Discoidin Domain Receptor 1 Receptor Tyrosine Kinase Induces Cyclooxygenase-2 and Promotes Chemoresistance Through Nuclear Factor-KB Pathway Activation

Research Article

Discoidin Domain Receptor 1 Receptor Tyrosine Kinase Induces Cyclooxygenase-2 and Promotes Chemoresistance through Nuclear Factor-KB Pathway Activation

Sanjeev Das,1 Pat P. Ongusaha,1 Yoon Sun Yang,2 Jin-Mo Park,1 Stuart A. Aaronson,3 and Sam W. Lee1

1Cutaneous Biology Research Center, Massachusetts General Hospital and Harvard Medical School, Charlestown, Massachusetts; 2Curonix, Inc., Seoul, South Korea; and 3Department of Oncological Sciences, Mount Sinai School of Medicine, New York, New York

Abstract matrix and in controlling adhesion and cell motility (8).DDR1 Discoidin domain receptor 1 (DDR1) is a receptor tyrosine signaling is essential for cerebellar granule differentiation (9), the kinase activated by various types of collagens and is known to development of adaptive immune responses (10), arterial wound play a role in cell attachment, migration, survival, and repair (11), and mammary gland development (12).A number of proliferation. However, little is known about the molecular studies have reported overexpression of DDR1 in lung, esophageal, mechanism(s) underlying the role of DDR1 in cancer. We breast, ovary, and pediatric brain cancers, suggesting a role for report here that DDR1 induces cyclooxygenase-2 (Cox-2) DDR1 in tumor progression (2, 7, 13–17).Moreover, the elevated expression resulting in enhanced chemoresistance. Depletion expression of DDR1 in a number of fast-growing invasive tumors of DDR1-mediated Cox-2 induction using short hairpin RNA has suggested that this matrix-activated RTK may be involved in (shRNA) results in increased chemosensitivity. We also show the proliferation and stromal invasion of tumors. Accumulating evidence indicates that cyclooxygenase-2 (Cox-2) that DDR1 activates the nuclear factor-KB (NF-KB) pathway and blocking this activation by an IKB superrepressor mutant plays an important role in the promotion of various proliferative results in the ablation of DDR1-induced Cox-2, leading to diseases, including cancer.Cox-2 catalyzes the synthesis of enhanced chemosensitivity, indicating that DDR1-mediated prostaglandins (18), and Cox-2 up-regulation as well as increased Cox-2 induction is NF-KB dependent. We identify the levels of prostanoids have been reported in a high percentage of upstream activating kinases of the NF-KB pathway, IKKB tumors (19).Prostanoids affect multiple mechanisms that have and IKK;, as essential for DDR1-mediated NF-KB activation, been implicated in carcinogenesis (19–21).For example, prosta- whereas IKKA seems to be dispensable. Finally, shRNA- glandin E2 can stimulate cell proliferation and angiogenesis while mediated inhibition of DDR1 expression significantly en- inhibiting apoptosis and immune surveillance (21).Furthermore, hanced chemosensitivity to genotoxic drugs in breast cancer overexpression of Cox-2 is sufficient to induce tumorigenesis in cells. Thus, DDR1 signaling provides a novel target for transgenic mice (22).Epidemiologic studies have also shown that therapeutic intervention with the prosurvival/antiapoptotic the use of nonsteroidal anti-inflammatory drugs, prototypic machinery of tumor cells. (Cancer Res 2006; 66(16): 8123-30) inhibitors of cyclooxygenases, is associated with a reduced risk of several malignancies (23). There is as yet relatively little information concerning DDR1 Introduction downstream signaling pathways or functions.It was shown that Discoidin domain receptor 1 (DDR1) belongs to a small family of the DDR1 is a direct p53 transcriptional target and that it can receptor tyrosine kinases that contain an extracellular domain of function as a survival effector in wild-type (wt) p53-containing 160 amino acids exhibiting strong homology to the Dictyostelium cells exposed to genotoxic stress (24), suggesting that inhibition of discoideum protein discoidin 1 (1).DDR1 was isolated as a novel DDR1 function may provide a novel approach to selectively member of the receptor tyrosine kinase (RTK) family (2) and was enhance therapy of such tumors.We undertook our present studies also identified as a gene overexpressed in breast cancer cells (3). in an effort to identify downstream signaling targets of DDR1, in Most RTKs are activated by soluble proteins present in the blood particular those that may contribute to the DDR1-induced or other body fluids.The DDR1 RTK is unusual in that it is prosurvival pathways.Our results identify a signaling pathway activated by the extracellular matrix protein collagen (4, 5), but mediated by DDR1 involving activation of nuclear factor-nB (NF- independent of the major cellular collagen receptor integrins (6). nB) and its downstream effectors, including Cox-2 and XIAP.These Various types of collagen have been identified as ligands capable of findings establish the molecular mechanisms underlying DDR1 activating DDR1 (4, 5).The activation process is unique in that it is protection against genotoxic stress. slow but is sustained over long periods with no down-regulation by endocytosis or receptor degradation (4, 5).DDR1 is widely expressed in human and mouse epithelial cells (7).It has been Materials and Methods found to be involved in cell interactions with the extracellular Cell lines and culture conditions. Human breast cancer cell lines MDAMB435 and T47D were cultured in DMEM containing fetal bovine serum (FBS; Invitrogen, Carlsbad, CA), 100 units/mL penicillin, and 100 Ag/mL streptomycin at 37jC.All mouse embryonic fibroblasts (MEF), including Requests for reprints: Sam W.Lee, Cutaneous Biology Research Center, Cox-2À/À MEF, IKKaÀ/À MEF, IKKhÀ/À MEF, and IKKgÀ/À MEF Massachusetts General Hospital, Building 149, 13th Street, Charlestown, MA 02129. (25–27) were also maintained in DMEM containing 10% FBS.For drug Phone: 617-726-6691; Fax: 617-643-2334; E-mail: [email protected]. f I2006 American Association for Cancer Research. treatment, cells were grown to 50% confluency before exposure to the doi:10.1158/0008-5472.CAN-06-1215 DNA-damaging agents, mitomycin C at a concentrations of 2.5 to www.aacrjournals.org 8123 Cancer Res 2006; 66: (16). August 15, 2006

