c-Abl regulates Early Growth Response Protein (EGR1) in ... - Nature

Oncogene (2005) 24, 8085–8092

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

ORIGINAL PAPERS

c-Abl regulates Early Growth Response Protein (EGR1) in response to oxidative stress Jeremy R Stuart1, Hidehiko Kawai1, Kelvin KC Tsai1, Eric Y Chuang2 and Zhi-Min Yuan*,1 1 Department of Genetics and Complex Diseases, Harvard School of Public Health, Boston, MA 02115, USA; 2Radiation Oncology Branch, National Cancer Institute, Bethesda, MD 20892, USA

c-Abl is a tyrosine kinase that can act as a regulator of cell growth and apoptosis in response to stress. Using cell lines expressing c-Abl in an inducible manner, we identified genes whose expression was regulated by c-Abl kinase activity. Microarray analysis indicated that Early Growth Response-1 (EGR1) gene expression is induced by c-Abl kinase activity, which was confirmed at the message and protein levels. Promoter mapping experiments revealed that c-Abl utilizes three distal serum response elements (SREs) in the EGR1 promoter, which are transactivated by mitogen/extracellular receptor kinase (MEK/ERK) signaling. PD 95089, a specific inhibitor of MEK/ERK signaling, attenuated c-Ablmediated upregulation of EGR1 expression in a dosedependent manner. Similar results were obtained by using a dominant-negative mutant of mitogen/extracellular kinase. Significantly, hydrogen peroxide-induced EGR1 expression appears to be mediated by c-Abl, as cells expressing dominant negative c-Abl, and c-Abl / murine embryonic fibroblasts, are completely defective in hydrogen peroxide-induced EGR1 expression. In addition, c-Abl-induced apoptosis is partially mitigated by EGR1 activity, as cells devoid of EGR1 expression undergo reduced rates of c-Abl-induced apoptosis. Together, these results indicate that c-Abl promotes the induction of EGR1 through the MEK/ERK pathway in regulating apoptotic response to oxidative stress. Oncogene (2005) 24, 8085–8092. doi:10.1038/sj.onc.1208953; published online 1 August 2005 Keywords: c-Abl; EGR1; oxidative stress; tyrosine kinase

Introduction c-Abl is a member of the nonreceptor tyrosine kinase family, a class of enzymes responsible for the transmission of cellular signals in response to stimuli. The c-abl *Correspondence: Z-M Yuan, Department of Genetics and Complex Diseases, Harvard School of Public Health, 665 Huntington Ave, HSPH-1 Room 508, Boston, MA 02115, USA; E-mail: [email protected] Received 24 May 2005; accepted 22 June 2005; published online 1 August 2005

gene encodes a 140 kDa protein that localizes to the nucleus and cytoplasm and whose transforming variants have been implicated in tumorigenesis. c-Abl has various functions, including regulation of growth cycle checkpoints. Overexpression of c-Abl inhibits cell growth and leads to G1 arrest (Goga et al., 1995), whereas cells expressing RNA interference (RNAi) oligonucleotides directed against c-Abl message exhibit shorter G1/S transitions (Daniel et al., 1995). c-Abl is also activated in response to various DNA-damaging agents. In response to ionizing radiation, the ataxiatelangiectasia-mutated protein (ATM) directly phosphorylates c-Abl, promoting p53-dependent mechanisms of apoptosis and cell cycle arrest (Baskaran et al., 1997). DNA-PK, a kinase involved in DNA repair, phosphorylates c-Abl in response to stress, leading to its activation (Kharbanda et al., 1997). Expression of wild type (WT), but not kinase defective c-Abl, is associated with an induction of apoptotic cell death (Yuan et al., 1997). In the cytoplasm, growth factors have been shown to stimulate kinase activity of c-Abl (Plattner et al., 1999), thereby linking c-Abl with growthpromoting as well as growth-inhibiting functions. The early growth response-1 (EGR1) protein is another growth-related protein that responds to stress. EGR1 is a zinc-finger transcription factor belonging to a group of proteins that quickly respond to stimuli, as stimulation with many environmental signals, including growth factors, hormones, and neurotransmitters, rapidly induces EGR1 gene expression (for review, see Thiel and Cibelli, 2002). Various target genes of EGR1 have been identified, such as p53, insulin-like growth factor, PDGF, TGF-beta, and PTEN (Khachigian et al., 1995; Khachigian et al., 1996; Nair et al., 1997; Svaren et al., 2000; Virolle et al., 2001). As a result, EGR1 has been implicated in growth and proliferation as well as in tumor suppression and apoptosis, depending on the cellular and environmental context. EGR1 expression is stimulated by cellular stress, such as DNA damage, hypoxia, and chemical treatment (Thiel and Cibelli, 2002). The mechanism of EGR1 induction in response to various agents has been partially elucidated. The promoter of EGR1 contains many CArG elements, also known as serum response elements (SREs), which were first identified to bind the serum response transcription factor (SRF) and Elk1 transcription factor in response to growth agents (Shaw et al., 1989). Other promoter

c-Abl regulates EGR1 JR Stuart et al

8086

a

c-Abl

80 60 40 20 0

Actin GADPH

24 hr

P-Tyr

48 hr

KA KR EV KA KR EV

EGR1 Actin

c H202 STI-571

Actin

d + +

KR Tet H202

+ +

+ + +

EV + + + EGR1

Abl

Flag c-Abl P-Tyr Actin

e

EGR1 -/H202 FBS

Results

b

KA KR 0 24 48 0 24 48 (hr)

