Increased proteasome subunit protein expression and ... - Nature

Oncogene (2009) 28, 3983–3996

& 2009 Macmillan Publishers Limited All rights reserved 0950-9232/09 $32.00 www.nature.com/onc

ORIGINAL ARTICLE

Increased proteasome subunit protein expression and proteasome activity in colon cancer relate to an enhanced activation of nuclear factor E2-related factor 2 (Nrf2) A Arlt1, I Bauer1, C Schafmayer2, J Tepel2, S Sebens Muërko¨ster1, M Brosch1, C Ro¨der2, H Kalthoff2, J Hampe1, MP Moyer3, UR Fo¨lsch1 and H Scha¨fer1 1 Department of Internal Medicine, University-Hospital Schleswig-Holstein (UKSH)-Campus Kiel, Kiel, Germany; 2Department of General and Thoracic Surgery, UKSH-Campus Kiel, Kiel, Germany and 3INCELL Corporation, LLC, Cimarron Path, San Antonio, TX, USA

An elevated proteasome activity contributes to tumorigenesis, particularly by providing cancer cells with antiapoptotic protection and efficient clearance from irregular proteins. Still, the underlying mechanisms are poorly known. In this study, we report that in colon cancer patients, higher proteasome activity was detected in tumoral tissue compared with surrounding normal tissue, and also that increased levels of proteasomal subunit proteins, such as S5a/PSMD4 and a-5/PSMA5, could be detected. Colon tumors showed higher nuclear levels of nuclear factor E2-related factor 2 (Nrf2), a transcription factor supposed to be involved in the control of proteasomal subunit protein expression. The induction or overexpression of Nrf2 led to stronger S5a and a-5 expression in the human colon cancer cell lines, Colo320 and Lovo, as well as in NCM460 colonocytes along with higher proteasome activity. The small interfering RNA (siRNA)-mediated Nrf2 knockdown decreased S5a and a-5 expression and reduced proteasome activity. Additionally, Nrf2-dependent S5a and a-5 expression conferred protection from tumor necrosis factor-related apoptosisinducing ligand (TRAIL)-induced apoptosis, an effect preceded by an increased nuclear factor (NF)-jB activation and higher expression of antiapoptotic NF-jB target genes. These findings point to an important role of Nrf2 in the gain of proteasome activity, thereby contributing to colorectal carcinogenesis. Nrf2 may therefore serve as a potential target in anticancer therapy. Oncogene (2009) 28, 3983–3996; doi:10.1038/onc.2009.264; published online 7 September 2009 Keywords: regulated proteolysis; ubiquitination; gene transcription; oxidative stress; tumorigenesis

Correspondence: Professor H Scha¨fer, Department of Internal Medicine, University-Hospital Schleswig-Holstein (UKSH)-Campus Kiel, Schittenhelmstr. 12, Kiel D-24105, Germany. E-mail: [email protected] Received 3 April 2009; revised 1 July 2009; accepted 25 July 2009; published online 7 September 2009

Introduction Regulated proteolysis by the ubiquitin–proteasome pathway (UPP) has a key role in both intracellular homeostatis and in the execution of a great number of signal-transduction pathways (Hershko and Ciechanover, 1998; Schubert et al., 2000; Conaway et al., 2002; Glickman and Ciechanover, 2002). An intact and properly regulated UPP is therefore an essential prerequisite for cell growth, cell viability and many other biological processes (McBride et al., 2003; Goldberg, 2007; Gudmundsdottir et al., 2007; Yi and Ehlers, 2007; van Tijn et al., 2008). Consequently, deregulation and dysfunctions in the UPP account for severe pathological conditions, such as immune defects, neurodegenerative disorders and cancer (Orlowski and Dees, 2003; Chen and Madura, 2005; Devoy et al., 2005; Mani and Gelmann, 2005; Sun, 2006; Dahlmann, 2007). The degradation of proteins once recruited to the UPP is executed by the 26S proteasome (26S-Pr), a large multiprotein complex representing the major site of nonlysosomal proteolysis in the eukaryotic cell (Coux et al., 1996; Baumeister et al., 1998; London et al., 2004). The 26S-Pr is located both in the cytosol and the nucleus and consists of a dimeric, barrel-shaped catalytical subunit— the 20S proteasome or 20S catalytical particle (20SCP)—flanked on each site by two cap-like regulatory subunits—termed 19S regulatory particles (19S-RP) (Coux et al., 1996; Baumeister et al., 1998). The assembly and disassembly of 26S-Pr, its maturation and the expression of more than 30 different proteasomal subunit proteins are tightly regulated processes (London et al., 2004; Murata et al., 2009) controlled by subunit self-processing, by the actions of maturation proteins and certain transcription factors and also by the cellular energy status (Schmidtke et al., 1996; Xie and Varshavsky, 2001; Leggett et al., 2002; Smith et al., 2005; Liu et al., 2006). With respect to cancer, more and more evidence accumulates that, beside amplifications in signal-transduction pathways upstream to the UPP, the proteasome activity itself is also elevated in tumor cells (Ren et al., 2000; Wyke et al., 2004; Chen and Madura, 2005). This finding is in line with the fact that tumor cells are much

Nrf2 and proteasomal activation in colorectal cancer A Arlt et al

3984

more sensitive to proteasome inhibitors than normal cells (Adams, 2004), thus indicating that tumor cells particularly depend on high proteasome activity. Such a condition probably reflects the need of tumor cells to get rid of many immature and irregular proteins that would be otherwise harmful to the cell (Schubert et al., 2000; Goldberg, 2003), as well as the need for maintaining certain signal-transduction pathways—that is, the IkB/ NF-kB pathway—that protect the tumor cell from apoptosis (Wang et al., 1999; Karin et al., 2002; Zhang et al., 2004). Accordingly, proteasome inhibition has meanwhile become a valuable and powerful tool in the treatment of certain types of cancer (Almond and Cohen, 2003; Goy et al., 2003; Adams, 2004); yet, other tumor entities turned out to be rather resistant to proteasome inhibitors. Nevertheless, recent studies (Chen et al., 2009; Pitts et al., 2009) showed that proteasome inhibition might also be effective for the treatment of colon cancer. Still, the mechanisms leading to the elevated proteasome activity in tumors are largely unknown, but recent data point to an involvement of the overexpression of proteasomal subunit proteins. For example, tumor samples from breast cancer patients show elevated expression level of certain proteasomal subunits when compared with normal tissue, along with higher proteasome activity (Chen and Madura, 2005). In melanoma, overexpression of the proteasomal subunit, S10, has been reported, emerging from gene amplification due to a chromosomal translocation (Ren et al., 2000) and accounting for an increased proteasome activity. In a recent study, the 20S-CP protein, PSMB7, was identified to be overexpressed in colon cancer tissue (Rho et al., 2008). In yeast, it has been shown that the de novo synthesis of proteasomal proteins is under tight control of the transcription factor, Rpn4, which is itself a UPPcontrolled protein and thus represent a feedback regulator of the 26S-Pr gene expression (Xie and Varshavsky, 2001; London et al., 2004). However, the mammalian homolog of Rpn4 has not been identified so far. Instead, other transcription factors seem to be involved, in particular, the antioxidative transcription factor, nuclear factor E2-related factor 2 (Nrf2) (Kwak et al., 2003; Hu et al., 2006; Kwak and Kensler, 2006; Arlt et al., 2007), which is also a target protein of 26SPr. Nrf2 is mainly involved in cellular phase II responses to oxidative stress and is under the control of the Keap1

