Regulation of normal cell cycle progression by flavin ... - Nature

5 downloads 113 Views 301KB Size Report
Jul 16, 2007 - of Medicine and Dentistry of New Jersey, New Jersey Medical School,. 185 South Orange Avenue, MSB – F451, Newark, NJ 07101-1709,.
Oncogene (2008) 27, 20–31

& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Regulation of normal cell cycle progression by flavin-containing oxidases P Venkatachalam1, SM de Toledo1, BN Pandey1, LA Tephly2, AB Carter2, JB Little3, DR Spitz4 and EI Azzam1 1

Department of Radiology, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, NJ, USA; Department of Medicine, University of Iowa, Roy J and Lucille A Carver College of Medicine, Iowa City, IA, USA; 3Center for Radiation Sciences and Environmental Health, Harvard School of Public Health, Boston, MA, USA and 4Free Radical and Radiation Biology Program, Department of Radiation Oncology, Holden Comprehensive Cancer Center, University of Iowa, Iowa City, IA, USA 2

Mechanisms underlying the role of reactive oxygen species (ROS) generated by flavin-containing oxidases in regulating cell cycle progression were examined in human and rodent fibroblasts. Incubation of confluent cell cultures with nontoxic/nonclastogenic concentrations of the flavoprotein inhibitor, diphenyleneiodonium (DPI), reduced nicotinamide adenine dinucleotide phosphate (NAD(P)H) oxidase activity and basal ROS levels, but increased proteolysis of cyclin D1, p21Waf1 and phospho-p38MAPK. When these cells were allowed to proliferate by subculture in DPI-free medium, an extensive G1 delay was observed with concomitant activation of p53/p21Waf1 signaling and reduced phosphorylation of mitogen-activated kinases. Compensation for decreased oxidant generation by simultaneous exposure to DPI and nontoxic doses of the ROS generators, c-radiation or t-butyl-hydroperoxide, attenuated the G1 delay. Whereas the DPI-induced G1 checkpoint was completely dependent on PHOX91, ATM and WAF1, it was only partially dependent on P53. Interestingly, G1 to S progression was not affected when another flavin-containing enzyme, nitric oxide synthase, was inhibited nor was it associated with changes in mitochondrial membrane potential. Proliferating cells treated with DPI also experienced a significant but attenuated delay in G2. We propose that ATM performs a critical function in mediating normal cellular proliferation that is regulated by nonphagocytic NAD(P)H oxidase enzymes activity, which may serve as a novel target for arresting cancer cells in G1. Oncogene (2008) 27, 20–31; doi:10.1038/sj.onc.1210634; published online 16 July 2007 Keywords: flavin-containing oxidases; NAD(P)H oxidase/ nitric oxide synthase; reactive oxygen species; cellular proliferation/G1 checkpoint/G2 checkpoint; ATM/p53/ p21Waf1/p38MAPK/cyclin D1

Correspondence: Dr E Azzam, Department of Radiology, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, 185 South Orange Avenue, MSB – F451, Newark, NJ 07101-1709, USA. E-mail: [email protected] Received 6 February 2007; revised 30 April 2007; accepted 25 May 2007; published online 16 July 2007

Introduction Cellular exposure to high levels of reactive oxygen species (ROS) produced by endogenous enzymatic reactions or induced by external agents damages DNA, proteins and lipids (Halliwell, 1996), and contributes to numerous human disorders (Droge, 2002). In contrast, low levels of ROS participate in signaling pathways that control essential cellular functions including proliferation (Burdon, 1995). ROS regulate expression of specific genes, modulate ion channel activities and mimic or affect intermediates (for example, second messengers) in signal transduction (Schulze-Osthoff et al., 1997). Hence, homeostatic cellular functions require tight control of the cellular redox environment (Spitz et al., 2004). Disruption of the balance between oxidant production and antioxidant capacity alters this environment, resulting in detrimental consequences to the cell (Schafer and Buettner, 2001; Spitz et al., 2004). Mitochondrial oxidative metabolism, nicotinamide adenine dinucleotide phosphate (NAD(P)H) oxidases and lipoxygenases are the major cellular sources of ROS (Droge, 2002). Most of the oxidants they generate and their byproducts are metabolized by antioxidants. However, ROS that escape antioxidant defense are believed to contribute to homeostatic regulation of redox signaling (Droge, 2002). Superoxide anions produced by the multicomponent NAD(P)H oxidase (a flavin-containing oxidase) were initially detected in phagocytes, and were found to be essential to their antimicrobial activity (Ushio-Fukai et al., 1996). Recently, similar ROS-generating oxidases have been shown to exist also in nonphagocytic cells, including fibroblasts (Babior, 1999). NAD(P)H oxidase affects signal transduction by growth factor receptors (Ammendola et al., 2002) and promotes proliferation in a variety of cell types (Irani and Goldschmidt-Clermont, 1998). It is thought that mitogenic signaling elicited by cytokines or growth factors induces intracellular ROS production via activation of the PI3K pathway resulting in stimulation of Rac1, which upregulates NAD(P)H oxidase activity (Bae et al., 1997). However, the overall mechanism(s) by which NAD(P)H oxidase (and/or

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

21

a

control 1

+0.5 µM (7 days)

0.5

Treatment

b 80 Labeled cells (%)

Exposure of MEFs to DPI results in reversible decrease in clonogenic survival While examining the role of flavin oxidases in the proliferative response of irradiated C3H-10T1/2 mouse embryo fibroblasts (MEFs), an apparent cell killing effect of the flavin-oxidase inhibitor, diphenyleneiodonium (DPI), was observed following prolonged exposure (Figure 1a). Cells treated with 0.5 mM DPI for 7 days were completely inhibited from proliferating. Removal of DPI-containing medium and feeding with fresh medium resulted in about 50% of the cells recovering to form colonies consisting of over 200 cells/colony 10 days later. These data suggest that inhibition of flavincontaining oxidases by chronic exposure to low concentrations of DPI induces reversible cell cycle arrests. They support the concept that ROS have a major role in regulating cellular proliferation (Murrell et al., 1990; Menon et al., 2003).

Surviving fraction

Results

- 0.5 µM (10 days)

on progression of normal human fibroblasts from G1 to S phase. AG1522 cells synchronized in G0/G1 by density inhibition of growth were incubated with 0.05 or 0.15 mM DPI for 20 h. After treatment, control and DPI-treated cells were subcultured in DPI-free medium containing [3H]-thymidine (1 mCi/ml) and allowed to progress through the cell cycle. Movement into S phase was monitored autoradiographically by measuring the cumulative labeling indices (CLI) at multiple time points up to 50 h after subculture. Compared with shamtreated cells, significant delays in progression into

+ 0.5 µM (7 days)

other flavin-containing oxidases or lipoxygenases) mediates cellular proliferation under normal endogenous oxidative conditions is not clear. Using genetic and biochemical approaches, here we describe molecular and biochemical events involved in the effects of flavoprotein enzyme activity on cellular progression from G1 to S and G2 to M phase. We investigate the involvement of: (1) the flavoproteins, NAD(P)H oxidase and nitric oxide synthase, (2) the cell cycle regulators, ATM, P53, Waf1, and (3) mitochondrial membrane potential (Dc) in signaling these effects. Regulation of key cell cycle controllers is examined, and direct evidence for involvement of ROS produced by flavin oxidases in cell cycle transitions is provided.

