A possible involvement of ion transporter in tumor necrosis factor alpha and cycloheximide-induced apoptosis of endothelial cells.

We examined the tumor necrosis factor alpha (TNFalpha)-induced apoptosis of vascular endothelial cells from the standpoint of ion channels. Cultured vascular endothelial cells from bovine carotid artery were used. Apoptosis was determined by a propidium iodide assay. Treatment of the endothelial cells with TNFalpha and cycloheximide for 6 h induced nuclear fragmentation in a TNFalpha dose-dependent manner (1-10 ng/ml). Concomitant treatment of endothelial cells with TNFalpha at a dose of 10 ng/ml and cycloheximide at a dose of 10 microg/ml elicited endothelial cell apoptosis as high as 23.4+/-4.1% at 6 h after administration. However, 10 ng/ml TNFalpha alone elicited a little apoptosis at 6 h after its administration (% apoptosis=4.1+/-0.8%). Cycloheximide (10 microg/ml) did not induce apoptosis at all. Concomitant treatment of endothelial cells with 1 mmol/l of 4,4-diisothiocyanatostilbene-2,2-disulfonic acid, which is a chloride bicarbonate exchanger blocker, partially inhibited the TNFalpha and cycloheximide-induced endothelial cell apoptosis. On the other hand, endothelial cell apoptosis due to TNFalpha and cycloheximide was completely inhibited by benzyloxycarbonyl-Asp-CH2OC(O)-2,6-dichlorobenzene (50 micromol/l), an inhibitor of caspase. Moreover, pyrrolidine dithiocarbanate, an inhibitor of nuclear factor kappa B (NF-kappaB), also suppressed endothelial cell apoptosis induced by TNFalpha and cycloheximide completely. These findings suggest that the endothelial cell apoptosis induced by TNFalpha and cycloheximide is closely related to not only chloride ions, but also both NF-kappaB and caspase activation. That is to say, there is a possibility that chloride ions or bicarbonate (pH) may play an important role in signal transduction such as NF-kappaB and caspase activation in the apoptosis induced by TNFalpha and cycloheximide.


Introduction
Apoptosis (programme d ce ll de ath) is a fundame ntal proce ss in embryogene sis, tis sue he mostasis, and immune syste m maturatio n. 1 Its dere gulation may have important implic ations for carc inogene sis and immune syste m disorde rs. It has been re porte d that apoptos is plays an important role in vario us pathophysiological c onditio ns such as hypox ia/re pe rfusion injury, radiatio n pne umonitis , etc . 2,3 Endothelial ce ll apoptos is has be en observe d w hen endothe lial cells are ex p ose d to basic fibroblast grow th fac tor depletion. 4 Lung injury such as pulmonar y e de ma due to radiation is based on e ndothelial ce ll ap optosis . 3 More ove r, lupus antic oagulant-induc ed apoptosis of endothelial c ells w ith re cognitio n of anne x in V has bee n re porte d. 5 It w as suggeste d that apoptosis of endothelial c ells may be re sponsible for the vasculitis associate d w ith syste mic lupus e rythe matos us. Thus, endothelial c ell apoptosis may c aus e the various pathologic al c onditio ns.
Tumor ne crosis fac tor a (TNFa ) is an induc er of apoptos is in vario us c ells as w e ll as an inflammatory cytokine . TNFa ac tivate d the Fas-associate d prote in w ith de ath domain (FADD) through the re c eptor of TNFa , subse que ntly caused the ac tivation of the caspase , re sulting in apoptosis. 6 TNFa re ce ptor binding prote in, TRAF2, elic ite d nuc le ar fac tor kappa B (NF-k B) ac tivation, w hich has an inhibitor y action on caspase ac tivity. 7 On the other hand, ce ramide , w hich is produc ed from sphigomye lin due to activate d sphingomyelinase by TNFa , can also induc e both NFk B and apoptos is. 8 In the pre sent study, w e show that TNFa can induc e apoptosis of endothe lial c ells c oncomitantly w ith c yclohex imide . This apoptosis w as inhibite d by an NF-k B inhibitor such as p yrrolidine dithioc arbonate (PDTC) or a c as pas e inhibitor. More ove r, w e w ill demonstrate that 4,4-diisothiocyanato stilbe ne -2,2-disulfonic acid (DIDS), a blocker of chloride bicarbonate ex change r, c an pre vent TNFa and cyclohex imideinduc ed apoptosis of e ndothe lial c ells.

