logy 7 (2007) 1414–1421www.elsevier.com/locate/intimp

International Immunopharmaco

Apoptosis of murine lupus T cells induced by the selectivecyclooxygenase-2 inhibitor celecoxib: Molecular

mechanisms and therapeutic potential

Peng Yang, Yan Zhang, Lv Ping, Xiao-Ming Gao⁎

Department of Immunology, Peking University Health Science Center, Peking University, Beijing 100083, ChinaKey Laboratory of Immunology, Ministry for Public Health, China

Received 9 May 2007; received in revised form 26 June 2007; accepted 27 June 2007

Abstract

Upregulation of cyclooxygenase (COX)-2 in T cells from patients with systemic lupus erythematosus (SLE) is associated withtheir resistance to functional inactivation (anergy) and to activation-induced cell death through apoptosis. It has been demonstratedthat celecoxib, a selective COX-2 inhibitor, can enhance apoptosis of human lupus T cells. The present study was undertaken toinvestigate whether COX-2 expression is also upregulated in T cells from the lupus-prone BXBS strain of mice and if murine lupusis modified by celecoxib. COX-2 expression was detected in splenic T cells from 6 month-old male BXSB mice (murine lupus Tcells) but not in T cells from 2 month-old male or 6-month-old female BXSB or in 6-month-old male C57BL/6 mice, indicating astrong correlation between COX-2 expression in T cells and lupus manifestation in mice. Celecoxib treatment induced apoptosis ofmurine lupus T cells in vitro, which was inhibited by z-VAD-fmk, a pan-caspase inhibitor. In the murine lupus T cells treated withcelecoxib, procaspases 3 and 9, but not procaspase 8, were activated. In addition, celecoxib treatment decreased the mitochondrialmembrane potential of murine lupus T cells. These data combine to suggest that celecoxib mainly uses the mitochondrial pathwayrather than FADD pathway to trigger apoptosis of COX-2 expressing murine lupus T cells. Intragastric administration of celecoxib(40 mg/kg/day for 60 days) in 6-month-old male BXSB mice effectively limited the production of serum antibodies againstdsDNA. Our data suggest that celecoxib may have a beneficial effect in treating autoimmune diseases such as SLE throughinducing apoptosis of autoreactive T cells.© 2007 Published by Elsevier B.V.

Keywords: COX-2; Celecoxib; Apoptosis; Lupus; T cells

1. Introduction

Production of autoantibodies against double-strandedDNA (dsDNA) is the hallmark of systemic lupus

⁎ Corresponding author. Department of Immunology, PekingUniversity Health Science Center, 38 Xueyuan Road, Beijing100083, China. Tel./fax: +86 10 82801156.

E-mail address: xmgao@bjmu.edu.cn (X.-M. Gao).

1567-5769/$ - see front matter © 2007 Published by Elsevier B.V.doi:10.1016/j.intimp.2007.06.013

erythematosus (SLE), an autoimmune disease involvingmultiple systems and organs [1–3]. Accumulatingevidence suggests that lupus T cells play a pivotal rolein the development of SLE [2–6], providing help forautoantibody production by B cells and/or facilitatingtissue damage in the end organs [2,7–10]. It has alsobeen documented that lupus T cells are resistant to theinduction of apoptosis [11–13]. Recent studies showthat anti-CD3 monoclonal Ab-activated human lupus T

1415P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

cells resist both anergy and activation-induced cell death(AICD) by upregulating the expression of cyclooxy-genase (COX)-2 [14], a molecule over-expressed inmany tumor cells and is considered at least partiallyresponsible for their increased resistance to programmedcell death [14–18]. Interestingly, celecoxib (4-[5-(4-methylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfon-amide), a selective COX-2 inhibitor, iscapable of inducing apoptosis of tumor cells and humanlupus T cells [14,19–23].