FoldInduction

elements have subsequently been identified that mitigate transactivation of the EGR1 promoter in response to stress (Sakamoto et al., 1991). However, the mechanism of EGR1 induction upstream of canonical-signaling pathways remains largely undefined. In this study, we show that c-Abl kinase activity results in increased expression of EGR1. c-Abl transactivates the EGR1 promoter by activating the mitogen/ extracellular receptor kinase (MEK/ERK)-signaling pathway, leading to SRE binding in the EGR1 promoter. We also demonstrate that c-Abl is necessary for EGR1 induction in response to oxidative stress and that some of the apoptotic functions of c-Abl are mediated by EGR1. Our results support a model in which c-Abl activates the MEK/ERK-signaling pathway in response to oxidative stress, leading to EGR1 expression and subsequent apoptosis.

Abl -/-

WT

+ + + 1 2 3 4 5 6 7 8 9 10 11 12 Abl EGR1

Oxidative stress activates c-Abl to induce EGR1 expression In order to identify potential novel effectors of c-Abl kinase, stable 293, U20S cell lines expressing either kinase active c-Abl (KA) or a kinase inactive mutant (KR) gene under the control of the tetracycline promoter were created. These cells were induced to express c-Abl for 0, 24, and 48 h, and Western blotting assessed c-Abl induction and corresponding kinase activity. c-Abl is effectively expressed in these cells after induction with tetracycline, and immunoblots of whole cell lysates with phosphotyrosine antibodies show a dramatic increase in c-Abl kinase activity as compared to noninduced or c-Abl (KR)-expressing cells (Figure 1a). Microarray analysis of these cells identified a number of genes whose expression patterns varied as a result of c-Abl kinase activity (data not shown). Among these genes, the gene encoding the transcription factor EGR1 was consistently induced by c-Abl in a kinasedependent manner. To verify the microarray results, and assess the extent of EGR1 induction, Q-RT-PCR was performed. Upon induction of c-Abl expression, EGR1 message was increased 50–60-fold when normalized to either actin or GADPH (Figure 1b). Western analysis showed that EGR1 protein levels were indeed increased in cells expressing c-Abl KA, but not in cells expressing c-Abl KR or empty vector controls (Figure 1b, bottom panels). To examine the biological relevance of the results obtained from c-Abl expression, we used a lossof-function approach to validate c-Abl-mediated upregulation of EGR1. Since c-Abl and EGR1 respond to similar types of cellular stress, such as cisplatin, IR, and hydrogen peroxide, we decided to test whether oxidative stress-induced c-Abl activation could result in induction of EGR1 expression. We first verified that hydrogen peroxide treatment could stimulate c-Abl kinase activity as reported by Sun et al. (2000). As expected, cells treated with 500 mM hydrogen peroxide showed an increase in c-Abl kinase activity as reflected by c-Abl Oncogene

Actin

Figure 1 c-Abl kinase activity results in elevated EGR1 expression in response to oxidative stress. (a) Stable 293 cells were induced to express kinase active (KA) or inactive (KR) c-Abl for the indicated time points. c-Abl expression and kinase activity was verified using anti-c-Abl and anti-phospho-tyrosine antibodies, respectively. (b) Q-RT-PCR and Western analysis of the cell lines in (a) to verify microarray results. Cells were induced for the indicated time points and Q-RT-PCR was performed using two reference genes (actin and GADPH) whose expression did not vary. EGR1 protein expression after c-Abl induction was determined by Western analysis in whole cell lysates derived from c-Abl-KA, KR or empty vector (EV)-expressing cells. (c) Activation of endogenous Abl after oxidative stress. WT fibroblasts were serum starved for 24 h, and treated with 500 mM H2O2 for 30 min prior to harvesting. Lysates were immunoprecipitated with c-Abl GST-Crk and blotted with the indicated antibodies to detect c-Abl expression and autophosphorylation. (d) Dominant-negative abrogation of c-Abl kinase activity abolishes EGR1 induction in response to oxidative stress. U20S cells were induced to express dominant-negative c-Abl-KR or EV in the presence of hydrogen peroxide treatment. Cells were induced 24 h prior to H2O2 treatment and cells were harvested 3 h later. (e) EGR1 induction in response to oxidative stress is absent in Abl null MEFs. Abl / , EGR1 / , or WT MEFs were treated for 3 h with 100 mM hydrogen peroxide or 2 h with 15% FBS prior to harvest. Lysates were run on SDS–PAGE and blotted with the indicated antibodies

autophosphorylation (Brasher and Van Etten, 2000; Tanis et al., 2003), which was completely blocked by the Abl inhibitor STI-571 (Figure 1c). Next, we asked whether hydrogen peroxide-activated c-Abl was associated with EGR1 induction. Towards this end, we employed U20S cells inducibly expressing c-Abl-KR, which can function as a dominant-negative regulator of c-Abl kinase activity (Yuan et al., 1997), to examine the dependence of EGR1 induction on the kinase activity of c-Abl. Significantly, hydrogen peroxide exposure resulted in considerable induction of EGR1 in vectorexpressing cells and in noninduced cells, but not in c-Abl-KR expressing cells (Figure 1d). To further confirm