5.0 Ag/mL, etoposide at concentrations of 25 to 80 Amol/L for the indicated induction was specific for the activated DDR1.In fact, activated time.At the time of drug addition, collagen IV (10 Ag/mL, Sigma, St.Louis, induced DDR1 as measured by tyrosine phosphorylation, upon MO) was also added to the cells.Adenoviruses expressing DDR1 (Ad-DDR1) tet-removal, was detected only in the presence of collagen (Fig.1 A, or green fluorescent protein (Ad-GFP) were generated, amplified, and titrated right).These results support the conclusion that activation of the as previously reported (24).Cells were grown to f50% to 70% confluency overexpressed DDR1 is required for its induction of Cox-2 and infected with recombinant adenovirus at multiplicity of infection (MOI) of 10 to 20 for the indicated time.Ad-GFP was used as control. expression. Transfections and luciferase assays. Transfections were carried out DDR1-mediated chemoresistance is Cox-2 dependent. The using LipofectAMINE 2000 (Invitrogen) according to instructions from the DDR1 kinase is activated/phosphorylated in response to genotoxic manufacturer.In transient transfection experiments, plasmid DNA was kept stress in wt-p53-containing cells and confers increased resistance constant with empty vector.For luciferase assays, cells were harvested to apoptosis (24).Cox-2 is also up-regulated in wt-p53-containing 24 hours posttransfection and luciferase activity was determined using Dual cells and is protective against genotoxic stress (30).Thus, we next Luciferase Reporter assay system (Promega, Madison, WI).The difference tested whether DDR1-induced Cox-2 played an important role in in transfection efficiency across samples was normalized by cotransfecting DDR1-mediated cell survival upon genotoxic stress.For this pRL-TK, which expresses Renilla luciferase. purpose, we infected wt- and Cox-2À/À MEFs with Ad-DDR1 or Apoptosis assay. Apoptosis was measured using the Cell Death control virus Ad-GFP at a MOI of 20, and measured the extent of Detection ELISA kit (Roche, Indianapolis, IN).Briefly, the cells were lysed and the amount of nucleosomes in the cytoplasmic fraction of the cell cell death at different time points following exposure to different lysates was measured using antihistone antibody and anti-DNA antibody chemotherapeutic agents.DDR1 overexpression resulted in a linked to peroxidase in a sandwich ELISA-based protocol.For calculating marked decrease in apoptosis in the presence of etoposide at relative nucleosome content, the absorbance of all the samples was two different concentrations (40 and 80 Amol/L) in wt-MEFs, but normalized with respect to that of untreated/untransfected cells. had no significant effects in Cox-2-null MEFs under the same Short hairpin RNAs. Vectors expressing short hairpin RNA (shRNA) conditions (Fig.1 B).Similar results were observed with another against Cox-2 (5¶-GCTCAACACCGGAATTTTT-3¶), luciferase (5¶- therapeutic agent mitomycin C.These findings supported the ¶ ¶ ¶ GTGTTGGGCGCGTTATTTC-3 ), DDR1 (5 -AATCTCCTTCATCTCTGAT-3 ), concept that DDR1-mediated survival or chemoresistance is Cox-2 ¶ ¶ and GFP (5 -GCAAGCTGACCCTGAAGTC-3 ) were generated using the dependent.To test the role of DDR1 in chemoresistance in human pBabe-U6-shRNA plasmid.Cells were transfected with LipofectAMINE 2000 cancer cells, we induced DDR1 expression in tet-off–regulated (Invitrogen) according to the protocol of the manufacturer in the presence of shRNA. MDAMB435-DDR1 and measured the extent of apoptosis upon the Electrophoretic mobility shift assay.Electrophoretic mobility shift addition of genotoxic drugs, etoposide or mitomycin C.As shown assays (EMSA) were done as described previously (28, 29).Briefly, after the in Fig.1C , DDR1 induction by tet removal or Ad-DDR1 infection indicated drug treatment, nuclear extracts were made using a Nuclear resulted in f2- to 3-fold reduction in the extent of apoptosis, upon Extraction kit (Panomics, Redwood City, CA).Ten micrograms of nuclear the addition of etoposide or mitomycin C, compared with similarly extract were incubated with 32P-labeled NF-nB consensus oligonucleotide treated control cells grown in the presence of tet or infected with (Santa Cruz Biotechnology, Santa Cruz, CA) in gel shift binding buffer Ad-GFP (Fig.1 C).To further validate the effect of Cox-2 induction (Promega).The DNA-protein complexes were then resolved by electropho- on DDR1-mediated survival, we used shRNA to suppress the resis through Novex 6% DNA retardation gel (Invitrogen).Control oligos expression of endogenous Cox-2 in human breast cancer cells. representing the consensus oligonucleotide unlabeled or unlabeled NF-nB Transfection of Cox-2 shRNA resulted in the suppression of Cox-2 mutant oligonucleotide (Santa Cruz Biotechnology) were added in 500-fold excess. expression in Ad-DDR1–infected MDAMB435 cells, compared with Western blot analysis. Western blots were done as described cells transfected with control shRNA vector (pBabe-U6-shRNA- previously (30).The following antibodies were used: DDR1 (Santa Cruz luciferase; Fig.2 A, left).Next, we examined the effects of Cox-2 Biotechnology), h-actin (Sigma), Bcl-2 (BD Transduction Laboratories, depletion on DDR1-mediated chemoresistance in DDR1-over- San Diego, CA), XIAP (Cell Signaling Technology, Danvers, MA), Cox-2 expressing MDAMB435 cells.As shown in Fig.2 A (right), the Ad- (BD Transduction Laboratories), and InB (Santa Cruz Biotechnology). DDR1–infected cells were more resistant to etoposide treatment compared with control Ad-GFP–infected cells.However, depletion of Cox-2 induction via Cox-2 shRNA transfection resulted in a Results dramatic increase of genotoxic stress–induced apoptosis in Ad- DDR1 induces Cox-2 expression. We have previously shown DDR1–infected cells (compare columns 3 and 6).The antiapoptotic that both DDR1 and Cox-2 promote cell survival upon p53- protein, Bcl-2, was also increased in response to ectopic expression dependent genotoxic stress (24, 30).To identify the downstream of DDR1, but Bcl-2 induction was drastically reduced by Cox-2 signaling targets of DDR1, in particular those that may contribute depletion (Fig.2 A), consistent with several previous reports that to the DDR1-induced prosurvival pathway, we initially investigated Bcl-2 expression is regulated through a Cox-2–dependent mech- the effects of DDR1 overexpression on Cox-2 expression.For this anism (22, 31).To further validate these results, we examined the purpose, we infected wt and Cox-2À/À MEF with recombinant Ad- effect of Cox-2 inhibition in a human breast cancer cell line, T47D, DDR1 or control Ad-GFP and measured levels of Cox-2 expression. containing high basal expression of both DDR1 and Cox-2 proteins As shown in Fig.1 A, ectopic DDR1 expression in wt-MEF resulted (Fig.2B, left ).Endogenous Cox-2 expression was significantly in induction of Cox-2 expression in the presence of its ligand suppressed by transfection with Cox-2 shRNA but not by luciferase collagen, whereas the control virus, Ad-GFP, had no effect.We also control shRNA (Fig.2 B).Moreover, depletion of Cox-2 resulted in a generated tetracycline-regulated (tet-off) DDR1 expression system reproducible decrease in Bcl-2 expression in T47D cells and also in MDAMB435 cells (MDAMB435-DDR1) and tested whether DDR1 caused a marked increase in apoptosis in the presence of also induced Cox-2 in a different cell type.The results showed that etoposide, compared with results with shRNA-transfected or following tet withdrawal, Cox-2 expression increased concomitant- nontransfected T47D cells (Fig.2 B, right).Thus, the high basal ly with increasing DDR1 expression and only in the presence of expression level of DDR1 in T47D cells seems likely to be a collagen IV (Fig.1 A, middle), suggesting that DDR1-mediated Cox-2 causative factor in the chemoresistant phenotype of these tumor