BXSB mice are a recombinant inbred strain derivedfrom a cross between C57BL/6 (H-2b) female and SB/Le(H-2b) male mice [24], and represent one of the mostextensively studied animal models for human SLE.Whilst mice below the age of 2 months of age appear tobe healthy, adult male BXSB mice (5 to 6 months of ageor older) show characteristics similar to human disease,such as production of autoantibodies against dsDNA andother nuclear antigens (e.g. nucleosome), an inflamma-tory reaction in the joints and the development ofnephritis in the later stages of the disease [25]. BXSBmice differ from other mouse models of SLE (e.g.NZB×NZWF1 mice) in that the incidence for thefemales to develop the disease is much lower thanmales. The Yaa (Y chromosome-linked autoimmuneaccelerating) gene of BXSB mice has been shown toaccelerate autoantibody production and disease manifes-tation in lupus-prone animals [26,27]. It has recently beenreported that B cells containing the Yaa locus areintrinsically biased toward nucleolar antigens becauseof increased expression of Toll-like receptor 7 (TLR7), asingle-stranded RNA-binding innate immune receptor[28,29]. The resulting over-expression of Tlr7 increasedin vitro responses to TLR7 signaling in all Yaa-bearingmale BXSB mice. In addition, B6.Sle1yaa CD4 T cellsshow expression changes in numerous cytokines andchemokines [29]. The present study was undertaken toinvestigate whether T cells from adult male BXSB mice(murine lupus Tcells) express COX-2 at an elevated leveland also to explore the possibility of treating murinelupus using celecoxib in vivo.

2. Materials and methods

2.1. Mice

BXSB (H-2b) mice were purchased from Jackson Laborato-ries, USA and maintained at the Experimental Animal Division,Peking University Health Science Center. C57BL/6 (H-2b) micewere provided by Unilihua Bioscience Center, Beijing, China.All animal experiments were carried out, with the permission ofBeijing Experimental Animal Management Authority, Beijing,China, at the animal facilities of this department.

2.2. Reagents

Celecoxib was a generous gift from Dr. Bang-Le Zhang(Pharmacology Department, the Fourth Military MedicalUniversity, China). A combination of infrared and ultravioletspectroscopy, element analysis and proton NMR confirmed thepurity of the celecoxib to be N99% (data not shown). Thecelecoxib was dissolved at 100 mM in DMSO and this stocksolution stored at −20 °C. Fluorescein isothiocyanate (FITC)-conjugated rat anti-mouse CD3 and phycoerythrin (PE)-conjugated hamster anti-mouse FasL mAbs were purchasedfrom PharMingen (San Diego, California, USA). PE-conju-gated mouse anti-mouse Fas mAbs and unlabeled purifiedanti-mouse FasL mAbs were purchased from eBioscience (SanDiego, California, USA). Antibodies against COX-2, β-actin,procaspase-3, procaspase-8 and procaspase-9 were fromLABVISION (Fremont, CA, USA), and horseradish peroxi-dase (HRP)-labeled goat anti-rabbit IgG from Zhongshan Bio.Co., China. The mitochondrial membrane potential probe 5,5′,6,6′-tetrachloro-1,1′ 3,3′-tetraethyl-benzimidazol-carbocya-nine chloride (JC-1) was obtained from Biotium (Hayward,CA, USA). Pan-caspase inhibitor (z-VAD-fmk) was purchasedfrom R&D (Minneapolis, MN, USA).

2.3. Isolation of T cells

Unless otherwise specified, T cells were derived from thepooled spleens from 5 mice. Single cell suspensions wereprepared from the spleens of BXSB and C57BL/6 mice and thecells labeled with FITC-conjugated rat anti-mouse CD3 mAbat 4 °C for 30 min. After washing three times, fluorescent cellswere sorted using a BD FACSAria™ cell sorter. The sortedcells were T cells with a purity of N95% as judged by flowcytometric analysis. Alternatively, splenic cell suspensionswere first adhered to plastic at 37 °C, 5% CO2 for 4 h and thenon-adherent cells then mixed with Dynal beads coated withanti-mouse B220 antibody (Dynal, Denmark). Followingincubation for 30 min at 4 °C with regular mixing, the bead-attached cells were removed using a magnet and the remainingcells harvested. The negatively selected cells were T cells witha purity of N90% as shown by flow cytometric analysis.

2.4. Cell culture

T cells were cultured in RPMI 1640 (Gibco, USA)supplemented with 10% (v/v) fetal calf serum (FCS) (Hyclone,USA), penicillin/streptomycin (100 U/ml), 2-mercaptoethanol(5×10−5 M) and L-glutamine (2 mM) in a humidifiedatmosphere of 5% CO2 at 37 °C.

