NADPH Oxidase-generated Hydrogen Peroxide Induces DNA Damage in Mutant FLT3-expressing Leukemia Cells*

Background: NADPH oxidase is a hydrogen peroxide-generating enzyme, involved in redox signaling in cancer. Results: NADPH oxidase-generated hydrogen peroxide increases DNA damage and genomic instability. Conclusion: NADPH oxidase-derived hydrogen peroxide mediates DNA damage in acute myeloid leukemia (AML). Significance: The mechanism of DNA damage generation is important for studying aggressive AML phenotypes. Internal tandem duplication of the FMS-like tyrosine kinase (FLT3-ITD) receptor is present in 20% of acute myeloid leukemia (AML) patients and it has been associated with an aggressive AML phenotype. FLT3-ITD expressing cell lines have been shown to generate increased levels of reactive oxygen species (ROS) and DNA double strand breaks (DSBs). However, the molecular basis of how FLT3-ITD-driven ROS leads to the aggressive form of AML is not clearly understood. Our group has previously reported that inhibition of FLT3-ITD signaling results in post-translational down-regulation of p22phox, a small membrane-bound subunit of the NADPH oxidase (NOX) complex. Here we demonstrated that 32D cells, a myeloblast-like cell line transfected with FLT3-ITD, have a higher protein level of p22phox and p22phox-interacting NOX isoforms than 32D cells transfected with the wild type FLT3 receptor (FLT3-WT). The inhibition of NOX proteins, p22phox, and NOX protein knockdowns caused a reduction in ROS, as measured with a hydrogen peroxide (H2O2)-specific dye, peroxy orange 1 (PO1), and nuclear H2O2, as measured with nuclear peroxy emerald 1 (NucPE1). These reductions in the level of H2O2 following the NOX knockdowns were accompanied by a decrease in the number of DNA DSBs. We showed that 32D cells that express FLT3-ITD have a higher level of both oxidized DNA and DNA DSBs than their wild type counterparts. We also observed that NOX4 and p22phox localize to the nuclear membrane in MV4–11 cells expressing FLT3-ITD. Taken together these data indicate that NOX and p22phox mediate the ROS production from FLT3-ITD that signal to the nucleus causing genomic instability.

FMS-like tyrosine kinase 3 (FLT3) 2 is a type III receptor tyrosine kinase important for hematopoiesis, normally expressed in early hematopoietic precursors (1,2). Internal tandem duplication (ITD) is a mutation in the juxtamembrane domain of the FLT3 that results in the loss of the autoinhibitory function and a constitutive activation of the receptor (1,3). FLT3-ITD mutation occurs in about 20% of AML cases and it is associated with a poor prognosis for patients (4,5). Constitutively activated FLT3-ITD kinase stimulates aberrant proliferative signaling pathways, including PI3K, RAS/ERK, and STAT5 pathways characterized by resistance to apoptosis, abnormal cell growth, and differentiation block (6 -10). It was also demonstrated that cells expressing the ITD-mutated FLT3 generate higher levels of ROS (11)(12)(13).
ROS, for example, H 2 O 2 , are considered to play an important role in cancers, including leukemia (14,15). Due to their DNA damaging properties, they are known to contribute to genomic instability. It was proposed that tumor cells must acquire some form of genomic instability, because the normal rate of mutation is insufficient to provide the number of mutations required for the oncogenic transformation (16 -18). Different oncogenes have been shown to induce a higher level of ROS and genomic instability, e.g. BCR-ABL, RAS (18 -21). However, little is known of how FLT3-ITD generates such as stress.
There are several proposed mechanisms of how genomic instability occurs in cancers. FLT3-ITD was demonstrated to activate alternative unfaithful DNA repair pathways that leads to increased levels of unrepaired DNA damage (22). Interestingly, it was also shown that increased efficiency of FLT3-ITDstimulated DNA repair contributes to drug resistance (23).