Reagents
Human re combinant TNFa w as purchase d from Endogen Inc . (NY) DIDS, 5-N, N-dime thyl amiloride , propidium iodide (PI), PDTC, N-acetyl cyste ine (NAC) and anth rac ene c arbox ylic ac id w ere purchased from Sigma Co. Ltd (St. Louis , MO). Benzylox yc arbonyl-Asp-CH 2 OC(O)-2,6-dic hlorobenzene (zD-dcb) w as purifie d as described pre viously. 9 Endothelial cell culture Endothe lial ce lls w e re isolate d from fre shly ex cise d bovine c arotid arte rie s as described pre viously. 10 Brie fly, e ndothelial c ells w ere obtaine d by lightly sc raping the intimal surfac e of longitudinally op ene d vessels. The ce lls w e re the n see de d into 60 mm dishes (Falc on Labw are , Divis ion of Bec ton Dickinson and Company, Linc oln Park, NJ) in a grow th me dium containing minimum essential me dium (MEM) (Gibc o Laboratorie s, Grand Island, NY) sup ple mente d w ith 20% heat-inac tivate d fetal calf se rum (FCS) (Inte rgen Co. Purc hase , NY), pe nic illin, and stre ptomycin (2% total me dium volume, Gibc o Laboratorie s) and subculture d in a 10% FCS-containing MEM. Culture s w ere maintaine d in a humidifie d inc ubator at 37°C unde r 95% air/5% CO 2 , and w ere ide ntifie d as e ndothelial ce lls from their typical c obblestone appe aranc e and their inc orporatio n activity of ac etylate d low -de nsity lipop rote in. For ex pe rime ntal use, the e ndothelial ce lls (passage leve ls unde r 20) w e re plate d in 24-w e ll dishe s (Corning Glass Works , Corning , NY). Ce lls w ere plate d in e ight-w ell slide chambers (Nunc Inc ., He rlev, Denmark) for PI staining s.

Apoptosis
Apoptosis w as dete rmine d from the pre senc e of nuc lear fragme ntatio n on PI staining . 11 Ce lls w ere fix ed w ith 70% alcohol for 10 min at 4°C. Afte r being w ashe d w ith phosphate -buffere d saline , the cells w ere tre ate d w ith PI solution (0.01 g/l) containing RNAs e (100 mg/l) for 30 min at 37°C. Follow ing the inc ubatio n, the nuc le ar fragmentation (ove r 2) w as observe d using a fluore sc enc e mic ros cope (Nikon mic rophoto-FX, Nikon Japan Co., Tokyo, Japan), and ce lls (e ndothelial ce lls, 100 c ells) w ere c ounte d in 10 diffe re nt fie lds chosen at rando m. The p ercentage of ce lls that unde rw ent apoptosis w as calc ulate d using the formula: % apoptosis = (apoptotic ce lls/total c ounte d c ells) 3 100.

Viability assay
Ce ll viability w as e valuate d using a 3-4,5-dime thylthiazol-2-yl-2,5-diphenylte trazolium bromide (MTT) assay. 12 Brie fly, e ndothelial c ells w e re see de d into 96-w ell dishes (Corning ). The ce lls w ere tre ate d w ith TNFa (1-10 ng/ml) and cyclohex imide (10 m g/ml) for 6 h, and the MTT solution (200 mg/l) w as the n adde d. Thirty minute s late r, dime thylsulfox ide w as adde d to dissolve the c ells. The fluore sc enc e at 570 nm w as me asure d. The ce ll survival rate w as ex pre sse d according to the formula: % survival rate = [(test value -blank value )/(max imal value -blank value )] 3 100, w here the blank value re pre sents the ce ll-fre e dish, and the max imal value re pre sents the fluore sc enc e from the total cells.

Statistical analysis
Data are ex pre ssed as the mean ± standard error of the me an (SEM). For comparis ons of the me an value s betw e en tw o subsets of data, Stude nt's t-te st w as used, employing the standard c rite rion of P< 0.01 or P< 0.05 to indic ate statis tic ally signific ant diffe re nc es.