2.5. Western blotting

Cells were lysed in buffer containing 50 mM Tris/HCl (pH7.4), 200 mM NaCl, 0.5% NP-40, 10% glycerol, 0.1 mMEDTA, 1 mM DDT, 0.4 mM phenylmethylsulfonyl fluoride(PMSF), 2 μg/ml aprotinin, 2 μg/ml pepstatin, and 1 μg/ml

1416 P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

leupeptin on ice for 30 min. Following centrifugation at13,000 rpm for 20 min, the supernatant was collected. Equalamounts of protein (50 μg) were separated by SDS-PAGE, andthe protein bands transferred onto nitrocellulose membranes(Amersham Biosciences, Sweden). Following blocking with5% non-fat dry milk at 4 °C overnight, the membranes wereincubated with primary antibody (2.5 μg/ml, 2 h at roomtemperature) then washed in Tris-buffered saline (TBS, pH8.0) containing 0.05% Tween 20 and subsequently incubatedwith HRP-labeled goat anti-rabbit IgG. The ECL system(Applygen, Beijing, China) was used to visualize the reaction.

2.6. Apoptosis analysis by flow cytometry

Splenic T cells from mice treated with celecoxib wereharvested and washed once with cold PBS followed byfixation with 70% cold ethanol at 4 °C for 16–18 h. The fixedcells were washed twice with PBS, and then treated with20 μg/ml RNaseA in PBS at 37 °C for 30 min. The reactionwas terminated on ice and the cells stained with propidiumiodide (50 μg/ml) in 1% Triton. The DNA content of the cells(104/sample) was analyzed immediately using FACScan(Becton Dickinson, USA). The sub-G1 population wasdesignated as being apoptotic cells.

2.7. Activation-induced cell death in T cells

Activation-induced apoptosis was performed as describedby Pahlavani and Vargas [30], with minor modification.Purified Tcells were added into a 24-well plate pre-coated withanti-mouse CD3 antibody (4 μg/ml) and cultured for 48 h. Thecells were then harvested, washed and added into a new platepre-coated with anti-CD3 mAb. Exogenous IL-2 (20 U/ml)

Fig. 1. Resistance of murine lupus T cells to activation-induced cell death6 month-old male BXSB mice (A) or C57BL/6 (B) mice were cultured for 48IL-2 was then added (20 U/ml final concentration) and the cells were cultapoptosis (sub-G1 phase cells). To assess COX-2 expression in murine lupus Tmale BXSB (Lane 1) or C57BL/6 mice (Lane 2), or 2 month-old male BXSB mSDS-PAGE gels followed by Western blotting with polyclonal rabbit anti-COX

was also added and the cells cultured for another 48 h beforethe T cells were harvested for analyzed for cell apoptosis.

2.8. Cell surface expression of Fas and FasL

Lupus T cells were cultured in medium with or withoutcelecoxib (50 μM) for 24 h and then washed 3 times in PBSprior to incubation with PE-conjugated anti-mouse Fas or FasLmAb on ice for 30 min. Following washing three times withPBS, cell surface fluorescence was detected using FACScan.

2.9. Analysis of mitochondrial membrane potential (Δψm)

To detectΔψm, after treatment of lupus T cells with variousconcentrations of celecoxib the cells were incubated with 5 μg/mlJC-1 at RT for 30 min in the dark. Following two washes withPBS the cells were analyzed by FACScan.

2.10. Effect of celecoxib in vivo

Male BXSBmice (6months old, 20–30 g) were fed 40mg/kgof celecoxib in 0.5% carboxymethylcellulose (CMC) via gavageeveryday, or as a control fed only 0.5% CMC. On days 0, 15, 30,45 and 60 the sera of individual mice in control and drug groupswere collected from the tails, and the level of anti-dsDNA anti-body in each sample measured by ELISA.

2.11. ELISA assays

96-well ELISA plates (Greiner, Germany) were coated withdsDNA (100μg/ml, 200μl/well) dissolved in carbonate buffer atpH 9 and incubated overnight at 4 °C. The coated wells werethen blocked with blocking solution (PBS containing 10% FCS,

and their enhanced COX-2 expression. T cells freshly purified fromh in 24-well plates pre-coated with anti-mouse CD3 mAb. Exogenous

ured for an additional 48 h followed by flow cytometric analysis forcells (C), lysates of T cells, purified from the splenocytes of 6 month-oldice (Lane 3), or 6 month-old female BXSB mice (Lane 4), were run on

-2 IgG antibody. These are representatives of 3 independent experiments.

Fig. 2. Celecoxib-induced apoptosis of BXSB Tcells. T cells, freshly purified from splenocytes of 6 month-old male BXSBmice, were cultured in thepresence (25 or 50 μM), or absence, of celecoxib for up to 24 h. The cells were sampled at different time points and stained with PI for FACS analysis.Histograms A, B and C show the percentages of sub-G1 phase (apoptotic) T cells in the different groups 24 h after incubation. The time course ofcelecoxib (50 μM)-induced apoptosis of murine lupus T cells is shown in D.