0 µM

40

0.05 µM 0.15 µM 0

DPI treatment inhibits NADP(H) oxidase activity and induces significant delays in cellular progression from G1 to S To investigate the generality of our observations in MEFs and to determine the control level at which flavincontaining oxidases regulate cell proliferation, we examined the effect of ROS generated by these enzymes

10

30

50

70

Time after subculture (h) MWM (kDa)

c

62 47

36

d 60000 0 µM

RLU

Figure 1 Flavin-containing oxidases and cell proliferation: effects on the G1 checkpoint. (a) Clonogenic survival of C3H 10T½ MEFs exposed to DPI. Confluent, density-inhibited cultures were exposed to 0.5 mM DPI for 2 h, following which they were subcultured in DPI-containing medium (0.5 mM) and incubated for 7 days. Alternatively, subcultured cells were incubated for 7 days in medium containing 0.5 mM DPI followed by incubation in DPI-free medium for an additional 10 days. (b) Confluent, densityinhibited AG1522 human fibroblast cultures were incubated for 20 h with 0, 0.05 or 0.15 mM DPI and subsequently subcultured in DPI-free medium. Cumulative labeling indexes were measured as a function of time after subculture: —, 0 mM; ’- - - -, 0.05 mM; ,yy, 0.15 mM; (c) western blot analysis of gp91phox in AG1522 density-inhibited fibroblasts; (d) kinetics of NAD(P)H oxidase activity, determined in relative luminescence units (RLU), in membrane fractions isolated from confluent AG1522 fibroblasts incubated for 20 h in the absence or presence of 0.05 or 0.15 mM DPI.

0.05 µM 0.15 µM

20000

0

4

8

Time (min) Oncogene

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

Oncogene

0.05

0.1 Gy (0.002 Gy/h)

1.35

0.1Gy (1 Gy/min)

4 Gy (1 Gy/min)

Figure 2 Effect of diphenyleneiodonium (DPI) on G2 to M transition: relative movement into G2/M phase of control and DPItreated (0.15 mM) AG1522 fibroblasts determined at different times after the initiation of bromodeoxyuridine (BrdU) labeling. The G2 delay is the time elapsed between the peaks of the relative movement curves: —, 0 mM DPI; ,yy, 0.15 mM DPI.

0.15 µM DPI

The DPI-induced G1 checkpoint occurs in the absence of DNA damage DNA damage is a strong inducer of cell cycle checkpoints (Little, 1968). To investigate whether the DPI-induced G1 delay (Figure 1b) is a consequence of induced DNA damage, formation of micronuclei was examined. Micronucleus frequency in AG1522 cells incubated with 0.15 mM DPI for 24 h was similar to that observed in sham-manipulated control cells (Figure 3). These data strongly suggest that the DPI-induced G1 checkpoint results from modulation of redox-sensitive events that control the G1/S traverse and is not due to DNA damage that results in micronucleus formation. This however does not exclude the occurrence of other types of DNA changes. To ascertain the sensitivity of the micronucleus assay to detect DNA damage, micronucleus formation was concurrently examined in cell cultures exposed to g-ray doses as low as 0.1 Gy (delivered at 1 Gy/min or 0.002 Gy/h) or to 4 Gy (1 Gy/min), a cytotoxic dose that induces a G1 delay similar in magnitude (11 h) (Azzam et al., 2000) to that observed in AG1522 cells treated with 0.15 mM DPI. Micronucleus frequency was significantly (Po0.01) increased in cells exposed to 0.1 Gy delivered acutely (Figure 3). As expected, protracting delivery of this small dose over 48 h resulted in sparing of induced DNA damage (de Toledo et al., 2006). In contrast, in cells exposed to 4 Gy, the micronucleus frequency was 50-fold greater than background (Figure 3). These results show that micronucleus

40

Control

G2 to M phase transition in DPI-treated normal human fibroblasts To determine whether flavin-containing oxidases also modulate transition through G2, the kinetics of G2 to M transition in sham-manipulated and DPI-treated normal human fibroblasts was examined. Proliferating AG1522 cells were labeled with bromodeoxyuridine (BrdU), treated with 0.15 mM DPI and movement of S-phase cells in the growth cycle was monitored. In contrast to the extensive DPI-induced G1 delay (11.75 h, Figure 1b), the DPI-induced delay in G2 to M transition was small (1 h75 min; Figure 2).

0 µM 0.15 µM

80

10 20 Time after BrdU addition (h)

Micronuclei frequency / per cell

S phase occurred in DPI-treated cells (Figure 1b). At 50% of the maximum CLI, transient G1 delays of 6.75 and 11.75 h were observed in cells from cultures exposed to 0.05 or 0.15 mM DPI, respectively. The flavin-containing enzyme NAD(P)H oxidase is a likely target of DPI. The western analyses in Figure 1c confirmed the presence of gp91phox, the catalytic subunit of NAD(P)H oxidase, in AG1522 fibroblasts. As expected, a protein of about 60 kDa reacted with an antibody specific to gp91phox. Importantly, enzyme activity kinetics data (Figure 1d) indicated that NAD(P)H oxidase is active in these cells, and incubation with 0.05 or 0.15 mM DPI for 20 h inhibits this activity. These results indicate that DPI-sensitive flavin oxidases (presumably membrane-bound NAD(P)H oxidase) are major mediators in cellular progression through G1.

Relative movement of BrdU labeled cells

22

0 Treatment Figure 3 Diphenyleneiodonium (DPI)-induced G1 checkpoint occurs in the absence of DNA damage. Micronucleus formation in AG1522 fibroblasts exposed to DPI or to g-radiation. Confluent density-inhibited cultures were either incubated for 20 h in growth medium containing 0 or 0.15 mM DPI or g-irradiated. Immediately after treatments, they were subcultured in DPI-free growth medium and DNA damage was measured by the micronucleus formation assay.

formation assay is highly sensitive to detect DNA damage. In further support of these data, clonogenic survival and in situ immunoblotting with anti-g-H2AX and p53-BP1 (not shown) showed that DPI treatment was neither cytotoxic nor clastogenic. Incubation of AG1522 cells with 0.05 or 0.15 mM DPI for 20 h resulted in 99.9% of the cells being trypan-blue negative indicating that plasma membrane integrity was not altered. Therefore, multiple endpoints show that inhibition of flavin oxidases by DPI induces a G1 delay that is not dependent on clastogenic/cytotoxic events.

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

23

The DPI-induced G1 checkpoint correlates with upregulation of p53 signaling An increase in p53 activity has been associated with induction of the G1 checkpoint by DNA-damaging agents (Kastan et al., 1991), and WAF1 has been shown to be an important mediator of p53 regulation of this checkpoint (Deng et al., 1995). To investigate whether the G1 delay induced in DPI-treated cells is also associated with activation of p53/p21Waf1 signaling, expression of members of this pathway was examined in confluent AG1522 fibroblasts incubated with DPI (0–0.2 mM, 20 h) that were subsequently subcultured in DPI-free medium. Western analyses indicated that 24 h after subculture, p53 and p21Waf1 levels were elevated in cells previously incubated with DPI (3.5- and 2.5-fold, respectively, at 0.15 mM DPI) (Figure 4). The levels of p34cdc2 whose expression under stress conditions is p53 dependent (Azzam et al., 1997; de Toledo et al., 1998) were downregulated (threefold at 0.15 mM DPI). Thus, similar to the DNA damage-induced G1 checkpoint, the DPI-induced G1 delay is also associated with 0.05 0.15 0.2 µM

0 p53 p21Waf1 p34cdc2 P-ERK1/2 P-p38 cyclin D1 ku70

Figure 4 The DPI-induced G1 checkpoint upregulates p53 signaling. Western blot analyses of cell cycle-regulated and stressresponsive proteins in DPI-treated AG1522 fibroblasts. Confluent, density-inhibited cells were incubated in medium containing DPI (0–0.2 mM) for 20 h. The cells were subsequently subcultured to lower density and harvested for analyses 24 h later. Immunoblotting with anti-ku70 antibody served as a loading control.