regulatory protein in a redox-dependent manner. In response to oxidative stress (reactive oxygen species (ROS) and glutathione (GSH) depletion) and electrophilic insults (that is, by antioxidants such as tert-butylhydroquinone (tBHQ) or sulforaphane) (Ishii et al., 2000; Huang et al., 2002; Wakabayashi et al., 2004; Cho et al., 2005), the E3-ligase adaptor, Keap1, dissociates from Nrf2, thereby releasing it from polyubiquitination and proteasomal degradation (Itoh et al., 1999). Nrf2 then enters the nucleus and induces antioxidant target genes and also a number of proteasomal proteins (Hu et al., 2006). Along with others, we have shown that many of the proteasomal subunit genes possess an antioxidant response element in their promoter, which is under the control of Nrf2 (Kwak et al., 2003; Kwak and Kensler, 2006; Arlt et al., 2007). In tumors, a link between Nrf2 signalling and proteasomal dysfunction has not been established yet, but it is conceivable that stress-adapted tumor cells constitutively activate Nrf2. In this study we screened colon tumor samples for proteasome activity and proteasome subunit protein expression on the one hand and for Nrf2 activity on the other hand. Particular attention was paid to the 19S-RP protein, S5a/PSMD4, and the 20S-CP protein, a-5/ PSMA5, expected to be involved in the exaggerated proteasome activity in cancer (Murata et al., 2009). Using human colon cancer cell lines and a human colonic epithelial cell line, the observed link between Nrf2 activation and proteasome activity was further verified in vitro.

Results Increased proteasome activity in colon cancer tissue Cell lysates of frozen tissue specimens from 21 colon cancer patients were analysed for proteasome activity using the N-succinyl-L-leucyl-L-leucyl-L-valyl-L-tyrosyl7-amido-4-methylcumarin substrate assay. As shown by the proteasome-specific (MG132-inhibitable) release of the fluorescent cleavage product, AMC, tumor tissues showed significantly (Po0.005) increased levels (mean±s.d.: 9700±4120, specific OD/mg protein) of proteasome activity as compared with normal colon tissue (mean±s.d.: 5250±2080, specific OD/mg protein) from the same patient (Figure 1a). In 17 out of 21 tissue samples (>80%), higher proteasome activity was

Figure 1 Screening of matched tumoral and normal tissue from colon cancer patients for proteasome activity, S5a and a-5 protein level and nuclear factor E2-related factor 2 (Nrf2) levels. (a) Tissue homogenates were adjusted for equal protein concentration and submitted to a proteasome assay using the N-succinyl-L-leucyl-L-leucyl-L-valyl-L-tyrosyl-7-amido-4-methylcumarin (Suc-LLVY-AMC) substrate. Proteasome specificity was determined by the difference of AMC formation in the absence and presence of 5 mg/ml MG132 and the results are expressed as specific AMC formation (fluorescence units)/mg protein. In the upper panel, results of individual samples from 21 patients (p1–p21) are shown along with tumor characteristics indicated below (according to Sobin and Wittekind, 2002): T, tumor status; N, node status; M, metastatic status; G, grading. The lower panel shows the mean±s.d. (n ¼ 21), *Po0.005. (b, c) Whole cell lysates or nuclear extracts were prepared from frozen tissue samples and adjusted for equal protein concentration. (b) S5a and a-5 protein level in whole cell lysates were analysed by western blotting using actin as control. (c) Nrf2 protein level in nuclear extracts from 21 patients were analysed by western blotting (upper panel) using Hsp90 as control, and antioxidant response element (ARE)-binding activity in nuclear extracts from 11 patients was determined using gel shift assay (lower panel); c 1,2 ¼ control: Colo320± hemagglutinin (HA)-Nrf2. Oncogene


3985

detected in the tumoral tissue than in the surrounding normal tissue. This included nine samples, in which the proteasome-specific AMC formation of the tumor was

more than twofold higher in comparison with the normal tissue, and seven samples in which AMC formation was 1.5- to 2-fold higher.

Normal 16000

Tumor

14000 12000 10000 8000 6000 4000 2000 0 p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14 p15 p16 p17 p18 p19 p20 p21 T: N: M: G:

3 2 1 2

3 2 0 2

3 1 1 3

4 0 1 2

2 1 0 2

3 2 1 2

3 2 1 3

4 2 1 3

3 1 0 2

3 1 0 2

3 0 1 3

2 0 0 3

3 0 0 2

4 2 1 2

2 0 1 3

2 1 1 2

3 2 1 3

2 1 1 2

3 2 1 3

3 2 0 2

3 2 1 3

16000

Specific AMC formation (FU/mg protein)

Specific AMC formation (FU/mg protein)

18000

12000 10000 8000 6000 4000 2000 0

P2 N T

P1 N T

P3 N T

P4 P5 N T N T

*

14000

P6 N T

normal P7 P8 N T N T

tumor P9 N T

P10 P11 P12 N T N T N T S5a cell lysate

α5 actin

P14 N T

P13 N T

P15 P16 N T N T

P17 P18 N T N T

P19 P20 N T N T

P21 N T S5a cell lysate

α5 actin

P1 N T

P2 N T

P3 N T

P4 P5 N T N T

P6 N T

P7 P8 P9 P10 P11 P12 N T N T N T N T N T N T Hsp90 Nrf2

P13 N T

P14 N T

P15 P16 N T N T

P17 P18 N T N T

nuclear extract

P19 P20 P21 N T N T N T Hsp90 Nrf2

c 1

P2 2 N T

nuclear extract

P3 P4 P5 P6 P7 P10 P12 P16 P19 P20 P21 N T N T N T N T N T N T N T N T N T N T N T < ARE-EMSA

Oncogene


3986

Proteasomal subunit proteins are overexpressed in colon cancer tissue The western blot analysis of total cell lysates prepared from tumor tissues revealed that S5a and a-5— representing 19S and 20S subunit proteins, respectively—are expressed at higher levels in the tumor when compared with normal tissue. With the exception of three cases, all colon cancer samples showed such an increase in S5a and a-5 protein levels (Figure 1b). Elevated nuclear localization of nuclear factor E2-related factor 2 (Nrf2) in colon cancer tissue Addressing the question of whether the increased proteasome activity and expression of S5a and a-5 relate to higher activation of Nrf2, nuclear extracts from colon cancer tissues were submitted to Nrf2 western blot analysis. As shown in Figure 1c (upper panel), nuclear levels of Nrf2 were elevated in tumoral tissue as compared with normal tissue. In >50% (n ¼ 11) of all cases, the level of Nrf2 was strongly increased. In six cases the level of Nrf2 was slightly higher and in two cases Nrf2 levels were almost equal, respectively. Two cases showed a slightly lower Nrf2 level in the tumoral tissue. In addition, 12 matched tissue samples of sufficient amount of nuclear protein from both normal and tumor tissue were also screened for nuclear Nrf2 activity using gel shift assay (Figure 1c, lower panel). Using a radiolabelled consensus antioxidant response element oligonucleotide as probe, elevated levels of Nrf2–DNA complexes were detected in 10 tumor samples as compared with normal tissue. This pattern (10 out of 12) strongly matches the corresponding results from western blot analysis, thus confirming higher Nrf2 activity in these cancer cells. Collectively, the majority of the tumor samples showing higher proteasome activity and S5a and a-5 expression fairly matches the increased status of nuclear Nrf2 (Supplementary Figure 1). There was a considerable correlation (r2 ¼ 0.70 and 0.63, respectively) between proteasome activity (see Figure 1a) and S5a or Nrf2 protein level (upon densitometric evaluation of western blot analyses, Figures 1b and c) and also between NRF2 and S5a level (r2 ¼ 0.77). Similar correlation was seen between a-5 level and proteasome activity and Nrf2, respectively (data not shown). Immunostaining for S5a and nuclear factor E2-related factor 2 (Nrf2) in tumor tissue sections In order to confirm that the higher expression levels of Nrf2 and S5a in the tumoral tissue derive from the adenocarcinoma cells, rather than from leukocyte infiltrations, immunohistochemistry analysis of frozen tissue sections was performed. As shown in Figure 2, strong staining of tumors for S5a and Nrf2 was detectable, and those tumoral areas immunopositive for S5a and Nrf2 were also positive for cytokeratin-20, indicating the epithelial origin of the bulk of S5a and Nrf2 immunoreactivity. Moreover, S5a and Nrf2 staining in tumor cells was detectable in both the nucleus and Oncogene