Cancer Res 2006; 66: (16). August 15, 2006 8124 www.aacrjournals.org

Figure 1. DDR1 induces Cox2 expression, and DDR-mediated chemoresistance is abolished in Cox-2 null cells. A, wt and Cox-2À/À MEFs were infected with Ad-DDR1 or the control Ad-GFP at 20 MOI in the presence of collagen IV (10 Ag/mL). Forty-eight hours postinfection, cell extracts were made and Western blotting was carried out for the indicated proteins (left). Right, tetracycline-regulated DDR1 expression (MDAMB435-DDR1 tet-off) induces Cox-2 expression after tet removal for indicated times only in the presence of collagen IV, and cell extracts were made and Western blotting was carried out for DDR1, Cox-2, and h-actin. At the indicated time after tet removal, cells were lysed and DDR1 proteins were immunoprecipitated. The levels of tyrosine phosphorylation were determined by Western blot (WB) analysis using PY20 phosphotyrosine antibody. The quantity of DDR1 proteins in the immunoprecipitate (IP) was determined by Western blotting using DDR1 antibody. B, DDR1 expression renders chemoresistance in a Cox-2–dependent manner. Wt- and Cox-2À/À MEFs were infected with Ad-DDR1 or the control Ad-GFP at 20 MOI, and 24 hours later, cells were treated with the indicated concentrations of the chemotherapeutic drugs, etoposide and mitomycin C, for 24 and 48 hours. At the same time, collagen IV was also added to the cells. At the indicated time points, cell death detection ELISA was carried out to determine the apoptosis rate as described in Materials and Methods. DMSO was treated as a choice of solvent. Columns, mean of three independent experiments with each sample in triplicate; bars, SD. C, DDR1 expression increases cell survival in response to genotoxic stress in MDAMB435 cells. MDAMB435 cells were infected with Ad-DDR1 or the control virus Ad-GFP at 10 MOI in the presence of collagen IV. Twenty-four hours later, indicated concentrations of etoposide and mitomycin C were added. Twenty-four hours after drug treatment, cell death detection ELISAwas carried out. Asimilar experiment was also carried out with MDAMB4 35-DDR1 tet-off cells, which were treated with etoposide in the presence of collagen IV with or without tet. cells.Taken together, these results show that DDR1, through a DDR1 on NF-nB DNA-binding activity.The DNA-binding activity Cox-2–dependent mechanism, can play an important role in the of NF-nB was significantly increased in cells infected with Ad- chemoresistant phenotype of tumor cells. DDR1 to a level comparable with that observed with tumor DDR1 activates the NF-KB prosurvival pathway. It is well necrosis factor-a, activator of this pathway, whereas there was no established that Cox-2 is a downstream target of the NF-nB NF-nB binding activity in Ad-GFP–infected control cells (Fig.3 B). pathway (32, 33).To explore whether DDR1 activates NF- nB The increased level of NF-nB DNA-protein complexes observed in signaling, we first examined its effects on NF-nB activity by response to DDR1 overexpression was abrogated when an excess using an NF-nB promoter–luciferase reporter construct.As shown of cold wt-NF-nB binding oligonucleotide was added to compete in Fig.3A , Ad-DDR1 infection activated NF-nB activity, whereas its binding, whereas the addition of excess of mutant oligo had no Ad-GFP infection of parental MDAMB435 cells had no effect on the effect on NF-nB DNA-binding activity (Fig.3 B).To further NF-nB promoter.Next, we did EMSA to examine the effects of determine the specific role of DDR1 in NF-nB–mediated Cox-2 www.aacrjournals.org 8125 Cancer Res 2006; 66: (16). August 15, 2006