Inhibition of flavin oxidases does not induce a G1 delay in ataxia telangiectasia cells The role of ATM in mediating the DNA damageinduced G1 checkpoint is well established (Kastan et al., 1992). To investigate whether the DPI-induced G1 checkpoint (Figure 1b) is ATM-dependent, confluent AG4405 human cells homozygous for mutant ATM were exposed to 0.05 or 0.15 mM DPI for 20 h and were subsequently stimulated to proliferate by subculture in DPI-free medium. Progression into S phase was measured by autoradiography. Cells treated with 0.05 mM DPI progressed from G1 to S at the same rate as control cells (Figure 5a). Strikingly, cells treated with 0.15 mM DPI progressed into S at a faster rate. At 50% of the maximum CLI, cells treated with 0.15 mM DPI entered S phase 4.4 h earlier than control or 0.05 mM-treated cells. In contrast, parallel experiments with DPI-treated AG1522 cells consistently revealed G1 delays similar to those in Figure 1b. Enzyme activity measurements (Figure 5b) revealed significant NAD(P)H oxidase activity in AG4405 cells (two- to threefold greater than in AG1522 normal cells), and incubation with DPI completely inhibited this activity. These data strongly suggest that the DPI-induced G1 checkpoint is ATMdependent and the atm protein regulates flavin-containing oxidase-dependent signaling events that determine duration of G1 to S transition. They show that ATM regulates G1 delays that are induced in absence of DNA breaks.

b 80 AG4405 (0 µM DPI)

40000 40

AG1522 (0 µM DPI)

RLU

Labeled cells (%)

a

upregulation of p53 signaling. In addition, inhibition of flavin-containing oxidases led to downregulation of active extracellular-regulated protein kinases 1/2 (ERK1/2) and p38 mitogen-responsive kinases in subcultured cells (Figure 4). While decrease in phosphorylated (P)-ERK1/2 was detectable at DPI concentrations >0.05 mM (1.8-fold at 0.15 mM), that in P-p38MAPK occurred at concentrations as low as 0.05 mM (10-fold). These results implicate downregulation of the mitogenactivated protein kinase (MAPK) pathway in the DPIinduced G1 checkpoint. Expression of cyclin D1 was slightly increased (1.5-fold; Figure 4).

0 µM 0.05 µ M 0.15 µ M

20000 AG4405 (0.15 µM DPI) AG1522 (0.15 µM DPI)

10

30

50

Time after subculture (h)

70

0

4

8

Time (min)

Figure 5 The role of ATM in regulation of the G1 checkpoint by flavin-containing oxidases. (a) Cumulative labeling index as a function of time after release from confluent, density-inhibited growth of AG4405 ataxia telangiectasia homozygous fibroblasts that were incubated in DPI for 20 h: —, 0 mM; ’- - - -, 0.05 mM; ,yy, 0.15 mM. (b) Kinetics of NAD(P)H oxidase activity, determined in relative luminescence units (RLU), in membrane fractions isolated from confluent AG4405 or AG1522 fibroblasts incubated for 20 h in the absence or presence of 0.15 mM DPI. Oncogene

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

24

(not shown). These results indicate that the DPI-induced G1 checkpoint is partially dependent on p53 and entirely dependent on p21WAF1. Together, the data in Figures 5 and 6 suggest that ATM mediates redox-sensitive G1 checkpoint through p21Waf1 as a result of modulation of p53 and other signaling molecule(s).

Mediation of G1 to S transition by flavin-containing oxidases is partially dependent on P53 and completely dependent on WAF1 ATM mediates the DNA damage-induced G1 checkpoint, predominantly, through downstream p53 signaling. To examine the role of p53 and its effector p21Waf1 (a universal inhibitor of cyclin-dependent kinases (cdk)) in flavin-containing oxidase(s)-mediated events that regulate G1 to S transition, cellular progression into S phase was examined in sham-manipulated and DPI-treated wild type (wt), p53/ and p21Waf1/ MEFs. Similar to normal human fibroblasts (Figure 1b), a G1 checkpoint was induced in DPI-treated wt MEFs (Figure 6a). Compared to sham-treated cells, transient G1 delays of 8 and 10 h were observed in wt MEFs incubated with 0.05 or 0.15 mM DPI, respectively. While the magnitude of the induced G1 checkpoint in normal human cells is DPI concentration-dependent (Figure 1b), the G1 delay in wt MEFs incubated in 0.15 mM DPI is only slightly greater than that induced following incubation with 0.05 mM DPI (Figure 6a). Compared to wt cells, incubation of p53/ MEFs with 0.05 or 0.15 mM DPI resulted in attenuated G1 delays (Figure 6b). As in wt cells, the induced delays were similar at both DPI concentrations (4.2 and 4 h delays in cells incubated with 0.05 and 0.15 mM, respectively) (Figure 6b). In contrast to wt (Figure 6a) and p53/ cells (Figure 6b), no G1 delay occurred in p21WAF1/ MEFs incubated with DPI (Figure 6c). Similar to human ataxia telangiectasia (AT) cells, incubation of p21WAF1/ MEFs with DPI stimulated rather than inhibited progression through G1. At 50% of the maximum CLI, cells incubated with 0.05 mM DPI entered S phase 2.5 h earlier than respective sham-manipulated controls. Interestingly, similar to AG4405 cells, NAD(P)H oxidase activity in p21WAF1/ MEFs was on average 1.6-fold greater than in wt cells

a

The DPI-induced G1 checkpoint is mediated through NAD(P)H oxidase and not nitric oxide synthase Depending on cellular context and drug concentration, DPI inhibits several flavoprotein oxidoreductases, including NAD(P)H oxidase (Cross and Jones, 1986) and nitric oxide synthase (Stuehr et al., 1991). To examine the role of NAD(P)H oxidase in DPI modulation of G1 to S transition, progression into S phase was investigated in DPI-treated wt and isogenic MEFs deficient in gp91phox, the plasma membrane catalytic subunit of NAD(P)H oxidase in fibroblasts (Li et al., 2001). In contrast to wt cells (Figure 6a), isogenic gp91phox/ MEFs were unaffected by DPI treatment: progression of cells pretreated with 0.05 or 0.15 mM DPI into S was similar to that of sham-treated cells (Figure 6d). These data strongly support involvement of NAD(P)H oxidase in G1 to S transition. Compared to wt MEFs, the faster rate of progression of gp91Phox/ MEFs into S phase may be due to cell passage number or to effects that result from knockout of PHOX91. To investigate the role of nitric oxide synthase in regulation of DPI-induced G1 checkpoint, G1 to S-phase progression was examined in AG1522 cells pretreated with nontoxic concentrations of NG-nitro-L-argininemethyl ester (L-NAME), a nonselective inhibitor of nitric oxide synthases, including the endothelial (NOS3) and inducible (NOS2) isoforms in skin fibroblasts (CalsGrierson and Ormerod, 2004). Incubation of confluent cultures for 4 or 20 h with 25 or 100 mM L-NAME

b

WT

40 Labeled cells (%)

p53 -/-

80

80

0 µM 0.05 µM

40

0 µM 0.05 µM

0.15 µM

0.15 µM 10

c

p21Waf1

30

50

70

10

d

-/-

80

gp91Phox

30

50

70

-/-

80

0 µM 0.05 µM

40

0 µM 0.05 µM

40

0.15 µM

0.15 µM 10

30

50

70

10

30

50

70

Time after subculture (h)

Figure 6 The role of p53, p21Waf1 and gp91phox in regulation of G1 to S transition by flavin-containing oxidases. Confluent/quiescent MEFs were incubated for 20 h with 0, 0.05 or 0.15 mM DPI and subsequently subcultured in DPI-free medium. Cumulative labeling indexes were measured as a function of time after subculture: J—, 0 mM; ’- - - -, 0.05 mM; ,yy, 0.15 mM. (a) wild-type (wt) control; (b) p53/; (c) p21Waf1/; (d) gp91phox/. Oncogene

∆ψ (Arbitrary units)

80

0 µM

40

25 µM

25

0.15 µM DPI

3

Control

Labeled cells (%)

a

0.05 µM DPI

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

2 1 0 Treatment

100 uM, 20 h

25 uM, 20 h

80

100 uM, 4 h

25 uM, 4 h

Cells in G1 (%)

b

50

0.15 uM, 20 h

10 30 Time after subculture (h)

in increase in the fraction of cells delayed in G1 (Figure 7b).