8087

To elucidate the mechanism by which c-Abl induces EGR1 expression, deletion constructs of the EGR1 promoter were synthesized and fused to a luciferase reporter. These promoter constructs were cotransfected with c-Abl-KA in 293T cells to map the region responsible for EGR1 induction by c-Abl. The fulllength and 480 EGR1 promoter constructs showed an approximate 10-fold increase in luciferase reporter activity whereas promoter constructs containing the 235 or 110 proximal nucleotides from the ATG start codon in the EGR1 promoter exhibited no additional reporter activity (Figure 2a). These results suggest that the three distal but not the three proximal SREs in the EGR1 promoter respond to c-Abl kinase activity. Further analysis of the EGR1 promoter using additional deletion constructs showed that the region containing the three distal SREs from the ATG start codon in the EGR1 promoter displayed elevated luciferase reporter activity similar to the full-length EGR1 promoter, while a promoter construct containing only the two proximal SREs showed no enhanced reporter activity (Figure 2b; compare 480/ 235 and 347). A summary of the promoter constructs and their corresponding luciferase reporter activity is shown in Figure 2c. These results indicate that the three distal SREs in the EGR1 promoter are necessary and sufficient for c-Ablmediated transcription of EGR1. c-Abl induces EGR1 expression by activation of the MEK/ERK pathway Previous studies have shown that the transcription factors Elk-1 and SRF cooperate to induce EGR1 expression via SRE binding in the EGR1 promoter, and that they are in turn activated by the MEK/ERK- and SAPK-signaling pathways (Kaufmann et al., 2001). In addition, activation of MEK Kinase 1 by c-Abl has been demonstrated in response to genotoxic agents (Kharbanda et al., 2000). These reports along with our promoter mapping data prompted us to investigate whether c-Abl promotes EGR1 expression by activation of the MEK/ERK-signaling pathway. Full-length EGR1 promoter construct fused to a luciferase reporter was cotransfected with c-Abl-KA alone or in the presence of

Fold-Induction

12 10 8 6 4 2 0

Baseline c-Abl KA c-Abl KR

-600 -480 -235 -110 12 10 8 6 4 2 0

-2 35

80

/-2

-6 00

-3 47

Baseline c-Abl KA c-Abl KR

35

b

-4

Serum response elements in the EGR1 promoter are necessary for c-Abl-mediated EGR1 transcriptional induction

a

Fold-Induction

these results, we tested EGR1 induction in Abl null MEFs in response to hydrogen peroxide treatment. The results of this experiment indicate that Abl / MEFs are completely defective in hydrogen peroxide-induced EGR1 expression as compared to their normal counterparts (Figure 1e). Intriguingly, these cells were able to respond to serum treatment (Figure 1e lane 8), suggesting that c-Abl mediates specific aspects of EGR1 regulation. Together, these results indicate that oxidative stress-induced upregulation of EGR1 is mediated by c-Abl.

c EB -600

S

1 Sp -500

P B E E Es p1 Fk M TA S R S R S N cA S R T A -400 -300

-200

-100

Mean FoldInduction -600 11 -480 10 -347 1.8 -235 1.5 -110 1.2 -480/-235 8.0

Figure 2 Serum response elements are necessary and sufficient for c-Abl-mediated transactivation of the EGR1 promoter. (a, b) EGR1 promoter constructs fused to a luciferase reporter were cotransfected with c-Abl-KA or -KR. Cell lysates were harvested 24 h later and assayed for luciferase reporter activity. Values are the result of three independent experiments performed in duplicate. (c) A map of the EGR1 promoter and corresponding induction values obtained for each promoter construct used in (a) and (b). SREs are boxed in white

increasing amounts of the MEK inhibitor PD 98059. Exogenous MEK kinase, a downstream effector of the MAPK pathway and potent inducer of EGR1 expression (Datta et al., 1993), was included as a positive control. PD 98059 attenuated c-Abl and MEK kinasemediated luciferase reporter activity in a dose-dependent manner, culminating in a 50% loss in reporter activity (Figure 3a). To confirm these results, we expressed a dominant-negative mutant of MEKK-1 (MEKK-1 DA) in the presence of c-Abl expression. In agreement with our hypothesis, dominant-negative expression of MEKK-1 diminished c-Abl and MEKK1 induction of EGR1 luciferase reporters in a dose-dependent manner, resulting in almost 70% loss in luciferase reporter activity (Figure 3b). To validate the results obtained from luciferase assays, EGR1 protein levels resulting from c-Abl kinase activity were assessed in the presence or absence of PD 98059 by induction of stable c-Abl-expressing cells. In both cases, addition of PD 98059 severely attenuated c-Abl-induced EGR1 expression without affecting c-Abl kinase activity (Figure 3c). These results indicate that c-Abl-induced EGR1 expression is mediated by activation of the MEK/ERK-signaling pathway. Oncogene