Figure 2. DDR1-mediated chemoresistance is Cox-2 dependent. A, depletion of Cox-2 induction represses Bcl-2 expression and reduces DDR1-mediated chemoresistance. MDAMB435 cells were transfected with Cox-2 shRNAor luciferase shRNAconstructs, then the cells were infected with Ad-GFP or Ad-DDR1 at 10 MOI. Twenty-four hours later, 40 Amol/L etoposide was added in the presence of collagen IV. Twenty-four hours after drug addition, cell extracts were made and Western blotting were carried out for the indicated proteins or subjected to cell death detection ELISA. Left, Western blot analysis using lysates prepared from MDAMB435 cells transfected with Cox-2 shRNAor control shRNA. Right, the pattern of apoptosis analysis by cell death ELISA. B, knockdown of basal Cox-2 expression suppresses Bcl-2 expression and enhances apoptosis in response to etoposide treatment. T47D cells were transfected with Cox-2 shRNAor luciferase shRNA constructs. Twenty-four hours later, 40 Amol/L etoposide was added in the presence of collagen IV. Twenty-four hours after drug treatment, cell extracts were made and Western blots were carried out for the indicated proteins or subjected to cell death detection ELISA.

induction, we tested whether the InBa superrepressor mutant activation, we examined the ability of DDR1 to activate NF-nB and abolishes Cox-2 expression induced by DDR1.MDAMB435 cells its targets in IKK knockout MEFs (i.e., IKKaÀ/À,IKKhÀ/À, and were transfected with the superrepressor expressing plasmid in the IKKgÀ/À MEFs).Each MEF was transfected with an NF- nB presence of Ad-DDR1 or Ad-GFP.Western blot analysis, as shown luciferase reporter construct (NF-nB-Luc) or control luciferase in Fig.3C , shows that the superrepresor InB markedly inhibited construct (pGL3) along with Ad-DDR1 or Ad-GFP.DDR1 enhanced Cox-2 expression induced by DDR1 (Ad-DDR1 infection), whereas NF-nB activity in wt-MEF and IKKaÀ/À MEF (Fig.5 A), but the transfection of the vector alone (pCDNA3) failed to do so.Taken activation of NF-nB by DDR1 was significantly diminished in both together, these results show that DDR1 increases NF-nB activity IKKhÀ/À and IKKgÀ/À MEFs (Fig.5A ).Ectopic DDR1 induced and that DDR1-mediated Cox-2 induction is NF-nB dependent. increased expression of NF-nB targets, Cox-2 and XIAP, in wt- Abrogation of NF-KB activity abolishes DDR1-mediated MEFs, and transfection of the superrepressor InB mutant into chemoresistance. Given the ability of DDR1 to enhance NF-nB wt-MEFs significantly inhibited the DDR1-mediated increase of activity, we investigated whether DDR1 actually exerts its Cox-2 and XIAP expression (Fig.5 B, left).Moreover, the DDR1- prosurvival effect via NF-nB activation.To assess the specificity mediated activation of NF-nB targets, Cox-2 and XIAP, was of this response, we examined whether abrogation of the NF-nB abolished in IKKhÀ/À MEFs (Fig.5 B, right).Taken together, these pathway by the superrepressor InB suppresses DDR1-mediated data show that IKKh and IKKg are essential for DDR1-mediated chemoresistance in DDR1-overexpressing breast cancer cells. NF-nB activation, but that IKKa is dispensable.To determine Ectopic DDR1 expression in MDAMB435 cells with little basal whether DDR1-mediated survival or chemoresistance also requires DDR1 expression significantly reduced the apoptotic response to the IKK subunits, we examined the antiapoptotic role of DDR1 in etoposide, compared with that of cells with ectopic GFP expression response to etoposide in IKK knockout MEFs.As seen in other cell (Fig.4 A).However, transfection of the superrepressor I nB resulted types, DDR1 overexpression suppressed etoposide-induced apo- in a significant reduction of DDR1-induced chemoresistance ptosis in wt-MEFs and IKKaÀ/À MEFs (Fig.5C and D).Moreover, (Fig.4A , columns 3 and 6).We also examined the effect of NF- nB abrogation of NF-nB pathway via the superrepressor InB mutant inhibition in T47D cells, which have a high basal level of DDR1 further enhanced the apoptotic response to etoposide in wt-MEFs. expression.As shown in Fig.4 B, transfection of the InB super- Of note, the DDR1-mediated antiapoptotic effect was not observed repressor resulted in a drastic increase in etoposide-induced in IKKhÀ/À or IKKgÀ/À MEFs (Fig.5D ).These results strongly apoptosis in a dose-dependent manner.In contrast, transfection of imply that DDR1 promotes cell survival via the activation of vector itself along with increasing doses of etoposide induced only NF-nB and its target genes and that IKKh and IKKg subunits of a marginal increase in apoptosis.These results indicate that DDR1 the IKK complex are required for mediating the prosurvival signal increases NF-nB activity and its activation plays an important role of DDR1 upon genotoxic stress. in DDR1-mediated chemoresistance in some breast cancer cells. Chemosensitization of breast cancer cells by depletion of IKKB and IKK; are required for DDR1-mediated NF-KB DDR1 expression. GiventhatDDR1isinvolvedinNF-nB activation and chemoresistance. The IKK complex (InB kinase) prosurvival signaling and chemoresistance, we sought to determine is the upstream activating kinase of the NF-nB pathway.It is whether DDR1 inhibition might be sufficient to promote chemo- composed of two catalytic subunits, IKKa and IKKh, which can sensitization in human breast cancer cells exhibiting high basal directly phosphorylate InB, whereas IKKg acts as the regulatory levels of DDR1.Inhibition of DDR1 expression by DDR1 shRNA in subunit.Activation of the canonical NF- nB pathway based on InB T47D cells in which DDR1 and Cox-2 are expressed at high levels degradation is mostly dependent on IKKh.To investigate whether significantly suppressed the levels of endogenous DDR1 and Cox-2 as these NF-nB upstream regulators affect DDR1-mediated NF-nB measured by immunoblotting (Fig.6 A).Moreover, depletion of