40

C

c

L-NAME C

Figure 8 Mitochondrial membrane potential (Dc) in confluent, density-inhibited AG1522 fibroblasts incubated for 20 h in medium containing diphenyleneiodonium (DPI) (0–0.15 mM). Cells were treated with JC-1 probe, harvested and Dc was determined by flow cytometry. Arbitrary units represent the ratio of mean red and green fluorescence (a shift from red to green fluorescence is indicative of mitochondrial membrane depolarization).

DPI

L-NAME

pRb

p34cdc2 Non-specific Figure 7 Diphenyleneiodonium (DPI) mediates the G1 checkpoint through an effect on NADP(H) oxidase and not through effects on nitric oxide synthase or on mitochondrial function. (a) Confluent, density-inhibited AG1522 fibroblast cultures were incubated for 4 h with 0 or 25 mM NG-nitro-L-arginine-methyl ester (L-NAME) and subsequently subcultured in L-NAME-free medium. Cumulative labeling indexes were measured as a function of time after subculture: —, 0 mM; }yy, 25 mM. (b) Determination by flow cytometry of the fraction of cells in G1 at 24 h after subculture in normal growth medium of control, L-NAME or DPI-treated density-inhibited AG1522 fibroblasts. (c) Western blot analyses of pRb and p34cdc2 in AG1522 fibroblasts harvested at 24 h following subculture of confluent cultures incubated with 25 mM LNAME for 4 h.

(Figures 7a and b) followed by subculture in control medium had no effect on G1 to S progression. Interestingly, a 4-h incubation in either concentration slightly decreased the fraction of cells in G1. This growth-stimulatory effect was supported by upregulation of phosphorylated pRb and p34cdc2 (Figure 7c). These data suggest that nitric oxide-generating flavoproteins do not participate in the DPI-induced G1 checkpoint. Note that concurrent flow cytometry analyses of DPI-incubated cells resulted, as expected,

The DPI-induced G1 checkpoint does not alter mitochondrial membrane potential In addition to effects on flavoprotein oxidases (Figures 1d and 5b), DPI treatment may also result in nonspecific effects on mitochondrial function, which may contribute to the induced G1 checkpoint (Figure 1b). An important parameter of mitochondrial functionality is the Dc (Gordon et al., 2000). We therefore examined whether DPI treatment of quiescent AG1522 cells correlates with changes in Dc. Analyses of JC-1-labeled densityinhibited AG1522 cells treated with 0, 0.05 or 0.15 mM DPI indicated no significant change in Dc (Figure 8). These results rule out the possibility that the G1-induced checkpoint in DPI-treated cells results from an effect on Dc. Other effects involving changes in oxygen consumption or ATP generation may have occurred. ROS generated by flavin-containing oxidases mediate cellular proliferation To provide evidence for participation of ROS generated by flavin-containing oxidases in cellular progression from G1 to S, intracellular ROS concentration was examined in AG1522 cells reacted with 20 ,70 -dichlorodihydrofluorescence-diacetate (DCFH-DA) dye. Fluorescence intensity measurements (Figure 9a) performed in parallel to cell cycle progression studies indicated that a 20-h incubation with 0.15 mM DPI decreased intracellular level of DCFH-DA-sensitive ROS (2.5-fold decrease, Po0.001). These data provide evidence that the G1 delay induced in DPI-treated cells (Figure 1b) is associated with decreased generation of ROS. The involvement of ROS in DPI-induced cell cycle delays was further supported when density-inhibited AG1522 cells incubated with DPI (0.05 mM, 20 h) were simultaneously exposed to the ROS generators, g-rays (1 Gy, 0.05 Gy/h) or tertiary-butyl-hydroperoxide (t-BOOH; 0.5 mM). When these cells were stimulated to proliferate by subculture, shorter G1 delays occurred in cells Oncogene

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

1000

b

Labeled cells (%)

Control

RFU

9000

DPI (0.15 µM, 20 h)

a

t-BOOH (100 µM, 15 min)

26

c

80

80

Control 0.05 µM DPI 1 Gy 1 Gy+ 0.05 µM DPI

40

Control 0.05 µM DPI 0.5 µM t-BOOH 0.5 µM t-BOOH + 0.05 µM DPI

40

0 10 Treatment

30

50

10

30

50

Time after subculture (h)

Figure 9 ROS generated by flavin-containing oxidases mediate cellular proliferation. (a) Measurement by flow cytometry of DCFHDA-sensitive ROS in AG1522 fibroblasts incubated in control medium or medium containing 0.15 mM diphenyleneiodonium (DPI) (20 h). Cells incubated in 100 mM t-butyl-hydroperoxide (t-BOOH) for 15 min served as positive control. Difference in relative fluorescence units (RFU) between control and DPI-treated samples was highly significant (Po0.0001, Student’s t-test). (b) Cumulative labeling index as a function of time after release from confluent, density-inhibited growth of AG1522 fibroblasts that were incubated with DPI, g-irradiated or simultaneously incubated with DPI and g-irradiated: —, 0 mM DPI; Jy, 0.05 mM DPI, 20 h; ’- - - -, 1 Gy, 5 cGy/h; &—, 1 Gy þ 0.05 mM DPI. (c) Cumulative labeling index as a function of time after release from confluent, density-inhibited growth of AG1522 fibroblasts that were incubated with DPI, t-BOOH or simultaneously incubated with DPI and t-BOOH: —, 0 mM DPI; Jyy, 0.05 mM DPI, 20 h; ’y, 0.5 mM t-BOOH, 20 h; &—, 0.5 mM t-BOOH þ 0.05 mM DPI.

preexposed to both DPI and either of the oxidizing agents than to DPI alone (Figures 9b and c). At 50% of the maximum CLI, 8.2, 2.4 and 1 h delays were observed in cells preexposed to 0.05 mM DPI, 1 Gy from g-rays or 0.5 mM t-BOOH, respectively. The G1 delay was attenuated from 8.2 to 5 h in cells preexposed simultaneously to DPI and g-rays (Figure 9b), and to 5.4 h in cells preexposed to DPI and t-BOOH (Figure 9c). None of the various treatments affected clonogenic survival (not shown). These data suggest that ROS generated by g-rays or t-BOOH substitute for effects of ROS generated by flavin-containing oxidases (for example, membrane-bound NAD(P)H oxidase), and support regulation of G1 to S progression by flavin-containing oxidase signaling. They are consistent with a role for low levels of hydrogen peroxide and its precursor the superoxide anion as signaling intermediates in cellular proliferation (Burdon, 1995). Inhibition of flavin-containing oxidases modulates expression of cell cycle-regulatory proteins in confluent normal fibroblasts The data in Figure 4 describe expression of stressresponsive proteins in proliferating cells that were previously incubated with DPI. To gain insight into the mechanism by which DPI-inhibition of flavoproteins regulates G1 to S transition, we also examined changes in expression/activity of cell cycle-regulatory proteins that occur during incubation with DPI. Densityinhibited AG1522 fibroblasts were incubated with DPI (0.5–0.2 mM, 20 h) and harvested for analyses at the end of the incubation period. Compared to sham-treated cells, no change or slight decrease (10%) in expression of p53 and p27Kip1 occurred in DPI-treated cells (Figure 10a). In contrast, p21Waf1 and cyclin D1, two main Oncogene

