NF-κB Potentiates Caspase Independent Hydrogen Peroxide Induced Cell Death

Background The pro-survival activity of NF-κB in response to a variety of stimuli has been extensively characterized. Although there have been a few reports addressing the pro-cell death role of NF-κB, the precise mechanism of NF-κB's pro-cell death function still remains elusive. Methodology/Principal Findings In the present study, we investigated the role of NF-κB in cell death induced by chronic insult with hydrogen peroxide (H2O2). Here, we show that NF-κB promotes H2O2 induced caspase independent but PARP dependent fibroblast cell death. The pro-death activity of NF-κB is due to the DNA binding activity of RelA, which is induced through IKK- mediated IκBα degradation. NF-κB dependent pro-survival genes, Bcl-2 and XIAP, were significantly repressed, while NF-κB dependent pro-death genes, TNFα and Fas Ligand, were induced in response to H2O2. Conclusions/Significance We discovered an unexpected function of NF-κB, in that it potentiates chronic H2O2 exposure induced cell death, and suggest that NF-κB mediates cell death through the repression of pro-survival genes and induction of pro-death genes. Since unremitting exposure of tissues to H2O2 and other reactive oxygen species can lead to several degenerative disorders and diseases, our results have important implications for the use of NF-κB inhibitors in therapeutic drug design.


Introduction
Mammalian cells are constantly exposed to reactive oxygen species (ROS), such as hydrogen peroxide (H 2 O 2 ). Exogenous ROS arise from irradiation (UV, X-ray, c-ray) and atmospheric pollutants, while endogenous ROS are mainly produced by the incomplete reduction of oxygen by cytochrome c during cellular respiration [1]. However, when the antioxidant capabilities of the cell are overwhelmed by ROS, a state of oxidative stress ensues, which can result in damage to DNA, proteins, and lipids [2]. Moreover, high and/or persistent levels of ROS result in aberrant cell death, which leads to aging and neurodegenerative disorders [3,4]. In particular, ROS induced fibroblast cell death can cause chronic obstructive pulmonary disease [5,6] as well as inadequate wound healing following myocardial infarction/reperfusion [7,8]. ROS induces cell death by modulating cell signaling pathways. A prominent signaling pathway involved in mediating the cell survival/cell death fate is the nuclear factor-kB (NF-kB) pathway [3].
NF-kB is a family of transcription factors, which are comprised of five family members: RelA/p65, RelB, c-Rel, nfkb1/p50, and nfkb2/p52, that form homo-or hetero-dimers in a combinatorial manner. In resting cells, the NF-kB dimers are retained in the cytoplasm by forming stable complexes with NF-kB inhibitor molecules, IkB (a/b/e). In the canonical activation pathway, stimulation with an extracellular stimulus, such as tumor necrosis factor a (TNFa), a pro-inflammatory cytokine, leads to phosphorylation of IkBa on serines 32 and 36 by IKK, the IkB kinase. This results in the ubiquitination of IkBa, which signals for the degradation of IkBa by the 26S proteasome. The freed NF-kB dimers can then translocate to the nucleus and activate transcription of their target genes [9,10,11].
Activation of NF-kB by exogenous H 2 O 2 has been found to be highly cell type dependent, in which NF-kB is activated in a variety of cell lines such as Jurkat T cells and HeLa cells [12,13], whereas NF-kB activation is inhibited in other cell lines such as murine neutrophils [14]. In cases where activation of NF-kB occurs, several mechanisms of NF-kB activation have been reported. While canonical activation of NF-kB via IKK-dependent IkBa degradation has been reported, other reports focus on an atypical mechanism of NF-kB activation in response to stimulation with H 2 O 2 [15,16]. This atypical mechanism involves an IKK independent mechanism and Tyr42 phosphorylation of IkBa, and only occurs in the absence of SHIP-1 [17,18]. The pathway of NF-kB activation in other cell lines, such as in mouse embryonic fibroblasts (MEFs), has yet to be delineated.