Cancer Res 2006; 66: (16). August 15, 2006 8126 www.aacrjournals.org

DDR1 modulated total InB protein level (Fig.6 A).Although transfection of control shRNA (luciferase shRNA) as well as vector alone (pBabe-U6-shRNA) had no effect on the levels of Cox-2, Bcl-2, and InB, cotransfection of the superrepressor InB mutant with either luciferase-shRNA or vector alone inhibited both Cox-2 and Bcl-2 expression (Fig.6 A).These data further argue that DDR1 is an upstream regulator of the NF-nB pathway.To directly assess the involvement of DDR1 in survival or chemoresistance of T47D tumor cells, we compared the levels of etoposide-induced apoptosis observed in these breast cancer cells transfected with a control shRNA construct (pBabe-luciferase-shRNA), vector alone (pBabe), or DDR1 shRNA (pBabe-DDR1-shRNA).As shown in Fig.6 B, depletion of DDR1 expression resulted in a significant increase in apoptosis with increasing etoposide concentration, whereas there was only a marginal increase in apoptosis with empty vector or luciferase shRNA-transfected cells.These findings establish that the reduced

Figure 4. DDR1-mediated chemoresistance is NF-nB pathway dependent in human breast cancer cells. A, inhibition of DDR1-mediated chemoresistance by an InB superrepressor. MDAMB435cells were transfected with an InB mutant (superrepressor) or control vector (pcDNA3), then infected with Ad-GFP or Ad-DDR1 at 10 MOI. Twenty-four hours later, 40 Amol/L etoposide were added to cells in the presence of collagen IV. Twenty-four hours after drug addition, cell death detection ELISAwas carried out to determine the pattern of apoptosis. B, InB superrepressor expression increases apoptosis in T47D cells. T47D cells were transfected with an InB superrepressor or vector (pcDNA3). Twenty-four hours later, etoposide was added. Twenty-four hours after drug addition, cell death detection ELISAwas carried out. Points, mean of three independent experiments with triplicate samples; bars, SD.

levels of endogenous DDR1 contributed directly to etoposide- induced sensitization of these tumor cells to apoptosis.Finally, we examined whether DDR1-mediated chemoresistance in T47D cells was NF-nB dependent.Suppression of Cox-2 expression or NF- nB activity by the superrepressor InB mutant significantly reduced apoptosis induced by etoposide in control shRNA or vector- transfected cells.Of note, DDR1 depletion by DDR1 shRNA failed to further increase the level of etoposide-induced apoptosis in the presence of the superrepressor (Fig.6 C).These results indicate that DDR1-mediated chemoresistance in breast cancer cells, which exhibit high basal levels of this receptor, is dependent upon activation of NF-nB pathway and consequent Cox-2 induction.

Discussion Figure 3. DDR1 activates NF-nB signaling. A, modulation of NF-nB activity by DDR1. MDAMB435 cells were transfected with vector (pGL3-basic) or It is well documented that Cox-2 can be induced by growth NF-nB-Luc. The cells were then infected with Ad-GFP or Ad-DDR1 at 10 MOI in factors and cytokines, inflammatory stimuli, and tumor promoters the presence of collagen IV. Forty-eight hours postinfection, luciferase activity (34).Some studies have provided evidence that oxidative and DNA was measured as described in Materials and Methods. B, increasing NF-nB DNA-binding activity by DDR1. MDAMB435 cells were treated with TNF-a damage stresses can induce Cox-2 expression (35, 36).However, the (positive control) or infected with Ad-GFP or Ad-DDR1 at 10 MOI. Twelve hours precise mechanisms of induction of Cox-2 in response to these after TNF-a addition or 48 hours postinfection, nuclear extracts were isolated. About 5 Ag nuclear extracts were incubated with 32P-labeled NF-nB oligo, cold stresses are unknown.Our previous studies showed that tumor NF-nB oligo, or its mutant, followed by gel electrophoresis. C, inhibition of suppressor p53 or DNA-damaging agents acting through p53 can DDR1-mediated Cox-2 induction by an InB superrepressor. MDAMB435cells induce Cox-2, which is able to counteract p53-mediated cell death were transfected with InB mutant (superrepressor), then infected with Ad-GFP or Ad-DD1 at 10 MOI. Forty-eight hours postinfection, cell extracts were (30).These studies indicated that Cox-2 is a p53 response gene that made and Western blots were carried out. is induced to alleviate the adverse effects arising from DNA www.aacrjournals.org 8127 Cancer Res 2006; 66: (16). August 15, 2006