Another origin of genomic instability is increased ROS production that causes excessive DNA damage. Sallmyr et al. (11) showed that FLT3-ITD-generated ROS are mediated by Rac1 GTPase, which is an essential component of the NOX complex. NOXs are one of the sources of ROS in cells. There are 7 isoforms of NOXs, NOX1-5, and DUOX1-2 that display remarkable differences in the recruitment of regulatory subunits (p22 phox , p47 phox , p67 phox , and Rac1/2), mechanisms of activation, and distinct subcellular localization. NOX1-4 require p22 phox for the correct functioning and stability of the complex (24). The role of NOXs in various processes of the cellular transformation, e.g. genomic instability, cell growth and survival, angiogenesis and metastasis, has been well established in recent years (25). Emerging work has suggested that NOX4-derived ROS may play a substantial role in genomic instability (26).
It was proposed that FLT3-ITD controls NOX through levels of the rate-limiting substrate NADPH (27). The same report demonstrated that NOX2 and NOX4 have been shown to play a role in migration and growth in FLT3-ITD expressing cells (27). FLT3-ITD activated NOX-produced ROS were also revealed to cause oxidation of tumor suppressor DEP-1 phosphatase (12).
Our group demonstrated that FLT3-ITD-stimulated ROS are mediated by maintenance of expression of p22 phox , a small membrane-bound NOX complex subunit, expression (13). We have also shown that p22 phox -mediated ROS are critical for phosphorylation of STAT5 (13).
In this report we demonstrated that FLT3-ITD causes increased levels of the nuclear H 2 O 2 that damages DNA. We showed that in p22 phox , NOX siRNA knockdowns caused a decrease in H 2 O 2 with a subsequent decrease in DNA damage in these cells. Here we propose that FLT3-ITD causes an increase in NOX and p22 phox protein levels that generate H 2 O 2 at the nuclear membrane. This H 2 O 2 diffuses to the nucleus where it oxidatively damages DNA contributing to genomic instability.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-The human leukemic cell lines, MV4-11 (homozygous for the FLT3-ITD mutation) and HL-60 (homozygous for the FLT3-WT), were all purchased from DSMZ (Braunschweig, Germany). The 32D cell line, stably transfected with FLT3-WT or FLT3-ITD, was a kind gift from Prof. Hubert Serve from Goethe University Frankfurt and Prof. Frank D. Bohmer from the Universitatsklinkium Jena. The cell lines were maintained in RPMI 1640 supplemented with 10% FBS, 1% penicillin/streptomycin, and 2 mM L-glutamine in a humidified incubator at 37°C with 5% CO 2 . For 32D cell lines, 10% WEHI-conditioned medium was added as a source of IL-3. FLT3-ITD was inhibited using PKC412 (50 nM; Tocris Biosciences, Bristol, UK) at the indicated times. NOX inhibition was achieved using diphenyleneiodonium (DPI; Sigma) at the indicated times and concentrations. Dimethyl sulfoxide was used as a vehicle. Stimulation of wild type FLT3 receptor was achieved by incubation of the 32D cell line transfected with FLT3-WT with recombinant human FLT3 ligand (100 ng/ml; number 300-19, PeproTech).
Measurement of Intracellular H 2 O 2 -Total intracellular H 2 O 2 was measured by incubating cells with 10 M cell-permeable H 2 O 2 probe PO1 (a kind gift from Dr. Chang, University of California Berkley) for 1 h in the dark. Cells were then briefly washed with PBS and immediately read by flow cytometry using FACSCalibur (BD Biosciences) and Cellquest Pro software (BD Biosciences). The mean fluorescent intensity of 10,000 events was determined. The measurement of nuclear H 2 O 2 was achieved using NucPE1 (a gift from Dr. Chang). The cells were incubated for 45 min at 10 M NucPE1 in the dark. The incubation was followed by washing and analysis by flow cytometry as explained above. Mitochondrial ROS were measured using MitoSOX probe. The cells were incubated with 5 M MitoSOX for 15 min in the dark. The incubation was followed by washing and analysis by flow cytometry.