The anti-cell death role of NF-kB has been extensively characterized. RelA deficient cultured cells undergo apoptotic cell death upon treatment with TNFa due to deficiencies in pro-survival and anti-oxidant gene transcription [19,20]. RelA deficiency also leads to embryonic lethality accompanied by massive apoptosis in the embryonic liver [21]. In response to a variety of other stimuli, such as ionizing radiation and chemotherapeutic drugs, RelA also appears to have an anti-apoptotic effect [22]. Finally, NF-kB suppression of apoptosis in cancer cells is a central event in cancer biology, as well as in chemoresistance of tumor cells [23]. However, there have also been a few scattered reports addressing the pro-cell death function of NF-kB in response to atypical NF-kB activators [24,25,26,27,28]. Yet, the mechanism by which NF-kB mediates a pro-cell death response remains elusive.
In the present study, we sought to define NF-kB's role in immortalized MEF cell death induced by chronic insult with H 2 O 2 .
Here we present evidence that unremitting exposure to H 2 O 2 induces a caspase independent but PARP dependent cell death and that NF-kB potentiates cell death through the DNA binding activity of RelA, which is induced through the canonical activation pathway. Given that NF-kB dependent pro-survival genes, Bcl-2 and XIAP, were significantly repressed, while NF-kB dependent pro-death genes, TNFa and Fas Ligand, were induced in response to H 2 O 2 , we suggest that NF-kB mediates its pro-cell death function through the repression of pro-survival genes and induction of pro-death genes.

Generation of oxidative stress
Cells were first grown to 95-98% confluency. GO was dissolved in 50 mM sodium acetate pH 5.1, and added to fresh media supplemented with 10% BCS, except in the cases where RNA was extracted, in which serum free media was used.

Measurement of ROS production
Measurement of intracellular ROS production was carried out as described in [30].The resulting supernatants were scanned with a Fluoromax-P instrument (J. Y. Horiba, Inc.) using a bandpass of 5 nm and ex 492 nm/em 526 nm. , where FI = fluorescence intensity of PI or CFSE of GO treated or untreated samples. A minimum of 2610 6 cells was used to label cells with 5 mM CFSE. Following GO treatment, each plate was then washed twice with 16DPBS containing 100 mg/ml CaCl 2 and Mg 2 Cl 2 , and treated with 10 mg/ml PI for 15 min in the dark. To remove adherent cells, cells were incubated with Puck's buffer (5.4 mM KCl, 0.14 M NaCl, 4.2 mM NaHCO 3 , 5.6 mM D-glucose dextrose, 10 mM Hepes, 1 mM EDTA, pH 7.4) for 15 min in the dark. For UV treated cells, floating cells were combined with the adherent cells. The fluorescence intensity for PI (ex 535 nm/em 617 nm) and CFSE (ex 492 nm/em 517 nm) were then taken with a Fluoromax-P instrument (J. Y. Horiba, Inc.) using a bandpass of 5 nm.

Molecular Biology
IkBa constructs were cloned into the retrovirus vector pBabepuro between the restriction sites EcoRI and SalI. Mutagenesis reaction was performed with the Stratagene Quickchange Mutagenesis kit. The following primers were used to clone the R35AY36A Tg: forward: 59-CGGGGCATGCGATTCGCCG-CAAAATGCGAGGGGCGC-39, reverse: 59-GCGCCCCTC-GCATTTTGCGGCGAATCGCATGCCCCG-39.

Retroviral transgenic system
293T cells were transfected using Lipofectamine 2000 (Invitrogen) containing 7 mg of the retroviral vector and 3 mg of pCl-Eco (Imgenex). Serum free DMEM containing the Lipofectamine and DNA mixture was removed 6 hrs later and replaced with DMEM supplemented with 10% FBS and 1xPSG. Cells were allowed to grow for 38-42 hrs post transfection. The media was then placed onto the nfkb 2/2 cells along with 8 mg/ml polybrene (Sigma). These cells were then grown for another 48 hrs before selection with 2.5 mg/ml puromycin (Calbiochem).