1417P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

200 μl/well) and washed with washing buffer (PBS containing0.05% Tween-20). Serum samples (100 μl/well), diluted 1:100with blocking solution, were added to each well and incubatedfor 90 min at 37 °C. After washes with washing buffer, HRP-labeled goat anti-mouse IgGAb (1:5000) was added to eachwell(100 μl/well). After 1 h incubation at 37 °C and subsequentwashes, 100μl of substrate solution (3mgOPD and 2μl H2O2 in5 ml phosphate–citrate buffer, pH 5.0) was dispensed into eachwell and the plates incubated for 10 min at RT before adding50 μl 12.5% H2SO4 to stop the reaction. Absorbancy of eachwell at OD492 nm was determined using an ELISA plate reader(Labsystems, Finland).

2.12. Statistical analysis

All experiments were repeated at least three times. Thedata shown are representative of these independent experi-ments. The χ2 test was used to compare the apoptotic levelsof T cells from different groups of mice, and the independent-samples T test was used to compare anti-dsDNA antibodylevels on day 60 between control and drug groups usingSPSS13.0 software.

3. Results

3.1. Elevated COX-2 expression in murine lupus T cells

Similar to human lupus Tcells, Tcells from6month-oldmaleBXSB male mice with murine lupus were more resistant toactivation-induced apoptosis compared with T cells fromC57BL/6 mice of the same sex and age (Fig. 1A and B). Freshlypurified Tcells from BXSB and C57BL/6 mice of 2 or 6 monthsof age were lysed and then assayed for COX-2 expression byWestern blotting using polyclonal anti-COX-2 antibodies. Thelevel of COX-2 expression in T cells from 6-month-old maleBXSBmice (murine lupus Tcells) was substantially higher thanthat from female BXSBmice of the same age (Fig. 1C).Minimalamounts of COX-2 were detectable in the T cells from 2-month-old BXSB or 6-month-old C57BL/6 male mice. These resultssuggest a strong correlation between COX-2 expression inperipheral T cells and lupus development in mice.

3.2. Celecoxib accelerates apoptosis of murine lupus T cells

In Fig. 2, T cells from BXSB mice were treated withcelecoxib (25 μM or 50 μM) for up to 48 h. Apoptosis of the

Table 1Apoptosis of T cells induced by celecoxib a

Treatment

Donor micefor theT cells

Mediumalone b

Celecoxib(50 μM) c

Celecoxib+anti-FasL d

Celecoxib+z-VAD-fmk e

BXSB male,6 months

33% 71% 70% 21%

C57BL/6 male,6 months

28% 43% ND f ND

BXSB male,2 months

30% 44% ND ND

BXSB female,6 months

22% 43% ND ND

a T cells were freshly purified from splenocytes of 2-month-old maleBXSB, 6-month-old female or male BXSB and male C57BL/6 mice.The purified T cells (2×106 cells/well) were cultured in completeRPMI 1640 medium (2 ml/well, 24-well plates) with, or without,celecoxib (50 μM final concentration) for 24 h and then stained with PIand analyzed on a FACS machine. The results are expressed aspercentage of sub-G1 phase T cells (apoptotic cells) in the culture.b There was no significant difference in percent T cells undergoing

spontaneous apoptosis in all 4 groups (PN0.05, χ2 test).c The percentage of T cells undergoing apoptosis in the group of 6-

month-old BXSB males was significantly more than that from theother three groups (71% vs. either 44% or 43%, Pb0.05, χ2 test).d Freshly purified T cells from 6-month-old male BXSB were

cultured in complete RPMI 1640 medium in the presence of anti-FasLmAb (10 μg/ml) and celecoxib (50 μM).e z-VAD-fmk was added (40 μM final concentration) to the T cell

culture 1 h prior to the addition of celecoxib.f Not determined.

Fig. 3. Activation of caspase cascade was induced by celecoxib inmurine lupus T cells. T cells, freshly purified from splenocytes of6 month-old male BXSB mice, were cultured in medium alone (0 μM)or in the presence of celecoxib at 25 μM or 50 μM for 24 h. The cellswere then lysed for subsequent Western blotting using polyclonalrabbit IgG antibodies against procaspase-3, procaspase-8 and procas-pase-9.

1418 P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

lupus T cells was demonstrable 8 h after the treatment, asevidenced by the accumulation of sub-G1 phase (apoptotic)cells in the culture. By contrast, T cells from 6-month-old maleC57BL/6, or female BXSB, or 2-month-old male BXSB micewere significantly less susceptible to celecoxib-inducedapoptosis (Table 1).

