Cyclooxygenase Is an Immediate-early Gene Induced by Interleukin-1 in Human Endothelial Cells*

The monokine interleukin-1 (IL- 1) inhibits endothe- lial cell growth and induces prostacyclin production in human endothelial cells. Since cyclooxygenase (Cox) is the rate-limiting enzyme in the synthesis of prosta- noids, we evaluated the ability of IL- 1 to stimulate Cox expression by human umbilical vein endothelial cells (HUVEC) in vitro. Our data demonstrate that 1) the Cox mRNA is expressed at low levels in untreated cells; 2) IL-la induces the Cox mRNA within 2 h, and this induction is sustained for more than 24 h; 3) IL-la induction is dose-dependent; 4) cycloheximide poten-tiates the induction of the Cox mRNA by IL-la while actinomycin D prevents the induction, and 5) IL-la also stimulates Cox production in a time-dependent fashion which correlates with the increase in prosta- cyclin synthesis. These data suggest that Cox is an immediate-early gene induced by IL-1 in HUVEC and may contribute to the regulation of the endothelial cell differentiation pathway in vitro.

The monokine interleukin-1 (IL-1) inhibits endothelial cell growth and induces prostacyclin production in human endothelial cells. Since cyclooxygenase (Cox) is the rate-limiting enzyme in the synthesis of prostanoids, we evaluated the ability of IL-1 to stimulate Cox expression by human umbilical vein endothelial cells (HUVEC) in vitro. Our data demonstrate that 1) the Cox mRNA is expressed at low levels in untreated cells; 2) IL-la induces the Cox mRNA within 2 h, and this induction is sustained for more than 24 h; 3) IL-la induction is dose-dependent; 4) cycloheximide potentiates the induction of the Cox mRNA by IL-la while actinomycin D prevents the induction, and 5) IL-la also stimulates Cox production in a time-dependent fashion which correlates with the increase in prostacyclin synthesis. These data suggest that Cox is an immediate-early gene induced by IL-1 in HUVEC and may contribute to the regulation of the endothelial cell differentiation pathway in vitro.
Endothelial cells comprise the lining of all blood vessels and contribute to organ physiology and to the pathology of human diseases. A few of the functions of the endothelial cell include the maintenance of vascular tone and the production of prothrombotic and anti-thrombotic activities among others (reviewed in Ref. 1). Inflammatory cytokines such as interleukin-l (IL-l) ' are potent modulators of endothelial cell function. IL-l inhibits the growth of endothelial cells (2) and alters the monolayer phenotype of the endothelial cell in vitro (3). Indeed, the endothelial cell assumes an elongated fibroblast-like phenotype in the presence of IL-l, a phenotype that is present during the very early stages of the endothelial differentiation pathway in uitro (4, 5). In addition, IL-l also induces the expression of tissue factor procoagulant activity (6-g), increases plasminogen activator inhibitor-l activity (9-12), and decreases tissue plasminogen activator activity (12). IL-l also induces the production of prostacyclin (13, 14), a potent vasodilator and inhibitor of platelet aggregation (15), and, as a result, may significantly contribute to the maintenance of vascular tone. Further, the induction of prostanoids (cyclooxygenase products) by IL-l in the local vasculature of the hypothalamus is considered to be the primary event regulating body temperature (16). Cyclooxygenase (Cox; EC 1.14.99.1) is the first enzyme of the pathway in which arachidonic acid is converted to prostacyclin and prostaglandins (17). The enzyme is localized primarily in the endoplasmic reticulum (18) and has been purified as a dimer of 70-kDa subunits (19,20). The half-life of Cox appears to be extremely short (tl,z <lo min) as estimated by in vitro (21) as well as ex viuo studies (22). In addition, the enzyme also displays a phenomenon of irreversible self-inactivation (23); it is estimated that after 1500 turnovers of product formation, the enzyme is irreversibly inactivated by a yet unidentified peroxide-like molecule (24). Recently human genomic Cox has been cloned and the exonintron format described (25).
De nouo expression of Cox is assumed to be instrumental in the control of prostanoid production.