In vitro IKK kinase assay
Kinase assays were carried out as described in [29] using 2.0 mg of recombinant GST-IkBa (1-54). The reaction was visualized by PhosphorImager (Molecular Dynamics) and quantitated by ImageQuant TL.

Annexin V staining
Cells were washed with PBS and resuspended in 100 ml of binding buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl 2 ). Alexa Fluor 647 conjugated Annexin V (BioLegend) and 7-AAD (BioLegend) were added according to the manufacturer's instructions, and the cells were incubated in the dark for 15 min at room temperature. The cells were analyzed using a Becton Dickinson FACSCalibur flow cytometer. Data analysis was performed with FlowJo software (Tree Star). Annexin V/7-AAD double-positive cells are necrotic and excluded from analysis.

Quantitative Real Time PCR (qPCR)
RNA was isolated using the RNeasy kit (QIAGEN) and further purified with DNase I digestion (QIAGEN). cDNA was synthesized from 1 ug of RNA using the Superscript III First Strand synthesis kit (Invitrogen) and Oligo(dT) primers. Product accumulation was monitored by SYBR Green (Kapa Biosystems) fluorescence with Eppendorf Mastercycler ep realplex. The relative gene expression levels were calculated using the 2 ' (-DDCT) method [32,33]. The p value was calculated using the Student's t test. The following primers were used: GAPDH Forward-AATGTGTCCGTCGTGGATCT Reverse-CATCGAAGGTGGAAGAGTGG, TNFa Forward-CC- All primers are provided in 59 to 39 fashion. Primers obtained from the Primer Bank [34,35,36] are given with the PrimerBank ID Accession number. GAPDH primers were obtained from [37], TNFa primers were obtained from [38], and Bcl-2 primers were obtained from [39]. XIAP primers were designed using Primer3.

Continuous H 2 O 2 exposure (via GO) to fibroblasts induces a caspase independent but PARP dependent cell death
In chronic obstructive pulmonary disease and inadequate wound healing following myocardial infarction/reperfusion, chronic insult of ROS to fibroblast cells can lead to aberrant cell death [6,7,8]. We attempted to simulate this condition by generating persistent oxidative stress in immortalized MEFs with glucose oxidase (GO). It has been shown that addition of GO to the media produces continuous levels of intracellular H 2 O 2 , as opposed to a bolus addition of H 2 O 2 [40], which we were also able to verify (J.Ho, data not shown). Intracellular ROS generation was verified by staining with DCFH-DA, a commonly used membrane permeable dye that fluoresces upon its oxidation to DCF by intracellular ROS [30]. Accordingly, ROS was produced in MEFs throughout the duration of treatment with 25 mU/ml GO (Fig. 1A). To roughly estimate the amount of intracellular H 2 O 2 that GO was producing, known concentrations of H 2 O 2 were also added in a bolus fashion to the cells and the intensity of fluorescence monitored. Based on the H 2 O 2 standard curve, we estimated that, within the first 3 hrs of treatment with 25 mU/ml GO, less than 50 mM of H 2 O 2 was being produced, and following 3 hrs, roughly 50-100 mM of H 2 O 2 was produced (Fig. 1A).
We next determined the effect of prolonged H 2 O 2 exposure (via 25 mU/ml GO) on MEFs. Significant cell death occurred within 5 hrs after GO treatment, as quantitated by the increase in propidium iodide (PI) incorporation, which correlates to a loss in membrane integrity. PI incorporation increased even further 6 hrs following treatment (Fig. 1B). Since an increase in oxidative stress can switch the cell death mode from a caspase dependent to a caspase independent cell death mode [41], we set out to determine whether GO generated H 2 O 2 involved caspases. Treatment with irreversible general caspase inhibitor, z-VAD-fmk, was unable to prevent cell death. In agreement with these results, immunoblots against the cleaved form of caspase 3, one of the main executioners of apoptosis, revealed that cleaved caspase 3 is not present during treatment with GO (Fig. 1C). This is in contrast to UV induced MEF cell death, where caspases mediate cell death, as seen by the inhibition of cell death upon treatment with zVAD-fmk and by the production of the active form of caspase 3 (Fig. 1B, C).