8088

a RLU

15 10 5 0

+

EGR1 – 600 c-Abl KA MEKK-1 PD 95089

RLU

b

+ +

+ +

+ +

+

+

+

+

+

+

30 25 20 15 10 5 0

EGR1 – 600 c-Abl KA MEKK-1

+ + + + +

+ + + + +

+ + + + + + + +

MEKK-1 DA

c Tet PD 95089

KA + + +

EV + + +

1 2 3 4

5 6 7 8 EGR1 Flag c-Abl

P-Tyr

Actin Figure 3 c-Abl activates the MEK/ERK signaling pathway to induce EGR1 expression. (a) 293T cells were transfected with indicated constructs and treated with increasing amounts of the MEK inhibitor PD 95089. At 6 h after treatment, cells were harvested and relative luciferase activity was assessed. (b) 293T cells were transfected with c-Abl-KA alone or in the presence of increasing amounts of dominant-negative MEKK-1 (MEKK-1 DA). Cells were harvested and luciferase activity assayed 24 h after transfection. MEKK-1 was included as a positive control in both (a) and (b). (c) Stable 293 cells were induced to express c-Abl-KA or empty vector (EV) for 24 h in the presence or absence of PD 95089 and blotted with the indicated antibodies

Apoptosis induced by c-Abl is partially mediated by EGR1 Given the fact that both c-Abl and EGR1 have been implicated in stress-induced apoptosis, c-Abl-induced EGR1 expression led us to test the possibility that c-Abl triggers apoptosis through upregulation of EGR1 expression. To accomplish this, WT and EGR1 null cells were treated with hydrogen peroxide in the presence or absence of STI-571 and assayed for apoptosis. In each Oncogene

cell type, hydrogen peroxide treatment increased apoptotic death, although EGR1 null cells appear less sensitive to hydrogen peroxide treatment (Figure 4a), which is consistent with the apoptotic function of EGR1. Significantly, in WT fibroblasts, inhibition of c-Abl kinase by STI-571 treatment abrogated hydrogen peroxide-mediated apoptosis, whereas in EGR1 null fibroblasts, STI-571 treatment had no significant effect (Figure 4a). To complement these results, c-Abl-KA or -KR was introduced into EGR1-null and WT MEFs via retroviral-mediated transfer. Apoptosis in these cells was assessed by FACS analysis and by visualizing fragmented nuclei. Consistent with the result derived from hydrogen peroxide treatment, activation of c-Abl is associated with significant apoptosis in WT cells, but not in EGR1-null cells (Figure 4b). Additionally, WT cells expressing Flag-tagged c-Abl KA had a greater fraction of apoptotic nuclei compared to all other cell lines (Figure 4c). To further substantiate the role for EGR1 in c-Ablmediated apoptosis, we employed RNA-interference (RNAi) to knockdown EGR1 expression in U20S cells that inducibly expressed c-Abl-KA. Cells in which RNAi was targeted against EGR1 message (U2OSKA 5.3) displayed almost no EGR1 expression, whereas cells stably expressing empty vector (U2OSKA GFP) showed elevated EGR1 levels in the presence of c-Abl induction (Figure 5a). Correspondingly, RNAi-treated cells exhibited reduced rates of apoptosis as compared to empty vector controls after 48 h of c-Abl induction (Figure 5b). Together, these results support that the apoptotic effects conferred by c-Abl kinase are mediated in part by EGR1.

Discussion c-Abl induces EGR1 expression via activation of the MEK/ERK signaling pathway We established a system in which c-Abl-KA or -KR is stably expressed in cells and placed under the control of an inducible promoter. When induced, these cells express c-Abl approximately fourfold over endogenous levels while showing a dramatic increase in tyrosine kinase activity. Using this system, we conducted microarray analysis and identified the EGR1 transcription factor to be induced as a result of c-Abl kinase activity. We then confirmed this observation at the message and protein level. Our initial microarray results showed EGR1 induction between two and three-fold, whereas Q-RT-PCR revealed a more dramatic induction in EGR1 transcript levels as a result of c-Abl activity. Discrepancies between these two values were anticipated, since microarrays lack the dynamic range of more quantitative methods and absolute differences between array data and validation results have been widely reported (Chuaqui et al., 2002). Western analysis suggests that EGR1 protein levels as a result of c-Abl kinase activity range between the values obtained from our array and Q-RT-PCR. Additional experiments using the proteosome inhibitor MG132 suggest that EGR1 is rapidly degraded (data not shown).