3.3. Celecoxib induces apoptosis of murine lupus T cellsthrough the caspase cascade

The pan-caspase inhibitor z-VAD-fmk (40 μM finalconcentration) effectively inhibited celecoxib-induced apo-ptosis of murine lupus T cells, suggesting the involvement ofthe caspase cascade in apoptosis of these cells (Table 1).Western blotting assays were subsequently carried out, usingantibodies specific for procaspases 3, 9 or 8, to investigate theactivation (proteolic cleavage) of relevant procaspases inmurine T cells before and after celecoxib treatment. As shownin Fig. 3, the amount of procaspases 3 and 9, but not ofprocaspase-8, decreased in murine lupus T cells followingcelecoxib treatment, suggesting that celecoxib accelerated Tcell apoptosis via the pathway of caspases 3 and 9 rather thanthat of caspase 8.

3.4. Antibodies against FasL fail to block celecoxib-inducedapoptosis of murine lupus T cells

Celecoxib treatment substantially increased the expressionof Fas or FasL on murine lupus T cells (Fig. 4), in agreementwith similar observations in human lupus T cells [14].However, anti-mouse FasL antibody was unable to blockcelecoxib-induced apoptosis of T cells (Table 1). This data,taken together with the observation that procaspase 8 was notactivated in celecoxib-treated T cells, indicate that the Fas-associated death domain (FADD) pathway is unlikely toconstitute the main route through which celecoxib triggersapoptosis in murine lupus T cells.

3.5. Celecoxib induces apoptosis of murine T cells via themitochondrial pathway

Several studies have shown that celecoxib could induceapoptosis by disrupting mitochondrial membrane potential oftumor cells [22,23].Whenmurine lupus Tcells were treatedwithcelecoxib for 24 h and then stained with JC-1 for determinationof mitochondrial membrane potential, a clear decrease ofmitochondrial membrane potential was observed (Fig. 5).

3.6. Effect of celecoxib on circulating anti-dsDNA IgGantibodies in BXSB mice

To assess the possibility of treating lupus with celecoxib invivo, 6-month-old male BXSB mice were intragastricallyadministered celecoxib (40 mg/kg/day, once daily) or vehiclecontrol only (n=10) for 60 days. Serum samples werecollected on days 0, 15, 30, 45 and 60 and analyzed for anti-dsDNA IgG levels using ELISA. The levels of serum IgGspecific for dsDNA increased with time in the control groupwhilst the levels in the celecoxib treated group remainedalmost unchanged (Fig. 6).

Fig. 5. Mitochondrial membrane potential (Δψm) changes in murine lupus T cells following celecoxib treatment. T cells from 6 month-old maleBXSB mice were treated with various concentrations of celecoxib (0 μM, 25 μM, 50 μM) for 24 h and then washed and treated with JC-1 (5 μg/ml)for 30 min followed by flow cytometric analysis. The monomeric form of JC-1 (green fluorescence) binds to mitochondrial membranes withdecreased potential, whilst the aggregated form (orange fluorescence) binds with mitochondrial membranes with a normal potential.

Fig. 6. Celecoxib prevents the age-related increase in serum anti-dsDNA antibody levels in lupus-prone mice. Male BXSBmice (6 months old, 10 pergroup) were treated with 0.5% carboxymethylcellulose (CMC) or with celecoxib (in CMC) at 40 mg/kg/day for 60 days. The animals were bled every15 days from the onset of the experiment and the sera assayed for IgG Abs against dsDNA using ELISA. The results are expressed as anti-dsDNA IgGindex (calculated as Day N/Day 0) with the OD reading at 492 nm on Day 0 as 1. Student's T test showed that the difference between the two groupson Day 60 was statistically significant (Pb0.05).

Fig. 4. Celecoxib-induced Fas and FasL expression in T cells. T cells from 6-month-old male BXSB mice were treated with (open histogram), orwithout (shaded histogram), celecoxib (50 μM) for 24 h in culture. The cells were then analysed using PE-labeled mAbs against Fas or FasL.

1419P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

1420 P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

4. Discussion

Animal models of SLE, such as the BXSB and(NZB×NZW)F1 strains of mice, are useful for the inves-tigation of disease mechanisms and for exploring thepotential of candidate therapeutic agents. In the presentstudy we demonstrate that murine lupus T cells fromBXSB mice express high levels of COX-2, as previouslyobserved for human lupus T cells [14]. More importantly,celecoxib, a selective COX-2 inhibitor, not only acceler-ated apoptosis of murine lupus T cells in vitro but alsoexhibited a potentially beneficial effect by preventing thedevelopment of enhanced anti-dsDNA IgG levels in thesera of aged male BXSB mice.