Caspase independent MEF cell death induced by bolus addition of H 2 O 2 occurs in a Poly(ADP-ribose) polymerase (PARP) dependent manner [42], which we were also able to confirm (Fig. 1D). PARP, a nuclear enzyme involved in the DNA damage response and cell death, becomes activated in response to DNA damage and attaches ADP-ribose units (PAR polymer) to itself and nuclear proteins [43]. We next sought to evaluate the contribution of PARP in cell death induced by continuous exposure to H 2 O 2 . Upon GO addition, cells pretreated with PARP inhibitor, DPQ, showed significant resistance to membrane permeabilization, as opposed to cells treated without DPQ (Fig. 1E). Given that PARP protein levels remained constant while PAR formation increased, as verified by western blot against PARP and PAR, we concluded that PARP was activated within 15 min of stimulation with GO. PAR formation was then eliminated upon pretreatment with DPQ (Fig. 1E). Thus, treatment of MEFs with GO results in a caspase independent but PARP dependent cell death.

NF-kB potentiates H 2 O 2 induced cell death
There have been extensive reports illustrating cooperativity between PARP and NF-kB. In response to DNA damaging agents such as ionizing radiation, PARP-1 has been shown to be an essential upstream mediator of NF-kB activation [44,45]. PARP-1 can also act as a direct co-activator of NF-kB [46]. Since we have shown that chronic insult with H 2 O 2 induces a PARP dependent cell death, we next wanted to determine whether NF-kB also plays any role in H 2 O 2 induced cell death. To examine this, we compared rates of cell death in wt and nfkb 2/2 (nfkb1 2/2 rela 2/2 crel 2/2 ) MEFs treated with 25 mU/ml GO. Although we expected that cell death would be enhanced in the absence of NF-kB due to its well established pro-survival activity, we found, to our surprise, that MEFs which lacked p50, RelA and c-Rel were more resistant to H 2 O 2 induced cell death than wt cells ( Fig. 2A).
To directly address the involvement of NF-kB in the delay of cell death, nfkb 2/2 cells were reconstituted, using a retroviral transgenic system, with either rela transgene (Tg), which is the major transactivating subunit of NF-kB, or a DNA defective binding mutant of rela (R35AY36A Tg). Both rela Tg and R35AY36A Tg reconstituted cells contained similar levels of RelA protein when compared to wt MEFs (Fig. 2B). Upon treatment with GO, rela Tg cells displayed increased incorporation of PI as opposed to nfkb 2/2 cells reconstituted with either empty vector (pBabe) or R35AY36A Tg (Fig. 2B). However the amount of PI incorporation in rela Tg cells was decreased compared to wt cells, indicating only partial rescue of the cell death phenotype. Our attempt to study cell death in relA Tg reconstituted in rela 2/2 cells was unsuccessful since reconstitution with relA did not rescue the cell death phenotype. This could be because of the transformation of the rela 2/2 cells with viral oncoproteins, which disrupts many transcriptional regulatory pathways, and resulted in the loss of this mechanism of modulating NF-kB function. Regardless, the lack of cell death in R35AY36A Tg in nfkb 2/2 cells indicates that NF-kB promoted cell death is dependent on NF-kB, in particular, RelA DNA binding activity. Additionally, we confirmed that the differences in the rate of cell death between wt and nfkb 2/2 cells reconstituted with either pBabe or rela Tg were primarily dependent on NF-kB. This was done by determining that all cell lines contained similar basal levels of p53, a well-known and potent contributor of cell death (Fig. S1).