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Cancer Research damage/genotoxic stress.A number of recent reports strongly increased apoptosis of wt-p53–containing cells in response to imply that enhanced Cox-2 expression plays a significant role in genotoxic stress (24).Hence, DDR1 seems to play a critical role in resistance to chemotherapy-induced apoptosis, by protecting determining the cellular outcome in response to genotoxic/ cancer cells against apoptosis (37).Moreover, increased Cox-2 therapeutic stress in wt-p53–containing cells.Our present studies levels are commonly detected in a variety of premalignant and show that DDR1 acts as a novel upstream regulator for Cox-2, and malignant tissues (19).Expression of Cox-2 in several types of that Cox-2 functions as an important mediator for DDR1- cancer, including breast cancer, correlates with poor prognosis, and mediated survival in the p53-dependent response to genotoxic Cox-2 enzyme inhibitors have been reported to reduce breast stress.Moreover, we showed that the mechanism by which DDR1 cancer incidence in humans (38, 39).Recent studies have shown causes increased Cox-2 expression is mediated through DDR1 that overexpression of Cox-2 in the mammary gland of CD1 activation of the NF-nB pathway, which induces increased levels of transgenic mice results in tumorigenic transformation as well as its known targets, including Cox-2 and XIAP.DDR1 up-regulation angiogenic stimulation (22, 40).Cox-2 is expressed by the tumor of Cox-2 as well as XIAP was shown to be dependent on endothelium in human colon, breast, prostate, lung, and other functional components of the NF-nB signaling.DDR1 activation cancers (41), whereas normal vascular endothelial cells express led to strong induction of NF-nB–mediated transcription, and only Cox-1 (42).In addition, several studies have shown that MEFs null for either IKKh or IKKg failed to induce Cox-2 in selective inhibition of Cox-2 activity suppresses angiogenesis response to DDR1 signaling.One well-known characteristic of in vitro and in vivo (43).Thus, it has been strongly suggested that cancer cytotoxic treatment is the activation of the transcription Cox-2 is a promising molecular target for cancer prognosis and factor NF-nB, which regulates cell survival (45).NF- nB activation treatment (44). suppresses the apoptotic potential of chemotherapeutic agents The wide range of genetic, environmental, and metabolic and contributes to chemotherapy resistance.It is also known that stimuli that activate p53 clearly distinguishes it from other many human tumors display constitutive NF-nB activation that known tumor suppressor genes.Our previous studies have shown allows malignant cells escape apoptosis (46).Our present findings that DDR1 is a direct p53 target gene and can be functionally show that DDR1 induces NF-nB activation, establishing this RTK activated/phosphorylated in a p53-dependent manner (24).We as an important upstream regulator for the NF-nB prosurvival also found that inhibition of DDR1 function resulted in strikingly pathway.

Figure 5. DDR1-mediated NF-nB activation requires IKKh and IKKg. A, IKKh and IKKg are required for DDR1-mediated NF-nB activation. WtÀ, IKKaÀ/À, IKKhÀ/À, and IKKgÀ/À MEFs were transfected with pGL3Basic or NF-nB-Luc along with 0.1 Ag tk-Renilla-Luc for normalization. Cells were then infected with Ad-GFP or Ad-DDR1 at 20 MOI. Forty-eight hours postinfection, luciferase activity was measured as described in Materials and Methods. B, DDR1-mediated induction of the NF-nB target genes, Cox-2 and XIAP, is suppressed by the superrepressor InB mutant (left). Wt-MEFs were infected with Ad-GFP or Ad-DDR1 at 20 MOI. Twenty-four hours later, 40 Amol/L etoposide were treated. After additional 24 hours, cell extracts were made and Western blots were carried out for the indicated proteins. DDR1 does not induce Cox-2 and XIAP expression in IKKh-null MEFs (right). IKKh-null MEFs were infected with Ad-GFP or Ad-DDR1 at 20 MOI. Twenty-four hours later, 40 Amol/L etoposide were treated. C and D, IKKh and IKKg are required for DDR1-dependent, NF-nB–mediated prosurvival. C, wt-MEFs were transfected with InB superrepressor, then infected with Ad-GFP or Ad-DDR1 at 20 MOI. Twenty-four hours later, MEFs were treated with 40 Amol/L etoposide. Twenty-four hours after drug addition, cell death detection ELISAwas carried out. D, WtÀ,IKKaÀ/À,IKKhÀ/À, and IKKgÀ/À MEFs were infected with Ad-GFP or Ad-DDR1 at 20 MOI. Twenty-four hours later, cells were treated with 40 Amol/L etoposide. Twenty-four hours after drug addition, apoptosis was measured by cell death detection ELISA. Columns (A, C, andD), mean of three independent experiments with each sample in triplicate; bars, SD.

Cancer Res 2006; 66: (16). August 15, 2006 8128 www.aacrjournals.org

Figure 6. Knockdown of endogenous DDR1 expression enhances chemosensitivity. A, modulation of NF-nB pathway proteins, Cox-2, Bcl-2, and InB. T47D cells were stably transfected with pBabe (vector), luciferase shRNA(pBabe-luciferase shRNA),or DDR1 shRNAconstruct (pBabe-DDR1 shRNA),and selected for s table clones under puromycin. The resistant clones were pooled and determined for reduction of DDR1 expression. DDR1 knockdown cells (T47-DDR1 shRNA) as well as control shRNAlines (T47D-luciferase shRNAand T47D-pBabe shRNA)were then transiently transfected with I nB superrepressor mutant or vector (pcDNA3), and total cell lysates were isolated for Western blot analysis using anti-DDR1, anti-Cox-2, anti-InB, anti-Bcl-2, or anti-h-actin antibodies. Note that depletion of DDR1 significantly reduced Cox-2 expression along with an increase of total InB level and that expression of InB mutant superrepressor inhibited Cox-2 expression. DDR1 knockdown or control T47D cells were also transfected with Cox-2 shRNAor GFP-shRNA.Forty-eight hours posttransfection, cell extracts were ma de and Western blots were carried out for the indicated proteins. B, effect of DDR1 depletion on genotoxic agent (etoposide)–mediated apoptosis. DDR1 knockdown T47D cells and control T47D cells (pBabe-shRNAand Luciferase shRNA)were treated with indicated concentrations of etoposide. Twenty-four hours after drug ad dition, apoptosis was measured by cell death detection ELISA. C, depletion of Cox-2 or InB superrepressor mutant antagonizes DDR1-mediated prosurvival. The same double-knockdown cells from (A) were also treated with etoposide (40 Amol/L) for 24 hours, then analyzed to determine the patterns of apoptosis by cell death detection ELISA.