Live Confocal Microscopy-The MV4-11 cells were incubated overnight on the poly-D-lysine (P4707; Sigma)-coated glass bottomed dishes (P35G-1.5-14-C; MatTek Corp., Ashland, MA). 1 h before imaging the cells were treated where indicated with 10 M DPI or PKC412 (50 nM; Tocris Biosciences, Bristol, UK). For siRNA experiments, following the transfection, the cells were plated overnight on the glass-bottomed dishes as indicated above. Confocal fluorescence live imaging of PO1-stained cells was performed as described previously in Woolley et al. (13) using a Zeiss LSM510 META confocal microscope fitted with ϫ63 1.4 plan achromat lens. Live imaging of the NucPE1 probe was carried out using excitation at 488 nm with an argon laser, and emission was collected using a META detector at about 520 nm. The Hoechst dye was incubated together with NucPE1 where indicated. When multiple staining of NucPE1 and PO1 was performed, the multitracking mode of scanning was applied for acquisition of the images. Image analysis was performed in Carl Zeiss Zen Light Edition.
Immunofluorescence-The immunofluorescence procedure was performed as described in Woolley et al. (13). Briefly, the cells were incubated overnight on the poly-D-lysine-coated glass coverslips. The next day, cells were fixed for 1 h in 3% PFA/PBS and permeabilized with 0.2% BSA, 0.05% saponin/ PBS. The cells were then incubated with 50 l of appropriate primary antibody diluted in 5% FBS/PBS for 1 h at room temperature in a humidified chamber. Following washing, secondary antibodies conjugated to Alexa Fluor 594 or 488 diluted in 5% FBS/PBS were incubated for 1 h in the dark. After washing the coverslips were mounted on the slides using Mowiol. Images were acquired using Flouview FV1000 Confocal Laser Scanning Microscope with a ϫ100 oil immersion objective. Images are represented as a single slice from a Z stack projection.
8-OHdG Assay-Oxidative damage was assessed using 8-hydroxy-2Ј-deoxyguanosine (8-OHdG) as a marker, as described in Ref. 28. Approximately 2 ϫ 10 6 cells were washed with PBS and fixed for 1 h in 3% paraformaldehyde/PBS. After washing, cells were treated with 2 M HCl for 20 min at room temperature. HCl was removed and the cells were treated with 0.1 M sodium borate, pH 8.5, for 2 min. Cells were then washed and permeabilized with 0.2% BSA, 0.05% saponin/PBS. Following three additional washes with 3% BSA/PBS, cells were incubated with primary antibody against 8-OHdG (ab26842; 1:50; Abcam) overnight at 4°C in a humidified chamber. The next day, the cells were washed three times with 3% BSA/PBS solution, and incubated with the secondary antibody conjugated to Alexa Fluor 594 for 1 h at room temperature. After removing the secondary antibody, cells were washed three times in PBS and analyzed using flow cytometry.
␥-H2AX Assay-DNA DSBs were measured using ␥-H2AX (phosphorylated H2AX histone) as a marker. Approximately 2 ϫ 10 6 cells were washed with PBS and fixed for 1 h in 3% PFA/PBS. Cells were then incubated in 70% ethanol at Ϫ20°C overnight. The next day, cells were washed and blocked for 15 min with 1% BSA/PBS. Following washing, cells were incubated in ␥-H2AX antibody conjugated to Alexa Fluor 488 (Cell Signaling Technology, number 9719) in 1% BSA/PBS solution at 4°C overnight. Cells were washed and analyzed by flow cytometry.
Western Blotting-The immunoblotting procedure was carried out as previously described by us in Woolley et al. (13). Briefly, the cells were lysed in RIPA buffer. Following the determination of protein concentration using the Bio-Rad Protein Assay (Bio-Rad), the proteins were resolved using SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose membrane and incubated with appropriate primary antibody. 24 h later, the membrane was incubated in the secondary antibody coupled with Alexa Fluor 680 or 800. The signal was detected with the Odyssey infrared imaging system (LI-COR Biosciences).
Statistical Analysis-the results are expressed as a percentage of control, defined to 100%. Values are mean Ϯ S.D. and are representative of three independent experiments. Data were statistically analyzed using Student's t test with p Ͻ 0.05.