Next, to examine the reason for the partial rescue of the cell death phenotype in response to GO, the response of the rela reconstituted cells to TNFa induced cell death, as well as TNFa induced DNA binding was inspected. Since it has been shown that de novo synthesis prevents the occurrence of apoptotic death upon treatment with TNFa [20], cells containing a transcriptionally active RelA should prevent cell death from occurring. Accordingly, upon TNFa stimulation, wt and rela Tg cells did not undergo cell death, as opposed to R35AY36A Tg and pBabe cells (Fig. 2C). EMSA results also show that NF-kB was not activated in both pBabe and R35AY36A Tg cells, in contrast to wt and, to a lower extent, rela Tg cells (Fig. 2D). The incomplete rescue of NF-kB activation in rela Tg cells might be due to the absence of NF-kB subunits, p50 and cRel. The p50:RelA heterodimer is the major NF-kB species, and both cRel and RelA homodimers have also been reported to play roles in the regulation of target gene expression [47]. Therefore, the absence of these dimers and the presence of only RelA homodimer alone might result in only suboptimal activation of some NF-kB target genes, and explains the partially functionally rescued relA reconstituted cells. Due to technical difficulties reconstituting p50 into nfkb 2/2 cells, we were unable to distinguish whether NF-kB's pro-cell death function is due to RelA homodimer or the p50:RelA heterodimer. However, from our data (Fig. 2B), it is clear that a RelA containing dimer is involved in promoting cell death.
The canonical NF-kB activation pathway is required for the pro-cell death function of NF-kB in response to chronic exposure to H 2 O 2 To further determine whether NF-kB acts as a promoter of cell death in response to H 2 O 2 , we compared the rates of cell death of ikba 2/2 MEFs which were reconstituted with either wt IkBa (wt ikba Tg) or a mutant IkBa, where both Ser 32 and Ser36, which are the IKK phosphorylation sites, are mutated to alanine (aa ikba Tg), thus preventing NF-kB activation. Interestingly, in order to achieve cell death in a timely fashion, 50 mU/ml GO instead of 25 mU/ml GO, was needed to induce cell death in both cell lines. This may be a reflection of cell line-specific characteristics or due to the over-expression of IkBa upon reconstitution in these cell lines (G.Ghosh, data not shown). Nevertheless, wt ikba Tg cells had a rapid rate of cell death compared to aa ikba Tg cells, where the canonical NF-kB activation pathway had been blocked (Fig. 3A). As expected, NF-kB activation was completely blocked in aa ikba Tg cells as demonstrated by the lack of NF-kB activation upon TNFa stimulation by EMSA analysis (Fig. 3B). These results further suggest that the resistance to death of aa ikba Tg cells is due to the lack of NF-kB activity. All together, these results strongly indicate that NF-kB plays a pro-cell death role in response to H 2 O 2 induced caspase independent cell death, and that the canonical activation pathway is required in mediating NF-kB's pro-cell death role.
We next examined whether NF-kB was activated in response to chronic exposure to H 2 O 2 . Indeed, as seen by EMSA analysis, NF-kB was activated in a prolonged manner upon addition of GO in MEFs (Fig. 3C). Given that NF-kB has been reported to be activated in both an IKK independent and dependent manner in response to H 2 O 2 , we then performed IKK activity assays, in which IKK was immunoprecipitated from GO treated MEFs followed by an in vitro kinase assay. IKK activity assays reveal that IKK is activated following one hour of treatment with GO (Fig. 3D). Accordingly, there is also concomitant degradation of IkBa (Fig. 3E). All together these results show that NF-kB is activated via the canonical pathway in MEFs in response to chronic exposure to H 2 O 2 .