COX-2 is the rate-limiting enzyme in the biosynthe-sis of prostanoids and is largely responsible forprostaglandin formation [31]. More importantly in thecontext of the present study, COX-2 is also an anti-apoptotic molecule. There is ample evidence suggestingthat the antiapoptosis effect of COX-2 is independent ofits enzymatic activity, as evidence by in vitro as well asin vivo experiments [20,21,32]. Nuclear localization ofCOX-2 suggested a PGE2-independent role in tran-scriptional regulation. The Cox-2 gene (and its regula-tory region) is located in the chromosome 1 region forlupus susceptibility in mice and humans [reviewed in[33]]. Although it has been shown that celecoxib blocksthe active site of COX-2 by a structure–functionrelationship [34], its ability to downregulate COX-2expression in cells may be of much greater importance.

Apoptosis is critical for homeostasis of the immunesystem. Our study provides insight into the mechanismsby which celecoxib accelerates apoptosis of lupus Tcells. COX-2 expression in T cells is strongly associatedwith disease manifestation in BXSB mice (Fig. 1) and itis known that this enzyme can contribute to cellularresistance to apoptosis [17]. Whilst human lupus T cellsonly express COX-2 after activation via the TCR [14], itappears to be constitutively expressed in ex vivoperipheral T cells from the aged male BXSB mice.Thus, the presence of COX-2 in the murine T cells couldat least partly explain the prolonged survival andexcessive proliferation of autoreactive T cells seen invivo. Although murine lupus T cells do not show lessspontaneous apoptosis in vitro compared with T cellsfrom healthy animals (Table 1 and Fig. 2), they seem tobe more resistant to AICD (Fig. 1A and B).

COX-2 promotes Tcell resistance to apoptosis throughtwo different pathways: (i) the FADD/Caspase-8 pathway;and (ii) the caspase 3 and 9 pathway [35,36]. Fas (CD95)is one of the death receptors mediating AICD and itsengagement with Fas ligand (FasL) triggers Fas oligo-

merization, leading to the recruitment of FADD protein.FADD then promotes the autoactivation of procaspase8 thereby inducing apoptosis by subsequent activation ofdownstream effector caspases [37,38]. Celecoxib treat-ment increased the surface expression of FAS and FasL inmurine T cells (Fig. 4), confirming an earlier report byKaser and colleagues [38]. Antibodies against FasL canblock the binding of FasL to Fas; the latter interactionnormally leading to cell apoptosis via the FADD/caspase-8 pathway. However, anti-FasL antibodies did not evenpartially block celecoxib-induced apoptosis in vitro(Table 1), indicating that COX-2 does not play animportant role in the resistance of murine lupus T cellsto apoptosis via the FADD/caspase-8 pathway. This issupported by the fact that celecoxib treatment did notlead to activation of procaspase-8 in murine lupus T cells(Fig. 3).

There is evidence suggesting that celecoxib inducesapoptosis of human tumor cells predominantly via theintrinsic mitochondrial pathway [18,22,23]. In thispathway, loss of mitochondrial membrane potential(Δψm) leads to the release of cytochrome c, Apaf-1 andprocaspase-9 to form a complex called the apoptosome,thereby generating active caspase-9 and active caspase-3and eventually leading to nuclear disintegration [39,40].We observed a decrease in mitochondrial membranepotential in murine lupus T cells following treatment withcelecoxib, supporting the notion that COX-2 may be ableto block the intrinsic mitochondrial pathway of apoptosisin T cells. Activation of procaspases 9 and 3 in celecoxib-treated murine lupus T cells also supports this view.

Taken together, the data show that celecoxib inducesapoptosis of murine lupus T cells via the mitochondrialpathway and can inhibit anti-dsDNA production inlupus-prone BXSB mice. It is therefore proposed thatcelecoxib, a selective COX-2 inhibitor, may havepotential for use as a novel therapeutic drug for thetreatment of SLE in human subjects.

Acknowledgements

This study was supported by grants from the NationalKey Basic Research (2007CB512406) Programs andNational “863” Programs (2006AA02Z495) of China.

References

[1] Klinman DM, Steinberg AD. Inquiry into murine and humanlupus. Immunol Rev 1995;144:157–93.