Inhibition of FLT3-ITD and NOX in MV4-11 Cell Line Reduced Total and Nuclear Endogenous H 2 O 2 -MV4-11
is a well established model AML cell line that expresses homozygous FLT3-ITD (Fig. 1A). Cells were treated with a commonly used, FLT3 inhibitor, PKC412 that is currently in clinical trials to treat AML patients, or with DPI, a flavoprotein inhibitor, which inhibits NOX proteins. We showed that both of these treatments result in a 20 -30% decrease in H 2 O 2 levels in MV4-11 when compared with vehicle (dimethyl sulfoxide)treated control (Fig. 1B). These data suggested that both FLT3-ITD and NOX proteins play a role in the generation of H 2 O 2 in these cells.
Having confirmed that FLT3-ITD and NOX are involved in the increase in total H 2 O 2 , we investigated if the levels of nuclear H 2 O 2 were affected by FLT3-ITD or NOX inhibition.
To study H 2 O 2 levels in the nucleus, we used a specific nuclear H 2 O 2 probe, NucPE1. The probe was shown to specifically respond to H 2 O 2 changes in the nucleus in vitro and in vivo (29). However, to confirm this occurrence in our cell line, we colocalized the NucPE1 fluorescence with nuclear tracker (Hoechst 33342) (Fig. 1C).
Using confocal microscopy, we show in Fig. 1D, that inhibition of FLT3-ITD with PKC412 resulted in a significant decrease in the nuclear H 2 O 2 (Fig. 1D). Quantification of this change using flow cytometry revealed a 30% decrease in nuclear H 2 O 2 following the PKC412 treatment (Fig. 1E). To determine whether NOX proteins were involved in generating H 2 O 2 that was detectable in the nucleus, we measured nuclear H 2 O 2 upon DPI treatment. Here we report that inhibition of NOX proteins resulted in a decrease of 35% in nuclear H 2 O 2 in MV4-11 ( Fig. 1, F and G). These data suggested that FLT3-ITD and NOX proteins are involved in the increase in nuclear H 2 O 2 .

32D Stably Transfected with FLT3-ITD Have higher Levels of Endogenous H 2 O 2 Than 32Ds Transfected with FLT3-WT-
The inhibition of FLT3 receptor or NOX protein caused a decrease in H 2 O 2 in 32D/FLT3-ITD cells, but not in 32D/ FLT3-WT cells. The 32D cell line was previously used as a model to show differences in signaling from wild type versus mutated FLT3 receptor (8) (Fig. 1A). Here we showed that these cells transfected with FLT3-ITD possessed 100% more H 2 O 2 than 32D cells transfected with the FLT3-WT receptor ( Fig.  2A). Moreover, inhibition of FLT3 or NOX resulted in the decrease in the cellular H 2 O 2 by ϳ20% specifically in cells expressing mutated FLT3-ITD, and not the wild type FLT3 (Fig.  2B). Similarly, we investigated the nuclear H 2 O 2 level in these cells. 32D cells expressing the FLT-ITD carried increased levels of nuclear H 2 O 2 when compared with the FLT3-WT expressing equivalents (Fig. 2C).
To examine if mitochondria were responsible for the differences in H 2 O 2 levels in transfected 32Ds, we used MitoSOX, a probe that specifically responds to ROS changes in mitochondria. When we compared the mitochondrial ROS levels in 32D/ FLT3-ITD and 32D/FLT3-WT, we saw no significant change between them (Fig. 2D).

FLT3-ITD-expressing Cells Have Higher Levels of DNA Oxidation and DNA DSBs Than Their Wild Type Equivalents-
The inhibition on the FLT3-ITD oncogene with PKC412 caused a decrease in DNA damage. Given that both FLT3-ITD and NOX are involved in the H 2 O 2 increase in the nucleus, we investigated if this H 2 O 2 caused DNA damage in the nucleus. First, we investigated the levels of DNA DSBs. ␥-H2AX has been used repeatedly as a marker of DNA DSBs. We examined ␥-H2AX levels in patient-derived cell lines carrying FLT3-WT (HL-60) or FLT3-ITD (MV4-11). We show here that MV4-11 cells possess about 50% more DNA DSBs than HL-60 (Fig. 3A).
