). Consistent with this correlation, superoxide steady-state levels were low as MnSOD (+/+) fibroblasts exited from the proliferative cycle, which was accompanied by an increase in hydrogen peroxide levels in quiescent growth (Fig. 5). This growth-state-related transition from superoxide to hydrogen peroxide was absent in MnSOD (+/–) and (–/–) fibroblasts, and accordingly these cells failed to exit from the proliferative cycle. The failure to exit from the proliferation cycle in MnSOD (+/–) fibroblasts was probably due to a slower exit from G2 because scavenging of superoxide with Tiron (or overexpressing MnSOD) decreased G2-delay and these cells subsequently exited from the proliferation cycle (Fig. 2C,D). The concept of a ‘ROS-switch’ regulating periodic events during cellular proliferation is further supported by a recent report (Tu et al., 2005). These authors demonstrated that budding yeast exhibits a metabolic redox-cycle consisting of a reductive nonrespiratory phase and an oxidative respiratory phase. This metabolic redox cycle coordinates with periods of gene expression regulating essential cellular and metabolic events. The initiation of cell cycle occurs very late during the oxidative phase and the reductive phase triggers expression of many of the genes required for DNA replication. Our results show an inverse correlation between the percentage of S-phase and extracellular hydrogen peroxide levels (Fig. 6A). Extracellular hydrogen peroxide levels were lower in exponential compared to quiescent MnSOD (+/+) MEFs. Interestingly, MitoSox-fluorescence was higher in exponential compared to quiescent MnSOD (+/+) MEFs (Fig. 6B). These results are consistent with results presented in Fig. 5 supporting the hypothesis that MnSOD activity and mitochondrially generated ROS levels could link cellular metabolism to proliferative and quiescent growth (Fig. 7).
Recent evidence suggests that ROS could reversibly modify the redox-state of specific cysteine in phosphatases (protein tyrosine phosphatases and dual specificity phosphatases) inactivating their activities, which could favour Ser/Thr phosphorylation-dependent signal pathways initiating proliferation (Finkel, 2003; Rhee et al., 2003; Tonks, 2005). While the source of ROS is not completely understood, it is suggested that ROS generated from membrane-localized NADPH oxidase could initiate the signalling cascades. Our results indicate that a cross-talk between MnSOD-activity (possibly via mitochondria-derived ROS-levels pathways) and cell cycle regulatory proteins could significantly influence the transitions between quiescent and proliferative growth. Consistent with this hypothesis MnSOD activity and ROS levels were found to be associated with changes in cyclin D1 and cyclin B1 protein levels. Cyclin D1 is believed to be the first cell cycle regulatory protein that responds to mitogenic stimuli facilitating cells’ entry into S-phase. Cyclin B1 is known to regulate progression from G2 to M. Both protein levels were higher in asynchronously growing exponential cultures of MnSOD (+/+) fibroblasts, and decreased as cells exited from the proliferative cycle (Fig. 4). Cyclin D1 and cyclin B1 protein levels remained high in MnSOD (+/–) fibroblasts, consistent with their failure to exit from the proliferative cycle (Figs 1 and 7). However, overexpression of MnSOD in MnSOD (+/–) fibroblasts decreased cyclin B1 protein levels. These cells managed to exit from the proliferative cycle, indicated by a decrease in superoxide steady-state levels and minimal change in cell number between 3 and 6 days postinfection (Fig. 4). Although the exact mechanisms of cellular redox environment regulating cell cycle proteins are not completely understood, we have recently shown an inverse correlation between MnSOD activity and cyclin D1 protein levels in NIH3T3 mouse fibroblasts that were treated with the thiol antioxidant, N-acetyl-L-cysteine (NAC). NAC treatment increased MnSOD protein levels and activity while decreasing cyclin D1 protein levels. NAC-induced decrease in cyclin D1 was abrogated in NIH3T3 fibroblasts overexpressing a mutant form of cyclin D1 (T286A). Since phosphorylation at T286 targets cyclin D1 for degradation, we speculated NAC-induced increase in MnSOD activity could modulate redox-sensitive pathways affecting T286-phosphorylation and subsequent degradation of cyclin D1 (Menon et al., 2007). These results support the hypothesis that the cellular redox environment coordinates cell cycle regulatory protein expression during transitions between quiescent and proliferative growth.