regulators of G1 progression, were significantly decreased (five- and sevenfold, respectively) (Figure 10a). Expression of P-p38MAPK and P-ERK1/2, members of MAPK pathways, was also decreased (threefold; Figure 10a); no change occurred in expression of total p38MAPK or ERK1/2. Interestingly, incubation with even 0.2 mM DPI did not alter expression of cyclin A, which functions primarily in S-phase progression and mitosis (Figure 10a). Collectively, these results show that ROS generated by flavin-containing oxidases regulate, in G0/G1 phase, the expression of G1-regulatory proteins, most notably cyclin D1. Protein degradation mediates downregulation of cell cycle regulatory proteins in DPI-treated cells The apparent decrease in protein expression during DPI-treatment of AG1522 cells (Figure 10a) could be due to transcription/post-transcription and/or translation/post-translation regulation. It may result from reduced expression, or post-translational modification that renders proteins unrecognizable by their respective antibodies. To investigate the role of proteolysis, confluent AG1522 cells were incubated with 0.05 or 0.15 mM DPI for 20 h in absence or presence of the proteasome/calpain inhibitor MG-132. The western analyses in Figure 10b confirm those in Figure 10a and show significant decreases in expression of p21Waf1, cyclin D1 and P-p38MAPK in DPI-treated cells. Simultaneous incubation with DPI and MG-132 restored expression of these proteins to levels similar or approaching those observed in sham-manipulated controls. Basal expression of p53 is regulated by the proteasome and calpain pathways (Maki et al., 1996). The appearance of similar intensity bands of ubiquitinated

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

27

a

0

0.05

0.15

0.2 µM

b

Control

p53

0

0.05

MG132 0.15

0

0.05

0.15 µM

p27Kip1 p21Waf1 p53

cyclin D1

p21Waf1 ERK 1/2 P-ERK 1/2

P-p38 cyclin D1

p38 P-p38

ku70

cyclin A ku70 Figure 10 Inhibition of flavin-containing oxidases targets cell cycle-regulatory proteins for degradation in confluent normal fibroblasts. (a) Western blot analyses of cell cycle regulated and stress-responsive proteins in confluent, density-inhibited AG1522 fibroblasts incubated for 20 h in medium containing DPI (0–0.2 mM). (b) Western blot analyses of cell cycle regulatory proteins in AG1522 fibroblasts incubated with 0.05 or 0.15 mM DPI for 20 h in the absence or presence of the proteasome/calpain inhibitor MG-132. Immunoblotting with anti-ku70 antibody served as a loading control.

forms of p53 in control and DPI-treated cells incubated with MG-132 (Figure 10b) indicates that the drug was effective in inhibiting proteolysis. These data show that the cellular redox environment significantly affects protein stability.

Discussion The molecular basis of cellular proliferation control by endogenously produced oxidants, the precise source(s) of the oxidants, and sensitivity of various phases of the cell cycle to their effects has not been clearly defined. This study shows that ROS generation by flavincontaining oxidases regulates normal cellular progression from G1 to S and to lesser extent from G2 to M. Extensive transient delays in G1 (11 and 8 h in human and rodent cells, respectively) occurred when densityinhibited fibroblasts incubated with DPI at concentrations as low as 0.05 mM for 20 h were allowed to proliferate by subculture in DPI-free medium (Figures 1b, 6a, 9b and c). Smaller delays (1 h) were observed in G2 (Figure 2) when proliferating cells were treated with DPI. The induced delays were neither a result of toxicity (not shown) nor clastogenicity (Figure 3). Delays in G1, of similar magnitude, were observed in AG1522 cells and wt MEFs exposed to doses of ionizing radiation that generate extensive DNA damage and result in toxicity (Azzam et al., 2000). Hence, normal cell cycle progression is highly sensitive to regulation by flavincontaining oxidases, and DNA damage is not a necessary prerequisite for checkpoint control. We confirm that DPI inhibits NAD(P)H oxidase activity (Figures 1d and 5b) and results in significant reduction in generation of DCFH-DA-reactive prooxidants (Figure 9a). Covalent binding of DPI to a 45 kDa membrane flavoprotein and to the low-potential

cytochrome b558 of NAD(P)H oxidase complex is thought to be responsible for this effect (Doussiere and Vignais, 1992). We provide strong support for involvement of ROS-modulated signaling events in DPIinduced cell cycle delays. Significant attenuation of the DPI-induced G1 checkpoint occurred when normal fibroblasts were simultaneously exposed to DPI and oxidants produced by mild doses of g-radiation or t-BOOH (Figures 9b and c). It is noteworthy that oxidants (H2O2 and its precursor O2 ) generated by the latter agents are also a product of NAD(P)H oxidases. At low concentrations, these oxidants are required participants of normal cellular proliferation (Burdon, 1995). The fact that both oxidizing agents did not fully eliminate the DPI-induced G1 delay is likely due to checkpoint control caused by their DNA-damaging effects (Little, 2000). While DPI may induce several signaling events that affect G1 to S progression, our data (Figures 9b and c) show that it primarily acts through inhibition of ROS generation by NAD(P)H oxidase (Figure 6d). Changes in Dc (Figure 8) or DPI-induced effects that operate through nitric oxide synthase are apparently not involved (Figures 7a and b). Inhibition of nitric oxide synthases with L-NAME stimulated rather than slowed G1 to S-phase progression (Figure 7b), probably as a result of interference in effects that normally occur from interactions between nitric oxide and ROSdriven proliferation control (Matsumoto et al., 2007). Immunoblotting analyses revealed that DPI effect is associated with downregulation of P-p38MAPK, P-ERK1/2, cyclin D1 and p21Waf1 in quiescent fibroblasts (Figure 10a). While proteolysis mediated by the proteasome/calpain pathways contributes to the downregulation (Figure 10b), other mechanisms (for example, transcriptional/post-transcriptional) cannot be excluded. These changes likely contribute to the DPI-induced delays in cell cycle transitions. Oncogene

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

28

Cell cycle progression is mediated by sequential activation of cyclins and cdk, with the increase in cyclin D expression being among the initial steps in the proliferative process. In association with their binding partners, cdk 4 and 6, D-type cyclins promote G0–G1 to S-phase transition by activating phosphorylation of pRb, thereby helping to cancel its growth-repressive function (Matsushime et al., 1992). Ample evidence supports the role of cyclin D in mitogenic signaling: for example mitogenic pathways (for example, RAS and ERK1/2) positively regulate expression of D-type cyclins (Lavoie et al., 1996). Hence, decreased expression of cyclin D1 in DPI-treated cells (Figure 10a) is consistent with decreased activity of ERK1/2 in these cells (Figure 10a) and induction of the G1 checkpoint (Figures 1b and 6a). Additional mechanisms resulting from decreased activity of MAPK pathways (p38MAPK and ERK1/2) (Figures 4 and 10) may have also contributed to the DPI-induced G1 checkpoint. While the role of ERK1/2 in proliferation is well established and works downstream of many stimuli, including growth factors and hormones (Yoon and Seger, 2006), p38MAPK has often been associated with stress responses (Dent et al., 2003); however, p38MAPK has also been shown to promote proliferation in a variety of cell types (Juretic et al., 2001; Frigo et al., 2006). Common (for example, c-jun, c-fos) and numerous distinct substrates have been identified as targets of ERK1/2 and p38MAPK activity (Ono and Han, 2000; Yoon and Seger, 2006). We show that cellular incubation with DPI modulates expression of p21Waf1. At substoichiometric levels, p21Waf1 is essential for cellular proliferation. It promotes the association of cdk4 with D-type cyclins (Cheng et al., 1999) and provides a localization signal for their nuclear import (Deng et al., 1995). Upon growth factor deprivation or in response to genotoxic stress, D cyclins are rapidly destroyed (Diehl et al., 1998); their degradation causes the release of p21Waf1 molecules from cdk4/6 complexes to arrest progression in G1, at least in part, by inhibiting cdk2 activity (Agami and Bernards, 2002). Interestingly, in DPI-treated quiescent fibroblasts, the decrease in cyclin D1 expression (Figure 10a) occurred in the presence of serum growth factors and absence of DNA damage (Figure 3). Whereas decreased pools of D-type cyclins may contribute to the DPI-induced G1 checkpoint (Figure 1b), they may also, along with decreased expression of p21Waf1, reflect entry of DPItreated density-inhibited cells into a more quiescent state. While downregulation of p21Waf1 (Figure 10a) may be due to DPI-induced biochemical events that target it for proteolysis, it may also be a consequence of feedback control mechanisms resulting from downregulation of cyclin D1. Specifically, it may also be an effect of decreased activity of p38MAPK (Figure 10a), which positively regulates its expression (Esposito et al., 2001). Overall, the DPI-mediated decrease in cyclin D1 and induction of the G1 arrest is consistent with similar observations in MEFs treated with N-acetyl-L-cysteine (NAC) (Menon et al., 2003). At high concentration (X2 : 1 ratio of p21Waf1 to cyclin/cdk complexes), p21Waf1 has well-characterized Oncogene