For the purpose of pinpointing DNA damage specifically to FLT3-ITD signaling, we again employed 32Ds transfected with wild type and mutated FLT3. We examined the level of oxidized DNA with 8-OHdG as a marker. 32D/FLT3-ITD cells had notably higher levels of oxidized DNA than 32D/FLT3-WT cells (Fig. 3B).
We also compared the levels of DNA DSBs in the transfected 32D cells. Cells transfected with FLT3-ITD had ϳ75% more DNA DSBs than cells transfected with FLT3-WT, measured by ␥-H2AX, and quantified by flow cytometry (Fig. 3C).
It was shown previously that FLT3 inhibitors reversibly inhibited FLT3 autophosphorylation (30). We investigated if FLT3-ITD-induced DNA damage could be reversed. First, we inhibited FLT3 (PKC412) or NOX (DPI) in 32D/FLT3-ITD and 32D/FLT3-WT cells. These treatments caused 25-40% reduction in DNA damage in cells possessing the mutated FLT3-ITD. The DNA damage in FLT3-WT-possessing 32D cells remained unchanged (Fig. 3D). Second, we treated MV4-11 cells with PKC412 and we noted a significant decrease in DNA oxidation. After removing PKC412 by washing and allowing a recovery period of 24 h from the treatment, the oxidized DNA returned to initial level. However, when DPI was included in the medium during the recovery time, the oxidized DNA did not return to its initial level (Fig. 3E). These results suggested that NOXs plays a role in oxidizing DNA in FLT3-ITD expressing cells.
The same experimental protocol was applied to measure the reversibility of DNA DSBs in FLT3-ITD expressing cells. Similarly to oxidized DNA, PKC412 treatment reduced the level of DNA DSBs. The 24-h recovery period from the drug resulted in the regeneration of DNA DSBs in these cells. However, the increase in the level of DNA DSBs was slightly higher than that in the control. This may be due to the fact that the PKC412 inhibitor may distinctly affect 8-oxoguanine/␥-H2AX generating and repairing pathways. Removal of the FLT3 inhibitor led to an immediate stimulation of the major oncogene signaling pathway in these cells. That may have led to a temporary upregulation/down-regulation of certain pathways that would result in the increase in the ␥-H2AX, e.g. DNA DSB repair. Importantly, when the cells were treated with DPI during the recovery time, restoration of the initial number of DNA DSBs was partially prevented (Fig. 3F). These data suggested that FLT3 and NOX-generated DNA damage could be reversed by inhibiting either of these proteins.
FLT3-ITD-driven p22 phox -mediated Nuclear H 2 O 2 , DNA Oxidation, and DNA DSBs Generation in FLT3-ITD Expressing Cells-Our group has shown that following the inhibition of FLT3-ITD with PKC412, the small membrane subunit of the NOX complex, p22 phox is post-translationally down-regulated (13) (Fig.  4A). It was reported previously that p22 phox is an essential component of the NOX complex functioning. Hence, we used specific p22 phox siRNA to knock down the protein and study the effects of p22 phox down-regulation (Fig. 4B). Using PO1 and NucPE1, we investigated the levels of total and nuclear endogenous H 2 O 2 . Knockdown of p22 phox resulted in a significant decrease in both total and nuclearH 2 O 2 (Fig.4C).Quantificationofthesedatausingflowcytometry revealed that specific p22 phox knockdown resulted in an almost 20% decrease in total and nuclear H 2 O 2 (Fig. 4D). To verify if the p22 phox -generated H 2 O 2 in the nucleus caused DNA damage, we utilized similar markers, 8-OHdG and ␥-H2AX. The p22 phox down-regulation caused a notable decrease in the level of oxidized DNA (Fig. 4E) and DNA DSBs (Fig. 4F). These data suggested that p22 phox mediates the FLT3-ITD-stimulated DNA damage.
NOX4-generated ROS Caused DNA DSBs in FLT3-ITD Expressing Cells-We have shown in Woolley et al. (13) that p22 phox -mediated ROS generation in MV4-11 takes place in the endoplasmic reticulum (ER). To date the only NOX isoform reported in the ER is NOX4. Hence, we knocked down NOX4 to determine whether NOX4 was the isoform that plays a role in generating H 2 O 2 in MV4-11. The down-regulation of the NOX4 expression resulted in ϳ30% reduction of total cellular H 2 O 2 in these cells and 20% reduction of nuclear H 2 O 2 , as measured by PO1 and NucPE1, respectively (Fig. 5, A and B). Furthermore, the drop in H 2 O 2 was accompanied by the decrease in the ␥-H2AX fluorescence, which corresponds to the number of DNA DSBs (Fig. 5C).
NOX4 and p22 phox Colocalized to the Nuclear Membrane in MV4-11-Given that NOX4 knockdown caused a decrease in both H 2 O 2 and DNA damage, we investigated if NOX4 localizes to the nucleus as previously shown in the literature in other cell types (31)(32)(33). We observed no clear NOX4 foci in the nucleus (Fig. 6A) and investigated if NOX4 colocalizes to the nuclear membrane, using NUP98 (nucleoporin) antibody (Fig. 6B). We saw that the most inner NOX4 foci colocalize to NUP98. Following that, we also confirmed that p22 phox similarly colocalizes to NUP98.
p22 phox Knockdown Caused a Decrease in DNA DSBs in 32D Cells Transfected with FLT3-ITD, but Not with FLT3-WT-To study the effects downstream of p22 phox in the 32D cell line in relationship to ITD/WT FLT3, first, we examined the protein levels of p22 phox in these cells in the presence/absence of IL-3 and PKC412 (Fig. 7A). Western blotting analysis of p22 phox revealed that the cells transfected with FLT3-ITD possessed a higher level of p22 phox than their wild type counterparts. Similarly to MV4-11 cells, the inhibition of FLT3 receptor with PKC412 over 24 h caused a partial decrease in p22 phox expression. IL-3 starvation resulted in a slight down-regulation of p22 phox . This is not surprising as IL-3 has been shown to activate the AKT pathway, which in turn has been implicated in stimulating p22 phox expression (34,35).
Specific p22 phox siRNA knockdown in 32D cells (Fig. 7B) have allowed us to investigate the effects of the p22 phox -dependent NOX complex in cells expressing wild type or mutant FLT3. Flow cytometric analysis of endogenous cellular H 2 O 2 , measured by PO1, showed that p22 phox knockdown caused an ϳ20% decrease in cellular H 2 O 2 , specifically in 32D cells expressing FLT3-ITD. These knockdowns did not reveal any statistically significant changes in 32D cells expressing FLT3-WT (Fig. 7C). To study the downstream effects of this p22 phox -generated H 2 O 2 , we examined the levels of DNA DSBs, using flow cytometric analysis of ␥H2AX fluorescence. Similarly, the siRNA knockdown of p22 phox proteins resulted in a reduction in the number of DNA DSBs in 32Ds transfected with constitutively active FLT3-ITD (Fig. 7D).
Knockdowns of p22 phox -dependent NOX Isoforms (NOX1, -2, and -4) Caused a Decrease in Endogenous H 2 O 2 and DNA Damage in 32D/FLT3-ITD Expressing Cells-Human FLT3-possessing myeloid cells have been previously shown to express NOX2, -4, and -5 (27). However, their murine counterparts have only been shown to express NOX1, -2, and -4 as they do not possess the NOX5 gene (27). We therefore compared the protein levels of these NOX isoforms in 32D/FLT3-WT and 32D/FLT3-ITD cells. Western blotting analysis revealed that FLT3-ITD-expressing cells possess increased protein levels of NOX1, NOX2,   . NOX4 and p22 phox colocalizes to the nuclear membrane in MV4-11. Cells were incubated on poly-D-lysine-coated coverslips for 16 h. The cells were then fixed in 3% PFA/PBS and followed by NOX4, NUP98, and p22 phox and Hoeschst staining. Images were acquired using confocal microscope. Brightness and contrast were adjusted. The scale bar represents 10 m. and NOX4 in comparison to their wild type counterparts (Fig.  8, A-C).
Following this observation we targeted these isoforms for siRNA transfection. Specific knockdowns of NOX1, -2, and -4 proteins were confirmed using Western blotting (Fig. 8, D-F).
Down-regulation of NOX2 or NOX4 protein caused ϳ20% decrease in the endogenous H 2 O 2 and a robust 30% decrease in DNA damage (Fig. 8, G and H). Conversely, NOX1 knockdown resulted only in a marginal decrease in both H 2 O 2 and DNA damage (Fig. 8, G and H). Interestingly, the change in DNA damage following the NOX2 or NOX4 knockdowns was greater than that observed for p22 phox (Fig. 8, G and H).

Stimulation of FLT3-WT with FLT3 Ligand (FL) Causes an Increase in p22 phox Protein Level, Followed by an Accumulation of Nuclear and Cellular H 2 O 2 and in Turn an Increase in the
Number of DNA DSBs-FLT3-ITD has been reported to induce differential signaling events than activated FLT3-WT. For example, in contrast to stimulated FLT3-WT, FLT3-ITD causes a strong activation of STAT5 (9). This may be a result of the distinct subcellular localizations of the two (9). Given that constitutively active FLT3-ITD stimulates generation of ROS and DNA damage, we investigated if the activated FLT3-WT had a similar effect. Following a 16-h incubation of the FL with 32D cells transfected with FLT3-WT, we observed an increase in p22 phox expression (Fig. 9A). The FL-stimulated increase in p22 phox level was similar to that induced by FLT3-ITD (Fig. 9A). Importantly, this change in p22 phox levels was followed by the 44% increase in the generation of H 2 O 2 , as measured by PO1 and about 20% increase in the generation of nuclear, as measured by NucPE1 (Fig. 9, B and C). Similarly to mutated FLT3, activated wild type FLT3 possessed about 50% higher number of DNA DSBs (Fig. 9D).

DISCUSSION
The FLT3-ITD oncogene is associated with an aggressive progression of AML (2). It is well established that the disease advances as the cancer cells become genomically unstable (16). Prior work has documented ROS contribution to leukemic phenotypes (extensively reviewed in Ref. 14). However, there has been little evidence of the mechanism of FLT3-ITD-stimulated ROS-driven genomic instability. FLT3-ITD was previously associated with increased ROS and DNA damage (11). However, the cellular source of DNA-damaging ROS in FLT3-ITD remains elusive.
In this study, we explored how FLT3-ITD stimulates H 2 O 2 production to cause DNA damage. We found that the NOX-p22 phox complex generates H 2 O 2 that causes DNA damage in the nuclei of FLT3-ITD-expressing cells. Importantly, to our knowledge, this is a first study that finds NOX to play a role in genomic instability in leukemia.
NOX-originated ROS have been documented to have multiple effects on leukemia cell proliferation and survival (13,19,27). Leukemic oncogenes have been shown to regulate either the expression of NOX components or the availability of their substrate. For example, overexpression of H-RAS has been shown to increase NOX4 expression (36). Interestingly, cells expressing FLT3-ITD have been reported to possess an increased concentration of NADPH (27). Our group also showed that inhibition of FLT3-ITD in MV4-11 cells leads to the degradation of p22 phox , a small membrane subunit of the NOX complex (13). In this report we demonstrate that 32D cells transfected with FLT3-ITD possess higher steady levels of p22 phox than their wild type counterparts. Furthermore, 32D/ FLT3-ITD cells exhibit higher NOX1, -2, and -4 protein levels than 32D cells expressing FLT3-WT.
To date, NOX proteins were shown to regulate the growth and migration of FLT3-ITD expressing cells (9). However, the authors did not report any significant changes in total ROS following the NOX4 knockdown, measured by 2Ј,7Ј-dichlorodihydrofluorescein diacetate (27). Although here, using a newly developed, specific H 2 O 2 probe, PO1, we show that NOX1, -2, and -4 knockdowns result in the decrease in the endogenous H 2 O 2 . We show that NOX2/4 siRNA knockdown causes a significant decrease in DNA DSBs. This indicates a novel NOX function in causing DNA damage in FLT3-ITD oncogenedriven AML.
Similar to what we observe in FLT3-ITD expressing cells, H-RAS-induced increased NOX4 expression is accompanied by an increase in the nuclear H 2 O 2 and DNA DSBs (36). We suggest that a similar mechanism operates in FLT3-ITD-expressing cells where increased levels of NOX proteins generate an ROS burst that causes DNA damage.
Apart from oncogene-regulated NOX4-stimulated oxidative DNA damage, NOX4 was shown to specifically play a role in the nucleus. For instance, in cardiomyocytes, NOX4-produced H 2 O 2 was also observed to specifically oxidize nuclear proteins, e.g. histone deacetylase (33). Also, NOX4, localized in the perinuclear space was also the source of nuclear superoxide generation in hepatocytes (38). Interestingly, exposure of mice to ionizing radiation increased the expression of NOX4 in hematopoietic stem cells (10,39). The inhibition of NOX in the ionizing radiation-treated mice, attenuated the ROS and DNA damage associated with the ionizing radiation (40).
Our group has shown that inhibition of FLT3-ITD results in a post-translational down-regulation of p22 phox (13). Here we show for the first time that the stimulation of FLT3-WT with FL results in an up-regulation of p22 phox , an increase in total and nuclear H 2 O 2 as well as DNA DSBs. It has been reported that FLT3 inhibitors were demonstrated to induce expression and release of FL (48). We suggest that the presence of FL in the bone marrow could lead to genomic instability through ROS generation.
We showed previously that p22 phox resides in the ER in FLT3-ITD where it regulates H 2 O 2 generation, which affects phosphorylation of STAT5 (13). Consistently with the earlier report, we show here that not only does p22 phox -regulate H 2 O 2 in the FLT3-ITD system but it also affects the H 2 O 2 level in the nucleus. We postulate that NOX proteins associated with p22 phox are essential for generation of H 2 O 2 that diffuses to the nucleus.
Additionally, p22 phox knockdown results in a decrease in oxidized DNA and DNA DSBs in FLT3-ITD expressing cells. This result is consistent with effects of p22 phox on DNA damage seen in H-RAS-transfected cells, where the knockdown of p22 phox resulted in a decrease in DNA DSBs (36). Together these data suggest that NOX-produced H 2 O 2 is directly responsible for causing DNA damage in the nucleus. This finding substantially expands the role of NOX in FLT3-ITD AML.
FLT3-ITD-stimulated ROS generation was demonstrated in several papers (11,13). However, we show for the first time, using confocal microscopy that FLT3-ITD directly affects nuclear H 2 O 2 , as measured by NucPE1. Furthermore, inhibition of FLT3-ITD or NOX causes a dramatic decrease in the level of nuclear H 2 O 2 .
Although FLT3 seems an attractive target in AML, the resistance arising to FLT3 inhibitors still remains a significant problem (5,8,49). Resistance to FLT3 inhibitors was demonstrated to be associated with mutations within the FLT3 gene (8,50,51). Therefore, it is of interest to study the mechanisms that lead to mutagenesis. 8-OHdG is one of the most persistent and mutagenic type of lesions. It was shown recently that patients in AML relapse possessed higher levels of 8-OHdG (52). We show here that FLT3-ITD is not only associated with increased levels of DNA DSBs, but also increased levels of 8-OHdG. Increased levels of 8-OHdG was shown to play a role in self-mutagenesis of BCR/ABL that led to imatinib resistance (53). We suggest that a similar mechanism can operate in FLT3-ITD. What is more, DNA DSBs were documented to arise from oxidative DNA damage in the S/G 2 M phase in BCR/ABL cells (54). Repair of 8-OHdG results in a removal of the oxidized base that produces a single strand break (55). When the latter one is encountered by the replication fork, it can result in a DSB (37). We suggest that a similar phenomenon may be occurring in the FLT3-ITD AML.
In conclusion, we suggest that NOX-p22 phox complex generates H 2 O 2 at the membrane in FLT3-ITD expressing cells (Fig.  10). This H 2 O 2 diffuses into the nucleus where it causes oxidation of DNA and DNA DSBs. FLT3-ITD increases the levels of H 2 O 2 through increases in both NOX and p22 phox protein levels. We speculate that NOX may be an attractive therapeutic target in FLT3-ITD-mutated AML cells, as NOX inhibition can decrease the level of genomic instability.