NF-kB-dependent survival genes are repressed whereas death promoting genes are induced by H 2 O 2 As a transcription factor, NF-kB's active participation in cell death is likely to be mediated through its target genes. Therefore, we set out to examine the expression pattern of several NF-kB target genes that are known to impact cell death or survival (Table 1). Using real time quantitative PCR (qPCR), we measured mRNA levels at 0 and 4 hrs after GO treatment in wt MEF and nfkb 2/2 cells reconstituted with empty vector (pBabe) or relA Tg. While the majority of these genes did not undergo changes in expression levels (Table 1), we identified 4 genes that underwent significant alterations: cell survival factors, Bcl-2 and XIAP, and cell death promoting factors, TNFa and Fas ligand (FasL). The significant reduction of Bcl-2 gene levels in both wt and rela Tg cells and not pBabe cells (Fig. 4) implies that Bcl-2 repression is due to the presence of RelA. Decreased Bcl-2 protein levels were also observed only in wt, and not nfkb 2/2 cells, 6 hrs following treatment with GO (Fig. S2), suggesting that changes in mRNA levels correspond to changes in protein levels. A similar trend in gene expression is also seen for X-linked inhibitor of apoptosis protein, XIAP. In contrast, TNFa expression was significantly induced in both wt and relA Tg cells, while not in pBabe cells. FasL was induced in only wt cells. These results suggest that TNFa, but not FasL, induction by H 2 O 2 is RelA dependent. The lack of FasL induction in rela Tg cells could be attributed to the partial rescue phenotype of the relA Tg cells (Fig. 2B).

Discussion
This study has discovered an unexpected function of NF-kB in that it promotes MEF cell death in response to chronic insult with H 2 O 2 . We have shown that intracellular H 2 O 2 was continuously produced in MEFs treated with GO and that this unremitting exposure to H 2 O 2 resulted in a caspase-independent but PARP dependent cell death. We also show that the pro-death activity of NF-kB is dependent on the DNA binding activity of RelA, which is induced through IKK-mediated IkBa degradation. The death promoting activity of NF-kB might be mediated by the down regulation of a subset of NF-kB dependent pro-survival factors and up-regulation of NF-kB dependent pro-death factors, as demonstrated by the repression of Bcl-2 and XIAP, and induction of TNFa and FasL.
Due to the repression of pro-survival factor Bcl-2 in wt and rela Tg cells, we propose that the pro-cell death function of NF-kB is primarily due to its transcriptional down regulation of Bcl-2, since it has been shown that in MNNG induced caspase independent but PARP-1 dependent MEF cell death, over-expression of Bcl-2 can delay cell death [42]. Bcl-2 over-expression delays cell death due to its ability to prevent translocation of apoptosis inducing factor, AIF, from the mitochondria to the nucleus. This translocation event is central in causing MNNG and H 2 O 2 induced MEF cell death [42]. Our proposed model is supported by two additional pieces of data. First, Bcl-2 protein levels decrease in wt MEFs as opposed to nfkb 2/2 MEFs after 6 hrs of treatment with 25 mU/ml of GO (Fig. S2), indicating that the changes in Bcl-2 mRNA level translates to changes in protein levels. Secondly, the fact that Bcl-2 over-expression can only delay cell death also fits with our proposed model, since nfkb 2/2 cells, which contain higher levels of Bcl-2 than either wt or rela Tg cells, eventually also succumb to cell death (J. Ho, data not shown). Additionally, our result, in which Bcl-2 is down regulated in a seemingly NF-kB dependent manner to promote PARP dependent fibroblast cell death, suggests a novel level of cooperativity between PARP and NF-kB. This is because PARP and NF-kB cooperativity has only been shown in instances where PARP-1 can act as a direct co-activator for RelA and p50, or as an upstream  mediator of NF-kB activation in response to various stimuli [44,45,46]. We have convincingly demonstrated that NF-kB has a pro-cell death function in response to chronic insult with H 2 O 2 in MEFs. Future experiments will be performed to determine whether this effect occurs in other cell types. Aside from MEFs, this type of caspase independent but PARP dependent cell death has only been reported to occur in cortical neurons, upon N-Methyl-Daspartic acid (NMDA) over-stimulation [42,48]. It is yet unknown whether NF-kB has any role in this NMDA induced, PARP dependent cell death in cortical neurons. Interestingly, it was shown that in HeLa cells, NF-kB promotes cell survival in response to a single bolus addition of low nanomolar concentration of H 2 O 2 [13]. This variation can be attributed to differences in cell type, as well as differences in the application of H 2 O 2 , and thus amount of intracellular H 2 O 2 , which, as previously stated, can switch the cell death mode. We propose that NF-kB can be anti-cell death in caspase dependent cell death induced by transient or low levels of H 2 O 2 , but that NF-kB has a pro-cell death function in caspase independent cell death induced by continuous or high levels of H 2 O 2 . Indeed, NF-kB's role in promoting cell survival and death is a complex event. This is demonstrated by genotoxic agent, VP16, induced cell death, in which both pro-and anti-apoptotic genes were induced in NF-kB dependent manner, such that the final outcome depended on a balance of the induction levels of pro-and anti-apoptotic genes [39].