TNFa -induced apoptosis of endothelial cells concomitantly with cycloheximide
We ex amine d the effe ct of TNFa (1-10 ng/ml) on endothelial ce lls. Contro l endothe lial c ells displaye d a typic al cobblestone morphology, but the ir tre atme nt w ith 1-10 ng/ml of TNFa for 6 h caused a little ce ll shrinkage . Concomitant tre atment w ith TNFa (1-10 ng/ml) and c yc lohex imide (10 m g/ml) c ause d se vere ce ll shrinkage for 6 h. To de te rmine w hether the c ell damage w as due to apoptosis or not, w e staine d the c ells w ith propidium iodide . The nuc lei of the contro l e ndothe lial c ells w ere round, w here as the TNFa (10 ng/ml) and cyclohex imide (10 m g/ml) tre ate d ce lls re veale d nuc lei that w e re fragmente d. The dose de pende nc y of the TNFa effec t on the apoptosis is illustrate d in Fig. 1. Cyclohex imide (10 m g/ml) give n alone affe cte d ne ither the morphology nor the apoptos is induc tion (Fig. 1). nuc lear fragmentation of the cells, as demons trate d in Fig. 2A. PDTC (1 mmol/l) give n alone affe cte d ne ither the morphology nor the apoptosis induc tion. PDTC suppre ssed the TNFa and cyclohex imide -induc e d endothelial c ell apop tos is in a dose -de pende nt manne r, as show n in Fig. 2A. PDTC also re ve rsed the decre as ed survival rate caused by TNFa and cycloheximide in a dose-depe nde nt manne r (Fig. 2B). PDTC give n alone did not influenc e the c ell survival rate . NAC, anothe r inhibitor of NF-k B, did not suppre ss the TNFa and cyclohex imide -induc e d apoptos is (NAC 1 mmol/l; % inhibition= 0.3±4.0%).

Prevention by the NF
Effect of the sphigomyelinase inhibitor, L-cycloserine, and the tyrosine phosphatase inhibitor, vanadate, on endothelial cell apoptosis induced by TNFa and cycloheximide Conc omitant tre atme nt of endothe lial c ells w ith L-c yclose rine (an inhibitor of sphingomyelinase, 1-10 mmol/l) did not inhibit the TNFa and cycloheximide -induc ed nuc le ar fragme ntatio n of the c ells, as demonstrate d in Fig. 3A. The phosp hatas e inhibitor, sodium vanadate (1 mmol/l), complete ly inhibite d the TNFa and c yclohex imide -induc ed apoptosis , as show n in Fig. 3B.
Protective effect of the caspase inhibitor, zD-dcb, on endothelial cell apoptosis induced by TNFa and cycloheximide Conc omitant tre atment of endothe lial ce lls w ith zDdcb (a c as pas e inhibitor, 50 m mol/l) c omple te ly inhibite d the TNFa and c yclohex imide -induc e d nuc lear fragmentatio n of the ce lls, as demonstrate d in Fig. 4A. zD-dc b suppre ssed the TNFa and cyclohex imideinduc ed endothe lial c ell apoptosis in a dose-depe ndent manne r, as show n in Fig. 4A. zD-dcb also re ve rse d the decre as ed survival rate c ause d by the TNFa and cyclohex imide in a dose-depe nde nt manne r (Fig.  4B).

Effects of ion transport inhibitors on endothelial cell apoptosis induced by TNFa and cycloheximide
Conc omitant tre atme nt of endothe lial c ells w ith DIDS (a chloride bicarbonate ex change r blocker, H. Fu jita et al.  Fig. 5A. On the othe r hand, the sodium proton antip orte r blocker, dime thyl amiloride (0.1 mmol/l), ex erte d no effect on the TNFa and cyclohex imide -induc e d apoptos is (Fig. 5B). The chloride ion p ump inhibitor, anthrac e ne carbox ylic acid (1 mmol/l), did not inhibit the TNFa and cyclohex imide -induc e d ap optosis of endothe lial c ells (Fig. 5C).

