binding (Acidic Fibroblast) Growth Factor - The Journal of Biological ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 266, No. 35, Issue of December 15, pp. 24059-24063,1991 Printed in U. S.A.

0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Cyclooxygenase Gene Expressionis Down-regulated by Heparinbinding (Acidic Fibroblast) Growth Factor-1 in Human Endothelial Cells* (Received for publication, April 17, 1991)

Timothy HlaS and Thomas Maciag From the Laboratory of Molecular Biology, Jerome H.Holland Laboratoryfor the Biomedical Sciences, The American Red Cross, Rockville, Maryland 20855

glandin synthesis, partlydue to theirreversible self-inactivating property of the enzyme which occurs aftera limited inflammatory prostanoids,is regulated by growth fac- amount of catalysis (4).The enzyme, once self-inactivated, is tors and cytokines. We have shown previously that the degraded rapidly (5) and the half-life of the enzyme activity cytokine interleukin- 1, an inhibitor of endothelial has been estimated to be in the order of minutes (6). Cox growth in vitroand stimulatorof prostacyclin produc- enzyme activity is regulated by a variety of cytokines, tumor tion, induces the expressionof the 3-kilobaseCox tran- promoters, and growth factors such as interleukin (1L)-la (71, script in cultured human umbilical vein endothelial epidermal growth factor (8),transforming growth factor-@(9) cells (HUVEC). In contrast, the endothelialcell mito- and phorbol myristic acetate(PMA)(6).Both IL-la and gen, heparin-binding (acidic fibroblast) growth factor1 (HBGF-1) inhibits the synthesis of prostacyclin in PMA are known to induce the synthesis of the Cox in human HUVEC. Inthisreport,we describetheeffect of umbilical vein endothelial cells (HUVEC, 6, 10-12), and the HBGF-1 on Cox mRNA expression in HUVEC. Cells potential intracellular activity of IL-la as an inhibitor of cultured in the presence of HBGF- 1express diminished human endothelial cell proliferation has been implicated as a Cox mRNA levels whereas quiescent cells maintained key regulatory feature of HUVEC senescence in vitro (13). in serum expressed a 7-fold higher level of the tran- These and other studies (reviewed in Ref. 14) suggest that script for Cox. Concomitantly, the level of the Cox factors involved in the regulation of vascular growth, inflamtranslation product and prostacyclin synthesis are also mation, or tissue repair induce an exaggerated level of prosreduced by HBGF-1.Further, HBGF- 1,in the presence tanoid synthesis by the up-regulation of cyclooxygenase. of heparin, down-regulates thelevels of the Cox tranThe endothelium is involved in many important physiologscript ina dose- and time-dependent manner. Theonset ical functions including thrombosis and hemostasis, control of action of HBGF-1 is slow, requiring up to 24 h to of vascular tone, inflammation, andangiogenesis (reviewed in depress the level of Cox mRNA. Lastly, the effective Ref. 15). Cultured human vascular endothelial cells express dose of HBGF- 1to depressCox mRNAlevels is similar many of the in vivo characteristics of the vascular endothelium to that required for mitogenesis, suggesting that cell in vitro and provide a valuable model system. Cultured HUproliferation may be required for the reduction of Cox expression in vitro.Thus, Cox may belong to a class of VEC exhibit a requirement for exogenous heparin-binding genes that are reversibly down-regulated during pe- (acidic fibroblast) growth factor (HBGF)-1 (16-18) and inthe absence of growth factor, fail to proliferate in vitro (18). riods of endothelial cell proliferation. HBGF-1 is recognized as aprototypic member of the heparinbinding (fibroblast) growth factor family of polypeptides (19). HBGF-1, like its structural homolog basic fibroblast growth factor, interactswith high affinity cell surface receptors which Arachidonic acid is converted into inflammatory and vasoactive prostaglandins, prostacyclin, and thromboxaneby the possess intrinsic tyrosine kinase activity (20). Thus, various cyclooxygenase pathway (1).In the vasculature, synthesis of intracellular polypeptides are phosphorylated on tyrosine resprostacyclin by the endothelial cells is importantin maintain- idues immediately after HBGF bindsto itsreceptors (19, 20). ing normal hemostasis and vascular tone (2). Prostaglandin In addition, the phosphatidyl inositol cycle and the protein synthesis is regulated by the activity of twokey enzymes, kinase C pathway may be activated as well (21). However, the namely, phospholipase and cyclooxygenase (3). Cyclooxygen- exact mechanisms bywhich HBGF-1 induces the cells to ase (Cox)' is thought to be the rate-limiting step in prosta- traverse through the cell-cycle remain unclear. The HBGF prototypes are the most potent inducers of HUVEC prolifer* This research was supported by National Institutes of Health ation in vitro, and this is thought to be the primary cellular Grants HL32348 and 35627 (to T. M,). The costs of publication of mechanism involved in the regulation of angiogenic phenomthis article were defrayed in part by the payment of page charges. ena in vivo (16). Studies on prostanoid synthesis in HUVEC This article must therefore be hereby marked "advertisement" in conducted by us (22), as well as others (23, 24), have demonaccordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Supported by a post-doctoral fellowship from the Arthritis Foun- strated that HUVEC grown in the presence of HBGF-1 prodation. duced diminished amounts of prostanoids. To examine the The abbreviations used are: Cox, cyclooxygenase; HBGF-1, hep- possibility that HBGF-1may down-regulate the expression of arin-binding growth factor-1; PMA, phorbol myristic acetate; IL-1, the Cox gene, we developed a highly sensitive reverse traninterleukin-1; HUVEC, human umbilical vein endothelial cells; RTscriptase-polymerase chain reaction (RT-PCR) assay to PCR, reverse transcriptase-polymerase chain reaction; PG, prostaglandin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; TNF, measure the steady-state levels of mRNA for Cox (12). We demonstrate that Cox mRNA levelsare depressed by HBGFtumor necrosis factor; FBS, fetal bovine serum.