antiproliferative function (Deng et al., 1995). When cells were released from DPI and stimulated to proliferate, the delayed transition from G1 to S phase was associated with upregulation of p53 signaling (Figure 4). Decreased levels of phosphorylated ERK1/2 and p38MAPK were also observed (Figure 4). Hence, the redox-induced G1 checkpoint follows a response similar to that induced by DNA-damaging agents (Little, 1994). These data demonstrate that cells pretreated with compounds that perturb their redox environment respond by a p53mediated mechanism that halts their proliferation upon release from these compounds. The arrest in G1 is presumably a time to permit replenishment of signaling molecules whose expression was modified during incubation with DPI and a return to homeostasis. By 24 h after release from DPI and subculture to lower density, the levels of cyclin D1 were slightly increased (Figure 4). This increase may be a compensatory mechanism to promote exit from the G1 delay and resumption of normal proliferation. While, the DNA damage-induced G1 checkpoint is ATM- and P53-dependent (Lavin and Khanna, 1999), that induced by DPI is only partially dependent on P53 (Figure 6b) but entirely dependent on ATM and WAF1 (Figures 5a and 6c). These data suggest that ATM is a component of a normal growth pathway that regulates progression from G1 to S through p53-dependent and -independent mechanisms that operate through p21Waf1. Regulatory pathways that involve isoforms of the p53 family members p63 and p73 may be involved. The latter regulate Waf1 mRNA expression (Dietz et al., 2002), and their participation may explain the partial p53 dependence in the observed DPI-induced G1 checkpoint (Figure 6b). Regardless, our data show that ATM is as responsive to change in the cellular redox environment as it is to DNA damage. Elucidating the molecular events that regulate atm activity in mediating the DPI-induced G1 checkpoint is of significant interest. Our preliminary data (not shown) indicate that it may involve mechanism other than autophosphorylation on serine 1981. Studies (Menon et al., 2003) in proliferating p53mutated MEFs incubated with NAC had shown that redox-sensitive progression from G1 to S is independent of p53. This difference with our observations (Figure 6b) may reflect the use of asynchronous cells (in contrast to quiescent cells) and distinctive signaling events emanating from cellular incubation with DPI or NAC. While treatment of MEFs with NAC failed to affect G2 progression (Menon et al., 2003), incubation of normal human fibroblasts with DPI slowed their progression in G2 (Figure 2). Interestingly, the DPI-induced G2 delay was shorter than the G1 delay (1 versus 11 h). These data highlight the sensitivity of molecular events in early G1 to redox regulation. This was supported by lack of DPIinduced changes in expression of cyclin A (Figure 10a), which is primarily involved in late G1, S and mitotic activities (Yam et al., 2002). In summary, this study has elucidated redox-modulated signaling events associated with cellular progression in G1 and to lesser extent in G2. Important

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

29

functions of NAD(P)H oxidase and ATM in normal proliferation were revealed. Distinct differences between human and rodent cell responses to DPI were detected. While the G1 delay in human cells was DPI-concentration dependent (Figure 1b), the G1 delay in MEFs was independent (Figure 6a), suggesting acute sensitivity of MEFs proliferation to redox regulation. We propose that the ability to arrest cell cycle progression in G1 by modulating the cellular redox environment may offer some therapeutic potential to manage tumor growth, particularly tumors that harbor cells with p53 mutations, which are prevalent in most human solid cancers (Cho and Vogelstein, 1992). Notably, many cancer cells show increased production of hydrogen peroxide and other oxidizing species, which is associated with cellular proliferation (Szatrowski and Nathan, 1991). Hence, inhibiting oxidant production by flavin oxidases in cancer cells under conditions that do not induce DNA double-strand breaks in these cells or the neighboring normal cells becomes an attractive adjuvant cancer therapy. Further, most tumors harbor proliferating cells that are distributed in the different phases of the growth cycle, which significantly impacts radiation sensitivity (Terasima and Tolmach, 1963). Blocking cells in G1, by inhibiting flavin-containing oxidases, could result in depletion of tumor cells from the relatively radioresistant S-phase compartment, thereby enhancing radiosensitivity and therapeutic response. This could complement radiotherapy of cancer cells, which cannot halt progression in G1 in the first cell cycle after exposure (Nagasawa et al., 1995).

Materials and methods Cell culture Human cells Skin fibroblasts from a normal human subject (AG1522) or from a patient with ataxia telangiectasia (AG4405) were obtained from the Coriell Institute (Camden, NJ, USA). The phenotypes were confirmed by their responses to ionizing radiation (de Toledo et al., 2000). Cells in passage 11 at a density of 1.2  105 cells/dish were seeded in 60-mm polystyrene dishes and maintained as described (Azzam et al., 2000; de Toledo et al., 2000). They were subsequently fed on days 5, 7 and 9 and treatments were started 48 h after the last feeding, when 95–98% of cells were in G0/G1. Mouse embryo fibroblasts C3H 10T1/2 clone eight cells were obtained from ATCC (Rockville, MD, USA). Wild-type, gp91phox/ (Li et al., 2001), p21Waf1–/– (Deng et al., 1995) and p53/ MEFs (established in our laboratory) were derived from the C57BL/6 strain, and cultured as described (Azzam et al., 1996, 2000).

Treatment of cells Confluent/quiescent cells were exposed to either DPI (0.05– 0.2 mM) or L-NAME (25 or 100 mM) or their respective diluents, t-BOOH (0.5 mM), or were irradiated at 371C with 4 Gy (1 Gy/ min) or 0.1 Gy (1 Gy/min or 0.002 Gy/h) from 137Cs-g-rays. Exposure of AG1522 to higher concentrations of the chemical inhibitors resulted in decreased clonogenic potential.