Discussion
TNFa and cycloheximide-induced apoptosis of endothelial cells was due to NF-k B activation and was different from TNFa -induced apoptosis TNFa has be en re porte d to be an induc er of apoptosis in some kinds of ce lls. This substanc e induc ed DNA ladde ring in tumor c ells, and the c ells displaye d a typic al morphologic al appe aranc e of apoptosis . 13 In the pre sent study, w e first demonstrate d apoptosis due to TNFa in e ndothelial ce lls from the vie w point of ion trans port. Apoptosis of vasc ular e ndothelial ce lls may play an important role in the deve lopme nt of inc re ased vascular perme ability and c apillary leak syndro me during syste mic inflammatory re sponse syndro me. Irradiatio n can also elic it a vascular pe rmeability inc re as e in rat lungs, re sulting from apoptosis of the pulmonar y endothe lial ce lls as judge d from the pathologic al finding s. 3 Rec ently, lupus antic oagulantinduc ed apoptosis of endothe lial c ells w ith re cognition of anne x in V has bee n re porte d. 5 It w as suggeste d that apoptosis of endothe lial c ells may be re spons ible for the vasc ulitis assoc iate d w ith syste mic lupus erythe matos us. The mechanism w he re by TNFa induc es apoptos is w as not fully unde rstood. Clarifying the me chanism w ill re ach the goal to tre at the vascular dis ease such as vasculitis. In ge ne ral, it is know n that TNFa elic ite d apoptosis through its re c eptor, 6 subsequently stimulating the death domain, FADD, and activating c as pas e activ ity, re sulting in ap optos is . 6 On the other hand, TNFa also activate s the sphyngomyelinase and produc es the ce ramide , w hich is capable of ac tivating the kinas e.