While ROS are known to regulate differentiation and cell death pathways both of which are irreversible processes, our results show mitochondria-derived ROS regulates cellular proliferation and quiescence, which are reversible processes essential for cell and tissue renewals. Furthermore, it is also well known that many cytokines produce ROS/reactive nitrogen species during the inflammation process. Cells respond to these fluxes in ROS/reactive nitrogen species by enhancing antioxidant enzymes gene expression including the expression of MnSOD. While MnSOD activity neutralizes ROS during inflammation, our results suggest that it could also regulate transitions between quiescent and proliferative growth. At present, it is unclear how the same ROS could regulate such diverse biological processes that include proliferation, quiescence, differentiation, inflammation, and cell death. It is hypothesized that location of ROS generation and ‘ROS threshold levels’ are two significant determinants of these diverse cellular processes. Our results indicate that mitochondrially generated ROS levels could be a major mediator linking cellular metabolism to the proliferative and quiescent growth states (Fig. 7).
Our results also suggest that MnSOD activity and mitochondria-derived ROS levels are critical for the maintenance of quiescence to prevent aberrant proliferation and subsequent exhaustion of normal cell proliferative capacity. Our results resemble recent reports of ROS regulating the proliferative capacity of hematopoietic stem cells (Tothova et al., 2007). FoxO transcription factor-deficient hematopoietic stem cells (HSCs) showed reduced ability to repopulate. Because FoxO is known to regulate antioxidant enzyme (MnSOD and catalase) gene transcription (Kops et al., 2002a,b), it was suggested that ROS could mediate the proliferative capacity of quiescent HSCs. Surprisingly, while treatment of FoxO-deficient HSCs with ROS scavenger, NAC, restored quiescent HSCs’ repopulating ability, this treatment also resulted in increased antioxidant enzyme levels. These previously published results suggest that ROS could activate additional pathways independent of FoxO to regulate antioxidant enzyme gene expression. We reported previously that NAC-treated MEFs increased MnSOD protein levels and activity independent of its mRNA levels (Menon et al., 2007). Our results from the present study showed significant increase in MnSOD activity (Fig. 1) during transition from the proliferative to quiescent growth, while there were no changes in MnSOD mRNA (data not shown) and protein levels (Fig. 1B). These results suggest that post-translational rather than transcriptional mechanisms could regulate MnSOD activity during transitions between quiescent and proliferative growth.
In summary, our results support the hypothesis that MnSOD activity regulates a ‘ROS-switch’ favouring a superoxide-signal regulating the proliferative cycle and a hydrogen peroxide-signal supporting quiescent growth. Higher levels of MnSOD activity were associated with quiescent growth while lower levels support proliferative growth. MnSOD-activity-regulated transitions between quiescent and proliferative growth were associated with changes in cyclin D1 and cyclin B1 protein levels. A loss in the redox-control of this regulation could have myriad cellular effects including aberrant proliferation, defects in cell-renewal, cancer and cellular aging (Fig. 7).
MnSOD wild-type (MnSOD +/+), heterozygous (MnSOD +/–), and homozygous knockout (MnSOD –/–) MEFs primary cultures were generously provided by Dr T. T. Huang (Stanford University). Normal human skin fibroblast (AG01522C) was purchased from Coriell. Fibroblasts were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum (FBS) and antibiotics. Monolayer cultures were grown at 37 °C in a humidified incubator with 5% carbon dioxide and 4% oxygen. All experiments were performed within passage 6. Cell cultures with S-phase equal or less than 5% were considered as quiescent growth. Cell population doubling time (Td) was calculated from the exponential portion of the growth curve using the following equation: Td = 0.693t ln−1 (Nt/N0), where t is time in days, and Nt and N0 represent cell numbers at time t and initial time, respectively.