[2] Shivakumar S, Tsokos GL,Data SK. Tcellα/β receptor expressingdouble-negative (CD4−/CD8−) and CD4+ T helper cells in humanaugment the production of pathogenic anti-DNA antibodies inassociation with lupus nephritis. J Immunol 1989;143: 103–12.

1421P. Yang et al. / International Immunopharmacology 7 (2007) 1414–1421

[3] Hoffman RW. T cells in the pathogenesis of systemic lupuserythematosus. Clin Immunol 2004;113:4–13.

[4] Ando AG, Sercare EE, Hahn BH. Mechanism of T and B cellcollaboration in the in vitro production of anti-DNA antibodies in theNZB×SWRmodel of lupusnephritis. J Immunol 1987;140: 2215–22.

[5] Lang TJ, Nguyen P, Papadimitriou JC, Via CS. Increased severityof murine lupus in female mice is due to enhanced expansion ofpathogenic T cells. J Immunol 2003;171:5795–801.

[6] Han SH, Li B, Chen YT, Gao XM. Isolation and functional analysisof auto-reactive T cells from BXSB mice with murine lupus. JAutoimmun 2002;19:45–54.

[7] Takeno M, Nagafuchi H, Kaneko S, Wakisaka S, Oneda K,Takeba Y, et al. Autoreactive T cell clones from patients withsystemic lupus erythematosus support polyclonal autoantibodyproduction. J Immunol 1997;158:3529–38.

[8] Busser BW, Cancro MP, Laufer TM. An increased frequency ofautoantibody-inducing CD4+ T cells in pre-diseased lupus-pronemice. Int Immunol 2004;16:1001–7.

[9] Rajagopalan, Zordan T, Tsokos GC, Datta SK. Pathogenic anti-DNA autoantibody-inducing T helper cell lines from patientswith active lupus nephritis: isolation of CD4−8− T helper celllines that express the gamma delta T-cell antigen receptor. ProcNatl Acad Sci U S A 1990;87:7020–4.

[10] Bagavant H, Fu SM. New insights from murine lupus: disassoci-ation of autoimmunity and end organ damage and the role of Tcells.Curr Opin Rheumatol 2005;17:523–8.

[11] BudagyanEG,BulanovaNI, SharovaMF,NikonovaML,StanislavAA,Yarylin. The resistance of activatedT-cells fromSLEpatients to apoptosisinduced by human thymic stromal cells. Immunol Lett 1998;60:1–5.

[12] KovacsB,VassilopoulosD,Vogelgesang SA,TsokosGC.DefectiveCD3-mediated cell death in activated T cells from patients withsystemic lupus erythematosus: role of decreased intracellular TNF-a.Clin Immunol Immunopathol 1996;81: 293–302.

[13] Sabzevari H, Propp S, Kono DH. G1 arrest and high expressionof cyclin kinase and apoptosis inhibitors in accumulatedactivated/memory phenotype CD4+ cells of older lupus mice.Eur J Immunol 1997;27:1901–10.

[14] XuL,ZhangL,YiY,KangHK,Datta SK.Human lupusTcells resistinactivation and escape death by upregulating COX-2. Nat Med2004;10:411–5.

[15] Johnsen JI, Lindskog M, Ponthan F, Pettersen I, Elfman L,Orrego A, et al. Cyclooxygenase-2 is expressed in neuroblasto-ma, and nonsteroidal anti-inflammatory drugs induce apoptosisand inhibit tumor growth in vivo. Cancer Res 2004;64:7210–5.

[16] Tucker ON, Dannenberg AJ, Yang EK, Zhang F, Teng L, Daly JM,et al. Cyclooxygenase-2 expression is up-regulated in humanpancreatic cancer. Cancer Res 1999;59:987–90.

[17] Nzeako UC, Guicciardi ME, Yoon JH, Bronk SF, Gores GJ.COX-2 inhibits Fas-mediated apoptosis in cholangiocarcinomacells. Hepatology 2002;35:552–9.

[18] Leng J, Han C, Demetris AJ, Michalopoulos GK, Wu T.Cyclooxygenase-2 promotes hepatocellular carcinoma cellgrowth through Akt activation: evidence for Akt inhibition incelecoxib-induced apoptosis. Hepatology 2003;38:756–68.

[19] OyamaK, Fujimura T, Ninomiya I,Miyashita T, Fushida S, Ohta T,et al. A COX-2 inhibitor prevents the esophageal inflammation–metaplasia–adenocarcinoma sequence in rats. Carcinogenesis2005;26:565–70.