Numerous studies have reported overexpression of DDR1 in lung, imply that in DDR1-overexpressing tumors, prosurvial signaling by esophageal, breast, ovary, and pediatric brain cancers (2, 7, 13–17). this newly worked out pathway may offer new tumor-specific Moreover, the elevated expression of DDR1 in a number of fast- therapeutic targets. growing invasive tumors has suggested an important role of this matrix-activated RTK in the proliferation and stromal invasion of Acknowledgments tumors (13, 15).In the present studies, we showed that a human Received 4/3/2006; revised 5/30/2006; accepted 6/13/2006. breast tumor line, T47D, which overexpressed DDR1, exhibited high Grant support: NIH grants CA097216, CA085214, and CA080058. constitutive levels of Cox-2.Moreover, inhibition of NF- nB by the The costs of publication of this article were defrayed in part by the payment of page InB superrepressor as well as shRNA inhibition of the expression of charges.This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Cox-2 or DDR1 itself all dramatically increased the apoptotic We thank the members of the Cutaneous Biology Research Center for helpful response of these tumor cells to genotoxic agents.These findings discussion and Sarah Boswell for critical reading of the manuscript.

References serve as nonintegrin collagen receptors.Mol Cell 1997;1: 10. Matsuyama W, Faure M, Yoshimura T.Activation of 25–34. discoidin domain receptor 1 facilitates the maturation 1. Di Marco E, Cutuli N, Guerra L, Cancedda R, De 6. Vogel W, Brakebusch C, Fassler R, Alves F, Ruggiero F, of human monocyte-derived dendritic cells through the Luca M.Molecular cloning of trkE, a novel trk-related Pawson T.Discoidin domain receptor 1 is activated TNF receptor associated factor 6/TGF-h-activated putative tyrosine kinase receptor isolated from normal independently of h(1) integrin.J Biol Chem 2000;275: protein kinase 1 binding protein 1h/p38a mitogen- human keratinocytes and widely expressed by normal 5779–84. activated protein kinase signaling cascade.J Immunol human tissues.J Biol Chem 1993;268:24290–5. 7. Alves F, Vogel W, Mossie K, Millauer B, Hafler H, 2003;171:3520–32. 2. Zerlin M, Julius MA, Goldfarb M.NEP: a novel Ullrich A.Distinct structural characteristics of discoidin 11. Hou G, Vogel W, Bendeck MP.The discoidin domain receptor-like tyrosine kinase expressed in proliferating I subfamily receptor tyrosine kinases and complemen- receptor tyrosine kinase DDR1 in arterial wound repair. neuroepithelia.Oncogene 1993;8:2731–9. tary expression in human cancer.Oncogene 1995;10: J Clin Invest 2001;107:727–35. 3. Johnson JD, Edman JC, Rutter WJ.A receptor tyrosine 609–18. 12. Vogel WF, Aszadi A, Alves F, Pawson T.Discoidin kinase found in breast carcinoma cells has an extra- 8. Bhatt RS, Tomoda T, Fang Y, Hatten ME.Discoidin domain receptor 1 tyrosine kinase has an essential role cellular discoidin I-like domain.Proc Natl Acad Sci domain receptor 1 functions in axon extension of in mammary gland development.Mol Cell Biol 2001;21: U S A 1993;90:5677–81. cerebellar granule neurons.Genes Dev 2000;14:2216–28. 2906–17. 4. Vogel W, Gish GD, Alves F, Pawson T.The discoidin 9. Ram R, Lorente G, Nikolich K, Urfer R, Foehr E, 13. Barker KT, Martindale JE, Mitchell PJ, et al. domain receptor tyrosine kinases are activated by Nagavarapu U.Discoidin domain receptor-1a (DDR1a) Expression patterns of the novel receptor-like tyrosine collagen.Mol Cell 1997;1:13–23. promotes glioma cell invasion and adhesion in associ- kinase, DDR, in human breast tumours.Oncogene 1995; 5. Shrivastava A, Radziejewski C, Campbell E, et al.An ation with matrix metalloproteinase-2.J Neurooncol 10:569–75. orphan receptor tyrosine kinase family whose members 2006;76:239–48. 14. Laval S, Butler R, Shelling AN, Hanby AM, Poulsom R, www.aacrjournals.org 8129 Cancer Res 2006; 66: (16). August 15, 2006