The expression of cyclooxygenase (EC 1.14.99.1,

Cox), a rate-limiting enzyme in the biosynthesis of

24059

24060

Suppression of Cyclooxygenase Gene Expression HBGF-1 by

1 in a dose- and time-dependent manner and thatcell proliferation may be necessary for this event to occur. EXPERIMENTAL PROCEDURES

were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions, transferred electrophoretically to nitrocellulose filters, and reacted with 2pg/ml of anti-ovine Cox IgG as described previously (12, 28). Prostacyclin Synthesis-HUVEC cultures were grown to confluence in 12-well cluster dishes, starved, and either treated or not treated with 10 ng/ml of HBGF-1 for 48 h. The cultures were washed and incubated in serum-free "199 containing increasing amounts (0-10 p ~ of) arachidonic acid (Sigma) for 10 min at 37 "C. The culture supernatantswere then analyzed for 6-keto-PGF1,, the stable hydrolytic product of prostacyclin, by radioimmunoassay (Amersham Corp.).

Cell Culture-Human umbilical vein endothelial cells were kindly provided by Dr. Michael Gimbrone (Harvard Medical School). Cells were grown in an environment containing Medium (M)-199,10% (v/ v) fetal bovine serum (FBS), 5 units/ml heparin (Upjohn), and 150 pg/ml of crude HBGF-1 as described previously (16). Cells between the passages of 4-10 were utilized in the experiments described herein. Confluent cultures were starved in "199 containing 10% FBS for 48 h before the experiments were initiated. Recombinant HBGF-la was purified by heparin-Sepharose chromatography and reversedRESULTS phase high performance liquid chromatography as described previExpression of Cyclooxygenase mRNA in Quiescent and Proously (25).Growth factors were added to "199 containing 10% FBS liferating HUVEC-For optimal growth in culture, HUVECs and 5 units/ml heparin. Purification of RNA-At indicated times, cells were rinsed twice were cultured in"199 containing FBS, heparin,and HBGFwith phosphate-buffered saline, lysed in 4 M guanidinium isothiocy- 1. Under these conditions, the population doubling time is anate solution, and RNA purified as described previously (26). Briefly, approximately 48 h (16). At confluence, the HUVEC populacell lysates were extracted with phenol and chloroform, isopropanol tion becomes growth arrested or quiescent when HBGF-1 was precipitated, and DNA digested to completion with DNase I (RNasefree) (Bethesda Research Laboratories) (B. R.L.). RNA was then removed from the medium for 72 h, and under these condipurified by phenol/chloroform extraction and ethanol precipitation. tions less than 20% of the cells synthesize DNA (29). Previous The ratio of absorbance of RNA a t 260 and 280 nm was checked for studies have shown that growth factor-treated HUVEC posa value of 22. The integrity of the RNA was assessed by agarose gel sess a reduced capacity to recover prostanoid synthesis after electrophoresis and ethidium bromide staining. irreversible inactivation of Cox by aspirin (22). In addition, RT-PCR-Total RNA (1pg) from HUVEC was reverse transcribed Weksler (24) and Hasegawa et al. (23) have demonstrated into cDNA by incubating with 200 units of maloney leukemia reverse transcriptase (B. R. L.) in a reaction buffer that contained 50 mM that HUVECs grown in the presence of crude HBGF-1 synTris. HCI, pH 8.0, 1mM dithiothreitol, 15 mM NaCl, 3 mM MgC12, 10 thesized less prostacyclin, PGEz and PGF2, than untreated units RNasin (Promega), 0.2 pg of random hexamers, 0.8 mM dNTPs cells. These studies indicated that Cox levels may be downand incubated for 1 h at 37 "C. The reaction was stopped by heating regulated in human endothelial cells treated with HBGF-1. a t 95 "C for 10 min, dilutedto 250 pl, and 5-p1 aliquots were used for However, direct measurement of the Cox mRNA or protein PCR amplification with either cyclooxygenase- or GAPDH-specific levels were not reported inthese studies. To determine primers. The sequences of the sense and antisense primers for Cox were: (Cox-1 sense) = 5'-TGC CCA GCT CCTGGC CCG CCG CTT- whether the reduced