the cytoplasm (Figure 2b). This staining pattern indicates that the rise of Nrf2 activity and the expression of proteasomal proteins mainly occur in the transformed colonic epithelium. Induction and overexpression of nuclear factor E2-related factor 2 (Nrf2) enhances proteasome activity in colon cancer cell lines and colon epithelial cells In order to strengthen the hypothesis that Nrf2 activation relates to proteasome activity, the colorectal cancer cell lines, Colo320 and LoVo, were treated with the well-established Nrf2 inducer, tBHQ (50 mM). At 4 and 8 h after tBHQ administration, the nuclear levels of Nrf2 and antioxidant response element/protein complexes increased in both cell lines (Figure 3a), indicating Nrf2 activation. A similar effect was seen when stimulating the established human colonic epithelial cell line, NCM460, with tBHQ. Again, the Nrf2 activity was elevated within 4–8 h of tBHQ treatment, as shown by western blot and gel shift assay (Figure 3a). After 24 h of tBHQ treatment, a significant increase in AMC formation was observed in all three cell lines (Figure 3b), indicating elevated proteasome activities. The proteasome activity was raised by 72, 67 and 130% in LoVo, Colo320, and NCM460 cells, respectively, when compared with untreated cells. Moreover, transient transfection of hemagglutinin (HA)-tagged Nrf2 complementary DNA into LoVo, Colo320 and NCM460 cells— as verified by western blot (Figure 4b)—also led to an increase in proteasome activity. As shown by proteasome assay (Figure 3c), the proteasome activity was raised by 76, 51 and 70% in Nrf2 overexpressing LoVo, Colo320 and NCM460 cells, respectively, as compared with the corresponding mock transfected controls. The Nrf2 induction and overexpression in colon cancer cell lines and colon epithelial cells increases the expression of S5a and a-5 To confirm an induction of S5a and a-5 expression upon Nrf2 activation, real-time PCR and western blot analysis (Figure 4a, upper and lower panels, respectively) were performed. These analyses revealed an increase in S5a and a-5 mRNA and protein level in LoVo, Colo320 or NCM460 cells, respectively, after tBHQ treatment (50 mM) for 4, 8 and 24 h. At 48 h after transient transfection of HA-Nrf2, increased protein level of S5a and a-5 in LoVo, Colo320 and NCM460 cells could also be noted, when compared with the corresponding mock transfected controls (Figure 4b). Silencing of nuclear factor E2-related factor 2 (Nrf2) expression in colon cancer cell lines and colon epithelial cells reduces proteasome activity To further validate the involvement of Nrf2 in the induction of proteasome activity, LoVo and Colo320, as well as NCM460 cells, were treated with Nrf2 small interfering RNA (siRNA) or control siRNA for 48 h, followed by administration of tBHQ. As shown by western blot analysis, the tBHQ-dependent increase in S5a and a-5 protein level was attenuated in Nrf2 siRNA-treated LoVo,


3987

S5 a

Isotype ctrl.

Nrf2

2nd Ab

Isotype ctrl.

CK20

S5 a

Isotype ctrl.

CK20

Nrf2

2nd Ab

Isotype ctrl.

Figure 2 Immunohistochemical detection of S5a and nuclear factor E2-related factor 2 (Nrf2) in colorectal tumor tissue. Cryosections of tumoral tissue from colon cancer patients were submitted to immunostainings (a) for S5a and Nrf2 as well as for cytokeratin 20 (CK20) for discrimination of cells of epithelial origin ( 40 magnification). Controls were performed either by incubation with the appropriate isotype-matched IgG control (Isotype ctrl.) as first antibody or by omitting the first antibody (2nd Ab). (b) The same panel of immunostainings of tumoral tisssue for S5a, Nrf2 and CK20 was compared at higher magnification ( 250), and controls were chosen as described above.

Colo320 and NCM460 cells, compared with cells treated with control siRNA (Figure 4c). Proteasome assays revealed (Figure 4d) that, compared with cells subjected to control siRNA treatment, proteasome activity in the

absence of tBHQ decreased in Nrf2 siRNA-treated LoVo, Colo320 and NCM460 cells by 20, 45 and 23%, respectively. In the presence of tBHQ, proteasome activity decreased by 34, 46, and 50% in Nrf2 siRNA-treated Oncogene


3988 LoVo

Colo320

NCM460 Hsp90 n.e. Nrf2

ARE-shift 4

0

0 4 8 50 µM tBHQ

LoVo

3 Specific AMC formation (FU/cell number)

8

2.5

4

*

8 (h)

NCM460

Colo320 *

2

0

*

*

*

*

1.5 1 0.5 0 tBHQ: 0

Specific AMC formation (FU/cell number)

2.4 2.0

8

24

0

8

24

0

8

24 (h)

HA-Nrf2 mock

1.6 1.2 0.8 0.4 0 LoVo

Colo320

NCM460

Figure 3 Nuclear factor E2-related factor 2 (Nrf2) confers increased proteasome activity to colon cancer cells and normal colonic epithelial cells. LoVo, Colo320 and NCM460 cells were incubated with 50 mM tert-butyl-hydroquinone (tBHQ) for the indicated periods. (a) Nuclear extracts (n.e.) were submitted to western blotting (upper panel) for the detection of Nrf2 and Hsp90 as control, or to gel shift assay (lower panel) using an antioxidant response element (ARE) consensus oligonucleotide as probe. (b) The tBHQ-treated cells were analysed by proteasome assays; mean±s.d. from four experiments, *Po0.05. (c) Lovo, Colo320 and NCN460 cells were transiently transfected with hemagglutinin (HA)-Nrf2 or the empty vector (mock) and analysed by proteasome assays 48 h later; mean±s.d., n ¼ 4.

LoVo, Colo320 and NCM460 cells, respectively, compared with control siRNA-treated cells. The silencing of S5a and a-5 expression abrogates tBHQ-induced proteasome activity in colon cancer cell lines and colon epithelial cells To assess whether the tBHQ-dependent and Nrf2mediated induction of proteasome activity involves Oncogene