NAD(P)H oxidase activity assay For plasma membrane isolation, cells were lysed in a buffer containing 50 mM Tris–HCl (pH 8.0) 10 mM ethylenediaminetetraacetic acid and protease inhibitors. Lysates were pulse sonicated on ice 30 times at 1 s intervals and centrifuged at 41C (3000 r.p.m., 3 min). The supernatant was transferred, centrifuged at 100 000 g (1 h), and the resulting pellet resuspended in lysis buffer. NAD(P)H oxidase activity was determined using a lucigenin-based assay. The reaction mixture contained 100 mg of protein, lucigenin (5 mM) and NADPH (100 mM), and luminescence was recorded every 30 s for 10 min. Viability/clonogenic survival Cellular viability/membrane integrity was analysed by trypanblue exclusion, and also by standard colony formation (Azzam et al., 2000). Survival values were corrected for the plating efficiency that ranged from 20 to 30%. Autoradiographic measurement of CLI Quiescent cells were suspended in medium containing [3H]thymidine (1 mCi/ml, specific activity 20 Ci/mmol), seeded at low density and incubated. At regular intervals, cells were fixed, and Kodak NTB2 nuclear emulsion was applied. After a 2-week exposure at 41C, the dishes were developed and stained (Azzam et al., 2000). To determine labeling indices, at least 1000 cells/dish were scored. Micronucleus formation Immediately after treatments, DNA damage was assayed by the micronucleus assay using the cytokinesis-block technique (Fenech and Morley, 1985). At least 1000 cells/treatment were examined, and micronuclei in binucleate cells were considered for analysis (Azzam et al., 2002). Mitochondrial membrane potential Dc was determined by flow cytometry with cell cultures incubated with the lipophilic cation JC-1 probe (Pandey et al., 2006). Cells treated with valimomycin (10 mM) to dissipate Dc served as controls. Intracellular ROS The fluorogenic substrate DCFH-DA was used to monitor intracellular generation of ROS (Hempel et al., 1999). As positive control, cells were treated with 100 mM t-BOOH (15 min, 371C). G2 checkpoint Confluent fibroblasts were subcultured, BrdU (10 mM) was added 17 h later and DPI (0.15 mM) added 2 h afterward. Immediately thereafter and at 2 h intervals, cells were harvested, fixed, stained with fluorescein isothiocyanate (FITC)-conjugated anti-BrdU antibody, counterstained with 7-amino-actinomycin D (7-AAD) and analysed by flow cytometry. Bivariate distributions of BrdU (FITC labeling) and DNA (7-AAD labeling) content were measured and percentage of cells with incorporated BrdU was determined. The rate of progression from S to G2 was calculated according to Begg et al. (1985). Western analysis Cell cultures were harvested immediately after treatment or subcultured (1 : 3 dilution) in fresh medium, and samples harvested at times thereafter. Proteins were extracted and analysed as described (de Toledo et al., 1998). Antibodies to p21Waf1 (Ab-1), p27Kip1 (Ab-6), p53 (Ab-6) and pRb (Ab-6) Oncogene

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

30 from EMD Biosciences (La Jolla, CA, USA); p34cdc2 (Sc-54) and Ku70 (Sc-1486) from Santa Cruz Biotechnology (Santa Cruz, CA, USA); p38 (9212), P-p38 (9211), ERK1/2 (9102) and P-ERK1/2 (9101) from Cell Signaling (Danvers, MA, USA); gp91phox (G95320), cyclin A (554174) and cyclinD1 (556470) from BD Biosciences (San Jose, CA, USA) and P-Ser-139 Histone H2AX (05–636) and P-Ser-1981 ATM (05– 740) from Upstate (Billerica, MA, USA) were used. Reaction with Ku70 or nonspecific antibody reaction was used to verify fractionation of equal amounts of sample. Inhibition of protein degradation MG132 (40 mM), a proteasome/calpain inhibitor, was added immediately after DPI supplementation, and remained until cell harvest. Statistical analysis Significance of difference in measured endpoints was determined by Student’s t-test. Experiments were repeated 2–5 times, and standard error of the mean is indicated in figures when greater than the size of data symbols.

Abbreviations BrdU, bromodeoxyuridine; CLI, cumulative labeling index; DCFH-DA, 20 ,70 -dichlorodihydrofluorescence diacetate; DPI, diphenyleneiodonium; ERK, extracellular-regulated-protein kinases; MAPK, mitogen-activated protein kinase; NAC, N-acetyl-L-cysteine; p53-BP1, p53-binding protein 1; RFU, relative fluorescence units; RLU, relative luminescence units; ROS, reactive oxygen species; t-BOOH, tertiary-butyl-hydroperoxide. Acknowledgements We thank Debkumar Pain, Prasad Neti, Ling Li and Mitchell Coleman for suggestions and assistance, and Philip Leder and Francis Miller for providing p21Waf1 and gp91phox knockout cells. Research grants CA92262 from the National Institutes of Health (EIA), FG02-02ER63447 and FG02-05ER64050 (EIA and DRS) and FG02-98ER62685 (JBL) from the US Department of Energy, and a VA Merit Grant and a Career Investigator Award from the American Lung Association (ABC) supported this investigation.

References Agami R, Bernards R. (2002). Convergence of mitogenic and DNA damage signaling in the G1 phase of the cell cycle. Cancer Lett 177: 111–118. Ammendola R, Ruocchio MR, Chirico G, Russo L, De Felice C, Esposito F et al. (2002). Inhibition of NADH/NADPH oxidase affects signal transduction by growth factor receptors in normal fibroblasts. Arch Biochem Biophys 397: 253–257. Azzam EI, de Toledo SM, Pykett MJ, Nagasawa H, Little JB. (1997). Cdc2 is down-regulated by ionizing radiation in a p53-dependent manner. Cell Growth Differ 8: 1161–1169. Azzam EI, de Toledo SM, Raaphorst GP, Mitchel RE. (1996). Lowdose ionizing radiation decreases the frequency of neoplastic transformation to a level below the spontaneous rate in C3H 10T1/2 cells. Radiat Res 146: 369–373. Azzam EI, de Toledo SM, Spitz DR, Little JB. (2002). Oxidative metabolism modulates signal transduction and micronucleus formation in bystander cells from alpha-particle-irradiated normal human fibroblast cultures. Cancer Res 62: 5436–5442. Azzam EI, de Toledo SM, Waker AJ, Little JB. (2000). High and low fluences of alpha-particles induce a G1 checkpoint in human diploid fibroblasts. Cancer Res 60: 2623–2631. Babior BM. (1999). NADPH oxidase: an update. Blood 93: 1464–1476. Bae YS, Kang SW, Seo MS, Baines IC, Tekle E, Chock PB et al. (1997). Epidermal growth factor (EGF)-induced generation of hydrogen peroxide. Role in EGF receptor-mediated tyrosine phosphorylation. J Biol Chem 272: 217–221. Begg AC, McNally NJ, Shrieve DC, Karcher H. (1985). A method to measure the duration of DNA synthesis and the potential doubling time from a single sample. Cytometry 6: 620–626. Burdon RH. (1995). Superoxide and hydrogen peroxide in relation to mammalian cell proliferation. Free Rad Biol Med 18: 775–794. Cals-Grierson MM, Ormerod AD. (2004). Nitric oxide function in the skin. Nitric Oxide 10: 179–193. Cheng M, Olivier P, Diehl JA, Fero M, Roussel MF, Roberts JM et al. (1999). The p21(Cip1) and p27(Kip1) CDK ‘inhibitors’ are essential activators of cyclin D-dependent kinases in murine fibroblasts. EMBO J 18: 1571–1583. Cho KR, Vogelstein B. (1992). Genetic alterations in the adenomacarcinoma sequence. Cancer 70: 1727–1731. Cross AR, Jones OT. (1986). The effect of the inhibitor diphenylene iodonium on the superoxide-generating system of neutrophils. Specific labelling of a component polypeptide of the oxidase. Biochem J 237: 111–116. Oncogene