8089

a Apoptotic Index

a

U20SKA GFP 5.3 + +

2 Tet

EGR1

1.5

Flag c-Abl

1

Actin 0.5 EGR1 -/-

+

H2O2 STI-571

+ +

+

+ +

1.6

b + +

Apoptotic Index

WT

b 0.8

KA

EGR1 -/-

WT

B

B

c

1024

1000/10000

100

0

0

G: AD 10000/10000

PHT5 IHT

1024

0

B PHT5 IHT

1024

0

0

PHT5 IHT

1024

Apoptotic Index

1 0.8 0.6

200

G: AD 10000/10000

B

0

1024 PHT5 INT

B 0

200

-

G: AD 10000/10000

B 0

Counts

100

Counts

Counts

1000/10000

Counts

100

100

KR

5.3

+

0

+ 200

G: AD 10000/10000

Counts

KR

G: AD 10000/10000

Counts

KA

-

Counts

200

PHT5 IHT

GFP

U20SKA Tet

0.4

c-Abl

0

1 0.8

0.6

0.2

0

1.2

Counts

Apoptotic Index

1

1.4

B

0

1024 PHT5 INT

0

B

0

1024 PHT5 INT

0

0

1024 PHT5 INT

Figure 5 Abrogation of EGR1 expression by RNAi attenuates c-Abl-mediated apoptosis. (a) Retroviral-mediated RNA-interference of EGR1 expression in U20S cells stably expressing c-Abl KA. Cells were induced to express c-Abl-KA for 48 h by tetracycline treatment and blotted with indicated antibodies. (b) FACS analysis of cells in (a) after induction of c-Abl KA for 48 h. FACS analysis was performed as previously described except that apoptosis rates were indexed to noninduced cells expressing the empty RNAi vector. A representative output of the FACS analysis is shown below. Results are from three independent experiments

0.4 0.2

0 c-Abl

KA

KR WT

KA

KR EGR1 -/-

Figure 4 c-Abl utilizes EGR1 to induce apoptosis in response to oxidative stress. (a) WT or EGR1 null cells were serum starved, pretreated with STI-571 for 6 h where indicated, and treated with 500 mM H2O2 for 24 h. Apoptosis rates were assayed by FACS analysis by measuring the percentage of cells in sub-G1 cell cycle, and rates were indexed to non-treated cells. (b) Retroviral expression of c-Abl-KA or -KR was performed in WT or EGR1 null MEFs and apoptosis was assayed by FACS analysis. A representative result of the FACS analysis is shown below the graph. Relative apoptosis rates were indexed to WT MEFs expressing c-Abl-KA. (c) EGR1-WT cells expressing c-Abl KA exhibit greater rates of apoptotic nuclei. Apoptosis rates from cells in (b) were assayed by visualizing fragmented or condensed nuclei. Approximately 500 cells were counted and the percentage of cells containing apoptotic nuclei was calculated. The rates of apoptosis were normalized to WT MEFs expressing c-Abl-KA. All results are from three independent experiments performed in duplicate

Thus, the EGR1 protein levels we observed may reflect dynamics between c-Abl-dependent synthesis and the innate, rapid proteolysis of this protein.

Promoter-mapping experiments revealed that the three distal SREs in the EGR1 promoter mediate c-Abl induction. These SREs are bound by the transcription factor SRF, acting in concert with ternary complex factors Elk1, Sap1 or Sap2 to induce EGR1 expression (Marais et al., 1993). Elk1 is phosphorylated by the c-Jun N-terminal protein kinase (JNK) and extracellular signal-regulated protein kinase (ERK), leading to enhanced DNA-binding activity, ternary complex formation, and SRE-mediated transcription (Kaufmann et al., 2001). Others have shown that c-Abl activates MEK kinase 1 (MEKK-1), an upstream effector of the SEKl-SAPK pathway, in response to DNA damage (Kharbanda et al., 2000), thus providing the potential link between c-Abl and induction of EGR1. Previous studies investigating EGR1 transcriptional induction revealed the importance of SREs in EGR1 response to stress. EGR1 activation in response to ionizing radiation, reactive oxygen species, hypoxia, hyperoxia, nitric oxide, cisplatin, and PMA have all been observed to depend on the CArG DNA elements in the EGR1 promoter (Christy and Nathans, 1989; Datta et al., 1992; Datta et al., 1993; Cibelli et al., 2002; Park Oncogene


8090

et al., 2002; Jones and Agani, 2003). However, the upstream signaling pathways activated by these stimuli seem to differ. Protein kinase C activation seems to be necessary for EGR1 induction in response to ionizing radiation, while EGR1 induction in response to oxidative stress seems to depend on ERK signaling (Hallahan et al., 1991; Jones and Agani, 2003). In either case, it seems that these different signaling cascades converge to promote Elk1-mediated transcription and induction of EGR1 via SRE binding. We showed that c-Abl-mediated induction of EGR1 expression involves activation of the MEK/ERK-signaling pathway. Treatment of the MEK inhibitor PD 98059 attenuated c-Abl-induced expression of EGR1 as measured by luciferase reporter activity and Western analysis. Although PD 98059 treatment resulted in only partial abrogation of c-Abl-dependent EGR1 expression, this is probably because the chemical binds only inactive MEKK-1 and thus cannot inhibit MEKK-1 already activated by c-Abl (Alessi et al., 1995). Another possibility is that c-Abl induction of EGR1 is mediated through activation of an additional MEK kinase, as these pathways are known to share overlapping functions (Houslay and Kolch, 2000; Stork and Schmitt, 2002). Regardless, treatment with a dominant-negative form of MEKK-1 almost completely blocked c-Ablmediated induction of EGR1 expression, consistent with previous data indicating that c-Abl activates this signaling pathway to induce EGR1 expression.