S5a and a-5 expression, both proteins were silenced by siRNA (Figure 5a). As shown by proteasome assay (Figure 5b), LoVo, Colo320 and NCM460 cells show decreased levels of tBHQ-induced proteasome activity when treated with S5a and a-5 siRNA. Although this decrease was moderate upon silencing of S5a in LoVo and Colo320 cells (by 20 and 17% of the tBHQ-induced AMC formation in cells treated with control siRNA, respectively), this effect was more pronounced in S5a siRNAtreated NCM460 cells (by 35% of the tBHQ-induced AMC formation in control cells) and in all cell lines when treated with a-5 siRNA (by 49, 42 and 41% of the tBHQinduced AMC formation in control cells, respectively). Nrf2-induced S5a and a-5 expression in colon epithelial cells confers protection from tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced apoptosis To address whether Nrf2-induced S5a and a-5 expression confers antiapoptotic protection, NCM460 cells were treated with 50 mM tBHQ for 24 h, followed by apoptosis induction using 0.1 mg/ml TRAIL. Caspase-3/7 assays (Figure 6a) and annexinV/propidium iodide assays (Figure 6b) revealed that tBHQ pretreatment significantly suppressed caspase-3/7 activation (Figure 6a, left panel) and the increase in annexinV staining (Figure 6b, left panel) after TRAIL treatment. Similarly, NCM460 cells showed lower TRAIL-induced caspase-3/7 activity and annexinV staining when overexpressing Nrf2 (Figures 6a and b, right panels), and TRAIL-induced cleavage of caspase-3, as detected by western blotting, was less pronounced in NCM460 cells if treated with tBHQ or overexpressing Nrf2 (Figure 6c). Conversely, tBHQ-treated NCM460 cells showed higher caspase-3/7 activity and annexinV staining, both in the absence and presence of TRAIL, when subjected to Nrf2, a-5 or S5a siRNA transfection (Figures 6d and e). Expectedly, the proteasome inhibitor, Mg132, greatly increased the sensitivity toward TRAIL in NCM460 cells (Supplementary Figure 3). Moreover, siRNA-mediated knockdown of Nrf2, a-5 or S5a enhanced TRAILinduced cleavage of caspase-3 in tBHQ-treated NCM460 cells (Figure 6f). Together, these data point to a protecting effect of Nrf2 against TRAIL-induced apoptosis, which depends on S5a and a-5 expression. TRAIL resistance through Nrf2-induced S5a and a-5 expression in colon epithelial cells involves an increased NF-kB activation and expression of antiapoptotic target genes. From the observation that Nrf2-induced proteasome activity contributes to the resistance against TRAIL and as TRAIL has been shown inducing NF-kB through the canonical pathway (Trauzold et al., 2001; Ehrhardt et al., 2003), we looked whether an altered NF-kB response relates to this effect. Interestingly, NCM460 cells showed a marked turnover of IkBa and an increased NF-kB activity in response to TRAIL treatment when overexpressing Nrf2 (Figure 7a). Moreover, expression of the antiapoptotic NF-kB target gene, cIAP1, was induced by TRAIL on mRNA and protein level more strongly in NCM460 cells transfected with Nrf2 (Figures 7b and c) than in mock transfected cells.


3989 α5

S5a LoVo

Relative mRNA content (normalized to actin)

3.5

Colo320

NCM460

3 2.5 2 1.5 1 0.5 0

4

0

8

24

LoVo

4 8 24 0 50 μM tBHQ (h)

4

0

8

24

NCM460

Colo320

Hsp90 S5a

w.c.l.

5 0

8

24

8

0

24

0

8

24 (h)

50 μM tBHQ LoVo

LoVo

NCM460

Colo320

Colo320

NCM460 Hsp90

Hsp90

n.e. Nrf2

S5a

Hsp90

5

S5a HA mock HA-Nrf2

mock HA-Nrf2 mock

HA-Nrf2

w.c.l

5 siRNA: c tBHQ: -

c Nrf2 Nrf2 + +

c -

c Nrf2 Nrf2 + +

c -

c Nrf2 Nrf2 + +

2.5

Specific AMC formation (FU / cell number)

control siRNA

Nrf2 siRNA

2

1.5 * 1

* *

*

0.5

0 tBHQ:

-

+ LoVo

+

-

+ Colo320

+

-

+ NCM460

+

Figure 4 Nuclear factor E2-related factor 2 (Nrf2)-dependent expression of S5a and a-5 in colon cancer cells and normal colonic epithelial cells. LoVo, Colo320 and NCM460 cells were incubated with 50 mM tert-butyl-hydroquinone (tBHQ) for the indicated periods. (a) Complementary DNA (cDNA) from total RNA was submitted to real-time PCR analysis (upper panel), and the data representing the mean of two independent experiments, or whole cell lysates (w.c.l.) were submitted to western blotting (lower panel) for the detection of S5a, a-5 and Hsp90 as control. (b) LoVo, Colo320 and NCM460 cells were transiently transfected with hemagglutinin (HA)-Nrf2 or the empty vector (mock) and submitted to western blotting for the detection of HA, S5a, a-5 and Hsp90 as control. (c, d) LoVo, Colo320 and NCM460 cells were treated with control or Nrf2 small interfering RNA (siRNA) for 48 h followed by incubation with 50 mM tBHQ. (c) Nuclear extracts (n.e.) or whole cell lysates (w.c.l.) from unstimulated or tBHQ-stimulated (8 h) cells were submitted to western blotting for the detection of Nrf2, S5a, a-5 and Hsp90 as control. (d) siRNA-treated cells were stimulated with tBHQ for 24 h and analysed using proteasome assays; mean±s.d., n ¼ 4; *Po0.05 versus control siRNA.

This effect was weaker with other NF-kB targets, such as Mcl1 and cIAP2. Upon tBHQ preincubation, the TRAIL induced NF-kB activation in NCM460 cells (Figure 7d)

and the increased expression of cIAP1—and to a lesser extent of cIAP2 and Mcl1—were diminished by Nrf2-, S5a- and a-5-siRNA treatment (Figures 7e and f). Oncogene


3990

Discussion In cancer, the UPP is affected in multiple ways, leading to an increased turnover rate of signalling molecules involved in the regulation of cell growth and apoptosis. As a hallmark of this condition, the activity of the proteasome itself is increased in tumor cells, an event Colo320

LoVo

NCM460 < Hsp90

< S5a siRNA:

c

S5a

c

S5a

c

S5a < Hsp90

< α5 siRNA:

α5

c

c

α5

c

α5

125 Colo320

Specific AMC formation (% of control)

LoVo 100

NCM460

* *

75

*

*

*

50 25

0 siRNA:

co S5a α5

co S5a α5

co S5a α5

Figure 5 Proteasome activity in colon cancer cells and normal colonic epithelial cells is impaired upon knockdown of S5a and a-5 expression. LoVo, Colo320 and NCM460 cells were treated with control (c or co), S5a or a-5 small interfering RNA (siRNA) for 48 h followed by incubation with 50 mM tert-butyl-hydroquinone (tBHQ) for 8 or 24 h. (a) Whole cell extracts from cells treated with tBHQ for 8 h were submitted to western blotting (upper panel, representative result) for the detection of S5a and a-5 or Hsp90 as control. (b) The siRNA-treated cells were stimulated with tBHQ for 24 h and analysed using proteasome assays; mean±s.d., n ¼ 4; *Po0.05 versus control siRNA.

that may result from overexpression of certain protein components of the 26S-Pr (Ren et al., 2000; Chen and Madura, 2005; Rho et al., 2008). For the first time, this study shows a relation between an increased activation of Nrf2 in cancer and the overexpression of proteasomal subunit proteins—that is, S5a and a-5—along with an elevated proteasome activity. Our study emphasizes these two proteasomal proteins as being likely involved in the exaggerated proteasome activity in colon cancer. While a-5 is an integral part of the 20S-CP and is required for adenosine triphosphatase-mediated threading of the substrate into the 20S-CP, the 19S-RP protein S5a has a particular role in the assembly of the 26S-Pr as it connects the lid and base during 19S-RP biosynthesis (Murata et al., 2009). Moreover, s5a is responsible for the recruitment of ubiquitinated substrates to the 26S-Pr (Elsasser et al., 2004; Kang et al., 2007). Thus, the finding that S5a—similar to a-5—is indispensable, not only from the proteasome activity in NCM460 or Colo320 cells, but also from the antiapoptotic protection of these cells, underscores its functional importance for both the UPP and UPP-mediated survival. Accordingly, a previous study identified S5a as an early target in the execution of apoptotis (Sun et al., 2004). In addition to the analysis of tumor samples from colon cancer patients, cell culture-based experiments, using Colo320 and LoVo colorectal cancer cells and the colonic epithelial cell line NCM460, confirmed Nrf2dependent induction of both proteasomal gene expression and proteasome activity. After exposure of these cells to electrophilic stress by administration of tBHQ, Nrf2 is released from its repression by Keap1 (Itoh et al., 1999), followed by an elevated mRNA and protein level of S5a and a-5 as well as by an increase in proteasome activity. Thus, both electrophilic and oxidative stress—conditions quite common during tumorigenesis—lead to persistent Nrf2 activation and thereby to an exaggerated proteasome formation and activation. Under such stress conditions, that is, during inflammation, the expression of proteasome proteins might be enhanced in a Nrf2-dependent manner, accounting for increased proteasome activity and UPP-mediated survival and growth of colon cancer cells. In line with this notion, very recent studies revealed that proteasome inhibition or knockdown of certain proteasomal subunit proteins impair the growth and survival of colorectal cancer cells (Chen et al., 2009; Pitts et al., 2009).