Replication deficient adenoviruses containing cytomegalovirus promoter driven cDNAs for mouse wild-type (AdMnSOD) and human mutant (AdmMnSOD) MnSOD were obtained from the University of Iowa DNA-Vector Core facility. The mutant form of MnSOD was generated by performing site-directed mutagenesis of histidine-26 to leucine. This substitution in amino acid resulted in a dominant negative function of the protein, which when overexpressed suppressed the activity of MnSOD (Zhang et al., 2006). Adenovirus infections were carried out using exponential and quiescent cultures following our previously published protocol (Sarsour et al., 2005). Monolayer cultures were infected with adenoviruses for 24 h in serum-free media followed by addition of regular media containing 10% FBS. A nonmodified vector (AdBgl II) was used as control for adenoviral infections. Transduction efficiency was measured by flow cytometry assay of AdGFP-infected cultures. Transgene expression was verified by immunoblotting and antioxidant enzyme activities measured by native-gel electrophoresis.
Immunoblotting and MnSOD native gel-activity assays
Equal amounts of total cellular proteins were separated by sodium dodecylsulfate–polyacrylamide gel electrophoresis and electro-transferred by semidry blotting onto a nitrocellulose membrane. Membranes were incubated with antibodies to MnSOD, cyclin D1 (PharMingen BD, San Jose, CA, USA), and cyclin B1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Immunoreactive bands were detected by an enhanced chemiluminescence kit and results were quantitated using AlphaImager 2000 software (Alpha Innotech, San Leandro, CA, USA). Actin protein levels were used for loading corrections. MnSOD activity was assayed by native gel-electrophoresis following our previously published protocol (Sarsour et al., 2005). Briefly, equal amounts of total cellular protein extracts were separated by 12.5% native gel-electrophoresis, and SOD activity determined by incubating the gel with nitroblue tetrazolum (2.43 mm) and riboflavin–TEMED (riboflavin 2.8 × 10−5 m and TEMED 28 mm). Achromatic bands were visualized and quantified using a computerized digital imaging system interfaced with AlphaImager 2000 software (Alpha Innotech).
Total cellular RNA was isolated and reverse transcribed using random-hexamer and reverse transcriptase. cDNAs were PCR amplified (ABI 7000 Real-Time PCR Sequence Detection System, Applied Biosystems, Foster City, CA, USA) using SYBR green and the following primer pairs: MnSOD forward: 5′-AGGCTCTGGCCAAGGGAGATGT-3′; reverse: 5′-CCACAGACACGGCTGTCAGCTT-3′; and 18S rRNA, forward: 5′-CCTTGGATGTGGTAGCCGTTT-3′; reverse: 5′-AACTTTCGATGGTAGTCGCCG-3′. MnSOD mRNA levels in each sample were normalized to 18S rRNA levels, and relative expression (2−ΔΔCT) calculated.
Electron paramagnetic resonance measurements of superoxide steady state levels
The EPR spin trapping technique with 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the spin trap was used to detect superoxide levels (Kalen et al., 2006). Monolayer cultures were washed with phosphate-buffered saline (PBS) and covered with 500 µL of chelated (iminodiacetic acid, sodium form, Sigma Chemicals, St. Louis, MO, USA) PBS containing 100 mm DMPO. Cells were then incubated at 37 °C for 15 min, and transferred to a flat cell for EPR measurement. The specificity of the superoxide signal was determined by pre-incubating the monolayer cells with CuZnSOD (1000 units mL−1) or Tiron (1 mm). EPR peak heights were calculated, and normalized to cell number.
Propidium iodide assay for DNA content
Ethanol-fixed cells were treated with RNase A and propidium iodide, and analyzed by flow cytometry for DNA content following our previously published protocol (Menon et al., 2003; Sarsour et al., 2005).
BrdU pulse-chase assay
Asynchronously growing exponential MEFs were pulse-labelled with 10-µm BrdU for 30 min and continued in BrdU-free culture media. Monolayer cultures were harvested by trypsin and fixed in ethanol. Isolated nuclei were incubated with anti-BrdU antibody (Immunocytometry Systems, San Jose, CA, USA) followed by incubation with Fluorescein Isothiocyanate (FITC)-conjugated goat antimouse IgG. Nuclei were treated with RNase A and counterstained with propidium iodide. FITC and propidium iodide fluorescence were analyzed on a FACScan flow cytometer (BD Biosciences) following our previously published protocols (Menon et al., 2003; Sarsour et al., 2005). Data from a minimum of 20 000 nuclei were acquired in list mode and processed using CellQuest Pro Software.
Measurement of ROS levels
Monolayer cultures were incubated with Hanks’ buffer salt solution (HBSS) containing 10 µm DHE (Invitrogen) for 45 min or 1 µg mL−1 DCFH (Invitrogen) for 15 min. Monolayers were trypsinized and resuspended in HBSS buffer containing 10% FBS. DHE and DCFH fluorescence were measured by flow cytometry. The flow cytometry measurements of DHE fluorescence was performed using 488-nm excitation laser and FL2 channel- 585 nm band pass emission filter. DCFH fluorescence was measured using 488-nm excitation laser and FL1 channel- 530 nm band pass emission filter. Mean fluorescence intensity was analyzed using Flowjo software. Auto-fluorescence of unlabeled cells was used for background fluorescence correction.
Mitochondrial superoxide measurements
Cell cultures in 60-mm dishes were washed three times with warm HBSS, and incubated with 3 µm MitoSOX® and 0.5 µm MitoTracker Green® (Invitrogen) for 15 min. Monolayer cultures were trypsinized and resuspended in HBSS buffer containing 10% FBS. Cell suspensions were washed once and filtered using nylon mesh. Flow cytometry measurements of MitoSOX and MitoTracker fluorescence were performed using 488-nm excitation laser and FL2 channel- 585/42 nm band pass emission filter for MitoSOX and FL1 channel- 530/30 nm band pass emission filter for MitoTracker. Mean fluorescence intensity was analyzed using Flowjo software. Auto-fluorescence of unlabeled cells was used for background fluorescence correction.
Measurement of extracellular hydrogen peroxide
Measurement of hydrogen peroxide released from quiescent and exponential MnSOD (+/+) MEFs was performed following previously published protocol (Panus et al., 1993). This method is based on the reaction of hydrogen peroxide with horseradish peroxidase (HRP) forming compound I, followed by a subsequent reaction where compound I with p-hydroxyphenyl acetic acid (pHPA) forms a stable fluorescent dimer, [pHPA]2. Monolayer cultures were washed with phenol red-free HBSS and incubated with 1.0 mL HBSS supplemented with glucose (6.5 mm), HEPES (1 mm), sodium bicarbonate (6 mm), pHPA (1.6 mm), and HRP (95 µg mL−1) at 37 °C for 1 h. The amount of hydrogen peroxide released into the buffer was measured spectrofluorometrically at 20-min intervals for 80 min using excitation and emission wavelengths of 323 and 400 nm, respectively. The fluorescent intensity of each sample was corrected for changes in pH. Hydrogen peroxide absorbance at 240 nm was used as standard.
Statistical analysis was done using the one and two-way analysis of variance with Tukey's honestly significant difference test. Results from at least n ≥ 3 with p < 0.05 were considered significant. All statistical analysis were done using SPSS computer software version 14.0 (SPSS Inc., Chicago, IL, USA).
We thank Dr Garry R. Buettner and the EPR core facility for assistance with EPR spectroscopy, Mr Burl E. Hess and the flow cytometry core facility for assisting with flow cytometry, and Mr Zhen Gao for assisting with the real-time RT-PCR assay. Funding from NIH CA 111365 and McCord Research Foundation supported this work.