(A) (B)
The ce ramide -activating kinas e stimulate s NF-k B ac tivation. 8 TNFa -induc ed apoptosis through the FADD is inde pe nde nt on NF-k B activatio n. 7 How e ve r, TNFa can induc e apop tos is through the NF-k B ac tivation in the c onc omitant pre se nce of c yclohex imide . 14 We che cke d the invo lve ment of NF-k B in TNFa and cyclohex imide -induc e d apoptosis (Fig. 2). In the endothelial ce lls from bovine c arotid arte ry, TNFa and cyclohex imide induc ed apoptos is through PDTCse nsitive NF-k B (Fig. 2). How eve r, anothe r inhibitor of NF-k B, NAC, did not supp re ss the apoptosis induc e d by TNFa and cyclohex imide . Bessho e t a l. also re porte d that etoposide -induc e d apoptosis of HL-60 ce lls and thymocyte s w as inhibite d by PDTC, but not NAC. 15 In our pre liminary data, hydrogen perox ide and cyclohex imide -induc ed apoptosis of e ndothelial ce lls w as inhibite d by e ithe r PDTC or NAC (data not show n). Thus, TNFa and cyclohex imide -induc e d apoptos is is diffe re nt from TNFa -induc ed apoptosis from the point of the invo lvement of NF-k B. Ne vertheless, e ndothelial c ell apoptos is induc ed by TNFa and cyclohex imide as w ell as TNFa alone , is re sponsible for the caspase activatio n (Fig. 4).
On the othe r hand, the re ason w hy cyclohex imide enhanc e s the TNFa -induc e d apoptos is re mains unc lear. While TNFa dire ctly c ause s apoptosis of tumor ce lls, normal ce lls are gene rally re sistant. How e ver, most re sis tant c ells, inc luding vasc ular endothelial cells, can be re nde re d susce ptible to TNFa by inhibiting RNA and prote in synthe sis. 16 This suggests that TNFa provide s a ce ll survival signal in additio n to a death signal. Ce ll survival fac tors induc ed by TNFa are BCl-2 analo gue , A1, FGF-1 and nitric ox ide (induc ible nitric ox ide synthas e). 17,18 Cyc lohex imide may inhibit the prote in synthe sis of their antiap optotic fac tors and be able to easily enhanc e the TNFa -induc ed apoptosis.
TNFa and cycloheximide-induced apoptosis was related to the tyrosine phosphatase, but not to the sphyngomyelinase TNFa also activate s the sphyngomyelinase and produc es the ceramide , w hich is capable of ac tivating the kinas e. The ce ramide -activating kinas e stimulate s NFk B ac tivation. 8 Ne ve rthe less, TNFa and cyclohex -H. Fu jita et al. imide -induc ed apoptosis w as not inhibite d by an inhibitor of sp hyngomyelinase (Fig. 3A), show ing that the apoptosis w as inde pende nt of the ac tivation of sphingomyelinase through the TNFa re ce ptor in our assay. Slow ik e t a l. al3o re porte d that TNFa -induc e d apoptos is w as not re late d to the ce ramide through the sphingomyelinase ac tivation. 19,20 In additio n, Modur e t a l. re porte d that ceramide produc tion in the endothelial ce lls tre ate d w ith TNFa did not ac tivate NF-k B. 21 The se re ports may sugge st the agre ement on our data. Tyros ine kinas e ac tivation, as w ell as a grow th fac tor, is a survival fac tor. Tyrosine kinas e inhibitors such as he rbimysin or ge niste in caused the apoptosis of leuke mic ce lls, w hich w as re scued by the tyros ine phosp hatas e inhibitor, sodium vanadate . 22,23 TNFa and cyclohex imide -induc e d apoptosis w as c omple te ly inhibite d by sodium vanadate , as show n in Fig. 3B. The me chanis m by w hich vanadate p re vente d e ndothelial ce lls from TNFa and c yclohex imide -induc ed apoptosis is unc lear. We sp eculate tw o possibilitie s: (1) vanadate may have an action on activatio n of tyrosine kinas e, re sulting in ce ll survival; (2) vanadate may c ause a marke d dec re ase in p53, the endoge nous apoptotic induc er, as described by the pre vious re port. 24 TNFa and cycloheximide-induced apoptosis was involved in the chloride bicarbonate exchanger In the pre sent study, w e atte mpte d to use a blocker of chloride bic arbonate ex change r to inhibit TNFa and cyclohex imide -induc e d apoptosis. In orde r to maintain cell shape, it is ne ce ssary to re gulate the osmotic pre ssure and ion influx . We first c ons ide re d the chloride channe l, sinc e this channe l is re sponsible for re gulatio n of the anionic p rope rtie s and intrac ellular pH. More ove r, chloride ion influx influenc es the shrinkage of ce lls. 15 A chloride bicarbonate ex change r is pre se nt in vario us kinds of c ells inc luding ventric ular myoc yte s, gastric ce lls, and re nal tubular ce lls. [26][27][28] DIDS, one of the chloride bicarbonate ex change r blockers, pre vente d TNFa and cycloheximide -induc ed apoptos is in e ndothelial c ells. In addition, anthrac ene c arbox ylic acid, w hich inhibits the chloride influx via the blockade of chloride channe l, did not affe c t the TNFa and cyclohex imide -induc e d apoptos is (Fig. 5C). There is one possibility that TNFa and cyclohex imide affe ct the func tion of the chloride bicarbonate ex change r, the re by e lic iting a change in the intrac e llular chloride ion le ve l as w ell as the intrac e llular bicarbonate le ve l. Apoptosis c ould then be induc ed by chloride ion e fflux , follow e d by a change in the intrac ellular p H. In our unpublishe d data, DIDS inhibite d the endothelial ce ll apoptosis and an inc re as e in pH, induc e d by stauro sporine , anothe r apoptos is induc e r. Stauro sporine change d the intrac e llular pH from 6.9 to 7.4 through a chloride bicarbonate ex change r (unpublished data). The change in pH doe s occ ur prior to the caspase activatio n (unpublished data). We spe culate that endothelial cell apoptosis induc ed by TNFa and cyclohex imide as w ell as stauros p orine may be due to the activatio n of a chloride bic arbonate ex change r; chloride efflux and bicarbonate influx (an inc re as e in pH). We conside r that pH change due to a chloride bicarbonate ex change r may be invo lved not in NF-k B activatio n, but caspase activatio n, although w e have no dire c t proof. On other hand, Li and Eastman 29 demonstrate d that inte rleukin-2-stimulate d lymphocyte s induc ed apoptosis of targe t ce lls via intrac ellular acidosis through the sodium proton trans porte r. We spe culate , the re fore , that one of the c ommon pathw ays of apoptosis may be via a sodium proton trans porte r. How e ve r, the sodium proton trans porte r blocker, dimethyl amiloride , did not affe ct the TNFa and c yclohex imide -induc ed apoptos is (Fig. 5B). In our assay syste m, a sodium proton trans porte r w as e vide ntly not re late d to TNFa and cyclohex imideinduc ed apoptosis .
Curre ntly, w e are atte mpting to e luc idate more pre c isely the me chanism re sponsible for the ap optosis , w hich is re late d to vario us ion pumps, inc luding an ex aminatio n of the role of chloride ion and intrac e llular pH.