Figure 6 Nuclear factor E2-related factor 2 (Nrf2)-induced S5a and a-5 expression in colon epithelial cells confers protection from tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced apoptosis. NCM460 cells (a–c) were left without tert-butylhydroquinone (tBHQ) or were treated with 50 mM tBHQ for 24 h, or were transfected with the empty vector or with the hemagglutinin (HA)-Nrf2 expression vector, followed by treatment with 0.1 mg/ml TRAIL for 16 h, or not. Apoptosis was determined by the caspase3/7 assay (a) or the annexinV assay (b) and data are expressed as mean±s.d. from four independent experiments; *Po0.05 versus untreated and mock transfected cells, respectively. (c) Caspase-3 cleavage in NCM460 cells was analysed by western blotting and representative results out of two independent experiments are shown. (d–f) Upon transfection with control-, Nrf2-, a-5- or S5a-small interfering RNA (siRNA) for 48 h and subsequent treatment with 50 mM tBHQ for 24 h, NCM460 cells were treated with 0.1 mg/ml TRAIL for 16 h, or not. Apoptosis was determined using the caspase-3/7 assay (d) or the annexinV assay (e) and data are expressed as mean±s.d. from four independent experiments; *Po0.05 versus control siRNA. (f) Caspase-3 cleavage in NCM460 cells was analysed by western blotting and representative results out of two independent experiments are shown. Oncogene


3991

The observation that Nrf2, proteasomal gene expression and proteasomal activity are inducible in nonmalignant colonic epithelial cells by electrophilic insults (such as by tBHQ treatment) indicates that this

condition might occur already in the intestinal epithelium during electrophilic and oxidative stress. As long as the Nrf2-dependent induction of proteasome activity does not confer growth advantages, the otherwise

NCM460 vehicle tBHQ

NCM460 9

vehicle tBHQ

mock HA-Nrf2

40

7 *

6 5 4

mock HA-Nrf2

35

8

% annexinV-pos cells

Casapse-3/-7 activity (n-fold)

10

*

3 2

30

*

25

*

20 15 10 5

1 0

0

vehicle

vehicle TRAIL vehicle TRAIL

TRAIL

NCM460 tBHQ mock

vehicle

vehicle

TRAIL

HA-Nrf2

Tubulin

pro-caspase 3 (35 kDa) cleaved Casp3 (17 / 20 kDa) TRAIL:

Casapse-3/-7 activity (n-fold)

9 8 7

+

-

+

NCM460 + tBHQ

+

*

5 4 3

40

*

* *

35 30 25 20 15

2

10

1

5

0

+

NCM460 +tBHQ

45 *

-

control siRNA Nrf2 siRNA S5a siRNA α5 siRNA

50

*

6

-

55

control siRNA Nrf2 siRNA S5a siRNA α5 siRNA

% annexinV-pos cells

10

-

0 vehicle

vehicle

TRAIL

TRAIL

NCM460 + tBHQ siRNA:

control

α5

s5a

Nrf2

Tubulin

pro-caspase 3 (35 kDa) cleaved Casp3 (17 / 20 kDa) TRAIL:

-

+

-

+

-

+

-

+ Oncogene


3992 NCM460 HA-Nrf2

mock

κBα Iκ 0.5

1

2 (h) NFκB

NCM460 / HA-Nrf2 Tubulin IκBα TRAIL: 0

0.5

1

3.5 3 2.5 2 1.5 1 0.5 0

2 (h)

cIAP1 TRAIL (1h) :

-

+

-

+

supershift:

-

-

-

-

+

cIAP2

Mcl1

+

p50 p65

NCM460 + 50 μM tBHQ

NCM460 mock

HA-Nrf2

mock

4

mRNA level (n-fold of basal)

Tubulin

TRAIL: 0

NCM460/ TRAIL (2h)

4.5

NCM460 / mock

siRNA: control

Nrf2

α5

s5a

HA-Nrf2

NCM460 + 50 μM tBHQ

Nrf2 >

Tubulin >

siRNA: control Nrf2

α5

s5a

S5a > cIAP1 >

α5 >

cIAP2 >

Tubulin >

Mcl-1 > TRAIL (4h):

TRAIL: 0

1

0

1

0

1

0

1 (h)

IκBα > -

4 mRNA level (n-fold of basal)

NFκB >

3.5 3

+

-

+

TRAIL: 0 0.5 1

0 0.5 1

0 0.5 1 0 0.5 1 (h)

NCM460 + tBHQ / TRAIL (2h) NCM460 + 50 μM tBHQ control siRNA Nrf2 siRNA

S5a siRNA α5 siRNA

siRNA:

control

Nrf2

α5

s5a

Tubulin > 2.5 cIAP1 >

2 1.5

cIAP2 >

1

Mcl1 >

0.5

TRAIL (4h):

-

+

-

+

-

+

-

+

0 cIAP1

cIAP2

Mcl1

Figure 7 Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) resistance through nuclear factor E2-related factor 2 (Nrf2)-induced S5a and a-5 expression in colon epithelial cells involves an increased nuclear factor (NF)-kB activation and expression of antiapoptotic target genes. NCM460 cells (a) were transfected with the empty vector or with the hemagglutinin (HA)-Nrf2 expression vector, followed by treatment with 0.1 mg/ml TRAIL for 1 h, or not. Cell lysates were analysed using western blotting for IkBa and tubulin (left panel) and nuclear extracts were submitted to NF-kB gel shift assay (right panel). The specificity of the NF-kB band was verified (arrow heads) by anti-p50 and anti-p65 supershifts. (b, c) Mock or HA-Nrf2 transfected NCM460 cells were left untreated or were treated with 0.1 mg/ml TRAIL, and the expression of cIAP1, cIAP2, and Mcl1 was analysed by (b) real-time PCR or by (c) western blotting using tubulin as control. (d) Upon transfection with control-, Nrf2-, a-5- or S5a-small interfering RNA (siRNA) for 40 h and subsequent treatment with 50 mM tert-butyl-hydroquinone (tBHQ) for 24 h, NCM460 cells were treated with 0.1 mg/ml TRAIL, or not. Cell lysates were analysed using western blotting for Nrf2, S5a, a-5, IkBa and tubulin (left panel) and nuclear extracts were submitted to NF-kB gel shift assay (right panel). (e, f) The siRNA-transfected and tBHQ-pretreated NCM460 cells were left untreated or were treated with 0.1 mg/ml TRAIL and the expression of cIAP1, cIAP2, and Mcl1 was analysed (e) by real-time PCR or (f) by western blotting using tubulin as control. In panels (b) and (e), data represent the mean±s.d., n ¼ 3.

protective effect of Nrf2 against potential mutagenic insults rather prevents tumorigenesis. By contrast, the Nrf2-dependent proteasome activation would favor Oncogene

tumor growth and survival if malignant transformation initiated (Hayes and McMahon, 2009). Recent reports revealed that Nrf2 is indeed involved in the protection of