[20] Gupta S, Adhami VM, Subbarayan M, MacLennan GT, Lewin JS,Hafeli UO, et al. Suppression of prostate carcinogenesis by dietarysupplementation of celecoxib in transgenic adenocarcinoma of themouse prostate model. Cancer Res 2004;64:3334–43.

[21] Shishodia S, Koul D, Aggarwal BB. Cyclooxygenase (COX)-2inhibitor celecoxib abrogates TNF-induced NF-kappa B activationthrough inhibition of activation of I kappa B alpha kinase and Aktin human non-small cell lung carcinoma: correlation withsuppression of COX-2 synthesis. J Immunol 2004;173: 2011–22.

[22] Wu T, Leng J, Han C, Demetris AJ. The cyclooxygenase-2 inhibitorcelecoxib blocks phosphorylation of Akt and induces apoptosis inhuman cholangiocarcinoma cells.MolCancer Ther 2004;3:299–307.

[23] Kern MA, Schubert D, Sahi D, Schoneweiss MM, Moll I, HauggAM, et al. Proapoptotic and antiproliferative potential ofselective cyclooxygenase-2 inhibitors in human liver tumorcells. Hepatology 2002;36:885–94.

[24] Murphy ED, Roth JB. New inbred strains. Mouse News Lett1978;58:51–2.

[25] Merino R, Fossati L, Lacour M, Lemoine R, Higaki M, Izui S. H-2-linked control of the Yaa gene-induced acceleration of lupus-likeautoimmune disease in BXSBmice. Eur J Immunol 1992;22:295–9.

[26] Izui S, Iwamoto M, Fossati L, Merino R, Takahashi S, Ibnou-ZekriN. The Yaa gene model of systemic lupus erythematosus. ImmunolRev 1995;144:137–56.

[27] Hogarth MB, Slingsby JH, Allen PJ, Thompson EM, Chandler P,Davies KA, et al. Multiple lupus susceptibility loci map tochromosome 1 in BXSB mice. J Immunol 1998;161:2753–61.

[28] Pisitkun P, Deane JA, Difilippantonio MJ, Tarasenko T,Satterthwaite AB, Bolland S. Autoreactive B cell responses toRNA-related antigens due to TLR7 gene duplication. Science2006;312:1669–72.

[29] Subramanian S, Tus K, Li QZ, Wang A, Tian XH, Zhou J, et al. ATlr7 translocation accelerates systemic autoimmunity in murinelupus. Proc Natl Acad Sci U S A 2006;103:9970–5.

[30] Pahlavani MA, Vargas DA. Aging but not dietary restrictionalters the activation-induced apoptosis in rat T cells. FEBS Lett2001;491:114–8.

[31] Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS,Van De Putte LB, et al. Cyclooxygenase in biology and disease.FASEB J 1998;12:1063–73.

[32] Lin MT, Lee RC, Yang PC, Ho FM, Kuo ML. Cyclooxygenase-2inducing Mcl-1-dependent survival mechanism in human lungadenocarcinoma CL1.0 cells. Involvement of phosphatidylinosi-tol 3-kinase/Akt pathway. J Biol Chem 2001;276: 48997–9002.

[33] Datta SK, Zhang L, Xu L. T helper cell intrinsic defects in lupusthat break peripheral tolerance to nuclear autoantibodies. J MolMed 2005;83:267–78.

[34] Kalgutkar AS, Zhao Z. Discovery and design of selectivecyclooxygenase-2 inhibitors as non-ulcerogenic, anti-inflamma-tory drugs with potential utility as anti-cancer agents. Curr DrugTargets 2001;2:79–106.

[35] Thornberry NA, Lazebnik Y. Caspases: enemies within. Science1998;281:1312–6.

[36] Hengartner MO. The biochemistry of apoptosis. Nature 2000;407:770–6.

[37] Wallach D, Varfolomeev EE, Malinin NL, Goltsev YV, KovalenkoAV, Boldin MP. Tumor necrosis factor receptor and Fas signalingmechanisms. Annu Rev Immunol 1999;17:331–67.

[38] Kaser A, Nagata S, Tilg H. Interferon alpha augments activation-induced T cell death by upregulation of Fas (CD95/APO-1) andFas ligand expression. Cytokine 1999;11:736–43.

[39] Irmler M, ThomeM, Hahne M, Schneider P, Hofmann K, SteinerV. Inhibition of death receptor signals by cellular FLIP. Nature1997;388:190–5.

[40] Green DR, Kroemer G. The pathophysiology of mitochondrialcell death. Science 2004;305:626–9.