radiation. Since c-Abl can activate the SEK-SAPK and MEK kinase cascades, activation of c-Abl may lie at the crux of EGR1 induction and work to activate multiple signaling pathways depending on the type of stress. Using a variety of approaches, we have shown that EGR1 induction resulting from c-Abl kinase activity results in apoptosis. STI-571 treatment of WT fibroblasts effectively blocked hydrogen peroxide-induced apoptosis, consistent with the results from Sun et al. (2000). However, in EGR1 null cells, STI treatment had little effect, indicating that c-Abl utilizes EGR1 expression to exert its apoptotic effects in response to oxidative stress. It is not surprising that EGR1 null cells are less sensitive to hydrogen peroxide treatment, since EGR1 is required for p53 transactivation (Krones-Herzig et al., 2003). Similarly, normal fibroblasts expressing c-Abl-KA undergo increased rates of apoptosis as compared to EGR1null fibroblasts. Although the baseline rates of apoptosis in fibroblasts expressing EGR1 are slightly elevated as compared to EGR1 null cells, there is a significant induction of apoptosis when c-Abl-KA is expressed. Knockdown of EGR1 expression by RNAi in the presence of c-Abl induction reduces apoptosis rates by approximately 50% as compared to cells with intact EGR1 expression, lending further support that EGR1 is crucial for c-Abl-mediated apoptosis. Together, our data have revealed a novel mechanism of EGR1 stress response and c-Abl-mediated apoptosis that could be relevant to the development and differentiation of various cell types.

EGR1 induction in response to stress is mediated by c-Abl Previous studies investigating stress induction of EGR1 have concentrated on mapping the promoter elements in the EGR1 promoter and subsequently identifying the signaling pathways responsible for EGR1 induction. Until now, no potential factors upstream of canonicalsignaling pathways have been identified and associated with EGR1 stress induction. We show that c-Abl is necessary for EGR1 induction in response to oxidative stress. Why does c-Abl trigger a mitogen-activated pathway, such as MEK/ERK, to induce EGR1 expression in response to stress? One argument simply states that MEK/ERK signaling is only one of many mechanisms that EGR1 has adopted to cope with stress, and that oxidative stress response may include signaling factors that overlap with mitogen-activated functions. Components of the cytoskeleton, such as actin, play an integral role in the response to oxidative stress while simultaneously functioning in growth signaling. c-Abl has been shown to associate with these factors, and it is possible that these interactions are important for c-Abl activation in response to oxidative stress. It will be interesting to investigate whether c-Abl is necessary for EGR1 induction in response to other forms of stress, such as ionizing radiation, nitric oxide, or cisplatin treatment, since the same promoter elements that respond to these factors also mitigate c-Abl induction of EGR1. It is not known how c-Abl is activated by oxidative stress, but it has been observed that ATM phosphorylates and activates c-Abl in response to ionizing Oncogene

Materials and methods Cells, cell culture, and treatments Human 293, U20S, inducible lines, and murine embryonic fibroblasts (MEFs) were grown in DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U penicillin per ml, 100 mg streptomycin per ml, and 2 mM L-glutamine. Inducible 293 and U2OS cells were treated with 1 mg/ml tetracycline for indicated time points to induce c-Abl expression. Prior to hydrogen peroxide treatment, cells were maintained in 0.5% FBS for 24 h to reduce EGR1 and catalase levels, and hydrogen peroxide was added to the medium at the indicated concentrations for 3 h to induce EGR1 expression. Cells were treated 8–10 h with the MAPK inhibitor PD 95809 (Sigma) 24 h after c-Abl expression. Cells were treated with a final concentration of 1 mM STI-571 for 6 h prior to hydrogen peroxide treatment. EGR1 WT and null cells were developed and provided by Dan Mercola (Krones-Herzig et al., 2003). Microarray and quantitative-real-time-PCR (Q-RT-PCR) analysis of inducible c-Abl expression cells Total RNA was isolated from approximately 2 107 cells at indicated time points using Trizol/chloroform extraction as directed by manufacturer’s instructions (Invitrogen). In all, 3 mg of mRNA was labeled with Cy3-dUTP and Cy5-dUTP (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and hybridized to the microarray chip (NCI ROSP 8K Human Array), containing 7680 human cDNA clones. After hybridization, the slides were washed in 2 SCC with 0.1% SDS, 1 SCC, 0.2 SCC, and 0.05 SCC and spin dried. Hybridized


8091 arrays were scanned at 10 mm resolution on a GenePix 4000A scanner (Axon Instruments, Inc., Foster City, CA, USA). The resulting images were analysed by GenePix Pro 3.0 software (Axon Instruments, Inc.). For Q-RT-PCR, total RNA isolated from Trizol/chloroform extraction was subjected to DNAse treatment (DNAfree; Ambion) and purified using the RNeasy Mini Kit (Qiagen). Primer sequence combinations for amplifying EGR1, b-actin, and GADPH targets will be provided upon request. Quantitative PCR amplifications were performed using the QuantiTect SYBR Green PCR Kit (Qiagen) and the Opticon Monitor Thermocycler (MJ Research, Inc.). EGR1 expression levels were calculated relative to that of the b-actin and GADPH housekeeping genes.

Reporter assays EGR1 promoter constructs fused to a luciferase reporter were transfected alone (0.25 mg) or with c-Abl-KA, c-Abl-KR, or MEKK-1 (0.5 mg). A maximum of 2 mg MEKK-1DA was transfected in dominant-negative experiments. Cells were harvested 24 h later or treated as indicated after 24 h. Luciferase assays were conducted using the Dual-Luciferase Assay System (Promega) and analysed with the LUMAT LB 9507 variable injector luminometer (EG & G Berthold). Reactions were performed in duplicate and reported data are the result of three independent experiments. Apoptotic assays and immunofluorescence

Establishment of 293 and U2OS inducible lines, retroviral infection, and RNA interference of EGR1 expression Phoenix amphoteric cells were transfected with pCDNA4 Flag c-Abl KA, KR, and EV (7 mg), pMSCV-FLAG c-Abl KA and KR (7 mg), pCG-gag-pol (5 mg), and pCG-VSVG (1 mg) (Dr R Mulligan, Harvard Medical School) by the calcium-phosphate method as described previously (Gu et al., 2000). For establishment of inducible cell lines, stable transfectants were chosen according to the manufacturer’s instructions (Invitrogen). Retroviral supernatant was harvested 48 h after transfection. EGR1-null and WT cells were infected by incubation with retroviral supernatants and polybrene (4 mg/ml) for 24 h followed by selection in puromycin (1.0 mg/ml EGR1-WT; 1.5 mg/ml EGR1-null)-containing medium for 5 days. Inducible 293 cells expressing Flag c-Abl-KA were targeted for EGR1 RNAi using the pSuper method (Oligoengine). The EGR1 RNAi sequence corresponded to nucleotides 917–933 downstream of the ATG start codon. Stable U20S KA cells expressing EGR1 RNAi were established using the previously described retroviral expression protocol except that the medium was replaced with 500 mg/ml neomycin to select cells. Transient transfection, Western analysis, and assessment of c-Abl kinase activation All cells were seeded 24 h before transfection. 293T cells were transfected in the presence of 0.5 ml of calcium phosphate coprecipitate containing DNA. After 24 h at 371C, medium was replaced and cells were either treated as indicated or incubated for another 24 h prior to harvest. Cells were rinsed with PBS and harvested in RIPA buffer (50 mM Tris, pH 7.4; 150 mM NaCl; 0.5% NP-40; 0.5% Brij-96; 0.1% SDS; 1 mM DTT; 10 mg of leupeptin and aprotinin per ml). Soluble proteins were separated by 10 or 12.5% SDS–PAGE, and transferred to nitrocellulose filters. After blocking with 5% milk in PBST (PBS in 0.05% Tween 20), the filters were incubated with the indicated antibodies overnight at 41C and analysed by chemiluminescence. Antibodies used in Westerns were a-EGR1 (#4152 Cell Signaling), a-actin (Sigma), a-Abl (K2 – Santa Cruz), a-P-tyr (4G10 – Sigma), and a-Flag (M5 – Sigma). To assess c-Abl activation resulting from hydrogen peroxide treatment, we affinity purified endogenous c-Abl using GST-conjugated Crk and blotted with antiphosphotyrosine antibodies to detect c-Abl autophosphorylation.

Cells were seeded at a density of 105 cells/60 mm dish prior to immunofluorescence staining. Cells were washed with PBS, fixed, and permeablized with 0.2% Triton X-100. Mouse antiFlag IgG (Sigma) was used to detect Flag c-Abl expressing cells, while rabbit anti-EGR1 IgG (sc-110, Santa Cruz) was used to differentiate EGR1-null and WT cells. Mouse Texas Red IgG secondary antibody conjugated was used to detect Flag proteins, whereas Rabbit IgG FITC secondary antibody was used to detect EGR1 protein. All secondary antibodies included DAPI stain to visualize nuclei. Cells were visualized using fluorescence microscopy. Apoptosis rates in these cell lines were measured by visualizing approximately 500 DAPIstained nuclei and calculating the percentage of cells that exhibited fragmented or condensed nuclei. Apoptosis rates for each cell type were indexed to EGR1-WT cells expressing c-Abl-KA. Results are the average of three independent experiments. Fluorescent-activated cell-sorting (FACS) analysis was used to assess apoptosis in EGR1-WT or null fibroblasts expressing c-Abl-KA or c-Abl KR, U2OS cells stably transfected with EGR1 RNAi or empty vector, and WT and EGR1 null cells treated with ST-571 and hydrogen peroxide. Cells undergoing FACS analysis were trypsinized, harvested, rinsed, and fixed in 70%EtOH prior to staining with 50 mg/ml propidium iodide containing 100 mg/ml Rnase A. A total of 10 000 cells were counted in each analysis. Experiments were performed in triplicate and the percentage of cells in the sub-G1 phase of the cell cycle was denoted as the apoptotic rate for each cell type. Apoptosis rates for each experiment were indexed as indicated. Acknowledgements We thank Dr Dan Mercola (Sidney Kimmel Research Center, San Diego, CA, USA) for providing EGR1 null and WT MEFs, Dr Kathleen Sakamoto (Department of Pediatrics, Division of Hematology/Oncology, School of Medicine, Los Angeles, CA, USA) for providing the EGR1 promoter, and Dr Anthony Koleske (Dept. Molecular Biophysics and Biochemistry, Yale University, New Haven, CT, USA) for providing Abl null MEFs. We also thank Claire Bailey (Center for Genomics Research, Harvard University, Cambridge, MA, USA) for microarray and Q-RT-PCR analysis, and Colleen Dionne for manuscript preparation and submission. This work was supported by NIH grant R29 CA85679.