3993

colon mucosa from dextran sulfate sodium-induced colitis in mice (Khor et al., 2006; Osburn et al., 2007) and, consequently, Nrf2 deficiency leads to a more severe inflammation and tissue damage, giving rise to cancer. Although these findings indicate an antitumor protective activity of Nrf2 that underlies new concepts of chemopreventive measures (Osburn and Kensler, 2008)—that is, by administration of sulphoraphane— the constitutive Nrf2 activation may be ‘highjacked’ by the tumor cells later on, thereby providing survival and growth advantages (Osburn and Kensler, 2008). In the latter condition, Nrf2-dependent proteasome activation might be essentially involved. This is supported by the observation that the overexpression or induction of Nrf2 in NCM460 cells confers protection against TRAIL-induced apoptosis, an effect not seen if S5a or a-5 expression was knocked down. Intriguingly, the protecting effect by Nrf2 involves an enhanced NF-kB activation by TRAIL, giving rise to a stronger expression of antiapoptotic target genes. Moreover, a close relation of enhanced Nrf2 activation and higher NF-kB activity was detectable in a line of tissue samples from colon cancer patients (Supplementary Figure 2). Thus, even though inhibitory interferences between Nrf2 and NF-kB have also been reported (Liu et al., 2008; Song et al., 2009), our findings point to a stimulatory interference reflecting a complex crosstalk of these transcription factors in carcinogenesis (Nair et al., 2008). Beside persisting oxidative and electrophilic stress in the tumoral environment resulting from inflammatory processes, an enhanced metabolism or the depletion of iron, certain mutations or altered DNA methylation patterns affecting the Keap1–Nrf2 pathway similarly contribute to the constitutive activation of Nrf2 (Singh et al., 2006; Nioi and Nguyen, 2007; Shibata et al., 2008a; Hayes and McMahon, 2009). In various tumor entities, Keap-1 mutations have been detected, accounting for Nrf2-dependent protection from anticancer drug toxicity and tumorigenesis (Nioi and Nguyen, 2007; Yamamoto et al., 2008; Shibata et al., 2008a). Even mutations in Nrf2 itself are under suspect (Hayes and McMahon, 2009; Shibata et al., 2008b). With respect to its important role in the control of the 26S-Pr, it now emerges that the constitutive activation of Nrf2 provides another growth advantage to tumor cells, in addition to the Nrf2-dependent resistance to anticancer drugs (Lau et al., 2008). Thus, the increased proteasomal activity resulting from Nrf2 activation might lead to profound protection from apoptosis, an accelerated cell growth and an increased removal of potentially harmful proteins. Deregulation of Nrf2 activation—that is, due to Keap1 deficiency—along with the induction of proteasomal gene expression may also compensate for the antitumor effects of proteasome inhibitors, thereby accounting for the failure of these agents in the treatment of certain types of cancer. In summary, our findings indicate that Nrf2 activity is elevated in colon cancer, accounting for the overexpression of proteasome subunit proteins and, thereby, for an increased proteasome activity. As the gain of

proteasome activity accompanies malignant transformation of tumor cells, Nrf2 in this way might contribute to tumorigenesis. Moreover, the role of Nrf2 in cancer prevention has to be critically considered, because the pharmacological induction of Nrf2 (that is, by antioxidants) might be advantageous (preventive) at an early state of tumorigenesis, whereas this would be deleterious at a later stage of tumorigenesis (Lau et al., 2008).

Materials and methods Cell lines and culture The Colo320 and LoVo cells (DSMZ, Braunschweig) were cultured in RPMI-1640 (PAA, Co¨lbe, Germany) containing 1% glutamine (Invitrogen/GIBCO, Karlsruhe, Germany) and 10% fetal calf serum (Biochrom, Berlin, Germany). Human NCM460 colonocytes (Moyer et al., 1996) were provided by INCELL Corp. (San Anatonio, TX, USA) and cultured in M3A medium (INCELL) containing 10% fetal calf serum. The cells were cultured at 37 1C in 5% CO2 and at 85% humidity. Tumor sample preparation Tissue samples were collected from colorectal cancer patients. The study was approved (A110/99) by the local ethics committee, and informed consent was obtained according to the Declaration of Helsinki. Frozen tissue specimens were homogenized (30 s, 15 000 r.p.m.) using the cryogenic ball mill MM400 (Retsch, Haan, Germany) and resuspended in the indicated buffers for N-succinyl-L-leucyl-L-leucyl-L-valylL-tyrosyl-7-amido-4-methylcumarin assay, western blot or gel shift assay (see below and in the Supplementary information). Fluorometric proteasome assay For the determination of proteasomal activity in living cells and tissue extracts, a fluorometric assay using the proteasome substrate, N-succinyl-L-leucyl-L-leucyl-L-valyl-L-tyrosyl-7-amido4-methylcumarin, was performed as described in the Supplementary information. Western blotting The preparation of samples, electrophoresis and immunoblotting is specified in the Supplementary information. Gel shift assay Gel shift assays were conducted as described (Arlt et al., 2003) using the g-32P-labeled consensus antioxidant response elementoligonucleotide 50 -CGCAGTCACAGTGACTCAGCAGAAT CTGAG-30 (Kepa and Ross, 2003), or a consensus NF-kBoligonucleotide (Promega, Mannheim, Germany). NF-kB supershifts were performed as described (Arlt et al., 2003). Immunohistochemistry Consecutive cryostat sections (6 mm) were mounted on uncovered glass slides, air-dried overnight at room temperature, fixed in chilled acetone (Merck, Darmstadt, Germany) for 10 min and air-dried again for 10 min. Then, slides were washed in phosphate buffered saline. To avoid nonspecific binding, sections were treated with 4% bovine serum albumine (Serva, Heidelberg, Germany) for 20 min, followed by incubation with the primary antibody (Abcam, Cambridge, UK). For detection of Nrf2, a monoclonal rabbit antibody, and for Oncogene


3994 detection of S5a and CK20, monoclonal mouse antibodies were used at 1:100 dilutions in 1% bovine serum albumine/ phosphate buffered saline. After primary antibody incubation (overnight, 4 1C), sections were washed three times in phosphate buffered saline and then treated with EnVision peroxidase conjugates (Dako, Hamburg, Germany) for 30 min. Afterward, sections were washed three times in phosphate buffered saline. Then, peroxidase substrate reaction was performed using the DAB peroxidase substrate kit (Vector Laboratories, Peterborough, UK) according to the manufacturer’s instructions. Afterward, sections were washed in water, counterstained in 50% haemalaun (Merck) and mounted with glycerol-gelatin. The same protocol was performed for negative controls, either omitting the first antibody or using an isotype-matched control antibody.

Caspase-3/7 and annexinV/propidium iodide assay Apoptosis induced by killer-TRAIL (Axxora/Alexis, Gru¨nberg, Germany) was determined by staining cells with annexinV/ propidium iodide (BD-Biosciences, Heidelberg, Germany) or by measurement of caspase-3/-7 activity (Promega), both according to the manufacturer’s instructions and as described (Sebens Muërko¨ster et al., 2008). All assays were performed in duplicates.

RNA preparation and real-time PCR Isolation of total RNA and reverse-transcription into singlestranded complementary cDNA was carried out as described (Arlt et al., 2007). The complementary DNA was subjected to real-time PCR (iCycler; Bio-Rad) using the SYBR-Green assay with gene-specific primers at a final concentration of 0.2 mM. The primer sequences and PCR conditions are provided in the Supplementary information.

Abbreviations

Complementary DNA transfection and small interfering (si) RNA treatment The cells were subjected to lipofection using Effectene (Qiagen, Hilden, Germany) and either 0.8 mg pMH-Nrf2-HA (kind gift from P Nioi, Schering-Plough Research Institute, NJ, USA) or the empty vector. After 16 h, medium was changed and culture continued for 8–24 h until further treatment. Gene silencing was performed using the HiperFect reagent (Qiagen) and 150 ng/well of either negative control siRNA (Qiagen) or Nrf2 (SI03246614, Qiagen), S5a (SI03019331, Qiagen) or a-5 siRNA (SI100043316, Qiagen). After 8 h, medium was changed and cells were cultured for 48 h until further treatment.