Indeed, increased Cox enzyme activity leads to increased prostaglandin synthesis in murine smooth muscle cells and osteoblastic ceils treated with epidermal growth factor (26,27), in 3T3 cells treated with platelet-derived growth factor (28,33), in human amnion cells treated with epidermal growth factor (29), in human umbilical endothelial cells treated with IL-2 (30) or phorbol 12-myristate 13-acetate (21), and in human dermal fibroblasts stimulated by IL-l (31). In human dermal fibroblasts, selective inhibitors of transcription and translation were recently used to suggest that transcriptional as well as post-transcriptional events are responsible for the induction of Cox synthesis by IL-l (32). However, the precise mechanism of Cox expression regulation has not been determined. To gain insights into the mechanism of IL-l induction of Cox in human endothelial cells, we have studied the expression of Cox transcript in human umbilical vein endothelial cells treated with IL-la.
In this report, we demonstrate that IL-la stimulates the expression of the Cox mRNA in a timeand dose-dependent fashion. This correlates with the kinetics of induction of Cox polypeptide and prostacyclin synthesis. Further, the regulation of Cox expression by IL-la appears to mimic the behavior of immediate-early response genes induced by polypeptide growth factors (34,35). Thus, we suggest that Cox is an immediate-early response gene in human endothelial cell and may be involved in the regulation of the human endothelial ceil differentiation pathway in uitro.  (42). The reaction cycles were as follows: 94 'C for 1 min, 55 "C for 2 min, 72 "C for 3 min for 40 cycles. The sequence of the sense and antisense primers for Cox is: 5'-GCT GGG AGT CTT TCT CCA ACG TGA G-3' and 5'-GGC AAT GCG GTT GCG GTA TTG GAA CT-3' (25

AND DISCUSSION
Since the monokine IL-1 induces prostacyclin synthesis in human endothelial cells (13, 14) and the enzymatic activity of Cox governs prostaglandin and prostacyclin production (17), we isolated a human Cox cDNA fragment (38) and used it to evaluate the expression of the Cox mRNA expression in HUVEC. Northern blot analysis of poly(A+) RNA prepared from HUVEC incubated for 4 h with IL-la (1 rig/ml) demonstrated that HUVEC express the Cox RNA at low levels. Indeed, the Cox transcript was not detectable in 5 pg of poly(A') RNA derived from untreated cells (Fig. lA, lane 1). However, following stimulation with IL-la, the Cox mRNA was readily detectable as a single band having an approximate size of 3 kb (lane 3). The low abundance of Cox mRNA was also confirmed when total RNA was reverse transcribed and enzymatically amplified using Cox-specific oligonucleotide primers (Fig. lB, lane 1). In contrast, HUVEC treated with IL-la in the presence or in the absence of cycloheximide (chx; 5 pg/ml) demonstrated a superinduction of the Cox mRNA levels (Fig. lA, lane 4), whereas chx alone did not appreciably alter the level of the Cox transcript (Fig. lA, lane 2). Because the induction of the Cox mRNA by IL-la could be due to an increase in the rate of transcription, stabilization of previously transcribed mRNAs, or a combination of both, the inhibition of transcription with actinomycin D, an inhibitor of RNA synthesis, was examined. The data shown in Fig. 1B demonstrate that actinomycin D (10 rg/ml) blocked IL-la induction of the Cox mRNA (Fig. 1B). Since the level of Cox expression is relatively low, we used the reverse transcription-polymerase chain reaction (RT-PCR) method to assay for the presence of Cox mRNA. The products of the amplification were analyzed on a 1.0% agarose gel and stained with ethidium bromide. The oligonucleotide primers used for the specific amplification of Cox predict the generation of a 0.7-kb product, and a 0.7-kb band was readily detected (Fig. 1B). The 0.7-kb band strongly hybridized with radiolabeled Cox cDNA in a Southern blot (data not shown), confirming the identity of the 0.7-kb band as Cox. Further, amplification of the same   Lanes 1 and 2, control; hmes 3 and 4, 0.1 ng/mI IL-ltu; lanes 5 and 6, 1 rig/ml IL-ltr.