References Alessi DR, Cuenda A, Cohen P, Dudley DT and Saltiel AR. (1995). J. Biol. Chem., 270, 27489–27494. Baskaran R, Wood LD, Whitaker LL, Canman CE, Morgan SE, Xu Y, Barlow C, Baltimore D, Wynshaw-Boris

A, Kastan MB and Wang JY. (1997). Nature, 387, 516–519. Brasher BB and Van Etten RA. (2000). J. Biol. Chem, 275, 35631–35637. Oncogene


8092 Christy B and Nathans D. (1989). Mol. Cell Biol., 9, 4889–4895. Chuaqui RF, Bonner RF, Best CJ, Gillespie JW, Flaig MJ, Hewitt SM, Phillips JL, Krizman DB, Tangrea MA, Ahram M, Linehan WM, Knezevic V and Emmert-Buck MR. (2002). Nat. Genet., 32 (Suppl), 509–514. Cibelli G, Policastro V, Rossler OG and Thiel G. (2002). J. Neurosci. Res., 67, 450–460. Daniel R, Cai Y, Wong PM and Chung SW. (1995). Oncogene, 10, 1607–1614. Datta R, Rubin E, Sukhatme V, Qureshi S, Hallahan D, Weichselbaum RR and Kufe DW. (1992). Proc. Natl. Acad. Sci. USA, 89, 10149–10153. Datta R, Taneja N, Sukhatme VP, Qureshi SA, Weichselbaum R and Kufe DW. (1993). Proc. Natl. Acad. Sci. USA, 90, 2419–2422. Goga A, Liu X, Hambuch TM, Senechal K, Major E, Berk AJ, Witte ON and Sawyers CL. (1995). Oncogene, 11, 791–799. Gu J, Chen D, Rosenblum J, Rubin RM and Yuan ZM. (2000). Mol. Cell Biol., 20, 1243–1253. Hallahan DE, Sukhatme VP, Sherman ML, Virudachalam S, Kufe D and Weichselbaum RR. (1991). Proc. Natl. Acad. Sci. USA, 88, 2156–2160. Houslay MD and Kolch W. (2000). Mol. Pharmacol., 58, 659–668. Jones N and Agani FH. (2003). J. Cell Physiol., 196, 326–333. Kaufmann K, Bach K and Thiel G. (2001). Biol. Chem., 382, 1077–1081. Khachigian LM, Lindner V, Williams AJ and Collins T. (1996). Science, 271, 1427–1431. Khachigian LM, Williams AJ and Collins T. (1995). J. Biol. Chem., 270, 27679–27686. Kharbanda S, Pandey P, Jin S, Inoue S, Bharti A, Yuan ZM, Weichselbaum R, Weaver D and Kufe D. (1997). Nature, 386, 732–735.

Oncogene

Kharbanda S, Pandey P, Yamauchi T, Kumar S, Kaneki M, Kumar V, Bharti A, Yuan ZM, Ghanem L, Rana A, Weichselbaum R, Johnson G and Kufe D. (2000). Mol. Cell Biol., 20, 4979–4989. Krones-Herzig E, Adamson E and Mercola D. (2003). Proc. Natl. Acad. Sci. USA, 100, 3233–3238. Marais R, Wynne J and Treisman R. (1993). Cell, 73, 381–393. Nair P, Muthukkumar S, Sells SF, Han SS, Sukhatme VP and Rangnekar VM. (1997). J. Biol. Chem., 272, 20131–20138. Park JO, Lopez CA, Gupta VK, Brown CK, Mauceri HJ, Darga TE, Manan A, Hellman S, Posner MC, Kufe DW and Weichselbaum RR. (2002). J. Clin. Invest., 110, 403–410. Plattner R, Kadlec L, DeMali KA, Kazlauskas A and Pendergast AM. (1999). Genes Dev., 13, 2400–2411. Sakamoto KM, Bardeleben C, Yates KE, Raines MA, Golde DW and Gasson JC. (1991). Oncogene, 6, 867–871. Shaw PE, Schroter H and Nordheim A. (1989). Cell, 56, 563–572. Stork PJ and Schmitt JM. (2002). Trends Cell Biol., 12, 258–266. Sun X, Majumder P, Shioya H, Wu F, Kumar S, Weichselbaum R, Kharbanda S and Kufe D. (2000). J. Biol. Chem., 275, 17237–17240. Svaren J, Ehrig T, Abdulkadir SA, Ehrengruber MU, Watson MA and Milbrandt J. (2000). J. Biol. Chem., 275, 38524–38531. Tanis KQ, Veach D, Duewel HS, Bornmann WG and Koelske AJ. (2003). Mol. Cell Biol., 23, 3884–3896. Thiel G and Cibelli G. (2002). J. Cell Physiol., 193, 287–292. Virolle T, Adamson ED, Baron V, Birle D, Mercola D, Mustelin T and de Belle I. (2001). Nat. Cell Biol., 3, 1124–1128. Yuan ZM, Huang Y, Ishiko T, Kharbanda S, Weichselbaum R and Kufe D. (1997). Proc. Natl. Acad. Sci. USA, 94, 1437–1440.