Our gene expression results are strongly supported by previous studies, which showed that cell death stimulation with daunorubicin, cisplatin, or p14 ARF over-expression in U2OS osteosarcoma cells resulted in RelA mediated transcriptional repression of prosurvival genes, Bcl-xL or XIAP [49]. These previous reports also showed that RelA acts in a dominantly transcriptionally repressive manner. Interestingly, we observe selective activation and repression of certain death and survival genes, in that TNFa was induced while Bcl-2 and XIAP were repressed in a seemingly NF-kB dependent manner (Fig. 4). There are currently more reports addressing NF-kB's singular transcriptional response in promoting cell death (either complete repression or induction of survival or death genes) [26,27,28] and significantly fewer reports describing NF-kB's mixed transcriptional response in promoting cell death (repression and induction of survival and death genes, respectively) [25]. This is the first report suggesting that NF-kB uses a mixture of transcriptional responses in promoting cell death upon ROS stimulation. It is unclear at this stage as to how NF-kB mediates repression of some target promoters and activation of others. However, given that oxidative stress is known to inactivate the cysteine active sites of cellular phosphatases, this would change the cellular phospho-protein state [18] and could potentially inhibit or enhance the recruitment of coactivators and corepressors. Additionally, RelA mediated repression of pro-survival genes has been reported to involve Thr505 phosphorylation of RelA, which can enhance the interaction between RelA and HDAC1 [49]. Further studies are still required to fully unravel this mechanism.
The induction and repression of cell death and survival genes, respectively, in response to H 2 O 2 also suggests that activation of NF-kB is required for mediating its pro-cell death response. Previous reports of NF-kB's pro-cell death function have been shown to depend on either basal [24,25] or activated NF-kB [49]. However, we cannot fully conclude that NF-kB activation is solely required to mediate the pro-cell death response, since pBabe reconstituted cells contain lower levels of basal NF-kB (Fig. 2D).
Our studies have also clearly shown that NF-kB is activated via the canonical pathway in MEFs, which is in contrast to reports where NF-kB activated in response to oxidative stress occurred via an atypical mechanism, involving an IKK independent mechanism and IkBa Tyr42 phosphorylation. The presence of SHIP1 has been reported to revert the mechanism of NF-kB activation from atypical to canonical in response to H 2 O 2 [18]. Thus, the presence of SHIP2 in MEFs, which is functionally similar to SHIP1 [50], supports our observed NF-kB activation mechanism.
Unremitting exposure of tissues to ROS can lead to pathological conditions, such as neurodegenerative disorders and chronic obstructive pulmonary disease [3,4,6]. Our result that NF-kB might play a role in promoting cell death adds another layer of complexity to therapeutic drug design and should be taken into consideration when NF-kB inhibitor pharmaceutical targets are used in treatment. Overall, this study shows that NF-kB dependent transcription is responsible for promoting H 2 O 2 induced cell death. Further experiments in future study will be done to explore the detailed molecular mechanism. Figure S1 wt and pBabe or rela Tg reconstituted nfkb 2/2 MEFs contain similar levels of basal p53. Cell lysate of wt and nfkb 2/2 MEFs reconstituted with pBabe or rela Tg were analyzed by western blotting against p53 and a-tubulin. (TIF) Figure S2 Bcl-2 protein levels significantly decrease in wt MEFs as opposed to nfkb 2/2 cells upon continuous exposure to H 2 O 2 . wt MEFs and nfkb 2/2 MEFs were either untreated or treated with 25 mU/ml GO for 6 hrs. Cell lysate was analyzed by western blotting against Bcl-2 and a-tubulin. (TIF)