Serial dilution of RNA samples followed by RT-PCR assay demonstrated that Cox RNA was induced approximately 6-fold by IL-1 and chx after 4 h of incubation. Together, these data suggest that the increase in Cox mRNA expression by IL-la may represent an early transcriptional event involved in endothelial cell activation. Nuclear run-on assays were used to determine whether the increase in Cox mRNA was regulated by a transcriptional or posttranscriptional mechanism.
However, these attempts were without success' due to the low abundance of the Cox mRNA in this system.
The effect of IL-la on the expression of the Cox mRNA was dependent upon the concentration of IL-la (Fig. 2, A and R). In addition, the potentiation of Cox mRNA expression by chx was observed at both concentrations of IL-la (Fig. 2, A  and B), an observation consistent with the suggestion that chx stabilizes the Cox mRNA. The kinetics of Cox mRNA expression in response to IL-la (1 rig/ml) in the presence and in the absence of chx were also examined. As shown in Fig. 3 the expression of the Cox mRNA. It is interesting to point out that the effect of chx on the delayed transcriptional shutoff of the c-fos gene occurs rapidly (l-4 h), although the stabilization of the c-jos mRNA is sustained (47). We argue that a similar mechanism could explain chx potentiation of the Cox transcript by IL-1 in HUVE cells.
To determine whether IL-la also stimulates Cox production in HUVEC, we analyzed the level of the Cox polypeptide by Western blot analysis. As shown in Fig. 4, an increase in the level of the 70-kDa Cox polypeptide was observed in extracts prepared from HUVEC incubated with IL-Lou (1 ng/ ml). Accumulation of Cox was observed at 6 h, being maximal at 16 h. No band was observed in the Western blot format when the antibody was preincubated with purified ovine Cox (data not shown). Table I shows the stimulation of PGIr synthesis by HUVEC treated with IL-1 (1 rig/ml). PGIr production was clearly induced after 8 h of incubation with IL-l and was maximal after 24 h. The kinetics of IL-1 induction of Cox mRNA, Cox polypeptide, and PGI, synthesis suggest that the IL-l effect may be mediated mainly by its effect on de novo Cox synthesis in HUVECs. It is noteworthy that the effect of IL-1 on Cox mRNA appeared after 2 h and was sustained thereafter, while the increases in the level of Cox protein and enzyme activity appeared later (6-8 h) IL-l Induces Cox mRNA in HUVEC synthesis was achieved, suggesting that platelet-derived growth factor-stimulated prostaglandin synthesis is not dependent upon de nouo Cox expression in 3T3 cells (33). The ability to detect the Cox transcript in total RNA samples extracted from untreated 3T3 cells demonstrates that Cox expression is relatively high in this cell line. In contrast, the low abundance of Cox mRNA in HUVEC could depend on the requirement for HBGF-1 in the cell culture system (36). Indeed HBGF-1 has been shown to reduce prostaglandin synthesis in HUVEC (47).
IL-l has been shown to inhibit endothelial cell growth and alter the morphology of the endothelial-cell monolayer from the traditional cobblestone phenotype to a fibroblast-like cell shape in vitro (3). The later phenotype resembles the morphology assumed by the endothelial cell during the very early stages of the endothelial cell differentiation pathway in vitro (4, 5). A characteristic of the regulation of transcription during the immediate-early stages of the cell cycle for a variety of immediate-early genes includes the sensitivity to superinduction by chx (35,(48)(49)(50)(51). Indeed, the rapid transcriptional activation of c-fos, c-jun, and c-myc in response to inductive signals and their superinduction by chx is well described (50) and occurs as a result of the inhibition of mRNA degradation as well as a prolongation of transcriptional shut-off. Thus, the superinduction of Cox by IL-l and chx in HUVEC suggests that Cox is an immediate-early human endothelial cell differentiation-response gene. Since Cox is a key regulatory enzyme for the synthesis of prostanoids (18) and many of the biological effects of IL-l, such as pyresis, inflammation, and vasodilation, are prostanoid-mediated (52), these data may provide new insights into the molecular mechanisms utilized by IL-1 to regulate prostanoid synthesis.