Ganesan TS.Isolation and characterization of an 25. Hu Y, Baud V, Delhase M, et al.Abnormal mor- 35. Steinauer KK, Gibbs I, Ning S, French JN, Armstrong epithelial-specific receptor tyrosine kinase from an phogenesis but intact IKK activation in mice lacking the J, Knox SJ.Radiation induces upregulation of cyclo- ovarian cancer cell line.Cell Growth Differ 1994;5: IKKa subunit of the InB kinase.Science 1999;284:316–20. oxygenase-2 (COX-2) protein in PC-3 cells.Int J Radiat 1173–83. 26. Li Z-W, Chu W, Hu Y, et al.The IKK a subunit of InB Oncol Biol Phys 2000;48:325–8. 15. Perez JL, Jing SQ, Wong TW.Identification of two kinase (IKK) is essential for nuclear factor-nB activa- 36. Adderley SR, Fitzgerald DJ.Oxidative damage of isoforms of the Cak receptor kinase that are coex- tion and prevention of apoptosis.J Exp Med 1999;189: cardiomyocytes is limited by extracellular regulated pressed in breast tumor cell lines.Oncogene 1996;12: 1839–45. kinases 1/2-mediated induction of cyclooxygenase-2. 1469–77. 27. Makris C, Godfrey VL, Kra¨hn-Senftleben G, et al. J Biol Chem 1999;274:5038–46. 16. Nemoto T, Ohashi K, Akashi T, Johnson JD, Hirokawa Female mice heterozygous for IKKg/NEMO deficiencies 37. Cao Y, Prescott SM.Many actions of cyclooxygenase- K.Overexpression of protein tyrosine kinases in human develop a dermatopathy similar to the human 2 in cellular dynamics and in cancer.J Cell Physiol 2002; esophageal cancer.Pathobiology 1997;65:195–203. X-linked disorder incontinentia pigmenti.Mol Cell 190:279–86. 17. Weiner HL, Rothman M, Miller DC, Ziff EB.Pediatric 2000;5:969–79. 38. Denkert C, Winzer KJ, Hauptmann S.Prognostic brain tumors express multiple receptor tyrosine kinases 28. Yamaoka S, Courtois G, Bessia C, et al.Complemen- impact of cyclooxygenase-2 in breast cancer.Clin Breast including novel cell adhesion kinases.Pediatr Neurosurg tation cloning of NEMO, a component of the InB kinase Cancer 2004;4:428–33. 1996;25:64–71. complex essential for NF-nB activation.Cell 1998;93: 39. Arun B, Goss P.The role of COX-2 inhibition in 18. Dannenberg AJ, Subbaramaiah K.Targeting cyclo- 1231–40. breast cancer treatment and prevention.Semin Oncol oxygenase-2 in human neoplasia: rationale and promise. 29. Ryo A, Suizu F, Yoshida Y, et al.Regulation of NF- nB 2004;31:22–9. Cancer Cell 2003;6:431–43. signaling by Pin1-dependent prolyl isomerization and 40. Chang SH, Liu CH, Conway R, et al.Role of 19. Dannenberg AJ, Altorki NK, Boyle JO, et al.Cyclo- ubiquitin-mediated proteolysis of p65/RelA.Mol Cell prostaglandin E2-dependent angiogenic switch in cyclooxygenase 2: a pharmacological target for the preven- 2003;12:1413–26. oxygenase 2-induced breast cancer progression.Proc tion of cancer.Lancet Oncol 2001;2:544–51. 30. Han JA, Kim JI, Ongusaha PP, et al.p53-mediated Natl Acad Sci U S A 2004;101:591–6. 20. Gupta RA, Dubois RN.Colorectal cancer prevention induction of Cox-2 counteracts p53- or genotoxic stress- 41. Masferrer JL, Leahy KM, Koki AT, et al.Antiangio- and treatment by inhibition of cyclooxygenase-2.Nat induced apoptosis.EMBO J 2002;21:5635–44. genic and antitumor activities of cyclooxygenase-2 Rev Cancer 2001;1:11–21. 31. Tsujii M, DuBois RN.Alterations in cellular ad- inhibitors.Cancer Res 2000;60:1306–11. 21. Gasparini G, Longo R, Sarmiento R, Morabito A. hesion and apoptosis in epithelial cells overexpressing 42. Leahy KM, Ornberg RL, Wang Y, Zweifel BS, Koki A, Inhibitors of cyclo-oxygenase 2: a new class of prostaglandin endoperoxide synthase 2.Cell 1995;83: Masferrer L.Cyclooxygenase-2 inhibition by celecoxib anticancer agents? Lancet Oncol 2003;4:605–15. 493–501. reduces proliferation and induces apoptosis in angio- 22. Liu CH, Chang SH, Narko K, et al.Overexpression 32. Schmedtje JF, Ji YS, Liu WL, DuBois RN, Runge MS. genic endothelial cells in vivo.Cancer Res 2002;2:625–31. of cyclooxygenase-2 is sufficient to induce tumori- Hypoxia induces cyclooxygenase-2 via the NF-nB p65 43. Gately S, Li WW.Multiple roles of COX-2 in tumor genesis in transgenic mice.J Biol Chem 2001;276: transcription factor in human vascular endothelial cells. angiogenesis: a target for antiangiogenic therapy.Semin 18563–9. J Biol Chem 1997;272:601–8. Oncol 2004;31:2–11. 23. Thun MJ, Henley SJ, Patrono C.Nonsteroidal anti- 33. Yamamoto K, Arakawa T, Ueda N, Yamamoto S. 44. Subbaramaiah K, Dannenberg AJ.Cyclooxygenase 2: inflammatory drugs as anticancer agents: mechanistic, Transcriptional roles of nuclear factor n B and nuclear a molecular target for cancer prevention and treatment. pharmacologic, and clinical issues.J Natl Cancer Inst factor-interleukin-6 in the tumor necrosis factor Trends Pharmacol Sci 2003;24:96–102. 2002;94:252–26. a-dependent induction of cyclooxygenase-2 in MC3T3-1 45. Wang CY, Mayo MW, Baldwin AS, Jr.TNF- a and 24. Ongusaha PP, Kim JI, Fang L, et al.p53 induction and cells.J Biol Chem 1995;270:31315–20. cancer therapy-induced apoptosis: potentiation by activation of DDR1 kinase counteract p53-mediated 34. Smith WL, DeWitt DL, Garavito RM.Cyclooxyge- inhibition of NF-nB.Science 1996;274:784–7. apoptosis and influence p53 regulation through a nases: structural, cellular, and molecular biology.Annu 46. Rayet B, Gelinas C.Aberrant rel/nfkb genes and positive feedback loop.EMBO J 2003;22:1289–301. Rev Biochem 2000;69:145–82. activity in human cancer.Oncogene 1999;18:6938–47.

Cancer Res 2006; 66: (16). August 15, 2006 8130 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Discoidin Domain Receptor 1 Receptor Tyrosine Kinase Induces Cyclooxygenase-2 and Promotes Chemoresistance through Nuclear Factor-κB Pathway Activation

Sanjeev Das, Pat P. Ongusaha, Yoon Sun Yang, et al.

Cancer Res 2006;66:8123-8130.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/16/8123

Cited articles This article cites 46 articles, 18 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/16/8123.full#ref-list-1

Citing articles This article has been cited by 15 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/16/8123.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/66/16/8123. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.