de Toledo SM, Asaad N, Venkatachalam P, Li L, Spitz DR, Azzam EI. (2006). Adaptive responses to low-dose/low-dose-rate gamma rays in normal human fibroblasts: the role of growth architecture and oxidative metabolism. Radiat Res 166: 849–857. de Toledo SM, Azzam EI, Dahlberg WK, Gooding TB, Little JB. (2000). ATM complexes with HDM2 and promotes its rapid phosphorylation in a p53-independent manner in normal and tumor human cells exposed to ionizing radiation. Oncogene 19: 6185–6193. de Toledo SM, Azzam EI, Keng P, Laffrenier S, Little JB. (1998). Regulation by ionizing radiation of CDC2, cyclin A, cyclin B, thymidine kinase, topoisomerase IIalpha, and RAD51 expression in normal human diploid fibroblasts is dependent on p53/p21Waf1. Cell Growth Differ 9: 887–896. Deng C, Zhang P, Harper JW, Elledge SJ, Leder P. (1995). Mice lacking p21CIP1/WAF1 undergo normal development, but are defective in G1 checkpoint control. Cell 82: 675–684. Dent P, Yacoub A, Fisher PB, Hagan MP, Grant S. (2003). MAPK pathways in radiation responses. Oncogene 22: 5885–5896. Diehl JA, Cheng M, Roussel MF, Sherr CJ. (1998). Glycogen synthase kinase-3beta regulates cyclin D1 proteolysis and subcellular localization. Genes Dev 12: 3499–3511. Dietz S, Rother K, Bamberger C, Schmale H, Mossner J, Engeland K. (2002). Differential regulation of transcription and induction of programmed cell death by human p53-family members p63 and p73. FEBS Lett 525: 93–99. Doussiere J, Vignais PV. (1992). Diphenylene iodonium as an inhibitor of the NADPH oxidase complex of bovine neutrophils. Factors controlling the inhibitory potency of diphenylene iodonium in a cellfree system of oxidase activation. Eur J Biochem 208: 61–71. Droge W. (2002). Free radicals in the physiological control of cell function. Physiol Rev 82: 47–95. Esposito F, Russo L, Chirico G, Ammendola R, Russo T, Cimino F. (2001). Regulation of p21waf1/cip1 expression by intracellular redox conditions. IUBMB Life 52: 67–70. Fenech M, Morley AA. (1985). Measurement of micronuclei in lymphocytes. Mutat Res 147: 29–36. Frigo DE, Basu A, Nierth-Simpson EN, Weldon CB, Dugan CM, Elliott S et al. (2006). p38 mitogen-activated protein kinase stimulates estrogen-mediated transcription and proliferation through the phosphorylation and potentiation of the p160 coactivator glucocorticoid receptor-interacting protein 1. Mol Endocrinol 20: 971–983.

Flavin oxidases regulate G1 to S transition P Venkatachalam et al

31 Gordon DM, Dancis A, Pain D. (2000). Mechanisms of mitochondrial protein import. Essays Biochem 36: 61–73. Halliwell B. (1996). Free radicals, proteins and DNA: oxidative damage versus redox regulation. Biochem Soc Trans 24: 1023–1027. Hempel SL, Buettner GR, O’Malley YQ, Wessels DA, Flaherty DM. (1999). Dihydrofluorescein diacetate is superior for detecting intracellular oxidants: comparison with 20 ,70 -dichlorodihydrofluorescein diacetate, 5(and 6)-carboxy-20 ,70 -dichlorodihydrofluorescein diacetate, and dihydrorhodamine 123. Free Radic Biol Med 27: 146–159. Irani K, Goldschmidt-Clermont PJ. (1998). Ras, superoxide and signal transduction. Biochem Pharmacol 55: 1339–1346. Juretic N, Santibanez JF, Hurtado C, Martinez J. (2001). ERK 1,2 and p38 pathways are involved in the proliferative stimuli mediated by urokinase in osteoblastic SaOS-2 cell line. J Cell Biochem 83: 92–98. Kastan MB, Onyekwere O, Sidransky D, Vogelstein B, Craig RW. (1991). Participation of p53 protein in the cellular response to DNA damage. Cancer Res 51: 6304–6311. Kastan MB, Zhan Q, el-Deiry WS, Carrier F, Jacks T, Walsh WV et al. (1992). A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia. Cell 71: 587–597. Lavin MF, Khanna KK. (1999). ATM: the protein encoded by the gene mutated in the radiosensitive syndrome ataxia-telangiectasia. Int J Radiat Biol 75: 1201–1214. Lavoie JN, L’Allemain G, Brunet A, Muller R, Pouyssegur J. (1996). Cyclin D1 expression is regulated positively by the p42/p44MAPK and negatively by the p38/HOGMAPK pathway. J Biol Chem 271: 20608–20616. Li WG, Miller Jr FJ, Zhang HJ, Spitz DR, Oberley LW, Weintraub NL. (2001). H2O2-induced O2. Production by a non-phagocytic NAD(P)H oxidase causes oxidant injury. J Biol Chem 17: 17. Little JB. (1968). Delayed initiation of DNA synthesis in irradiated human diploid cells. Nature 218: 1064–1065. Little JB. (1994). Failla memorial lecture. Changing views of cellular radiosensitivity. Radiat Res 140: 299–311. Little JB. (2000). Radiation carcinogenesis. Carcinogenesis 21: 394–404. Maki CG, Huibregtse JM, Howley PM. (1996). In vivo ubiquitination and proteasome-mediated degradation of p53(1). Cancer Res 56: 2649–2654. Matsumoto H, Hamada N, Takahashi A, Kobayashi Y, Ohnishi T. (2007). Vanguards of paradigm shift in radiation biology: radiationinduced adaptive and bystander responses. J Radiat Res (Tokyo) 48: 97–106. Matsushime H, Ewen ME, Strom DK, Kato JY, Hanks SK, Roussel MF et al. (1992). Identification and properties of an atypical

catalytic subunit (p34PSK-J3/cdk4) for mammalian D type G1 cyclins. Cell 71: 323–334. Menon SG, Sarsour EH, Spitz DR, Higashikubo R, Sturm M, Zhang H et al. (2003). Redox regulation of the G1 to S phase transition in the mouse embryo fibroblast cell cycle. Cancer Res 63: 2109–2117. Murrell GA, Francis MJ, Bromley L. (1990). Modulation of fibroblast proliferation by oxygen free radicals. Biochem J 265: 659–665. Nagasawa H, Li CY, Maki CG, Imrich AC, Little JB. (1995). Relationship between radiation-induced G1 phase arrest and p53 function in human tumor cells. Cancer Res 55: 1842–1846. Ono K, Han J. (2000). The p38 signal transduction pathway: activation and function. Cell Signal 12: 1–13. Pandey BN, Gordon DA, Pain D, Azzam EI. (2006). Normal human fibroblasts exposed to high or low dose ionizing radiation: differential effects on mitochondrial protein import and membrane potential. Antioxid Redox Signal 8: 1253–1261. Schafer FQ, Buettner GR. (2001). Redox environment of the cell as viewed through the redox state of the glutathione disulfide/ glutathione couple. Free Radic Biol Med 30: 1191–1212. Schulze-Osthoff K, Bauer M, Vogt M, Wesselborg S, Baeuerle PA. (1997). Reactive oxygen intermediates as primary signals and second messengers in the activation of transcription factors. In: Forman Hj and Cadenas E (eds). Oxidative Stress and Signal Transduction. Chapman & Hall: New York, pp 239–259. Spitz DR, Azzam EI, Li JJ, Gius D. (2004). Metabolic oxidation/ reduction reactions and cellular responses to ionizing radiation: a unifying concept in stress response biology. Cancer Metastasis Rev 23: 311–322. Stuehr DJ, Fasehun OA, Kwon NS, Gross SS, Gonzalez JA, Levi R et al. (1991). Inhibition of macrophage and endothelial cell nitric oxide synthase by diphenyleneiodonium and its analogs. Faseb J 5: 98–103. Szatrowski TP, Nathan CF. (1991). Production of large amounts of hydrogen peroxide by human tumor cells. Cancer Res 51: 794–798. Terasima R, Tolmach LJ. (1963). X-ray sensitivity and DNA synthesis in synchronous populations of Hela cells. Science 140: 490–492. Ushio-Fukai M, Zafari AM, Fukui T, Ishizaka N, Griendling KK. (1996). p22phox is a critical component of the superoxide-generating NADH/NADPH oxidase system and regulates angiotensin IIinduced hypertrophy in vascular smooth muscle cells. J Biol Chem 271: 23317–23321. Yam CH, Fung TK, Poon RY. (2002). Cyclin A in cell cycle control and cancer. Cell Mol Life Sci 59: 1317–1326. Yoon S, Seger R. (2006). The extracellular signal-regulated kinase: multiple substrates regulate diverse cellular functions. Growth Factors 24: 21–44.

Oncogene