Statistics Data represent the mean±s.d. and were analysed by Student’s t-test or by Pearson correlation. A P-value o0.05 was considered as statistically significant.

LLVY-AMC, N-succinyl-L-leucyl-L-leucyl-L-valyl-L-tyrosyl7-amido-4-methylcumarin; Nrf2, nuclear factor E2-related factor-2; tBHQ, tert-butyl-hydroquinone; UPP, ubiquitin– proteasome pathway; 26S-Pr, 26S-proteasome; 20S-CP, 20S catalytical particle; 19S-RP, 19S regulatory particle.

Conflict of interest The authors declare no conflict of interest. Acknowledgements The authors thank Maike Gromann and Dagmar Leisner for technical support, and the German Research Society (DFG/SFB415-A13) and the German Cluster of Excellence ‘Inflammation at Interfaces’ for funding.

References Adams J. (2004). The development of proteasome inhibitors as anticancer drugs. Cancer Cell 5: 417–421. Almond JB, Cohen GM. (2003). The proteasome: a novel target for cancer chemotherapy. Leukemia 16: 433–443. Arlt A, Kruse ML, Breitenbroich M, Gehrz A, Koc B, Minkenberg J et al. (2003). The early response gene IEX-1 attenuates NFeˆB activation in 293 cells, a possible counterregulatory process leading to enhanced cell death. Oncogene 22: 3343–3350. Arlt A, Minkenberg J, Kruse ML, Grohmann F, Fo¨lsch UR, Scha¨fer H. (2007). Immediate early gene-X1 interferes with 26 S proteasome activity by attenuating expression of the 19 S proteasomal components S5a/Rpn10 and S1/Rpn2. Biochem J 402: 367–375. Baumeister W, Walz J, Zu¨hl F, Seemu¨ller E. (1998). The proteasome: paradigm of a self-compartmentalizing protease. Cell 92: 367–380. Chen L, Madura K. (2005). Increased proteasome activity, ubiquitinconjugating enzymes, and eEF1A translation factor detected in breast cancer tissue. Cancer Res 65: 5599–5606. Chen W, Hu XT, Shi QL, Zhang FB, He C. (2009). Knockdown of the novel proteasome subunit Adrm1 located on the 20q13 amplicon inhibits colorectal cancer cell migration, survival and tumorigenicity. Oncol Rep 21: 531–537. Cho HY, Reddy SP, Debiase A, Yamamoto M, Kleeberger SR. (2005). Gene expression profiling of NRF2-mediated protection against oxidative injury. Free Radic Biol Med 38: 325–343. Oncogene

Conaway RC, Brower CS, Conaway JW. (2002). Emerging roles of ubiquitin in transcription regulation. Science 296: 1254–1258. Coux O, Tanaka K, Goldberg AL. (1996). Structure and functions of the 20S and 26S proteasomes. Annu Rev Biochem 65: 801–847. Dahlmann B. (2007). Role of proteasomes in disease. BMC Biochem 8(Suppl 1): S3. Devoy A, Soane T, Welchman R, Mayer RJ. (2005). The ubiquitinproteasome system and cancer. Essays Biochem 41: 187–203. Ehrhardt H, Fulda S, Schmid I, Hiscott J, Debatin KM, Jeremias I. (2003). TRAIL induced survival and proliferation in cancer cells resistant towards TRAIL-induced apoptosis mediated by NF-kappaB. Oncogene 22: 3842–3852. Elsasser S, Chandler-Militello D, Mu¨ller B, Hanna J, Finley D. (2004). Rad23 and Rpn10 serve as alternative ubiquitin receptors for the proteasome. J Biol Chem 279: 26817–26822. Glickman MH, Ciechanover A. (2002). The ubiquitin-proteasome proteolytic pathway: destruction for the sake of construction. Physiol Rev 2: 373–428. Goldberg AL. (2007). Functions of the proteasome: from protein degradation and immune surveillance to cancer therapy. Biochem Soc Trans 35: 12–17. Goldberg AL. (2003). Protein degradation and protection against misfolded or damaged proteins. Nature 426: 895–899. Goy A, Hart S, Pro B, McLaughlin P, Younes A, Dang N et al. (2003). Report of a phase II study of proteasome inhibitor bortezomib


3995 (VELCADE) in patients with relapsed or refractory indolent or aggressive lymphomas. Blood 102: 180a. Gudmundsdottir K, Lord CJ, Ashworth A. (2007). The proteasome is involved in determining differential utilization of double-strand break repair pathways. Oncogene 26: 7601–7606. Hayes JD, McMahon M. (2009). NRF2 and KEAP1 mutations: permanent activation of an adaptive response in cancer. Trends Biochem Sci 34: 176–188. Hershko A, Ciechanover A. (1998). The ubiquitin system. Annu Rev Biochem 67: 425–479. Hu R, Xu C, Shen G, Jain MR, Khor TO, Gopalkrishnan A et al. (2006). Gene expression profiles induced by cancer chemopreventive isothiocyanate sulforaphane in the liver of C57BL/6J mice and C57BL/6J/Nrf2 (-/-) mice. Cancer Lett 243: 170–192. Huang HC, Nguyen T, Pickett CB. (2002). Phosphorylation of Nrf2 at Ser-40 by protein kinase C regulates antioxidant response elementmediated transcription. J Biol Chem 277: 42769–42774. Ishii T, Itoh K, Takahashi S, Sato H, Yanagawa T, Katoh Y et al. (2000). Transcription factor Nrf2 coordinately regulates a group of oxidative stress-inducible genes in macrophages. J Biol Chem 275: 16023–16029. Itoh K, Wakabayashi N, Katoh Y, Ishii T, Igarashi K, Engel JD et al. (1999). Keap1 represses nuclear activation of antioxidant responsive elements by Nrf2 through binding to the amino-terminal Neh2 domain. Genes Dev 13: 76–86. Karin M, Cao Y, Greten FR, Li ZW. (2002). NF-kB in cancer: From innocent bystander to major culprit. Nat Rev Cancer 2: 301–310. Kang Y, Chen X, Lary JW, Cole JL, Walters KJ. (2007). Defining how ubiquitin receptors hHR23a and S5a bind polyubiquitin. J Mol Biol 369: 168–176. Kepa JK, Ross D. (2003). Differential expression of the antioxidant response element within the hNQO1 promoter in NSCLC versus SCLC. Biochem Biophys Res Commun 311: 446–453. Khor TO, Huang MT, Kwon KH, Chan JY, Reddy BS, Kong AN. (2006). Nrf2-deficient mice have an increased susceptibility to dextran sulfate sodium-induced colitis. Cancer Res 66: 11580–11584. Kwak MK, Kensler TW. (2006). Induction of 26S proteasome subunit PSMB5 by the bifunctional inducer 3-methylcholanthrene through the Nrf2-ARE, but not the AhR/Arnt-XRE, pathway. Biochem Biophys Res Commun 345: 1350–1357. Kwak MK, Wakabayashi N, Greenlaw JL, Yamamoto M, Kensler TW. (2003). Antioxidants enhance mammalian proteasome expression through the Keap1-Nrf2 signalling pathway. Mol Cell Biol 23: 8786–8794. Lau A, Villeneuve NF, Sun Z, Wong PK, Zhang DD. (2008). Dual roles of Nrf2 in cancer. Pharmacol Res 58: 262–270. Leggett DS, Hanna J, Borodovsky A, Crosas B, Schmidt M, Baker RT et al. (2002). Multiple associated proteins regulate proteasome structure and function. Mol Cell 10: 495–507. Liu GH, Qu J, Shen X. (2008). NF-kappaB/p65 antagonizes Nrf2– ARE pathway by depriving CBP from Nrf2 and facilitating recruitment of HDAC3 to MafK. Biochim Biophys Acta 1783: 713–727. Liu CW, Li X, Thompson D, Wooding K, Chang TL, Tang Z et al. (2006). ATP binding and ATP hydrolysis play distinct roles in the function of 26S proteasome. Mol Cell 24: 39–50. London MK, Keck BI, Ramos PC, Dohmen RJ. (2004). Regulatory mechanisms controlling biogenesis of ubiquitin and the proteasome. FEBS Lett 567: 259–264. Mani A, Gelmann EP. (2005). The ubiquitin-proteasome pathway and its role in cancer. J Clin Oncol 23: 4776–4789. McBride WH, Iwamoto KS, Syljuasen R, Pervan M, Pajonk F. (2003). The role of the ubiquitin/proteasome system in cellular responses to radiation. Oncogene 22: 5755–5773. Moyer MP, Manzano LA, Merriman RL, Stauffer JS, Tanzer LR. (1996). NCM460, a normal human colon mucosal epithelial cell line. in vitro Cell Dev Biol Anim 32: 315–317. Murata S, Yashiroda H, Tanaka K. (2009). Molecular mechanisms of proteasome assembly. Nat Rev Mol Cell Biol 10: 104–115.