synthesis of prostanoids in HUVECs 3', (Cox-1 antisense) = 5'-GTG CAT CAACACAGG CGC CTC treated with HBGF-1 was due to the inhibition of Cox gene expression, we measured the steady-state levels of the Cox TTC-3', hfl (sense) = 5'-GCT GGG AGT CTT TCT CCAACG TGA G-3', hrl (antisense) = 5'-GGC AAT GCG GTT GCG GTA mRNA. The low abundance of the Cox mRNA necessitates TTG GAA CT-3'. The primer pairs Cox-1 sense/antisense or hfl/ the use of poly(A)+ RNA preparations for the successful h r l were utilized and were found to yield the amplified product of the detection by Northern blot analysis. As shown in Fig. 1, the expected sizes, 0.3 and 0.7 kilobases, respectively. Southern blot analysis using a human Cox cDNA probe (27) confirmedthe identity Cox cDNA probe detected a major band at 3 kilobases and a of the amplified products. The sense and antisense primersused for less prominent 5.3-kilobase band. Quiescent cells contained GAPDH were as follows: 5'-CCA CCC ATG GCA AAT TCC ATG 7-fold more Cox mRNA than HBGF-1-treated HUVEC, inGCA-3' and 5'-TCT AGA CGG CAG GTC AGG TCC ACC-3'. All dicating that HBGF-1 down-regulates the Cox mRNA levels. three pairs of primers were derived from two different exons. PCR In order to assess the Cox mRNA levels with relative ease, was performed in a reaction buffer (50 p l ) containing 50 mM Tris. we used the RT-PCR method, which is more sensitive than HCI, pH 8.0, 1.5 mM MgClz, 10 mM KCI, 0.2 mM dNTPs, 0.5 pg each of sense and antisense primers, and 2.5 units of cloned Taq polym- Northern blot procedures. Total RNA waspurified from either erase (Perkin-ElmerCetus). For quantitative purposes, the reactions quiescent or HBGF-1-treated cells and analyzed for the levels were amplified by 25 repetitive cycles of denaturation a t 94 "C for 1 ofCox mRNA. Aliquots of amplified cDNA were taken at min, annealing a t 58 "C for 2 min, and extension a t 72 "C for 3 min. cycles 20-30, analyzed by agarose gel electrophoresis, subUnder these conditions, the yield of the amplified product was linear jected to Southern blot analysis, and the respective bands with respect to the amount of input RNA for the specific pair of primers. Furthermore, the assay can resolve a 2 fold difference in the + HBGFl amount of input RNA. Amplified cDNAs were separated by agarose gel electrophoresis, Southern blotted with the radiolabeled human Cox-1 cDNA probe as described in the Northern blotting section, and autoradiographed. Quantitation of the bands were achieved by liquid scintillation counting of the respective area on the membrane. Northern Blot Analysis on HUVEC Poly(A)+RNA-Poly(A)+ RNA was purified by affinity chromatography on oligo(d)-T cellulose. Two cox pg of poly(A)+RNA was separated on formaldehyde agarose gels, and Northern blot analysis was conducted as described previously (26). The filters were hybridized under high stringency (65 "C, 20% formamide) using radiolabeled hCox-1 or GAPDH cDNA probes. The GAPDH human Cox-1 cDNA probe extended from amino acid 1 to 378, as numbered by Yokoyama and Tanabe (50). FIG.1. Northern blot analysis of Cox transcript expression Western Blot Analysis-HUVEC monolayers (15-cm dishes) were washed and scraped into ice-cold phosphate-buffered saline solution in proliferating and quiescent HUVEC. Poly(A)' RNA (2 pg) and pelleted. Cox was solubilized from cell pellets by sonication in a from cultured HUVEC maintained in media containing 10%FBS and buffer that consisted of 50 mM Tris.HC1, pH 8.0, 10 mM EDTA, 1 5 units/ml heparin, with or without 10 ng/ml of HBGF-1 for 2 days mM phenylmethylsulfonyl fluoride, 1.5 mM diethyldithiocarbamate, was analyzed by Northern blot analysis using the human Cox or and 1%(v/v) Tween-20. The sonicate was centrifuged a t 100,000 X g GAPDH probes. The filters were hybridized and washed under high for 30 min to remove insoluble material, and the supernatants were stringency and exposed to x-ray film. Arrows indicate 28 S and 18 S analyzed by Western blot analysis. Solubilized cell extracts (200 pg) RNA size markers.

-

om

Suppression of Cyclooxygenase Gene Expression by HBGF-1 quantified by liquid scintillation counting. As shown in Fig. 2, linear accumulation of amplified cDNA was observed between cycles 24-30. These results suggest thatthe gene expression of the Cox transcript is down-regulated in HBGF1-treated HUVECs. The RT-PCR data were in agreement with the data obtained by poly(A)+ Northern blot analysis. HBGF-1-treated cells contained 6-7-fold less Cox mRNA at all cycles examined. Effect of HBGF-1 on Cyclooxygenuse mRNA Levels in Quiescent HUVEC-It is known that HBGF-1 is a potent mitogen for HUVECs, thus thedown-regulation of the Cox transcript may be linked to the cell cycle. The ability of HBGF-1 to down-regulate the Cox transcript also occurs in a dose-dependent fashion. As shown in Fig. 3, as little as 0.1-1 ng/ml of HBGF-1, in the presence of heparin, reduced the level of Cox mRNA significantly. Further,the concentration of HBGF-1 required to depress Cox gene expression was similar to thatrequired to induce DNA synthesis (16, 19). To determine the time dependence of the HBGF-1 effect on the expression of CoxmRNA, quiescent cultures of HUVEC were treated with 10 ng/ml of HBGF-1 and5 units/ml of heparin for different periods of time (24-96 h), and the total RNA from these cultures was assayed for a qualitative estimate of the steady-state Cox mRNA level.As shown in Fig. 4, the level of CoxmRNA remained consistently higher in quiescent cultures whereas HBGF-1 treatment significantly reduced the level of the Cox transcript. The difference in Cox mRNA steady-state levels between HBGF-1-treated and untreated cells was maintained up to96 h. HBGF-1 treatment of quiescent HUVECs for less than 20 h did not result in a depression of Cox mRNA levels (data notshown). Effect of HBGF-1 on Immunoreactive Cyclooxygenuse Levels in HUVEC-To determine whetherthe effect of HBGF-1 on HUVECs resulted in the reduced levels of the Cox translation product, we examined the steady-state levels of the Cox protein by immunoblot analysis of HUVEC proteins. Quiescent cultures of HUVECs were either treated or not treated with 10 ng/ml HBGF-1 and 5 units/ml heparin for 24 and 72

00

EZ E R

B -i

m.

I.DXO!

I m 8 .

?ma! Iwa,

!I

Z 2f

Xgg

nglml HBGFl cox

B

0

0.1

10

1

DoaHBOFI (ne/M

FIG. 3. Dose dependence of Cox mRNA suppression by purified HBGF-1. Quiescent cultures of HUVEC were treated with indicated amounts of purified HBGF-1 in "199 containing 10%FBS and 5 units/ml heparin for 24 h. Total RNA was then purified and analyzed for the presence of Cox (Cox sense and antisense primers) and GAPDH by the RT-PCR method as described. PCR amplification was run for 25 cycles. Products were separated on a 1.2% agarose gel, Southern blotted with the respective cDNA probes, and autoradiographed for 15 min ( A ) . The radioactivity from the bands were quantitated by liquid scintillation counting and the ratio of COXto GAPDH bands were plotted ( B ) .

cox

GAPDH 1

1

DAY

1

2

3

14

1

2

" "

HBGF-1

+,+=+=+_I

i

3

4

rvzT

" " " "

(kb) 2.0

-

0.5 0.4

-

A cycles

--

0 .1 10 1

A

24061

FIG. 4. Sustained suppression of Cox mRNA by HBCF-1 in HUVEC. Quiescent cultures of HUVEC were either treated or not treated with 10 ng/ml of HBGF-1 in "199 which contains 10%FBS +HBGF 1 ..HBGF 1 and 5 units/ml heparin. At indicated days, cells were harvested, total RNA was purified and analyzed for the presence of Cox (hrl and hfl primers) and GAPDH for 25 cycles by the RT-PCRmethod. Aliquots pl) of the amplification were separated by 1.2%agarose gel (10 I " . , . , electrophoresis and stained with ethidium bromide for qualitative , H6GF.l. , comparison.

/I"'

"

"-E."-

/

h, and the relative levels of the Cox polypeptide were determined by Western blot analysis with an anti-Cox antiserum c@=l (28). As shown in Fig. 5, the 70 Kd immunoreactive level of FIG. 2. RT-PCR analysis of Cox mRNA expressionin pro- the Cox translation product was higher in the cells that were liferating and quiescentHUVEC. Panel A , quiescent cultures of not treated with HBGF-1. These data suggest that HBGF-1 HUVEC were either treated or not treated for 48 h with 10 ng/ml of HBGF-1 in "199 which contains 10% FBS and 5 units/ml heparin. suppressed the levels of both Cox protein and mRNA in HUVECs. Total RNA was then purified and 1 pg was used in the RT-PCR Effect of HBGF-1 on Prostacyclin Synthesis by HUVECassay for the Cox transcript using Cox-1 sense and antisense primers. At indicated cycles of PCR amplification, 10-pl aliquots were taken, To determine whether HBGF-1 also reduced Cox activity in separated on a 1.2% agarose gel, and Southern blotted with the Cox- HUVECs, quiescent cultures were treated with 10 ng/ml of radiolabeled cDNA probe. The autoradiogram was exposed for 15 HBGF-1 for 48 h. Control and HBGF-1-treated cells were min. Panel B, the autoradiogram from panel A was usedto locate the positions of the bands on the nylon filter, the respective areas were subsequently stimulated with increasing amounts of arachicut out, and radioactivity was determined by liquid scintillation donic acid for 10 min and prostacyclin synthesis (6-ketocounting. The yield of the Cox cDNA wasplotted with respect to the PGFJ was then measured. As shown in Fig. 6, the HBGF-1treated HUVECs synthesized significantly less (40% of concycle number. 0 1

.

22

24

m

28

SI

24062

Suppression of Cyclooxygenase Gene Expression by HBGF-1 24 hr. 7 2 hr.

(40) suggested that theactivity of Cox is restored by de m v o synthesis within 8-24 h in HUVEC. These studies reported that HBGF-treated proliferating cells exhibited diminished . capacity to recover Cox activity compared to quiescent cells (22). While these and other studies (23, 24) indicated that a 7 0 kd Cox enzyme activity is down-regulated by crude preparations of HBGF-1, the mechanism by which HBGF-1 suppresses Cox activity is unclear. Because cultured HUVECs express low levels of the Cox transcript, we developed a highly sensitive RT-PCR assay to measure Cox mRNA levels fromtotal RNA preparations. The FIG.5. Effect of HBGF-1 on Cox polypeptide levels. Quies- results obtained from the RT-PCR assay are in agreement cent cultures of HUVEC were either treated or not treated with 10 with those from Northern blot analysis of poly(A)+ RNA. ng/ml of HBGF-1 for 24 or 72 h, harvested, and analyzed for the levels of Cox protein by Western blot analysis using anti-CoxIgG as This assay is able to resolve a 2-fold difference in the amount of input RNA, and we utilized this assay to demonstrate that described under “Experimental Procedures.” IL-la induces the expression of Cox mRNA as animmediateearly gene (12). In contrast, proliferating HUVECs that are --- I I grown in media containing HBGF-1 express a diminished level of Cox mRNA, suggesting that HBGF-1 may directly down-regulate the mRNA for Cox. Indeed, HUVECs treated with purified HBGF-1 in the presence of heparin exhibited a dose- and time-dependent reduction in Cox mRNAlevels. The concentration of HBGF-1 required to suppress Cox mRNA wassimilar to thatrequired to induce HUVEC proliferation in vitro (16,18). In addition, these data areconsistent with the repression of the Cox translation product and enzyme FIG. 6. Effect of HBGF-1 on Cox enzyme activity in HU- activity by HBGF-1. Further, the kinetics ofCoxmRNA VEC. Confluent cultures of HUVEC were maintained in medium suppression (224 h) correlate with the HUVEC population 199,10% FBS, 5 units/ml of heparin with or without 10 ng/ml HBGF- doubling time. Thus, it is possible that down-regulation of 1 for 48 h. Duplicate cultures were then washed and incubated in Cox mRNA by HBGF-1 is associated with the proliferative plain medium containing increasing amounts of arachidonic acid for activity of the growth factor. 10 min. The culture supernatants were then analyzed for the presence Endothelial cell proliferation occurs during angiogenesis of 6-keto-PGF1, as described under “Experimental Procedures.” associated with wound healing processes as well as in pathotrol) prostacyclin. This datasuggests that Cox activity is also logic situations such as rheumatoid arthritis (41). HBGF-1 protein has been detected insitu by immunocytochemical suppressed in HBGF-1-treated HUVEC. methods in angiogenic tissues including fibroblasts, epidermis, blood vessels,and macrophages (41). Thus, it is possible DISCUSSION that enhanced production of HBGF-1 may limit the local The regulation ofCox enzyme activity in the vascular production by the proliferative endothelium of prostacyclin, endothelium is thought to be under the control of cytokines a potent vasodilator and an inhibitor of platelet aggregation. such as IL-la, a potent inhibitor of human endothelial cell It is also of interest that HBGF-1 has been shown to inhibit growth (30). IL-la has been shownto induce the synthesisof the expression of plasminogen activator inhibitor-1, and plasprostacyclin (7), stimulate the de mu0 synthesis of Cox pro- minogen activator inhibitor-1, like Cox, is inducible by IL-la tein (10,11), andto induce the expression of the Cox mRNA (42,43). The relationship between HBGF-1 and IL-la as as an immediate-early gene (12). Tumor necrosis factor inducers of HUVEC growth and differentiation, respectively, (TNF), a cytokine that mimics many of the biological activi- is especially interesting since both polypeptides lack a signal ties of IL-la in human endothelial cells, also stimulates sequence for secretion (44,45), andboth exhibit similar prostacyclin synthesis in HUVECs (31). In addition, the crystallographic structures (46,47). In addition, intracellular tumor promoter, PMA, stimulates prostacyclin synthesis and forms of both polypeptides have beenimplicated as regulators induces the de m u 0 synthesis of Cox protein in HUVEC (6) of proliferation (25) and senescence (13), and both polypepand fibroblasts (10). It is noteworthy that IL-1, TNF, and tides as well as their family members have beenidentified in PMA are potent antagonists of growth factor-stimulated the nucleus (41, 48, 49). Whether these events are related to endothelial cell proliferation (32, 33). However,PMA and the induction and repression of HUVEC Cox mRNA levels TNF induce an angiogenic response in vivo (32,34). IL-1 and by IL-la and HBGF-1 is presently not known. Thus, Cox TNF are also able to promote morphological changes in the may belong to a class of genes, the expression of which is endothelial phenotype that resemble the early stages of the reversibly down-regulated during the proliferation of vascular non-terminal differentiation pathway in vitro (35, 36). endothelial cells. Whereas considerable attention has been focused on the Acknowledgments-We thank Drs. W. H. Burgess and R. Friesel ability of cytokines to stimulate prostanoid synthesis, relatively little is known about mediators that suppress prosta- for critical reading of the manuscript and S. Young and K. Wawzinski glandin synthesis. Such mechanisms are assumed to be a part for expert secretarial support. of the natural anti-inflammatory system. Glucocorticoidsare REFERENCES perhaps the best known natural inhibitors of prostanoid syn1. Needleman, P., Turk, J., Jakshick, B. A., Morrison, A. R., and thesis (37) and appear to exert their actions primarily by Lefiowith, J. B. (1986) Annu. Rev. Biochem. 55,69-102 inhibiting the de novo synthesis of the Cox protein (11, 38, 2. Moncada, S., Higgs, E. A., and Vane, J. R. (1977) Lancet 1, 1839). Early studies on the recovery ofCox enzyme activity 20 after irreversible inhibition by aspirin (22) or indomethacin 3. Smith, W. L. (1986) Annu. Reu. Physiol. 48,251-262

- +- +

“ ”

“

HBGF-1

Suppression of Cyclooxygenase Gene Expression by HBGF-1 4. Smith, W.L., and Lands, W. E. M. (1972) Biochemistry 1 1 , 3276-3285 5. Chen, Y.P., Bienkowski, M. J., and Marnett,L. J. (1987) J. Bid. Chem. 262,16892-16899 6. Wu, K. K., Hatzakis, H., Los, S. S., Seong, D. C., Sanduja, S. K., and Tai, H. H. (1988) J. Bid. Chem. 263,19043-19047 7. Rossi, V., Breviario, F., Ghezzi, P., Dejana, E., and Mantovani, A. (1986) Science 2 2 9 , 174-176 8. Bailey, J. M., Muza, B., Hla, T., and Salata, K. (1985) J. Lipid. Res. 2 6 , 54-61 9. Ristimaki, A., Ylikovkala, O., and Viinikka, L. (1990) Arteriosclerosis 10,653-657 10. Raz, A., Wyche, A., Siegel, N., and Needleman, P. (1988) J. Biol. Chem. 263,3022-3028 11. Raz, A., Wyche, A., and Needleman, P. (1989) Proc. Natl. Acad. Sci. U. S. A. 8 6 , 1657-1661 12. Maier, J. A.M., Hla, T., and Maciag, T. (1990) J. Biol. Chem. 265,10805-10808 13. Maier, J. A. M., Voulalas, P., Roeder, D., and Maciag, T. (1990) Science 249,1570-1574 14. Smith, W. L. (1989) Biochem. J. 259,315-324 15. Ryan, U. (1988) The Endothelial Cell, CRC Press, Inc., Boca Raton, FL 16. Maciag, T., Hoover, G . A., Stemmerman, M. B., and Weinstein, R. (1981) J. Cell. Biol. 9 1 , 420-426 17. Burgess, W. H., Mehlman, T., Marshak, D. R., Fraser, B. A., and Maciag, T. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , 7216-7220 18. Maciag, T., Mehlman, T., Friesel, R., and Schreiber, A. B. (1984) Science 225,932-935 19. Burgess, W. H., and Maciag, T. (1989) Annu. Reu. Biochem. 5 8 , 575-606 20. Lee, P. L., Johnson, D. E., Cousens, L. S., Fried, A., and Williams, L. T. (1989) Science 245,57-60 21. Burgess, W. H., Dionne, C. A., Kaplow, J., Mudd, R., Friesel, R., Zilberstein, A., Schlessinger, J., and Jaye, M. (1990) Mol. Cell. Biol. 10,4770-4777 22. Hla, T., and Bailey, J. M. (1989) Prostaglandin Leukotrienes Essent. Fatty Acids 3 6 , 175-184 23. Hasegawa, N., Yamamoto, M., and Yamamoto, K. (1988) J. Cell. Physiol. 137,603-607 24. Weksler, B. B. (1990) J. Cell. Physiol. 1 4 2 , 514-522 25. Imamura, T., Engleka, K., Zhan, X., Tokita, Y., Forough, R., Roeder, D., Jackson, A., Maier, J., Hla, T., and Maciag, T. (1990) Science 2 4 9 , 1567-1570 26. Hla, T., and Maciag, T. (1990) J. Biol. Chem. 265,9308-9313 27. Smith, W.L., DeWitt, D.L., Kraemer, S. A., Andrews, M. J., Hla, T., Maciag, T., and Shimokawa, T. (1990) Adu. Prostaglandin Thromboxane Leukotriene Res. 20,14-21

24063

28. Hla, T., Farrell, M., Kumar, A., and Bailey, J. M. (1986) Prostaglandins 32,829-845 29. Gay, C., and Winkles, J. (1989) J. Bid. Chem. 2 6 5 , 3284-3292 30. Norioka, K., Hara, M., Kitani, A., Hirose, T., Hirose, W., Harigai, M., Suzuki, K., Kawakami, M., Tabata, H., Kawagoe, M., and Nakamura, H. (1987) Biochem.Biophys.Res.Commun. 145, 969-975 31. Endo, H., Akahoshi, T. K., and Kashiwazaki, S. (1988) Biochem. Biophys. Res. Commun. 1 5 6 , 1007-1014 32. Frater-Schroeder, M., Risau, W., Hallman, R., Gautschi, P., and Bohlen, P. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,5277-5281 33. Doctrow, S., and Folkman, J. (1987) J. Cell Biol. 1 0 4 , 679-687 34. Morris, P. B., Hida, T., Blackshear, P. J., Klintwork, G. K., and Swain, J. L. (1988) Am. J.Physiol. 23,6318-6322 35. Montesano, R., and Orci, L. (1987) J. Cell. Physiol. 130, 284291 36. Maciag, T. (1990) in Important Advances in Oncology, (DeVita, V. T., Hellman, S., Rosenberg, S. A., eds) pp. 85-100, L.B. Lippincott, Philadelphia, PA 37. Masferrer, J. L., Zweifel, B. S., Seibert, K., and Needleman, P. (1990) J. Clin. Znuest. 8 6 , 1375-1379 38. Pash, J. M., and Bailey, 3. M. (1988) FASEB J. 2 , 2613-2618 39. Fu, J., Masferrer, J. L., Seibert, K., Raz, A,, and Needleman, P. (1990) J.Biol. Chem. 265,16737-16740 40. Gorodetsky, G., Hla, T., and Bailey, J. M. (1985) Pharmacol. Res. Comm. 17,447-455 41. Sano, H., Forough, R., Maier, J. A. M., Case, J. P., Jackson, A., Engleka, K., Maciag, T., and Wilder, R. L. (1990) J. Cell Bid. 110,1417-1426 42. Bevilacqua, M. P., Schleef, R. R., Gimbrone, M. A., Jr.,and Loskutoff, D. J. (1986) J. Clin. Invest. 7 8 , 587-591 43. Konkle, B. A., Kollros, P. R., and Kelley, M. D. (1990) J. Biol. Chem. 265,21867-21873 44. Jaye, M., Howk, W., Burgess, W. H., Ricca, G. A., Chin, I. M., Ravera, M., O'Brien, S., Modi, W. S., Maciag, T., and Drohan, W. N. (1986) Science 233,541-545 45. Lomedico, P. T.,Gubler, U., Hellmann, C. P., Dukovich, M., Giri, J., Pan, Y. E., Collier, K., Semionow, R., Chua, A. O., and Mizel, S. B. (1984) Nature 312, 458-462 46. Brunger, A. T. (1988) J.Mol. Bid. 203,803-810 47. Zhu, X., Komiya, H., Chirino, A., Faham, S., Fox, G. M., Arakawa, T., Hus, B. T., and Rees, D. C. (1991) Science 251,90-93 48. Bouche, G . N., Gas, H., Prats, V., Baldin, J. P., Tauber, J., Teissie, J., and Amalric, F. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,6770-6774 49. Grenfell, S., Smithers, N., Miller, K., andSolar, R. (1989) Biochem. J. 264,813-817 50. Yokoyama, C.,and Tanabe, T. (1989) Biochem.Biophys.Res. Commun. 165,888-894