Nair S, Doh ST, Chan JY, Kong A-N, Cai L. (2008). Regulatory potential for concerted modulation of Nrf2- and Nfkb1-mediated gene expression in inflammation and Carcinogenesis. Br J Cancer 99: 2070–2082. Nioi P, Nguyen T. (2007). A mutation of Keap1 found in breast cancer impairs its ability to repress Nrf2 activity. Biochem Biophys Res Commun 362: 816–821. Orlowski RZ, Dees EC. (2003). The role of the ubiquitinationproteasome pathway in breast cancer: applying drugs that affect the ubiquitin-proteasome pathway to the therapy of breast cancer. Breast Cancer Res 5: 1–7. Osburn WO, Karim B, Dolan PM, Liu G, Yamamoto M, Huso DL et al. (2007). Increased colonic inflammatory injury and formation of aberrant crypt foci in Nrf2-deficient mice upon dextran sulfate treatment. Int J Cancer 121: 1883–1891. Osburn WO, Kensler TW. (2008). Nrf2 signaling: an adaptive response pathway for protection against environmental toxic insults. Mutat Res 659: 31–39. Pitts TM, Morrow M, Kaufman SA, Tentler JJ, Eckhardt SG. (2009). Vorinostat and bortezomib exert synergistic antiproliferative and proapoptotic effects in colon cancer cell models. Mol Cancer Ther 8: 342–349. Ren S, Smith MJ, Louro ID, McKie-Bell P, Bani MR, Wagner M et al. (2000). The p44S10 locus, encoding a subunit of the proteasome regulatory particle, is amplified during progression of cutaneous malignant melanoma. Oncogene 19: 1419–1427. Rho JH, Qin S, Wang JY, Roehrl MH. (2008). Proteomic expression analysis of surgical human colorectal cancer tissues: up-regulation of PSB7, PRDX1, and SRP9 and hypoxic adaptation in cancer. J Proteome Res 7: 2959–2972. Schmidtke G, Kraft R, Kostka S, Henklein P, Fro¨mmel C, Lo¨we J et al. (1996). Analysis of mammalian 20S proteasome biogenesis: the maturation of beta-subunits is an ordered two-step mechanism involving autocatalysis. EMBO J 15: 6887–6898. Schubert U, Antoń LC, Gibbs J, Norbury CC, Yewdell JW, Bennink JR. (2000). Rapid degradation of a large fraction of newly synthesized proteins by proteasomes. Nature 404: 770–774. Sebens Muërko¨ster S, Rausch AV, Isberner A, Minkenberg J, Blaszczuk E, Witt M et al. (2008). The apoptosis-inducing effect of gastrin on colorectal cancer cells relates to an increased IEX-1 expression mediating NF-kappa B inhibition. Oncogene 27: 1122–1134. Shibata T, Kokubu A, Gotoh M, Ojima H, Ohta T, Yamamoto M et al. (2008a). Genetic alteration of Keap1 confers constitutive Nrf2 activation and resistance to chemotherapy in gallbladder cancer. Gastroenterology 135: 1358–1368. Shibata T, Ohta T, Tong KI, Kokubu A, Odogawa R, Tsuta K et al. (2008b). Cancer related mutations in Nrf2 impair its recognition by Keap1-Cul3 E3 ligase and promote malignancy. Proc Natl Acad Sci USA 105: 13568–13573. Singh A, Misra V, Thimmulappa RK. (2006). Dysfunctional KEAP1-NRF2 interaction in non-small-cell lung cancer. Plos Med 3: e420. Smith DM, Kafri G, Cheng Y, Ng D, Walz T, Goldberg AL. (2005). ATP binding to PAN or the 26S ATPases causes association with the 20S proteasome, gate opening, and translocation of unfolded proteins. Mol Cell 20: 687–698. Sobin LH, Wittekind CH, International Union Against Cancer (UICC). (2002). TNM Classification on Malignant Tumours, 6th edn. Wiley-Liss Publications: New York. Song MY, Kim EK, Moon WS, Park JW, Kim HJ, So HS et al. (2009). Sulforaphane protects against cytokine- and streptozotocin-induced b-cell damage by suppressing the NF-kB pathway. Toxicol Appl Pharmacol 235: 57–67. Sun XM, Butterworth M, MacFarlane M, Dubiel W, Ciechanover A, Cohen GM. (2004). Caspase activation inhibits proteasome function during apoptosis. Mol Cell 14: 81–93. Sun Y. (2006). E3 ubiquitin ligases as cancer targets and biomarkers. Neoplasia 8: 645–654. Oncogene


3996 Trauzold A, Wermann H, Arlt A, Schu¨tze S, Scha¨fer H, Oestern S et al. (2001). CD95 and TRAIL receptor-mediated activation of protein kinase C and NF-kappaB contributes to apoptosis resistance in ductal pancreatic adenocarcinoma cells. Oncogene 20: 4258–4269. van Tijn P, Hol EM, van Leeuwen FW, Fischer DF. (2008). The neuronal ubiquitin-proteasome system: murine models and their neurological phenotype. Prog Neurobiol 85: 176–193. Wakabayashi N, Dinkova-Kostova AT, Holtzclaw WD, Kang MI, Kobayashi A, Yamamoto M et al. (2004). Protection against electrophile and oxidant stress by induction of the phase 2 response: fate of cysteines of the Keap1 sensor modified by inducers. Proc Natl Acad Sci USA 101: 2040–2045. Wang CY, Cusack Jr JC, Liu R, Baldwin Jr AS. (1999). Control of inducible chemoresistance: enhanced anti-tumor therapy through increased apoptosis by inhibition of NF-kappaB. Nat Med 5: 412–417.

Wyke SM, Russell ST, Tisdale MJ. (2004). Induction of proteasome expression in skeletal muscle is attenuated by inhibitors of NF-kappaB activation. Br J Cancer 91: 1742–1750. Xie Y, Varshavsky A. (2001). RPN4 is a ligand, substrate, and transcriptional regulator of the 26 S proteasome: a negative feedback circuit. Proc Natl Acad Sci USA 98: 3056–3061. Yamamoto T, Suzuki T, Kobayashi A, Wakabayashi J, Maher J, Motohashi H et al. (2008). Physiological significance of reactive cysteine residues of Keap1 in determining Nrf2 activity. Mol Cell Biol 28: 2758–2770. Yi JJ, Ehlers MD. (2007). Emerging roles for ubiquitin and protein degradation in neuronal function. Pharmacol Rev 59: 14–39. Zhang HG, Wang J, Yang X, Hsu HC, Mountz JD. (2004). Regulation of apoptosis proteins in cancer cells by ubiquitin. Oncogene 23: 2009–2015.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene